Contribution of Conserved ATP-Dependent Proteases of Campylobacter jejuni to Stress Tolerance and Virulence

Abstract

In prokaryotic cells the ATP-dependent proteases Lon and ClpP (Clp proteolytic subunit) are involved in the turnover of misfolded proteins and the degradation of regulatory proteins, and depending on the organism, these proteases contribute variably to stress tolerance. We constructed mutants in the lon and clpP genes of the food-borne human pathogen Campylobacter jejuni and found that the growth of both mutants was impaired at high temperature, a condition known to increase the level of misfolded protein. Moreover, the amounts of misfolded protein aggregates were increased when both proteases were absent, and we propose that both ClpP and Lon are involved in eliminating misfolded proteins in C. jejuni. In order to bind misfolded protein, ClpP has to associate with one of several Clp ATPases. Following inactivation of the ATPase genes clpA and clpX, only the clpX mutant displayed the same heat sensitivity as the clpP mutant, indicating that the ClpXP proteolytic complex is responsible for the degradation of heat-damaged proteins in C. jejuni. Notably, ClpP and ClpX are required for growth at 42°C, which is the temperature of the intestinal tract of poultry, one of the primary carriers of C. jejuni. Thus, ClpP and ClpX may be suitable targets of new intervention strategies aimed at reducing C. jejuni in poultry production. Further characterization of the clpP and lon mutants revealed other altered phenotypes, such as reduced motility, less autoagglutination, and lower levels of invasion of INT407 epithelial cells, suggesting that the proteases may contribute to the virulence of C. jejuni.

The intestinal gram-negative pathogen Campylobacter jejuni is the principal cause of bacterial food-borne gastroenteritis in humans worldwide (9, 20). Infection with C. jejuni usually causes watery to bloody diarrhea, with fever, nausea, and vomiting that may last for 7 to 10 days (61). Generally, infection with C. jejuni is uncomplicated, but notably, it is the most frequent antecedent to peripheral neuropathies, such as Guillain-Barré and Miller-Fisher syndromes (44, 65).

C. jejuni can be recovered from water, insects, and soil; however, the principal reservoir of C. jejuni is the alimentary tracts of mammalians and birds, and the most important source of infections in the industrialized nations is believed to be the consumption and handling of poultry meat products (20). C. jejuni primarily colonizes the mucous layer overlying the intestinal tract of poultry, and as many as 10⁹ CFU are recovered from the cecum (14). Thus, the bacterium appears to have adapted very effectively to this ecological niche. In particular, the optimal growth temperature of C. jejuni is similar to the avian body temperature (42°C), and colonization of the deep crypts of the ceca near the surfaces of the epithelial cells (5), where low levels of oxygen are present from the host cell metabolism, provides the microaerophilic environment that is optimal for growth of C. jejuni (36). Thus, in contrast to most other food-borne bacterial pathogens, C. jejuni does not multiply in refrigerated food products, as it only grows in a very narrow temperature interval (30°C to 47°C) (10, 59).

During transmission from the primary reservoirs to the consumer, C. jejuni is exposed to a changing environment, such as fluctuations in temperature and availability of oxygen, to which the bacterium has to adapt to in order to survive. One of the key consequences that environmental changes impose on living cells is the production of nonnative, misfolded proteins, which tend to unfold and form toxic protein aggregates (23). The cell responds by increasing the synthesis of a set of highly conserved chaperones and proteases, which play important roles in bacterial survival by either refolding or degrading stress-damaged proteins. Lon and the Clp proteolytic complexes are ATP-dependent proteases that degrade stress-damaged proteins in an energy-dependent manner. Lon consists of four identical subunits, each having a highly charged N-terminal domain, a centrally located ATP-binding domain, and a proteolytically active C-terminal domain (27, 28). However, while Lon contains both the proteolytic and ATPase activities within the same polypeptide chain, different subunits are responsible for the ATPase, substrate-binding activities, and proteolysis in the Clp family. The Clp complex consists of a proteolytic chamber composed of two heptameric rings of the ClpP subunit forming a barrel-like structure that has weak peptidase activity itself but, upon association with Clp ATPase subunits, increases its proteolytic activity on specific substrates severalfold (56, 60). Clp ATPases assemble in ring-shaped structures composed of six subunits positioned in either one or both ends of the proteolytic component (32, 63). They determine substrate specificity, unfold, and translocate specific substrates into the proteolytic chamber of ClpP, but independently of ClpP, the Clp ATPases have remodeling activities characteristic of molecular chaperones themselves (50, 63, 64). In the gram-negative model bacterium Escherichia coli, two Clp ATPases can associate with ClpP, namely, ClpA and ClpX, while in the gram-positive model bacterium Bacillus subtilis, either ClpX, ClpC, or ClpE can associate with ClpP (22, 27).

