Transcriptional Regulation of the CmeABC Multidrug Efflux Pump and the KatA Catalase by CosR in Campylobacter jejuni

Abstract

CosR is an essential response regulator in Campylobacter jejuni, a major food-borne pathogen causing enteritis worldwide. A transcriptomic analysis performed in this study discovered 93 genes whose transcriptional levels were changed >2-fold due to the repression of CosR expression by antisense peptide nucleic acid. The identified CosR-regulated genes are involved in various cellular functions, such as energy production, protein synthesis and folding, flagellum biogenesis, and lipid metabolism. Interestingly, 17 of the 93 CosR-regulated genes (18.3%) are predicted essential genes, indicating that CosR may participate in the regulation of vital biological processes in C. jejuni. In particular, CosR knockdown increased the transcriptional levels of cmeA, cmeB, and cmeC genes, whose protein product (CmeABC) is an important determinant conferring multidrug resistance in Campylobacter. Negative regulation of cmeABC by CosR was verified by quantitative real-time PCR (qRT-PCR) and P_cmeABC::lacZ assay. The results of electrophoretic mobility shift assays (EMSAs) and DNase I footprinting assays demonstrated that CosR directly binds to the cmeABC promoter. Another notable finding is that CosR regulates the transcription of katA, the sole catalase gene in C. jejuni. Further characterization with qRT-PCR, the catalase enzyme assay, EMSA, and DNase I footprinting assays successfully demonstrated that CosR affects the katA transcription and the catalase activity by direct interactions with the katA promoter. The findings in this study clearly demonstrated that CosR regulates resistance mechanisms in C. jejuni by controlling the expression of genes involved in oxidative stress defense and extrusion of toxic compounds out of the cell.

INTRODUCTION

Campylobacter jejuni is a leading bacterial cause of human gastroenteritis worldwide. The symptoms of campylobacteriosis include severe abdominal cramp, watery or bloody diarrhea, and in some rare cases Guillain-Barré syndrome as a postinfection complication (21). Although C. jejuni infection is usually self-limiting and doesn't require antibiotic treatment, antibiotics are warranted for severe cases of campylobacteriosis (30). However, the increasing prevalence of Campylobacter resistant to medically important antibiotics threatens public health. Among multiple mechanisms of antibiotic resistance, drug efflux pumps are considered the major cause of multidrug resistance in pathogenic bacteria (23). In Campylobacter, a few multidrug efflux pumps have been identified (1, 17, 27), and CmeABC is considered the most important multidrug efflux pump in this pathogenic bacterium, giving resistance to different classes of antibiotics and bile salts (27).

Aerobiosis inevitably produces reactive oxygen species (ROS), such as the superoxide anion (O₂⁻), hydrogen peroxide (H₂O₂) and hydroxyl radicals (HO^•) as by-products, and ROS can damage intracellular materials (16). Various defense and regulatory mechanisms exist to protect the cell from oxidative stress. C. jejuni contains only a single type of catalase (KatA) and superoxide dismutase (SodB) (2), whereas Escherichia coli harbors two types of catalase (KatG and KatE) and three kinds of superoxide dismutase (SodA, SodB, and SodC) (6). In enterobacteria, such as E. coli and Salmonella, oxidative stress response is regulated by the SoxRS and OxyR transcription factors, which sense superoxide and peroxide stresses, respectively (6, 16). Exposure to redox-cycling drugs activates SoxR by the oxidation of [2Fe-2S] centers, and the activated SoxR stimulates the expression of SoxS, which then upregulates oxidative stress defense genes (13, 24, 46). OxyR controls the expression of about 40 genes mostly involved in peroxide stress resistance (16). Neither the oxyR nor the soxRS homologue has been found in the genome sequence of C. jejuni (34). Notably, C. jejuni is the first Gram-negative bacterium which has been reported to harbor PerR, a non-OxyR-dependent regulatory system to control peroxide stress genes (42). The C. jejuni PerR regulates the peroxide stress genes including katA (catalase) and ahpC (alkyl hydroperoxide reductase); thus, inactivation of PerR renders C. jejuni hyper-resistant to hydrogen peroxide due to derepression of peroxide stress genes (42).

A recent proteomics study done by Garénaux et al. demonstrated that the protein level of CosR (Cj0355c) is significantly reduced by exposure to paraquat, a superoxide-generating drug, suggesting that CosR may be associated with the regulation of oxidative stress response in C. jejuni (9). CosR is an OmpR-type response regulator and essential for the viability of C. jejuni (9, 35). CosR homologues are found mostly in epsilonproteobacteria, such as Campylobacter, Helicobacter, and Wolinella (15), and the CosR homologue in Helicobacter pylori (HP1043) is also essential for bacterial viability (4). This suggests a common critical role played by CosR and its homologues in sustaining bacterial viability. In our previous study, we overcame the lethality problem resulting from a knockout mutation of the essential gene cosR by using antisense peptide nucleic acid (PNA), which was designed to specifically inhibit CosR, and demonstrated that CosR regulates the expression of oxidative stress proteins, such as SodB, Dps, and AhpC (15).

