Regulation of Oxidative Stress Response by CosR, an Essential Response Regulator in Campylobacter jejuni

CosR (Campylobacter oxidative stress regulator; Cj0355c) is an OmpR-type response regulator essential for the viability of Campylobacter jejuni, a leading foodborne pathogen causing human gastroenteritis worldwide. Despite importance, the function of CosR remains completely unknown mainly because of cell death caused by its knockout mutation. To overcome this technical limitation, in this study, antisense technology was used to investigate the regulatory function of CosR by modulating the level of CosR expression. Two-dimensional gel electrophoresis (2DGE) was performed to identify the CosR regulon either by suppressing CosR expression with antisense peptide nucleic acid (PNA) or by overexpressing CosR in C. jejuni. According to the results of 2DGE, CosR regulated 32 proteins involved in various cellular processes. Notably, CosR negatively regulated a few key proteins of the oxidative stress response of C. jejuni, such as SodB, Dps, Rrc and LuxS, whereas CosR positively controlled AhpC. Electrophoretic mobility shift assay showed that CosR directly bound to the promoter region of the oxidative stress genes. DNase I footprinting assays identified 21-bp CosR binding sequences in the sodB and ahpC promoters, suggesting CosR specifically recognizes and binds to the regulated genes. Interestingly, the level of CosR protein was significantly reduced by paraquat (a superoxide generator) but not by hydrogen peroxide. Consistent with the overall negative regulation of oxidative stress defense proteins by CosR, the CosR knockdown by antisense rendered C. jejuni more resistant to oxidative stress compared to the wild type. Overall, this study reveals the important role played by the essential response regulator CosR in the oxidative stress defense of C. jejuni.


Introduction
Campylobacter jejuni is a Gram-negative foodborne pathogen belonging to the e-subdivision of Proteobacteria [1,2]. As a microaerophilic bacterium, C. jejuni is highly sensitive to atmospheric oxygen [3]. Bacterial growth in the presence of oxygen inevitably generates reactive oxygen species, such as the superoxide anion (O 2 2 ), hydrogen peroxide (H 2 O 2 ) and hydroxyl radical (HO?), which can cause severe damage to biological molecules including nucleic acids and proteins [4,5]. To detoxify reactive oxygen species, C. jejuni utilizes various oxidative stress defense proteins, such as superoxide dismutase (SodB), catalase (KatA), alkyl hydroperoxide reductase (AhpC) and DNA-binding protein from starved cells (Dps) [6][7][8][9]. While all these oxidative stress proteins play a critical role in oxidative stress defense, as the sole superoxide dismutase in C. jejuni, SodB is the most important enzyme in the detoxification of O 2 2 , contributing to chicken colonization and intracellular survival of C. jejuni [9][10][11].
Two-component regulatory systems are one of the most important signal transduction mechanisms to modulate gene expression in bacteria [12]. The genome of C. jejuni NCTC 11168 contains only a small number of regulatory genes including those encoding three sigma factors, seven histidine kinases and 12 response regulators [13]. Several two-component regulatory systems that have been studied in C. jejuni to date are associated with bacterial motility, animal colonization and bile acid resistance [14][15][16][17][18][19]. Cj0355c, designated CosR (Campylobacter oxidative stress regulator) in this study, is a putative response regulator essential for C. jejuni viability, as its knockout mutation is lethal [17,20]. A proteomic study investigating protein expression profiles under oxidative stress showed that CosR expression is reduced after exposure to paraquat (O 2 2 ), suggesting that CosR may be involved in the oxidative stress response of C. jejuni [20]. Although CosR has been expected to play a critical role in sustaining the viability of C. jejuni, the inability to generate its knockout mutant due to lethality has limited functional characterization of this essential response regulator [17,20].
Antisense regulation is a natural gene regulation mechanism mediated through modulation of mRNA stability, transcription, or translation in eukaryotes and prokaryotes [21,22]. This natural process can be exploited by using synthetic antisense agents in biological research, which have significant versatility [23][24][25]. Binding of antisense agents to mRNA inhibits target gene expression by either inducing RNase H-mediated mRNA degradation or sterically interfering with ribosomal binding [23,25]. Among antisense agents that have been developed thus far, peptide nucleic acid (PNA) is often utilized for antisensemediated gene silencing because of its stability and high affinity for nucleic acids [23,26,27]. While antisense technology is frequently used in eukaryotic systems, it is seldom used in prokaryotes for gene expression control despite its usefulness. However, several recent studies have demonstrated that antisense technology can successfully control gene expression in bacteria, such as Escherichia coli, Staphylococcus aureus and C. jejuni [28][29][30][31]. As a novel approach to overcome lethality resulting from knockout mutagenesis, we performed antisense knockdown of CosR to characterize its function. Overall protein expression changes incurred by either CosR knockdown or overexpression demonstrate that CosR plays an important role in the regulation of oxidative stress response in C. jejuni.