Even though ATP-dependent proteases are structurally and functionally conserved, their biological roles vary substantially among bacteria. In the gram-negative bacterium E. coli, a mutant lacking clpP and lon, as well as a third protease gene (hslU), grows perfectly well at both 30 and 42°C (58). In contrast, ClpP is required for growth at high temperature (54°C) in the gram-positive B. subtilis (43), while an insertional mutation in the lonA gene does not lead to any obvious phenotypical change (47). ClpP is also required for growth at high temperature in other members of the Firmicutes, such as Lactococcus lactis, Staphylococcus aureus, and Listeria monocytogenes, and in fact, none of these organisms carries a Lon homologue (17, 18, 21, 37). In organisms carrying both a Lon and a Clp protease, their individual contributions to the degradation of stress-damaged proteins have been examined only in E. coli and B. subtilis, and the results show clear differences. In E. coli, the majority of misfolded protein is degraded by Lon (24, 39, 49); in contrast, ClpCP performs a similar function in B. subtilis (37). Interestingly, a Salmonella enterica serovar Typhimurium mutant lacking ClpP is slightly heat sensitive and has a reduced ability to degrade nonnative proteins (57), suggesting that the role of Lon in the degradation of nonnative proteins may not be conserved among the Enterobacteriaceae.

C. jejuni belongs to the Epsilonproteobacteria and is only distantly related to members of the clade Gammaproteobacteria, including E. coli and S. enterica serovar Typhimurium. Examination of the C. jejuni genome reveals that it resembles E. coli in carrying a homologue of Lon, as well as a ClpA and a ClpX homologue of Clp ATPase, whereas the regulators controlling the heat shock response are similar to those normally encountered in gram-positive bacteria (1, 52). These differences suggest that knowledge obtained from previous studies of protein damage response may not be applicable to C. jejuni. However, as our ability to limit the transmission of C. jejuni through the food production chain is highly dependent on a detailed understanding of the stress tolerance of the organism, we investigated the roles of the ClpP and Lon proteases in degrading heat-damaged and abnormally folded proteins in C. jejuni, as well as their impacts on the physiology of this food-borne organism. In addition, we examined the impacts of ClpP and Lon proteases on the virulence of C. jejuni, since studies of other pathogenic bacteria, such as S. aureus, S. enterica serovar Typhimurium, and L. monocytogenes, have revealed that the Clp proteolytic complex and Lon also play central roles in controlling the levels of regulatory proteins important for virulence (15, 18, 21, 42, 53, 55).

MATERIALS AND METHODS

Bacterial strains and growth conditions.

C. jejuni NCTC11168 was supplied by the National Collection of Type Cultures. C. jejuni strains were routinely grown on blood agar base II (Oxoid) supplemented with 5% calf blood (hereafter called base II) or in brain heart infusion broth (BHI) (Difco) at 37°C in a microaerobic environment provided by CampyGen (Oxoid). E. coli DH5α [λ⁻ φ80dlacZΔM15 Δ(lacZYA-argF)U169 recA1 endA1 hsdR17(r_K⁻ m_K⁻) supE44 thi-1 gyrA relA1] was grown in Luria-Bertani broth or on Luria-Bertani agar (Difco). When appropriate, media were supplemented with ampicillin (100 μg/ml), chloramphenicol (20 μg/ml), or kanamycin (50 μg/ml).

Transformation.

C. jejuni NCTC11168 was transformed by electroporation essentially as described by Wassenaar et al. (62). To produce competent cells, C. jejuni NCTC11168 was harvested from overnight-incubated agar plates with 2 ml ice-cold wash buffer (272 mM sucrose, 15% glycerol) and subjected to four repeated centrifugations (4°C; 10,000 rpm; 10 min) and wash steps. After the final centrifugation, the cells were resuspended in 1/10 volume wash buffer, resulting in a concentration of approximately 10⁹ CFU/ml. Cell volumes of 50 μl were electroporated (1.80 kV; 200 Ω; 25 F) with 1 to 5 μl plasmid DNA. Immediately after electroporation, 1 ml of recovery broth (10% glycerol, two-thirds Mueller-Hinton broth, one-third brucella broth) was added and the cells were plated on nonselective plates. After incubation overnight at 37°C in a microaerophilic environment, cells were harvested with recovery broth and plated on selective plates. E. coli DH5α was transformed by standard methods.