To obtain a better understanding of the CosR regulon at the transcriptomic level, in the present study, we performed an extensive transcriptomic analysis with a DNA microarray and report that CosR regulates expression of key determinants of stress resistance in C. jejuni, including the CmeABC multidrug efflux pump and the KatA catalase.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

C. jejuni NCTC 11168 was used in the present study. A CosR-overexpression strain was constructed by integrating cosR and a chloramphenicol cassette into the chromosome according to the method described in our previous study (15). C. jejuni strains were grown at 42°C on Mueller-Hinton (MH) media (Difco) in a microaerobic condition generated by Anoxomat (Mart Microbiology BV, Lichtenvoorde, Netherlands). Kanamycin (50 μg ml⁻¹) or chloramphenicol (10 μg ml⁻¹) was occasionally added to culture media where required. Broth cultures were microaerobically grown with shaking at 180 rpm.

Preparation and treatment of CosR-specific PNA (CosR-PNA).

PNAs used in this study were commercially synthesized by Panagene (Daejeon, Korea). The design and use of CosR-PNA was reported in our previous study (15). Based on the genome sequence of C. jejuni NCTC 11168 (34), briefly, 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, and conjugated with the permeabilizing oligopeptide (KFFKFFKFFK) (10). The optical density of bacterial culture was adjusted to ∼0.07 at 600 nm (∼10⁷ CFU ml⁻¹), and CosR-PNA (1.5 μM) was added to the bacterial cultures. C. jejuni was grown for 8 h in the culture conditions described above.

Transcriptomic analysis. (i) Preparation of total RNA for a DNA microarray.

C. jejuni cells were grown at 42°C in MH broth to the mid-exponential phase for ∼8 h with shaking, and the culture media were supplemented with 1.5 μM CosR-PNA for CosR-knockdown. The total RNA was extracted with TRIzol (Invitrogen) according to the manufacturer's instructions from three independent bacterial cultures. After DNase treatment with a Turbo DNA-free kit (Ambion), the total RNA was purified using acid-phenol solution. The integrity of purified total RNA was measured with a Bioanalyzer 2100 RNA Nano kit (Agilent Technologies, USA).

(ii) cDNA synthesis and microarray hybridization.

The synthesis of target cDNA probes and hybridization were performed as described previously (47). RNA samples (30 μg) were reverse transcribed using random primers and Superscript II kit (Invitrogen, USA) according to the manufacturer's instructions. After purification, labeling reactions were performed with a Bioprime labeling kit (Invitrogen) in a volume of 50 μl with a modified deoxynucleoside triphosphate pool containing 120 μM (each) dATP, dGTP, and dCTP; 60 μM dTTP; and 60 μM Cy5-dUTP (for CosR knockdown) or Cy3-dUTP (for the wild type). Labeled targets were subsequently purified by using a Qiaquick PCR cleanup kit (Qiagen). Labeled cDNAs were mixed together and hybridized to an assembled C. jejuni subsp. jejuni NCTC 11168 custom 3× 15K microarray slide with 97% coverage of the genome (MYcroarray, Ann Arbor, MI) at 50°C for 16 h using a hybridization oven (Agilent). The hybridized microarrays were washed according to the MYcroarray washing protocol. Finally, microarrays were spin dried and stored in the dark until scanned.

(iii) Image acquisition and analysis.

The hybridization images were analyzed by GenePix Pro 6.0 (Axon Instruments, CA). The average fluorescence intensity for each spot was calculated, and the local background was subtracted. All data normalization and selection of fold-changed genes were performed using GeneSpring 7.3.1 (Agilent). Intensity-dependent normalization (Lowess) was performed, where the ratio was reduced to the residual of the Lowess fit of the intensity-versus-ratio curve. The averages of normalized ratios were calculated by dividing the average of normalized signal channel intensity by the average of normalized control channel intensity. The entire set of microarray data was deposited to Gene Expression Omnibus (GEO) with the accession numbers GSE40201 and GPL15955.

Quantitative real-time PCR (qRT-PCR).

Total RNA was isolated with the RNeasy minikit (Qiagen) from C. jejuni cultures incubated as mentioned above in transcriptional analysis. After the removal of residual DNA contamination from the total RNA solution using the Turbo DNA-free 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 an iCycler optical module (Bio-Rad). The amplification program consisted of one cycle at 95°C for 5 min, followed by 40 cycles at 95°C for 30 s, 50°C for 30 s, and 72°C for 30 s. The mRNA levels of each gene were normalized to the expression level of Cjr01 encoding 16S rRNA. The relative amount of cDNA was calculated according to the 2^−ΔΔCT method (29). The sequences of primers for the expression of katA, cmeABC, and Cjr01 are presented in Table 1.

Table 1.

Primers used in this study

Primer	Sequence (5′–3′)a
Amplification and EMSA
cmeA_PF_F(BamH)	AAATGTTTTTCTAAATGGATCCAATAGCTCC
cmeA_PF_R(Xba)	AGTAAAAGCACAACATCTAGAGCTAAAATAG
katA_PF_F(Xba)	TAAAACAGCTCTAGAAGGAGTGATTTC
katA_PF_R(Xba)	TGAATTTTGGTTATCATCTAGAATGTTTCC
katA_ESIC_F	CGCAGGCGCAAAAGGACCTTTAC
katA_ESIC_R	TCAGGGAATTTATAAGCATCGCGGATG
cmeA_ES_F	ATGAAAAAAAATGCAGAAAAACTTGCTGTTC
cmeA_ES_R	TATTTTTGGTGCTTCTTCTTTGCTGC
cmeA_ESIC_F	CTGTAACAACCATGAGTGCTAAATCTGAAG
cmeA_ESIC_R	TAGACTAGCTTTTGAATTGTTAAATGTAGCAAG
qRT-PCR
katA_F	GCTTGAAAAACTTGCTCATC
katA_R	TTTGGTATAAGCACTTAAGTC
cmeABC_F	GCTTTAGGTGTTGTGCTTTT
cmeABC_R	ATGGTTGTTACAGGTTGAGG
Cjr01_F	TCGAACGATGAAGCTTTTAG
Cjr01_R	TTGTCCTCTTGTGTAGGG

^aUnderlining indicates the enzyme recognition sites.