Amino acid sequence analysis of CosR and genomic organization of cosR flanking regions
CosR is an OmpR-type response regulator essential for the viability of C. jejuni [17,20]. Basic Local Alignment Search Tool (BLAST) results showed that CosR is highly conserved among Campylobacter species, including C. jejuni, C. coli, C. fetus, C. hominis, C. curvus, C. lari and C. concisus (data not shown). Interestingly, CosR homologs with high amino acid sequence similarities were found primarily in e-proteobacteria, such as Helicobacter and Wolinella ( Figure 1A). Compared with C. jejuni CosR, the Nterminal receiver domain of H. pylori HP1043 exhibited lower sequence similarity (38% identity with five gaps over 112 amino acids) than the highly conserved C-terminal DNA-binding domain (83% identity over 111 amino acids; Figure 1A). The conserved aspartate residue D8 in the CosR homologs is replaced with a lysine in HP1043, and four amino acid residues in the receiver domain (51 st ,54 th amino acid residues in CosR) are deleted in HP1043 ( Figure 1A). Although an asparagine residue (N51) in CosR is substituted for a conserved aspartate residue present in the CosR homologs of Wolinella and Arcobacter, other conserved aspartate residues (e.g. D8, D56, and D58) may serve as an alternative phosphorylation site ( Figure 1A). Genes encoding the sensor kinase and response regulator of the two-component regulatory system are often positioned adjacent to each other; however, the potential cognate sensor kinase was not found near cosR ( Figure 1B). In C. jejuni, the cosR gene is located downstream of ferredoxin (fdxB) and upstream of dihydroneopterin aldolase (folB; Figure 1B).

Selective response of CosR to O 2 2 stress
According to a recent proteomic study performed by Garénaux et al. [20], the CosR protein level was reduced by exposure to paraquat (O 2 2 ). Consistent with the report, treatment of C. jejuni with paraquat significantly decreased the level of CosR protein based on western blotting (Figure 2A). To assess if CosR expression can be altered in response to peroxide stress, western blot analysis was performed after exposure to H 2 O 2 . Interestingly, H 2 O 2 stress did not make any differences in the CosR level even at the concentrations causing similar viability changes as paraquat ( Figure 2B). These findings showed that CosR expression changes responding to paraquat (O 2 2 ).

Modulation of CosR expression with antisense PNA
Due to the essentiality for C. jejuni viability, it has not been possible to knockout the cosR gene, because its knockout mutation leads to cell death [17,20]. As an alternative in this study, antisense-mediated gene silencing was utilized to knockdown cosR expression. Western blot analysis confirmed that the level of CosR expression was reduced in response to the increase in the concentration of antisense CosR-specific PNA (CosR-PNA) without affecting overall protein expression based on the results of SDS-PAGE ( Figure 3A). The viability of C. jejuni was slightly reduced with 2 mM CosR-PNA and significantly decreased with 4 mM CosR-PNA ( Figure 3B). A PNA oligomer with the same number of bases as CosR-PNA, but lacking significant sequence similarity to genes in the C. jejuni genome, was used as a negative control. This PNA control did not reduce bacterial viability even at 4 mM ( Figure 3B). These results demonstrated that antisensemediated gene knockdown specifically controlled the level of CosR.

Identification of the CosR regulon
To identify the CosR regulon, protein expression profiles were analyzed after CosR knockdown and CosR overexpression. The effective concentration of CosR-PNA was determined by altering the treatment time and the CosR-PNA concentration to suppress CosR expression without significant viability changes. Based on the results of western blot analysis and viability tests (Figure 3), 1.5 mM CosR-PNA was determined to be the most appropriate concentration to knockdown CosR. Alternatively, the effect of CosR overexpression was evaluated with a strain that harbors an extra copy of cosR in the chromosome. Quantitative real-time PCR (qRT-PCR) and western blot analysis confirmed cosR overexpression at the transcriptional and translational levels in the strain ( Figure S1). Analysis of the protein expression profiles in the CosRoverexpression strain provided complementary evidence for the protein expression changes observed in the CosR-knockdown condition. Altered CosR levels by knockdown or overexpression influenced the expression of 32 proteins involved in various cell functions, including macromolecule biosynthesis, energy metabolism, respiration, heat shock response, regulation and oxidative stress defense ( Figure 4 and Table 1). CosR appeared in the twodimensional gel as three spots located parallel with different pI values (Figures 4 and 5A), presumably due to protein modifications such as phosphorylation. Based on the results of two-dimensional gel electrophoresis (2DGE), we found that the level of CosR was reduced by PNA knockdown and elevated in the CosRoverexpression strain (Figures 4 and 5A), confirming successful control of the CosR level in both CosR knockdown and overexpression conditions. CosR knockdown resulted in a 2.1fold decrease in the protein level of PckA (phosphoenolpyruvate carboxykinase), an essential protein in C. jejuni [32], whereas CosR overexpression increased PckA protein levels by 3-fold ( Figure 4 and Table 1). CosR affected the expression level of proteins associated with copper tolerance and gene regulation ( Figure 4 and Table 1); CosR negatively regulated Cj1516, a homolog of the multicopper oxidase CueO in E. coli, and CprR, which is another essential response regulator in C. jejuni [18,33]. Notably, CosR regulated the expression of several important proteins of oxidative stress defense in C. jejuni; CosR negatively regulated SodB, Dps, LuxS and Rrc, whereas AhpC was positively regulated by CosR (Figures 4, 5A and Table 1). Electrophoretic mobility shift assay (EMSA) was performed with rCosR to determine if CosR directly binds to the promoter regions of sodB, dps, luxS and ahpC. The results showed that rCosR bound to the promoter regions, and binding of rCosR to the DNA resulted in concentration-dependent mobility shift ( Figure 5B). Unlabeled target DNA fragments effectively competed with the labeled DNA probes of sodB, dps, luxS and ahpC, but the internal coding regions did not ( Figure 5B), indicating specific binding of CosR to the target promoter regions. These results demonstrate that CosR is a pleiotropic regulator, particularly modulating the expression of oxidative stress genes in C. jejuni.