DNA manipulations.

Extraction of chromosomal DNA from C. jejuni was performed using Fast Prep DNA (Bio101), with the modification that cells were incubated in phosphate buffer, pH 7.0, containing lysozyme (10 mg/ml) and 20% sucrose for 80 min at 37°C prior to the purification. Recombinant plasmid DNA from E. coli was isolated by using Qiagen columns as recommended by the supplier (Qiagen, Hilden, Germany). Biolabs supplied restriction endonuclease enzymes, DNA polymerase Klenow fragment, T4 DNA polymerase, T4 DNA ligase, and buffer systems. All enzymes were used as recommended by the supplier. PCRs were performed using Bioline Taq DNA polymerase and buffer supplied by DNA Technology.

Construction of C. jejuni mutants.

The plasmids used are listed in Table 1. C. jejuni NCTC11168 chromosomal DNA was used as a template for amplification. For construction of the clpP mutant, an 802-bp DNA fragment containing the 5′ end of the clpP gene and upstream sequences (primers CP1 and CP2Ps) and an 869-bp fragment carrying the 3′ end of the clpP gene and downstream sequences (primers CP3Ps and CP4) (Table 2) were amplified. In a second round of SOEing PCR, the clpP fragments were joined using the splicing by overlap extension PCR method (30), creating a PCR fragment containing an in-frame deletion of 459 bp in clpP and introducing a PstI site between the upstream and downstream fragments. The ΔclpP PCR fragment was cloned in the TOPO TA cloning vector pCR2.1 (Invitrogen), resulting in plasmid pLB220. Subsequently, the ΔclpP PCR fragment of pLB220 was cloned into the EcoRI site of pGEM-7Zf(+) (Promega). Finally, the cat gene obtained from pRY109 (67) was cloned into the PstI site of pLB222, resulting in pLB226, in which the cat gene was transcribed in the same direction as the clpP gene.

TABLE 1.

Plasmids used in this study

Plasmid	Relevant genotype or description	Reference or source
pCR2.1	TA cloning vector	Invitrogen
pGEM-7Zf(+)	E. coli cloning vector	Promega
pBluescript KS(+)	E. coli cloning vector	Stratagene
pRY109	cat	67
pBCα3	aphA-3	7
pLonAB	pCR2.1::PCR lon upstream	This study
pLonCD	pCR2.1::PCR lon downstream	This study
pLonABcat	pCR2.1::PCR lon upstream::cat	This study
pMTA3	pBluescript KS(+)::aphA-3	This study
pMTA7	pMTA3::Δlon::cat	This study
pMTA9	pCR2.1::SOEing PCR ΔclpA	This study
pMTA91	pGEM7::ΔclpP::aphA-3	This study
pLB218	pCR2.1::SOEing PCR ΔclpX	This study
pLB220	pCR2.1::SOEing PCR ΔclpP	This study
pLB222	pGEM7::ΔclpP	This study
pLB223	pGEM7::ΔclpX	This study
pLB226	pGEM7::ΔclpP::cat	This study
pLB228	pGEM7::ΔclpX::cat	This study

TABLE 2.