Measurement of catalase activity.

The catalase activity was measured with a method based on the development of colorimetric precipitation as described previously (44). After separation of protein samples on an 8% polyacrylamide gel under native conditions, the gel was washed in three changes of distilled water for a total of about 45 min. The gel was then transferred and immersed in 0.003% H₂O₂ solution on a rotating turntable for 10 min at room temperature in the dark. The H₂O₂ solution was removed, and the gel was briefly rinsed with distilled water and placed in the negative staining solution containing 2% ferric chloride and 2% potassium ferric cyanide with gentle rocking over a light until the gel turned uniformly greenish blue due to the precipitation of ferric chloride and potassium ferric cyanide mediated by H₂O₂.

P_cmeABC::lacZ fusion assay.

The promoter and partial coding region of cmeABC were amplified with the primers cmeA_PF_F(BamHI) and cmeA_PF_R(XbaI) (Table 1) and cloned into pMW10 (45), which contains the promoterless lacZ gene. The plasmid was mobilized to C. jejuni by conjugation, and β-galactosidase assays were performed as described previously (45).

EMSA.

Electrophoretic mobility shift assay (EMSA) was performed as described previously (15). rCosR was produced in E. coli BL21(DE3) with pET15b::cosR and purified under the native conditions using Ni²⁺ affinity chromatography (15). The DNA fragments containing the promoter region of katA and cmeA were amplified by PCR with the primer pairs katA_PF_F(Xba)/katA_PF_R(Xba) and cmeA_ES_F/cmeA_ES_R, respectively (Table 1). The PCR products were purified from an agarose gel using a gel extraction kit (Qiagen) and labeled with [γ-³²P]ATP (GE Healthcare). After the elimination of the unincorporated radioisotope using a MicroSpinTMG-25 column (GE Healthcare), the 0.2 nM ³²P-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 37°C for 15 min in 10 μl of the gel-shift assay buffer [20 mM HEPES (pH 7.6), 1 mM EDTA, 10 mM (NH₄)₂SO₄, 5 mM dithiothreitol, 0.2% Tween 20, 30 mM KCl, 0.1 μg of poly(dI-dC)]. For the competitive EMSA, unlabeled target DNA probes and internal coding regions of each gene were prepared by gel extraction after PCR amplification with the specific primer pairs listed in Table 1. The reaction mixtures were resolved in a 6% polyacrylamide gel, and the radiolabeled DNA fragments were visualized using the BAS2500 system (Fuji Film).

DNase I footprinting assay.

DNase I footprinting assays were performed as described elsewhere (8). Briefly, DNA fragments containing the katA and cmeABC promoter region were amplified by PCR using a ³²P-labeled primer [katA_PF_R(Xba) and cmeA_PF_R(Xba)] and an unlabeled primer [katA_PF_F(Xba) and cmeA_ES_F], respectively (Table 1), and were purified from the agarose gel with a gel extraction kit (Qiagen). Binding reaction of rCosR to the katA and cmeABC promoters was performed at 37°C for 10 min in 40 μl of the gel-shift assay buffer (see above) containing 10 mM MgCl₂. The reaction mixture was treated either with or without 0.1 or 0.5 U of DNase I (TaKaRa), and the reactions were stopped by the addition of 200 μl of ice-cold stop solution (0.4 M sodium acetate, 2.5 mM EDTA). After the purification with 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 ³²P-labeled primer used to amplify DNA fragments for DNase I digestion.

RESULTS

Effects of CosR knockdown on transcriptomic profiles.

Transcriptomic profile changes under the CosR-knockdown condition were analyzed with DNA microarrays. Because of the cell death problem resulting from a knockout mutation in the essential gene cosR, microarray analysis was performed with RNA samples collected under the CosR-knockdown condition where the protein level of CosR was significantly reduced by gene knockdown with CosR-PNA as described in our previous study (15). The results of microarray analysis revealed that 480 genes were regulated either positively or negatively by CosR knockdown >1.5-fold with statistical significance (P < 0.05) (see Tables S1 and S2 in the supplemental material), and among these genes >2-fold changes were observed in 93 genes involved in various cellular processes, such as energy production, transcription, protein synthesis, motility, secondary metabolite biosynthesis, and stress defense (Tables 2 and 3). Interestingly, CosR knockdown affected the transcription of 17 genes which were reported to be essential in C. jejuni (31, 38), suggesting that CosR is associated with regulation of vital cellular processes sustaining C. jejuni's viability (Tables 2 and 3; essential genes are indicated in boldface and by underlining). qRT-PCR also confirmed that the effect of CosR knockdown on the expression of the essential genes (see Fig. S1 in the supplemental material). Also, CosR knockdown affected the expression of several genes associated with flagellar biogenesis (Table 3); CosR knockdown upregulated flgD (encoding flagellum basal body rod modification protein), flgE (flagellum hook protein), flgL (flagellum hook-associated protein), and fliK (putative flagellum hook-length control protein). Based on the genome sequence of C. jejuni (34), the flgD, flgE, and flgK genes are encoded in a single polycistronic operon (data not shown). The transcriptional level changes in flgL and fliK were confirmed by qRT-PCR (see Fig. S2A in the supplemental material). CosR knockdown increased bacterial motility compared to the wild type, whereas CosR overexpression reduced motility (see Fig. S2B in the supplemental material).