Determination of CosR binding sequence
Among the oxidative stress proteins identified by 2DGE, CosR negatively and positively regulated SodB and AhpC, respectively ( Figure 5A), and bound to the promoter region of the encoding genes ( Figure 5B). DNase I footprinting assays were performed to determine the CosR binding sites in the promoter region of sodB and ahpC genes. DNase I footprinting results were interpreted based on the previous studies of the sodB and ahpC promoter sequences in C. jejuni [6,9]. The results showed a region immediately downstream of the 210 region of the sodB promoter, which spans the leader sequence and the start codon of sodB, was protected from DNase I treatment ( Figure 6A). Two CosR binding sites were found in the ahpC promoter, including the 237 to 217 region overlapping the 235 region of the ahpC promoter and the +1 to +21 region that is between the 210 region and the ahpC start codon ( Figure 6B). The nucleotide sequences of the three CosR binding sites obtained from the DNase I footprinting assays were aligned to predict the consensus CosR binding sequence ( Figure 6C). The alignment revealed a 21-bp CosR-binding  [60,74], the highly conserved residues in the receiver domain are indicated by arrowheads. The aspartate residue which differs from the highly conserved aspartate residues in the other CosR homologs is marked with a star. (B) The analysis of cosR flanking region shows the lack of its cognate sensor kinase in its vicinity. doi:10.1371/journal.pone.0022300.g001 sequence (tttaAanAaAAaTtAtgaTTt, 'n' represents any nucleotide, lowercase letters indicate less conserved residues, and capital letters are highly conserved residues; Figure 6C), suggesting that CosR binds to a specific target site in the CosR-regulated genes. Putative CosR binding sites were also identified in the regulatory region of dps and luxS genes based on homology to the CosRbinding sequence ( Figure 6D).

Regulation of SodB by CosR
Although SodB is a key enzyme contributing to the detoxification of O 2 2 , its regulation has not been well understood in C. jejuni. Since the results of 2DGE, EMSA and DNase I footprinting assays indicated that CosR may regulate sodB transcription (Figures 4, 5, and 6A), qRT-PCR and primer extension assays were performed to further confirm the regulation of sodB transcription by CosR. In both assays, antisense inhibition of CosR upregulated sodB transcription, whereas CosR overexpression reduced the level of sodB transcription compared to the wild type ( Figures 7A and 7B). To examine if the changes in SodB expression would affect its enzyme activity, the superoxide dismutase colorimetric assay was performed. Consistently, the superoxide dismutase activity was enhanced by CosR knockdown and attenuated by CosR overexpression ( Figure 7C). These results demonstrated that CosR regulates sodB transcription and affects the SodB enzyme activity.  Effect of CosR on oxidative stress resistance Because CosR regulates several important oxidative stress genes in C. jejuni, susceptibility to oxidative stress was measured at different CosR levels. Viable counts were determined before and after 1 hr exposure to oxidant. The C. jejuni strain overexpressing CosR was more susceptible to paraquat and H 2 O 2 compared to the wild type, whereas CosR knockdown by antisense rendered C. jejuni more resistant to oxidative stress ( Figure 8). These results showed that CosR significantly contributes to the oxidative stress resistance of C. jejuni.