Primers used in this study

Primer	Sequence
CP1	5′-CCAGGTTTTGAAGATGGTA-3′
CP2Ps	5′-CTGCAGGCGTGAATAAATGTCATAAC-3′a
CP3Ps	5′-ACATTTATTCACGCCTGCAGGCACAAGAAGCTAAAGAATATG-3′a,b
CP4	5′-GGGATGCCTGAAGGACTAAG-3′
CP-up	5′-CCTGAAATCAAACTTGAGGG-3′
CP4B	5′-AAAGCGATAGCCAAATC-3′
Lon-A-F	5′-CAATCTTAGTGCTAGTGAGTGG-3′
Lon-B-R	5′-CAAATCACAAATCCTAGACGCATC-3′
Lon-C-F	5′-GGGTGATGTAATGAAAGAATCAGCT-3′
Lon-D-R	5′-CGCAGAAGTTGCAAACCAATAC-3′
Lon-up	5′-TAGGGCATTTAACAGCTAGGG-3′
Lon-down	5′-CAAAGCGGTAGTGCTATTGG-3′
ClpA-A-F	5′-CCAACAGAACTCATGCAAGAAG-3′
ClpA-B-R	5′-GACGTCAAACAGTACATGCTCACAGG-3′Cc
ClpA-C-F	5′-GAGCATGTACTGTTTGACGTCCGCATAATTTCTGAGGAAATAGG-3′c,d
ClpA-D-R	5′-GCTTACCATGCTTTCACCAAAG-3′
ClpA-upstream	5′-CAATGGCAGCTCTAATGATCG-3′
ClpA-downstream	5′-GCTAAAGCGATGTCATAAGCG-3′
CX1	5′-CATCCAAGTGCGGTGATT-3′
CX2Ps	5′-CTGCAGAGACAGGAATTCCACGC-3′a
CX3Ps	5′-GGAATTCCTGTCTCTGCAGACCAAAGAAGTTGTTAAAGA-3′a,e
CX4	5′-CCACTTACTTGATCGCGTCC-3′
CX-up	5′-CACAGGCATAGATGGAGC-3′
CX-down	5′-CGCTTCTCTTACATCTTCGC-3′
Cat-up	5′-GTCCTGAACTCTTCATGTCG-3′
Cat-down	5′-CGTGGACAAGCTTTGAAGG-3′
Km-F2	5′-CAGCTTATCATCGATAAACCCAGC-3′
Km-R	5′-GCGATGAAGTGCGTAAGAACATAG-3′

^aUnderlined sequence, PstI restriction enzyme site.

^bBoldface sequence, sequence complementary to parts of primer CP2Ps.

^cUnderlined sequence, AatII restriction enzyme site.

^dBoldface sequence, sequence complementary to parts of primer ClpA-B-R.

^eBoldface sequence, sequence complementary to parts of primer CX2Ps.

PCR fragments containing the 5′ end of the lon gene and upstream sequences (1,040 bp; primers Lon-A-F and Lon-B-R) or the 3′ end of the lon gene and downstream sequences (1,151 bp; primers Lon-C-F and Lon-D-R) were cloned in the Topo TA cloning vector pCR2.1 (Invitrogen), resulting in plasmids pLonAB and pLonCD. After the orientations of the fragments were determined, the cat gene obtained from pRY109 (67) was cloned into the BamHI site of pLonAB, resulting in pLonABcat. The DNA fragment containing the 5′ end of the lon gene and upstream sequences, as well as the cat gene, was excised from pLonABcat using SacI, the ends were made blunt with Klenow DNA polymerase, and it was subsequently digested with XbaI. The resulting fragment was then cloned in pLonCD digested with NotI, the ends were made blunt with Klenow DNA polymerase, and it was subsequently digested with XbaI. Finally, the Δlon::cat fragment carrying a 1,414-bp internal deletion of the lon gene was excised with SacI and XhoI and ligated to pMTA3 digested with SacI and XhoI to generate plasmid pMTA7. Plasmid pMTA3 contains a 1,477-bp aphA-3 fragment cloned into the SacI site of pBluescript KS (Stratagene) following treatment with Klenow DNA polymerase. The aphA-3 fragment was obtained by PCR using pBCα3 (7) as the template and primers Km-F2 and Km-R.

The clpA mutant of C. jejuni was constructed by amplifying a 524-bp DNA fragment containing the 5′ end of the clpA gene and an upstream sequence (primers ClpA-A-F and ClpA-B-R) and a 559-bp fragment carrying the 3′ end of the clpA gene and a downstream sequence (primers ClpA-C-F and ClpA-D-R). In a second round of PCR, the clpA fragments were joined by using the spliced overlap extension method (30) to create a PCR fragment containing an in-frame deletion of 1,908 bp in clpA and with an AatII site between the upstream and downstream fragments. The ΔclpA PCR fragment was then cloned in the Topo TA cloning vector pCR2.1 (Invitrogen), and a cat gene obtained from pRY109 (67) was cloned into the AatII site, resulting in plasmid pMTA9.

For construction of the clpX mutant, an 800-bp DNA fragment containing the 5′ end of the clpX gene and upstream sequences (primers CX1 and CX2Ps) and a 755-bp fragment carrying the 3′ end of the clpX gene and downstream sequences (primers CX3Ps and CX4) were amplified. In a second round of PCR, the clpP fragments were joined using the spliced overlap extension method (30) to create a PCR fragment containing an in-frame deletion of 1,188 bp in clpX and introducing a PstI site between the upstream and downstream fragments. The ΔclpX PCR fragment was cloned in the Topo TA cloning vector pCR2.1 (Invitrogen), resulting in plasmid pLB218. Subsequently, a 1.35-kb HindIII ΔclpX PCR fragment of pLB218 was cloned into the HindIII site of pGEM-7Zf(+) (Promega). Finally, the cat gene obtained from pRY109 (67) was cloned into the PstI site of pLB223, resulting in pLB228, in which the cat gene was transcribed in the same direction as the clpX gene.