Table 2.

Genes downregulated by CosR knockdown^a

Functional category	Gene	Locus tag	Description	Fold changeb	P
Translation, ribosomal structure, and biogenesis	tgt	Cj1010	Queuine tRNA-ribosyltransferase	–2.114	0.01636
	valS	Cj0775c	Valyl-tRNA synthetase	–2.732	0.017
Transcription	greA	Cj0287c	Transcription elongation factor GreA	–2.008	0.0053
DNA replication, recombination, and repair	ruvB	Cj1362	Holliday junction DNA helicase RuvB	–2.222	0.00859
Cell division and chromosome partitioning	mrp	Cj1606c	Putative ATP/GTP-binding protein	–4.975	0.015
Posttranslational modification, protein turnover, and chaperones	groES	Cj1220	Cochaperonin GroES	–13.138	0.00456
	groEL	Cj1221	Chaperonin GroEL	–7.583	0.00168
	Cj0954c	Cj0954c	Putative DnaJ-like protein	–2.921	0.00341
	hypB	Cj0623	Hydrogenase isoenzyme formation protein	–2.433	0.015
	hypC	Cj0624	Hydrogenase isoenzyme formation protein	–2.475	0.017
	hypD	Cj0625	Hydrogenase isoenzyme formation protein	–2.618	0.004
	hypE	Cj0626	Hydrogenase isoenzyme formation protein	–2.439	0.009
	trxB	Cj0146c	Thioredoxin reductase	–2.398	0.002
Inorganic ion transport and metabolism	Cj1613c	Cj1613c	Putative pyridoxamine 5′-phosphate oxidase	–2.410	0.01488
	Cj1658	Cj1658	Putative iron permease	–2.066	0.02772
	p19	Cj1659	Periplasmic protein p19	–2.747	0.025
Energy production and conversion	Cj0037c	Cj0037c	Putative cytochrome c	–2.790	0.01208
	Cj0203	Cj0203	Putative citrate transporter	–3.525	0.00698
	Cj1153	Cj1153	Putative periplasmic cytochrome c	–2.053	0.01988
	napA	Cj0780	Nitrate reductase catalytic subunit	–2.179	0.009
	napB	Cj0783	Periplasmic nitrate reductase small subunit	–2.262	0.01099
	napG	Cj0781	Quinol dehydrogenase periplasmic component	–2.326	0.024
	napH	Cj0782	Quinol dehydrogenase membrane component	–2.188	0.009
	sucC	Cj0533	Succinyl-CoA synthetase subunit beta	–2.146	0.027
	sucD	Cj0534	Succinyl-CoA synthetase alpha chain	–2.146	0.034
Amino acid transport and metabolism	dapD	Cj1605c	Putative 2,3,4,5-tetrahydropyridine-2-carboxylate N-succinyltransferase	–3.155	0.01173
	ilvH	Cj0575	Acetolactate synthase 3 regulatory subunit	–2.070	0.007
	ilvI	Cj0574	Acetolactate synthase 3 catalytic subunit	–2.110	0.009
Nucleotide transport and metabolism	Cj0594c	Cj0594c	Putative DNA/RNA nonspecific endonuclease	–2.410	0.02564
	Cj0898	Cj0898	Putative histidine triad (HIT) family protein	–2.163	0.00825
	purH	Cj0953c	Bifunctional phosphoribosylaminoimidazolecarboxamide formyltransferase/IMP cyclohydrolase	–2.392	0.01
	purL	Cj0955c	Phosphoribosylformylglycinamidine synthase II	–2.793	0.003
Coenzyme metabolism	Cj0436	Cj0436	Putative pyridoxamine 5′-phosphate oxidase	–2.456	0.01122
Lipid metabolism	acpP	Cj0441	Acyl carrier protein	–7.463	0.00875
	fabF	Cj0442	3-Oxoacyl-(acyl carrier protein) synthase II	–6.526	0.00383
	accA	Cj0443	Acetyl-CoA carboxylase carboxyltransferase subunit alpha	–4.065	0.00599
Secondary metabolites biosynthesis, transport, and catabolism	fabG	Cj0435	3-Ketoacyl-(acyl-carrier-protein) reductase	–2.567	0.01624
General function prediction only	amaA	Cj1363	Acid membrane antigen A	–2.156	0.00711
	Cj0773c	Cj0773c	Putative binding-protein dependent transport system permease	–3.072	0.00907
	Cj0774c	Cj0774c	ABC transporter ATP-binding protein	–2.990	0.00898
Function unknown	Cj0573	Cj0573	Putative GatB/Yqey family protein	–2.134	0.01623
	Cj0593c	Cj0593c	Putative integral membrane protein	–2.079	0.00487
	Cj0776c	Cj0776c	Putative periplasmic protein	–2.632	0.015
	Cj0880c	Cj0880c	Hypothetical protein	–2.070	0.02268
	Cj1725	Cj1725	Putative periplasmic protein	–2.252	0.01401
	ctsR	Cj1475c	Hypothetical protein	–2.008	0.01739
	glpT	Cj0292c	Pseudogene	–2.012	0.00746

^aUnderlined and boldfaced genes are essential for the viability of C. jejuni, based on reports by Stahl and Stintzi (38) and Metris et al. (31), respectively. CoA, coenzyme A.