Discussion
Given the limited presence of only three sigma factors in C. jejuni, the 12 two-component response regulators are thought to be important for the gene regulation of this bacterium [13]. Indeed, several two-component response regulators in C. jejuni have been reported to significantly contribute to the virulence features of this bacterial pathogen, such as bacterial motility, colonization and bile resistance [14,16,17,34]. Among the 12 response regulators, the response regulator CosR is known to be essential for the viability of C. jejuni [17,20], suggesting that CosR may regulate critical cellular processes in C. jejuni. In this study, the results of 2DGE analysis suggested that CosR is a pleiotropic regulator involved in various cellular processes, such as macromolecule biosynthesis, metabolism, gene regulation and oxidative stress response ( Figure 4 and Table 1). Interestingly, CosR regulated the expression of several important oxidative stress proteins in C. jejuni, including Dps, LuxS, Rrc, AhpC and SodB (Figures 4, 5A, and Table 1). Dps is an iron-binding protein contributing to oxidative stress defense by sequestering free intracellular iron available for the Fenton reaction, and its mutation makes C. jejuni more susceptible to H 2 O 2 than the wild type [8]. LuxS is S-ribosylhomocysteinase and its mutation downregulates ahpC and tpx in the presence of H 2 O 2 by an unknown mechanism, rendering C. jejuni susceptible to H 2 O 2 and cumene hydroperoxide [35]. Rrc is a non-heme iron protein that is extremely sensitive to reactive oxygen species [36]. AhpC confers resistance to cumene hydroperoxide [6,10].
Notably, CosR regulated SodB, a key enzyme of oxidative stress resistance in C. jejuni. Unlike other bacteria that often possess multiple forms of superoxide dismutase (SOD), such as Mncofactored SOD (SodA), Fe-cofactored SOD (SodB), and Cu and Zn-cofactored SOD (SodC) in E. coli [5], C. jejuni harbors only SodB [9]. As the sole SOD of C. jejuni, SodB plays an important role in C. jejuni's resistance to oxidative stress and survival in various conditions such as foods, chicken guts and epithelial cells [9,10]. In this study, we demonstrated that CosR negatively regulates sodB expression. CosR bound to a region downstream of the 210 element of the sodB promoter ( Figure 6A). In many other bacteria, Fur (ferric uptake repressor) is directly or indirectly involved in SOD regulation. For example, Fur downregulates sodA, but upregulates sodB through suppression of the small regulatory RNA RyhB that inhibits sodB translation in E. coli [37][38][39]. Even in the closely related bacterium H. pylori, Fur regulates sodB expression by directly binding to its promoter [40]. However, a fur mutation in C. jejuni does not affect SodB activity [41], suggesting C. jejuni may possess unique mechanisms for sodB regulation. To the best of our knowledge, CosR is the first regulator shown to control sodB expression in C. jejuni.
Among the 32 CosR-regulated proteins identified in this study, 12 proteins were positively regulated by CosR ( Figure 4 and Table 1). To address if CosR binds to the positively regulated genes, we performed EMSA and DNase I footprinting assays targeting ahpC, which is the sole oxidative stress gene positively regulated by CosR in this study. EMSA demonstrated that CosR bound to the ahpC promoter ( Figure 5B) and DNase I footprinting assays identified two CosR binding sites in the ahpC promoter region ( Figure 6B). Whereas ahpC is often regulated by the peroxide-sensing activator OxyR in many bacteria [5,42], C. jejuni lacks OxyR; instead, PerR (peroxide regulon repressor) negatively regulates ahpC expression in C. jejuni [41]. Due to the irondependence of PerR, ahpC expression is suppressed in iron-rich conditions [6,41]; however, overall mechanisms for ahpC regulation remain unknown in C. jejuni. It would be interesting to carry out future studies to examine how PerR and CosR may coordinately regulate ahpC expression in C. jejuni.  Table 1. doi:10.1371/journal.pone.0022300.g004 In this work, we demonstrated that CosR binds to the promoter regions of sodB and ahpC and regulates their expressions differentially. At this point, the molecular mechanism of differential regulation by CosR remains unexplained; however, the regulation mechanisms of OmpR reported in other bacteria may provide valuable insight into this OmpR-type response The spot number corresponds to that indicated in Figure 4. Fold changes of multiple protein spots are separated by a slash in the same order from left to right as regulator in C. jejuni. In E. coli and Salmonella, the function of OmpR is influenced by phosphorylation, since phosphorylation significantly affects OmpR binding to DNA and its interaction with OmpR and RNA polymerase [43][44][45][46]. As a global regulator in E. coli, OmpR is reported to modulate expression of 125 genes positively and negatively [47]. In association with the sensor kinase EnvZ, particularly, OmpR regulation of major porin genes ompC and ompF may be the most well-defined regulatory role played by this response regulator. EnvZ has both kinase and phosphatase activities, phosphorylating OmpR at high medium osmolarity and dephosphorylating the phosphorylated form (OmpR-P) at low osmolarity. Four (F1, F2, F3, and F4) and three (C1, C2, and C3) OmpR binding sites are present in the promoters of ompF and ompC, respectively. At high osmolarity, OmpR-P binds to the upstream site F4 and interacts with OmpR-P in F1, F2 or F3, leading to formation of a loop by OmpR-OmpR interactions and thereby repressing ompF, whereas binding of OmpR-P to C2 and C3 at high osmolarity promotes interactions with RNA polymerase and activates ompC [48]. As shown in the porin gene regulation in E. coli, OmpR negatively and positively regulates porin genes depending on osmolarity and phosphorylation. In Salmonella, the mechanism of OmpR regulation has been relatively less extensively investigated compared to E. coli. OmpR activates SsrAB, a two-component regulatory system regulating the expression of the type III secretion system of Salmonella pathogenicity island 2 (SPI2) by binding to the promoter regions of ssrAB [49,50]. While the unphosphorylated OmpR binds to only a single spot in the ssrA promoter, OmpR-P binds to multiple sites in the ssrA and ssrB promoters [49][50][51]. As reflected in OmpR regulation of ssrAB and porin genes, the regulation mechanism of OmpR cannot be explained without considering phosphorylation. If phosphorylation influences the binding affinity of CosR to the target promoters as OmpR, the phosphorylated form of CosR may bind to more spots in the sodB and ahpC promoters than identified in this work. Similar to OmpR, presumably, CosR phosphoryla-tion would affect interactions with RNA polymerase and/or the CosR protein in the vicinity, either activating or repressing gene expression. This model may explain differential regulation of sodB and ahpC by CosR; however, all these possibilities can be proved after investigating the following research questions, such as if CosR is phosphorylated, whether CosR is an orphan response regulator or CosR has its cognate sensor kinase, how phosphorylation affects CosR binding to target promoters, and if CosR interacts with RNA polymerase and the CosR protein. These series of studies still await future investigations. A previous proteomic study reported that the level of CosR was reduced by paraquat (O 2 2 ) [20]. Consistently, western blot analysis in this study confirmed that the CosR level was significantly decreased by exposure to paraquat (Figure 2A). However, H 2 O 2 did not change the CosR level ( Figure 2B), suggesting CosR may respond selectively to O 2 2 stress. CosR repression by antisense significantly affected the viability of C. jejuni ( Figure 3B); however, the level of CosR was not reduced by exposure to H 2 O 2 despite substantial viability changes ( Figure 2B). These results suggest that killing of C. jejuni under oxidative stress is not necessarily mediated by CosR depending on the type of oxidative stress, though CosR is essential to the viability of C. jejuni. The CosR reduction in response to O 2 2 stress would derepress the oxidative stress genes that are negatively regulated by CosR, such as sodB and dps. Consequently, sodB derepression may confer resistance to O 2 2 stress because SodB converts O 2 2 to H 2 O 2 . As the sole superoxide dismutase present in C. jejuni, SodB is the only enzyme able to detoxify O 2 2 in C. jejuni. In addition, the derepression of Dps by CosR reduction may enhance the sequestration of free iron and reduce the production of hydroxyl radicals generated from O 2 2 and H 2 O 2 by the Fenton and Haber-Weiss reactions. In fact, there is no enzyme capable of detoxifying hydroxyl radicals, though hydroxyl radicals can damage most biomolecules [5]. In contrast to SodB and Dps, the AhpC level will be decreased by CosR reduction, because AhpC is positively  [18]. According to a proteomic analysis performed by Holmes et al. [52], AhpC was significantly upregulated in ironlimited conditions, whereas SodB was downregulated. The differential regulation of SodB and AhpC would achieve efficient defense against various stresses of different reactive oxygen species, though its molecular mechanism has not been understood. As key enzymes involved in oxidative stress resistance, AhpC and SodB act on different substrates. While SodB consumes O 2 2 , AhpC detoxifies organic hydroperoxides and can scavenge only low physiological concentrations of H 2 O 2 [53]; thus, AhpC provides resistance to organic hydroperoxides but not to H 2 O 2 in sensitivity tests in E. coli [53,54]. Similarly, ahpC mutations change the susceptibility of C. jejuni to neither menadione (a O 2 2 generator) nor H 2 O 2 [6,10]. The level of CosR protein was decreased by exposure to O 2 2 (Figure 2A), and subsequently the CosR reduction would derepress the expression of SodB and Dps, which are important oxidative stress defense proteins to overcome the imposed O 2 2 stress, but may reduce the expression of AhpC that cannot contribute to the detoxification of O 2 2 . The overall role of CosR in oxidative stress resistance was evaluated by performing sensitivity assays at different CosR levels.
Results of these sensitivity assays demonstrated that the CosR knockdown by antisense increased resistance to both O 2 2 and H 2 O 2 , whereas CosR overexpression decreased the resistance (Figure 8). The effect of CosR on the oxidative stress resistance of C. jejuni would result from cumulative changes in multiple oxidative stress resistance factors regulated by CosR; CosR knockdown upregulates the expression of a few oxidative stress resistance proteins. Additionally, SodB is known to confer resistance to both peroxide (e.g. H 2 O 2 and cumene hydroperoxide) and O 2 2 (e.g. manadione) in C. jejuni [9,10], whereas AhpC and KatA are involved in defense only against cumene hydroperoxide and H 2 O 2 , respectively [10]. Therefore, the upregulation of SodB by CosR knockdown may contribute to the resistance to both paraquat and H 2 O 2 ( Figure 8). Iron participates in the Fenton reaction and plays an important role in generating reactive oxygen species; thus, the iron uptake regulator often regulates oxidative stress defense genes, and oxidative stress regulators also control the expression of genes involved in iron homeostasis [10,52]. To investigate if CosR may affect iron levels in C. jejuni, we determined intracellular iron levels using Ferene-S, a chromogenic iron chelator [55,56]. The results showed that both CosR knockdown and overexpression did not change intracellular iron levels (data not shown); however, this does not mean that CosR is functionally independent of other iron regulators. Since the Fur regulon proteins p19 and AhpC were downregulated in the CosR knockdown condition (Table 1), CosR likely interacts with Fur directly or indirectly. This possibility needs to be examined in future research.
CosR homologs are found primarily in the e-proteobacteria, such as Helicobacter and Wolinella ( Figure 1A). CosR exhibits high (60%) amino acid sequence similarity to H. pylori HP1043 and other homologs ( Figure 1A). Interestingly, HP1043 is also one of two essential two-component response regulators in the closely related bacterium H. pylori [57,58]. CosR expressed under the control of the HP1043 promoter can substitute for HP1043 in H. pylori [59], suggesting functional similarity between these two proteins. Compared to other CosR homologs, however, HP1043 has unique differences particularly in its N-terminus, including the substitution for and deletion of highly conserved aspartate residues, which are potential sites of phosphorylation ( Figure 1A). The alteration of these residues suggests that phosphorylation of HP1043 may differ from that of other CosR homologs. In fact, HP1043 is not phosphorylated by the small molecule phosphodonor acetyl phosphate in vitro and is known to have a phosphorylation-independent action [23,60]. Although it has not been understood if CosR is phosphorylated, the conformational heterogeneity of CosR observed in the 2DGE analysis suggests that CosR may be phosphorylated within C. jejuni ( Figures 5A). Since no potential histidine kinase was identified near the cosR gene, it remains unknown whether CosR would be phosphorylated by a yet-unidentified histidine kinase or by low molecular weight phosphodonors such as acetyl phosphate in C. jejuni. Acetyl phosphate is the intermediate produced in the enzymatic conversion between acetyl coenzyme A and acetate, and can autophosphorylate a number of two-component response regulators [61]. According to the genome sequence, C. jejuni possesses pta (Cj0688) and ackA (Cj0689), encoding phosphotransacetylase (Pta) and acetate kinase (AckA), respectively [13], which are responsible for the enzymatic reactions producing acetyl phosphate. Details of CosR phosphorylation will be investigated in future studies.
To characterize essential genes, conditional mutants have been constructed in E. coli, Salmonella, and Helicobacter [62][63][64]; however, construction of conditional mutants is not technically feasible in C. jejuni, because tools for genetic manipulation are less well developed in this fastidious bacterium. Thus, as a novel approach to overcome this technical limitation in investigating an essential gene, antisense-mediated gene silencing was used in this study to knockdown the expression of CosR. Antisense technology is a useful genetic tool in eukaryotic systems, but is seldom used in prokaryotic studies [26,27]. Jeon and Zhang recently reported that antisense successfully controlled the expression of C. jejuni CmeA, a periplasmic component of the CmeABC efflux pump [30]. In the present study, 8-hr PNA treatment suppressed CosR expression   Figure 3A), but treatment longer than 12 hrs reduced the effect on CosR knockdown, as assessed by western blotting (data not shown). Considering the high stability of PNA, these findings may be due to significant increase in bacterial population, which may reduce the ratio of PNA to bacteria. For that reason, application of antisense technology in prokaryotes should consider rapid bacterial growth compared to relatively slowly growing eukaryotic cells. High CosR-PNA concentrations significantly reduced the viability of C. jejuni, confirming the essentiality of CosR for bacterial survival ( Figure 3B). As expected, CosR is involved in the regulation of critical metabolic pathways in C. jejuni. The results of 2DGE demonstrated that PckA, an essential enzyme of central carbon metabolism, was positively regulated by CosR ( Figure 4 and Table 1). C. jejuni lacks 6-phosphofructokinase and is therefore unable to utilize exogenous glucose to produce pyruvate [13]. Instead, pyruvate is supplied by other metabolites that are primarily derived from amino acids. PckA converts oxaloacetate to phosphoenolpyruvate, which is then converted to pyruvate by pyruvate kinase [32]. Because C. jejuni does not have phosphoenolpyruvate synthase, PckA is the only way to produce phosphoenolpyruvate, and is thus essential for C. jejuni growth [32]. Although it has not been understood whether CosR regulates PckA directly or not, this regulation may partly explain why CosR is essential for C. jejuni viability. CosR also regulated CprR, another essential response regulator in C. jejuni. While the role of CprR has not been elucidated due to the inability to construct its mutant, its cognate sensor kinase CprS was shown to be involved in biofilm formation, oxidative stress resistance, and chicken colonization [18]. Interestingly, a CprS mutation decreased CosR expression, suggesting a functional relationship between these two essential response regulators; however, detailed mechanisms for sustaining the viability of C. jejuni by CosR still remain largely unknown.
In the present study, our findings provided new insights into the regulatory mechanisms for oxidative stress resistance in C. jejuni, an important human pathogen, clearly demonstrating that CosR regulates the expression of several key oxidative stress genes, particularly sodB, which is the sole superoxide dismutase gene in C. jejuni. In addition, we showed that antisense-mediated gene silencing is a useful approach to investigate essential genes in bacterial species in which genetic tools have not been well developed. To our knowledge, this is the first proteomic study to characterize an essential regulator in bacteria using antisense. The same technical approach will be useful to the investigation of essential genes in C. jejuni and other bacteria as well.