^bThe minus symbol indicates downregulation by CosR knockdown.

Table 3.

Genes upregulated by CosR knockdown^a

Functional category	Gene	Locus tag	Description	Fold change	P
Translation, ribosomal structure and biogenesis	Cj0722c	Cj0722c	Putative DNA methylase	2.036	0.00677
Posttranslational modification, protein turnover, and chaperones	Cj1365c	Cj1365c	Putative secreted serine protease	2.168	0.00613
	dsbA	Cj0872	Putative protein disulphide isomerase	2.404	0.02011
	htpG	Cj0518	Heat shock protein 90	2.079	0.01
Cell envelope biogenesis, outer membrane	dgkA	Cj0257	Diacylglycerol kinase	2.186	0.00238
	glmS	Cj1366c	Glucosamine–fructose-6-phosphate aminotransferase	2.012	0.01338
	pglF	Cj1120c	UDP-GlcNAc C4,6 dehydratase	2.138	0.028
Cell motility and secretion	flgD	Cj0042	Flagellar basal body rod modification protein	2.159	0.01308
	flgE	Cj0043	Flagellar hook protein	2.156	0.01035
	flgL	Cj0887c	Flagellar hook-associated protein FlgL	2.228	0.00678
	fliK	Cj0041	Putative flagellar hook-length control protein	2.157	0.00755
	lspA	Cj0361	Lipoprotein signal peptidase	2.091	0.012
Inorganic ion transport and metabolism	Cj0519	Cj0519	Putative rhodanese-like domain protein	2.185	0.00455
	Cj0522	Cj0522	Putative Na+/Pi cotransporter protein	2.154	0.00264
	pstS	Cj0613	Putative periplasmic phosphate binding protein	2.067	0.02
Energy production and conversion	rrc	Cj0012c	Non-haem iron protein	2.242	0.002
	sdhB	Cj0438	Putative succinate dehydrogenase iron-sulfur protein	2.159	0.006
	sdhC	Cj0439	Putative succinate dehydrogenase subunit C	2.068	0.009
Amino acid transport and metabolism	aroK	Cj0387	Shikimate kinase	2.827	0.01923
	putA	Cj1503c	Putative proline dehydrogenase/delta-1-pyrroline-5-carboxylate dehydrogenase	2.028	0.008
	putP	Cj1502c	Putative sodium/proline symporter	2.442	0.009
Nucleotide transport and metabolism	pyrC	Cj0259	Dihydroorotase	2.075	0.0096
Coenzyme metabolism	dfp	Cj0822	Bifunctional phosphopantothenoylcysteine decarboxylase/phosphopantothenate synthase	2.191	0.03033
	folD	Cj0855	Bifunctional 5,10-methylene-tetrahydrofolate dehydrogenase/5,10-methylene-tetrahydrofolate cyclohydrolase	2.121	0.01026
Secondary metabolites biosynthesis, transport, and catabolism	cmeA	Cj0367c	Periplasmic fusion protein CmeA	2.485	0.01579
	cmeB	Cj0366c	Inner membrane efflux transporter CmeB	2.294	0.01059
	cmeC	Cj0365c	Outer membrane channel protein CmeC	2.127	0.01653
General function prediction only	Cj0236c	Cj0236c	Putative integral membrane protein	2.089	0.0283
	Cj0256	Cj0256	Putative sulfatase family protein	2.078	0.00544
	Cj0770c	Cj0770c	Putative NLPA family lipoprotein	2.371	0.01032
	Cj0771c	Cj0771c	Putative NLPA family lipoprotein	2.408	0.01623
	Cj0772c	Cj0772c	Putative NLPA family lipoprotein	2.358	0.00361
	engA	Cj0386	GTP-binding protein EngA	2.080	0.00462
Function unknown	Cj0040	Cj0040	Hypothetical protein	2.306	0.04341
	Cj0069	Cj0069	Hypothetical protein	2.147	0.00597
	Cj0070c	Cj0070c	Hypothetical protein	2.156	0.04309
	Cj0243c	Cj0243c	Hypothetical protein	2.362	0.0106
	Cj0249	Cj0249	Hypothetical protein	2.012	0.03032
	Cj0391c	Cj0391c	Hypothetical protein	2.061	0.01112
	Cj0428	Cj0428	Hypothetical protein	2.190	0.00859
	Cj0520	Cj0520	Hypothetical protein	2.168	0.01622
	Cj0742	Cj0742	Pseudogene	2.454	0.04945
	Cj0873c	Cj0873c	Hypothetical protein	2.070	0.01804
	Cj1242	Cj1242	Hypothetical protein	2.419	0.00765
	Cj1501	Cj1501	Hypothetical protein	2.146	0.00568
	Cj1650	Cj1650	Hypothetical protein	2.948	0.00732

^aUnderlined and boldfaced genes are essential for the viability of C. jejuni, based on reports by Stahl and Stintzi (38) and Metris et al. (31), respectively.

Negative regulation of cmeABC by CosR.

CmeABC is a key determinant in conferring resistance to a broad range of antimicrobials and toxic compounds in Campylobacter (18, 27). The microarray results showed that the transcriptional levels of all cmeA, cmeB, and cmeC genes were increased by CosR knockdown >2-fold (Table 3). qRT-PCR and P_cmeABC::lacZ fusion assay showed that the transcriptional level of cmeABC was upregulated by CosR knockdown and reduced in the CosR-overexpressing strain (Fig. 1A and B), demonstrating the negative regulation of cmeABC expression by CosR. Since the binding of bile salts to CmeR, a repressor of cmeABC, derepresses cmeABC expression due to conformational changes in CmeR (26), the P_cmeABC::lacZ fusion assay carried out in the presence of cholic acid to examine whether the regulation of cmeABC by CosR would be affected by bile salts (Fig. 1B). Cholate (1 mg/ml) significantly increased the transcriptional level of cmeABC; however, the changes in the Miller units at different CosR levels were not affected by cholic acid compared to those without cholic acid (Fig. 1B). The results suggest that the regulation of cmeABC by CosR is independent of bile salts and presumably of CmeR as well, although the effect of bile salts on CosR expression is not yet known. Interaction between CosR and the cmeABC promoter was examined by performing EMSA, and the results showed that CosR bound to the cmeABC promoter (Fig. 1C). The PCR fragment from an internal coding region did not compete with the labeled probe of the cmeABC promoter; however, unlabeled cmeABC probes effectively competed with the labeled cmeABC probes (Fig. 1C). These results showed that CosR directly interacts with the cmeABC promoter and regulates cmeABC expression as a repressor.

Fig 1.

Negative regulation of cmeABC by CosR. (A) qRT-PCR analysis of cmeABC transcription. C. jejuni strains were cultured at different CosR levels at 42°C for 8 h under microaerobic conditions to extract total RNA. The CosR-knockdown condition was generated by supplementing 1.5 μM CosR-PNA to the medium. The results of three independent experiments are expressed as means and standard deviations. (B) P_cmeABC::lacZ fusion assay in the absence or presence of cholate (CA). Cholate (1 mg/ml) was added to the culture medium, and C. jejuni strains were grown for 6 h prior to the assay. The “–” symbol indicates the cultures without cholate. (C) Binding of rCosR on the cmeABC promoter. Radiolabeled DNA probes of the cmeABC promoter were incubated with rCosR at different concentrations, and unlabeled DNA probes (U.P) and internal coding regions (I.C) of cmeABC were used as a competitor.

Regulation of katA by CosR.

Consistent with our previous report, CosR affected the expression of oxidative stress genes. CosR-knockdown upregulated rrc (2.2-fold; Table 3), dps (1.5-fold; see Table S2 in the supplemental material), and sodB (1.7-fold; see Table S2 in the supplemental material) and downregulated katA (1.7-fold; see Table S1 in the supplemental material) and p19 (2.7-fold; Table 2). KatA is the sole catalase in C. jejuni and plays an important role in Campylobacter's resistance to H₂O₂ and persistence within macrophages (7). The results of qRT-PCR showed that the transcriptional level of katA was reduced by CosR knockdown and was increased by CosR overexpression compared to that of the wild type (Fig. 2A). To determine whether the transcriptional changes in katA would affect the enzyme activity, the catalase activity was measured with a catalase assay, which is based on the formation of colorimetric precipitate, ferrichloride-ferricyanide complex, via the reaction between H₂O₂ and ferricyanide (44). Consistent with the transcriptional changes, the catalase activity was decreased by CosR knockdown and increased in the CosR-overexpression strain (Fig. 2B). Interaction between CosR and the katA promoter was determined by EMSA, and the results showed that CosR bound to the katA promoter (Fig. 2C). These results successfully demonstrated that CosR binds to the katA promoter and positively regulates katA transcription and affects the catalase activity in C. jejuni.

Fig 2.

Positive regulation of katA by CosR. (A) qRT-PCR analysis of katA transcription. The results are expressed as the means and standard deviations of three independent experiments. (B) Catalase activity at different CosR levels. In these experiments, CosR-knockdown was achieved by adding 1.5 μM CosR-PNA to culture medium. The results show the representative of three independent experiments. The symbols “–” and “+” indicate knockdown and overexpression, respectively. (C) Binding of rCosR to the katA promoter. Radiolabeled DNA probes of the katA promoter were incubated with rCosR at different concentrations, and unlabeled DNA probes (U.P) and internal coding regions (I.C) of katA were used in the competition assay.

Determination of CosR binding sites in the katA and cmeABC promoters.