Materials and Methods
Bacterial strains and culture conditions C. jejuni NCTC 11168, the first genome-sequenced Campylobacter strain, was used for this study [13]. C. jejuni was grown at 42uC on Mueller-Hinton (MH) media (Difco) under microaerobic conditions generated by Anoxomat TM (Mart Microbiology B.V, Lichtenvoorde, Netherlands). Culture media were occasionally supplemented with kanamycin (50 mg ml 21 ) or chloramphenicol (10 mg ml 21 ) where required. Broth cultures were microaerobically grown with shaking at 180 rpm after adjusting the absorbance at 600 nm to approximately 0.07.

Preparation and treatment of CosR-PNA
The PNAs used in this study were conjugated with the oligopeptide (KFFKFFKFFK) to improve the permeability of PNA into bacterial cells [29], and were commercially synthesized by Panagene (Daejeon, Korea). Since the start codon and Shine-Dalgarno motif are known to be the primary target for antisense inhibition [65], a 16-mer PNA (CATTTGTTCTATCCTT) was designed to reverse-complementarily bind to the leader sequence spanning the ribosomal binding site and the start codon of cosR based on the genome sequence of C. jejuni NCTC 11168 [13]. Another PNA oligomer (Cont-PNA; TATCTCTCTCTGTATT), which possesses the same number of bases as CosR-PNA but does not exhibit significant homology to cosR and other C. jejuni genes, was used as a control. C. jejuni was treated with PNA as described previously by Jeon and Zhang [30]. Briefly, an overnight culture of C. jejuni on MH agar plates was collected and resuspended in MH broth to an optical density of approximately 0.07 at 600 nm (approximately 1610 7 CFU ml 21 ). PNAs were added to broth cultures and C. jejuni was grown for 8 hrs under the culture conditions described above.

Construction of a CosR-overexpressing strain
A C. jejuni strain that overexpresses cosR was constructed by chromosomal integration of an extra copy of cosR using the previously-reported methodology of chromosomal integration of a gene into a non-coding spacer region of rRNA gene clusters [66]. Briefly, pFMB is a pUC19 derivative carrying an rRNA gene cluster and a kanamycin resistance cassette, and was constructed according to the method described by Karlyshev and Wren [66]. A DNA fragment containing cosR and its upstream flanking region was amplified with CosRF_Xba and CosRR_Xba primers ( Table 2). The PCR product was digested with XbaI and cloned into pFMB. The constructed plasmid was delivered to the bacterial cell by electroporation, and quantitative real time polymerase chain reaction (qRT-PCR) and western blotting was performed to confirm the increased expression of CosR.
Purification of rCosR rCosR was produced in E. coli BL21 (DE3) with the pET15b vector (Novagen). The cosR gene was PCR-amplified using primers CosRHis(NdeI)-F and CosRHis(BamHI)-R (Table 2). After digestion with NdeI and BamHI, the PCR product was cloned into pET15b that had been digested with the same enzymes. E. coli BL21 (DE3) carrying plasmid pET15b::cosR was grown to an optical density of approximately one at 600 nm. After induction with 0.5 mM IPTG for 3 hrs, rCosR was purified under the native conditions using Ni 2+ affinity chromatography. The purified rCosR was used to generate polyclonal antibodies against rCosR in rabbits commercially by AbFrontier (Seoul, Korea).

SDS-PAGE and western blotting
After growing for 8 hrs in antisense experiments, bacterial cells were disrupted by sonication and insoluble cell debris was removed by centrifugation at 10,0006 g for 10 min. Bacterial samples under oxidative stress were prepared by following the same culture conditions used in the oxidative sensitivity test (5 hr culture and 1 hr exposure to oxidative stress agents). After adjusting the sample amount by the Bradford assay (BioRad), the equal amount of protein was boiled for 3 min in the loading buffer (0.05 M Tris-HCl pH 8, 1.6% SDS, 25% glycerol, 5% 2mercaptoethanol, 0.003% bromophenol blue). After electrophoresis in a 12% polyacrylamide gel in Tris-Glycine buffer, gels were either stained with Coomassie Blue or blotted to PVDF membrane (Immobilon-P, Millipore). The blotted membrane was incubated in blocking solution (Tris-buffered saline with 5% skim milk and 0.05% Tween 20) for 1 hr and was probed with the primary antibody (rabbit anti-rCosR, 1:200 dilution) for 1 hr. After three washings, the membrane was incubated with the secondary antibody (goat anti-rabbit IgG-HRP, 1:5,000, Santa Cruz Biotech.). After washing, the signals were developed with WEST-ZOL plus western blot detection system (Intron Biotechnology, Korea).

Proteomic analysis by 2DGE
A 2DGE analysis was performed according to a method reported previously [67]. C. jejuni cells were grown in MH broth, which was occasionally supplemented with 1.5 mM CosR-PNA for CosR knockdown, to the mid-exponential phase for approximately 8 hrs with shaking and harvested by centrifugation at 3,0006g for 15 min. The pellet was resuspended and incubated in sample preparation solution (7 M urea, 2 M thiourea, 4% CHAPS, 0.5% IPG buffer, and 40 mM DTT) for 5 min. Following centrifugation at 32,0006 g for 20 min at 4uC to remove insoluble proteins and large macromolecular complexes, the supernatant was precipitated with 36 volume of 100% acetone for 2 hrs at 220uC. After additional centrifugation, the protein pellet was resuspended in rehydration buffer (7 M urea, 2 M thiourea, 2% CHAPS, 0.5% IPG buffer, 0.02% bromophenol blue). Protein concentration was determined by the Bradford assay (BioRad). For the firstdimension separation, 200 mg of protein from each sample was separated by isoelectric focusing using 13 cm immobilized pI gradient strips ranging from 3 to 10 (GE Healthcare). For the second dimensional separation, proteins were separated according to their molecular weight by SDS-PAGE using 12% acrylamide gels. Gels were visualized by Colloidal Coomassie Blue staining (Invitrogen). The protein expression profiles were analyzed from the multiple gel images with Progenesis SameSpots (Nonlinear Dynamics). Though the 2-fold cut-off is often used in 2DGE analysis, 1.5-fold cut-off was adopted in this study, since the protein expression changes of this work resulted from knockdown, not from complete gene inactivation by gene knockout. A 1.5-fold change was set as a threshold during image analysis to determine proteins of different expression levels. Protein spots showing different expression levels were excised from stained gels and followed by in-gel digestion according to Mann's method [68]. Gel slices were washed twice with water and acetone-water (1:1) solution. To shrink gel pieces, acetonitrile was added and washed with 0.1 M NH 4 HCO 3 . After removal of all liquid by a Speedvac concentrator, gel pieces were reduced and alkylated with 10 mM DTT and 55 mM iodoacetamide, respectively. The gel pieces were dried again and the protein was digested by trypsin at 37uC overnight in digestion buffer (50 mM NH 4 HCO 3 , 5 mM NaCl, 12.5 mg ml 21 trypsin). The volume of the supernatant containing the generated peptides was reduced by the Speedvac concentrator. Peptide mass fingerprinting was done with a Voyager-DETM STR Biospectrometry Workstation (PerSeptive Biosystems, Framingham). Proteins were identified by searching for peptide mass fingerprints with the programs MS-FIT (UCSF Mass Spectrometry Facility at http://prospector.ucsf.edu/prospector/mshome. htm) and MASCOT (Matrix Science, London, UK at http:// www.matrixscience.com) supplemented with an option for Campylobacter in the NCBI database.

qRT-PCR
Total RNA was isolated using the RNeasy Mini Kit (Qiagen) from C. jejuni cultures incubated as mentioned above in the proteomics analysis. After DNase treatment of the total RNA solution using the Turbo DNA-free TM kit (Ambion), cDNA was synthesized using the Omniscript RT kit (Qiagen) and random hexamers (Invitrogen). Quantification of cDNA was carried out using IQ SYBR Green PCR Supermix (Bio-Rad), and real-time amplification of PCR product was analyzed using the iCycler TM Optical Module (Bio-Rad). The amplification program consisted of one cycle at 95uC for 5 min, followed by 40 cycles at 95uC for 30 sec, 50uC for 30 sec, and 72uC for 30 sec. mRNA levels of sodB were normalized to the expression level of Cjr01 encoding 16s rRNA. The relative amount of cDNA was calculated according to the 2 2DDC T method [69]. The sequences of the sodB (sodB-RT-F and sodB-RT-F) and Cjr01 (Cjr01-F and Cjr01-R) primers are presented in Table 2.