The results of microarray analysis, qRT-PCR, and EMSA demonstrated that CosR regulated katA and cmeABC by binding to their promoters (Fig. 1 and 2). DNase I footprinting assays were carried out to further determine the CosR binding sites in the katA and cmeABC promoter regions (Fig. 3). The assay found a region (from −37 to −17) overlapping with the −35 element of the katA promoter was protected from DNase I treatment (Fig. 3A and C). The CosR binding site in the katA promoter doesn't overlap with the predicted PerR binding site (−104 to −86) (43). We identified two putative holo-Fur binding sites (−51 to −65 and −91 to −105) in the katA promoter based on the consensus holo-Fur binding site (TGATAAT-T-ATTATCA) (5). Although one of the holo-Fur binding sites overlapped with the putative PerR binding site, the CosR binding site does not overlap with the holo-Fur binding sites (Fig. 3C). CosR also bound to an upstream region of the CmeR binding site in the cmeABC promoter (Fig. 3B and D) (25). Nucleotide sequences of the CosR binding sites in the katA and cmeABC promoters were aligned with the previously reported CosR binding sequences in the sodB and ahpC promoters (15), and the multiple alignment revealed a consensus CosR binding sequence (ttTaAanaAAAaTTAagaTTT; lowercase and capital letters indicate less and highly conserved residues, respectively, and “n” represents any nucleotide [Fig. 4]), which is almost identical to the consensus CosR binding sequence reported in our previous study (tttaAanAaAAaTtAtgaTTt) (15).

Fig 3.

Determination of the CosR binding sites in katA and cmeABC promoters by DNase I footprinting. The CosR binding regions in the promoter regions of katA (A) and cmeABC (B) are indicated by dotted lines; the promoters are based on previous reports (11, 25). Fur binding sites are marked with open boxes based on the consensus sequence for holo-Fur binding site (5). The CmeR binding site was reported previously by Lin et al. (25). The PerR binding region (C) in the katA promoter and the CmeR binding region (D) in cmeABC promoter are marked with boldface letters and arrows. The start codon, the transcriptional start site (+1), and the −10, −16, and −35 elements are underlined. The regions protected by rCosR from DNase I cleavage are indicated by a black background.

Fig 4.

Alignment of CosR binding sequences. Nucleotide sequences of the CosR binding sites for the katA (BkatA) and cmeABC (BcmeABC) promoters were determined from Fig. 3 and compared to the CosR binding sequences in the sodB (BsodB) and ahpC (BahpC1 and BahpC2) promoters reported in our previous study (15). Highly conserved nucleotides are shaded on black background, and identical nucleotides are indicated in capital letters. “n” means any nucleotide.

DISCUSSION

CmeABC is the major multidrug efflux pump contributing to Campylobacter's resistance to a variety of toxic compounds, antimicrobials, and bile salts (27, 28). The cmeABC mutant is highly susceptible to clinically important antibiotics and cannot colonize the chicken intestines due to elevated susceptibility to bile salts (28). CmeABC is a resistance-nodulation-cell division (RND)-type efflux pump consisting of the periplasmic protein CmeA, the inner membrane protein CmeB, and the outer membrane protein CmeC encoded in the polycistronic operon cmeABC (25, 27). The results of DNA microarray analysis in the CosR-knockdown condition demonstrated a significant change in the transcriptional levels of all three genes in the cmeABC operon (Table 3). The microarray results were further supported by qRT-PCR and P_cmeABC::lacZ assays (Fig. 1A and B). CmeR has thus far been the only transcriptional regulator involved in the control of cmeABC expression (26). Although bile salts are a substrate of the CmeABC efflux pump, bile salts can also affect the CmeR regulation (26). Interaction of bile salts with the CmeR protein causes conformational changes in CmeR and results in the derepression of cmeABC expression (26). To investigate whether CosR interferes with the function of CmeR on cmeABC expression, a P_cmeABC::lacZ assay was carried out in the presence of cholic acid. The results showed that cholic acid increased the transcriptional level of cmeABC (Fig. 1B), indicating that cmeABC expression was derepressed by the interaction between CmeR and bile salts; binding of bile salts to CmeR reduces the binding affinity of CmeR to the cmeABC promoter (26). However, CosR knockdown or overexpression caused similar changes in the Miller units of the P_cmeABC::lacZ assay regardless of the presence or absence of cholic acid, indicating that CosR did not affect the derepression of cmeABC by bile salts; however, it is unknown whether bile salts would affect the function or expression of CosR. Also, the CosR binding site discovered by DNase I footprinting assay was located 17 bp upstream of the CmeR binding site in the cmeABC promoter (Fig. 3B and D), suggesting that binding of CosR and CmeR to the cmeABC promoter may not spatially interfere with each other.

Lin et al. reported that a mutation of cmeR resulted in a >6-fold increase in the level of P_cmeABC transcription and increased the MICs of ciprofloxacin (2-fold), cefotaxime (2-fold), and erythromycin (4-fold) (25). Recently, it was reported that salicylate, a nonsteroidal anti-inflammatory compound, induces cmeABC transcription by inhibiting CmeR binding to the cmeABC promoter (37). An ∼2-fold increase in the level of cmeABC transcription at 100 μg of salicylate/ml did not change MICs of antibiotics except for only a moderate change in ciprofloxacin (2-fold) but enhanced bacterial growth in the presence of antibiotics at sub-MICs (37). Similarly, neither CosR knockdown nor overexpression produced noticeable changes in the MICs of several antibiotics tested here, including ciprofloxacin, erythromycin, and tetracycline (data not shown). Based on previous reports and our findings in the present study, both CmeR and CosR function as repressors for cmeABC, and it would be interesting to investigate whether CosR and CmeR would interplay in the regulation of cmeABC under certain conditions.