Primer extension assay
Primer extension assays were performed as described previously [70]. Briefly, C. jejuni was grown in MH broth under the same conditions described in the proteomics analysis. The total RNA was purified with Trizol reagent (Invitrogen) according to the manufacturer's instructions. Purified RNA was resuspended in sterile distilled RNase-free water and the RNA concentration was determined by measuring the optical density of the solution at 260 nm and 280 nm using NanoVue (GE Healthcare). Primer extension analysis was carried out as follows. Ten pM of the sodB_R_PriEx primer (Table 2) was labeled with 32 P at the 59-end by ten units of T4 polynucleotide kinase (Invitrogen) and 80 mCi of [c-32 P] dATP for 30 min at 37uC. The labeling mixture was heated at 70uC for 10 min and purified with MicroSpin TM G-25 The mixture was incubated at 42uC for 70 min and treated with 100 U of ribonuclease T1 (Invitrogen) at 37uC for 15 min. The sample was ethanol precipitated after the addition of 1.4 ml of 5 M NaCl with 2.5 volumes of absolute ethanol and then washed with 75% ethanol. Each sample was resuspended with 6 ml of formamide dye and 4 ml of TE (pH 8.0) buffer, and then denatured at 90uC for 3 min. The samples were resolved on 6% polyacrylamide-8M urea gels, which were then dried and exposed to Imaging plate and the reverse transcription signals were analyzed using BAS 2500 (Fuji-film). The primer was also used for sequencing the sodB promoter region using SequiTherm EX-CEL TM II DNA Sequencing system (Epicentre).
EMSA EMSA was performed as described elsewhere with some modifications [71]. The DNA fragments containing the promoter region of sodB, dps, luxS, and ahpC were amplified by PCR with the primer pairs of SodB_F and SodB_R, Dps_F and Dps_R, LuxS_F and LuxS_R, AhpC_F and AhpC_R, respectively ( Table 2). The amplified DNA was purified from an agarose gel using a gel extraction kit (Qiagen) and labeled with [c-32 P] ATP (GE Healthcare). The unincorporated radio isotope was removed with a MicroSpin TM G-25 column (GE Healthcare). The 0.2 nM of 32 P-labeled DNA probe was incubated with the purified rCosR protein at different concentrations (0, 0.8, 1.6, 2.4, and 3.2 nM) at 37uC for 15 min in 10 ml of the gel-shift assay buffer (20 mM HEPES (pH7.6), 1 mM EDTA, 10 mM (NH 4 ) 2 SO 4 , 5 mM DTT, 0.2% Tween 20, 30 mM KCl, 0.1 mg poly (dI-dC)). For the competitive EMSA, unlabeled target DNA probes and internal cording regions of each gene were prepared by gel extraction after PCR amplification with specific primer pairs listed in Table 2. The reaction mixtures were resolved in a 6% polyacrylamide gel, and the radiolabeled DNA fragments were visualized using the BAS2500 system (Fuji Film).

DNaseI footprinting assay
A DNaseI footprinting assay was performed following a method described previously [72]. DNA fragments containing the sodB and ahpC promoter region were amplified by PCR using a 32 P-labeled primer (SodB_R_PriEx, AhpC_ES_R) and an unlabeled primer (SodB_F_PriEx, AhpC_FP_F) respectively (Table 2), and were purified from the agarose gel with a gel extraction kit (Qiagen). Binding of rCosR to the 32 P-labeled sodB and ahpC promoter was performed at 37uC for 10 min in 40 ml of the gel-shift assay buffer (see above) containing 10 mM of MgCl 2 . The reaction mixture was treated either with or without 0.1 U or 0.5 U DNase I (Takara), the reactions were stopped by the addition of 200 ml of ice-cold stop solution (0.4 M NaOAc, 2.5 mM EDTA) and the DNA products were purified by phenol extraction and ethanol precipitation. The digested DNA fragments were separated by electrophoresis in 6% polyacrylamide-8 M urea gels alongside sequencing ladders that were generated with the same P 32 -labeled primer used to amplify DNA fragments for DNaseI digestion.

Analysis of SodB enzyme activity
The SodB enzymatic activity was quantified as described elsewhere [73]. The protein samples were separated on a 12% polyacrylamide gel under native conditions. The gel was then placed into a staining solution prepared by mixing 20 ml gel staining solution A (0.05 M potassium phosphate (pH 7.8), 1 mM EDTA, 0.25 mM nitroblue tetrazolium chloride) and 1 ml gel staining solution B (0.5 mM riboflavin) in subdued light for 5 min. The gel was removed from the staining solution and placed under a fluorescent desk lamp until the gel turned uniformly purple.

Oxidative stress sensitivity assay
To determine the sensitivity to oxidative agents, C. jejuni was cultured to the early-mid log phase in MH broth (approximately 5 hrs), then adjusted to an optical density of 0.2 at 600 nm (approximately 3610 7 cfu/ml). The aliquots of bacterial cells were exposed to the final concentration of 1 mM of paraquat and H 2 O 2 at 42uC under microaerobic conditions for 1 hr. After exposure, cultures were serially diluted and plated on MH agar plates for viable counts.