It is an interesting finding that CosR is involved in the regulation of both oxidative stress and drug efflux pump. A similar example can be found in SoxRS, a well-known oxidative stress regulator in E. coli. SoxRS not only controls oxidative stress defense but also affects antibiotic resistance by upregulating the AcrAB multidrug efflux pump (22). The E. coli AcrAB is functionally coupled to the outer membrane protein TolC and significantly contributes to E. coli's resistance to a wide range of toxic compounds, including antibiotics, detergents, and dyes (32). Increased levels of AcrAB-TolC resulting from elevated soxR expression by mutations render E. coli and Salmonella resistant to antibiotics (12, 19). The removal of toxic compounds by efflux pumps may contribute to the detoxification of cells, and presumably this is why some oxidative stress regulators control the expression of oxidative stress genes and efflux pumps as well. To the best of our knowledge, CosR is the first reported transcriptional regulator that controls the expression of genes involved in oxidative stress resistance and a multidrug efflux pump in C. jejuni.

Another important finding of the present study is that CosR positively regulates katA expression at the transcriptional level and affects the catalase activity in C. jejuni (Fig. 2). As the sole catalase in Campylobacter, KatA plays an important role in oxidative stress resistance and intracellular survival in the macrophage (7). Like Gram-positive bacteria, C. jejuni uses PerR as a repressor to control peroxide resistance genes including katA and ahpC (33, 42). Iron is also a regulatory factor of katA expression; iron represses katA, ahpC, and perR as well (3, 14). Although the effect of iron on perR expression results from negative perR autoregulation (20), the iron responsiveness of katA was eliminated not by inactivation of either perR or fur but only by the double mutation of perR and fur, indicating that both PerR and Fur participate in the control of katA expression (42). In fact, it was recently shown that holo-Fur binds to the katA promoter and negatively regulates katA expression (5). In contrast to the negative regulation of katA expression by PerR and holo-Fur, CosR activated katA expression like the E. coli OxyR. The CosR binding site is separated from the predicted PerR and Fur binding sites in the katA promoter (Fig. 3C), suggesting that CosR binding to the katA promoter may not interfere with PerR and Fur binding. The regulation of katA expression involves PerR, Fur, iron, and CosR, and further studies are required to explain how these multiple regulatory factors interplay to coordinate the expression of the sole catalase gene katA. The regulatory network of CosR is summarized in Fig. 5.

Fig 5.

Schematic representation of CosR regulatory network in C. jejuni. The detoxification flow of oxidants by oxidative stress resistance enzymes (gray boxes) is drawn with dashed lines, and the transcriptional regulators (CosR, Fur, PerR, and CmeR) are boxed in ovals. The regulation by CosR is indicated with double-strand lines, and positive and negative regulations by CosR are marked with arrowheads and flat-line heads, respectively. CosR is indispensable to Campylobacter's viability, although its regulation mechanism is not known. This is indicated with a dashed double-strand line.

Although the E. coli OxyR is subject to negative autoregulation, OxyR acts mostly as an activator of the peroxide defense genes (39). The OxyR binding sites in the promoters of positively regulated genes typically overlap the −35 region and extend upstream from the promoter regions (41). This unique positioning of OxyR binding sites in the promoter region stimulates the interaction between OxyR and the α subunit of the RNA polymerase, activating gene expression (40). The region protected by CosR from DNase I treatment in the katA promoter overlaps with the −35 region of the katA promoter sequence (Fig. 3A and C). In a previous study, we showed that CosR positively regulates AhpC, and one of the CosR binding sites also overlaps the −35 region in the ahpC promoter (15). These findings suggest that CosR may interact with the RNA polymerase for activation similar to the E. coli OxyR. To elucidate the CosR regulation, in addition, further studies are required to determine whether CosR has its cognate sensor kinase or is an orphan response regulator and whether CosR is phosphorylated or not, since the phosphorylation status can affect the DNA binding affinity of a response regulator. It has been reported that HP1043, the CosR homologue in H. pylori, functions in a phosphorylation-independent manner (36); however, it remains unknown whether CosR is phosphorylated or not.

The transcriptomic analysis described here discovered 93 genes whose transcriptional levels were changed >2-fold by CosR knockdown (Tables 2 and 3). The functions of the identified genes are involved in a wide range of cellular processes, such as energy production, lipid metabolism, motility, amino acid transport, and drug efflux; however, 21.5% of the genes (20 of 93) have not been fully characterized. The DNA microarray analysis in the present study identified 53% of the genes that were reported in our previous proteomic study, such as greA, groEL, rrc, p19, and htpG (15). Consistent with our previous proteomic analysis (15), sodB, rrc, and dps were upregulated by CosR knockdown (Table 3 and see Table S2 in the supplemental material); however, ahpC was not downregulated by CosR knockdown in a DNA microarray analysis despite obvious expression changes at the protein level in our previous study (15), suggesting that posttranscriptional factors may also affect AhpC expression. Interestingly, ca. 18.3% of the CosR-regulated genes (17 of 93) are predicted essential genes in C. jejuni (31, 38). Although it is not certain whether CosR affects the expression of these essential genes directly or indirectly, CosR appears to be closely involved in the regulation of vital biological processes in C. jejuni. Given that CosR regulates genes related to resistance to oxidative stress and toxic compounds, CosR may participate in the overall stress resistance and survival of Campylobacter. Future studies will fill in the knowledge gaps regarding the environmental stimuli affecting the regulatory function of CosR and the coordination of regulatory factors of oxidative stress defense in Campylobacter.