C. jejuni NCTC11168 was transformed with either pLB226 (ΔclpP::cat), pLB228 (ΔclpX::cat), pMTA9 (ΔclpA::cat), or pMTA7 (Δlon::cat), and in all cases, several chloramphenicol-resistant colonies were isolated. Chromosomal DNA was isolated from four different chloramphenicol-resistant colonies from each transformation, and these DNA isolates were used as templates in PCRs to verify that the mutations were transferred by a double-crossover event to the chromosome of C. jejuni NCTC11168. Primers that annealed to sequences upstream and downstream of the region cloned in pLB226 (primers CP-up and CP-down), pLB228 (primers Cx-up and Cx-down), pMTA9 (primers ClpA-upstream and ClpA-downstream), and pMTA7 (primers Lon-up and Lon-down) were combined with primers cat-up and cat-down annealing internally in the cat gene. In each case, PCR fragments were obtained verifying that the deletions carrying the cat gene were transferred to the chromosome of C. jejuni. Double-crossover events were verified by PCR using primers CP-up and CP-down, Cx-up and Cx-down, ClpA-upstream and ClpA-downstream, or Lon-up and Lon-down, showing that the wild-type alleles of clpP, clpX, clpA, and lon, respectively, were absent from the chromosome. Furthermore, none of the chloramphenicol-resistant colonies obtained from transformation using plasmid pMTA7 were found to be kanamycin resistant, showing that double crossover had occurred. One of each of the following mutants was used: C. jejuni ΔclpP::cat (LB1277), C. jejuni ΔclpX::cat (LB1263), C. jejuni ΔclpA::cat (MTA11), and C. jejuni Δlon::cat (MTA21). To construct a clpP lon double mutant of C. jejuni, the cat gene of plasmid pLB226 (ΔclpP::cat) was exchanged with an aphA-3 gene obtained by PCR using pBCα3 (7) as a template and primers Km-F2 and Km-R. The aphA-3 fragment and pLB226 digested with PstI were blunted using T4 DNA polymerase and ligated to form pMTA92. After transformation of C. jejuni Δlon::cat with pMTA92 carrying the ΔclpP::aphA-3 fragment, a single transformant (LB1313) that was resistant to both chloramphenicol and kanamycin was isolated. Introduction of the mutations was confirmed by PCR performed as for the single mutants.

Northern blotting.

Total RNA was isolated from exponential cultures of C. jejuni NCTC11168 and LB1277 (optical density at 600 nm [OD₆₀₀] ≈ 0.4) grown at 37°C by using a NucleoSpin RNA II kit (Machery Nagel) according to the manufacturer's instructions. Total RNA was quantified by spectrophotometry (λ = 260), 5 μg of RNA from each sample was loaded on a 1% agarose gel, and the RNA was separated in 10 mM sodium phosphate buffer as described by Pelle and Murphy (46). Blotting and hybridization were performed according to the method of Arnau et al. (2). A def-specific probe was obtained by labeling a 390-bp internal fragment by using [³²P]dCTP and a Ready-to-Go DNA-labeling bead system from Amersham. The ProbeQuantTM G-50 Micro Columns system from Amersham was used to separate the marked probes from unincorporated radioactive nucleotide.

Growth at different temperatures and in the presence of puromycin.

C. jejuni strains were grown overnight on base II agar plates at 37°C under microaerophilic conditions. Bacterial cells were harvested using BHI, and the OD₆₀₀ was adjusted to 0.1. Tenfold serial dilutions were made, and 10 μl of each dilution was spotted onto three base II agar plates and incubated under microaerobic conditions for 3 days at 37, 42, or 44°C. In some experiments, 6 μg/ml puromycin was added to the base II agar, and the plates were incubated for 5 days at 37°C under microaerobic conditions.

Aggregated proteins.

C. jejuni strains were grown overnight on base II agar plates at 37°C under microaerophilic conditions. Cells were harvested using BHI and were transferred directly either to 100 ml ice-cold BHI. For determination of aggregated proteins, the protocol developed by Tomoyasu et al. (58) was used. A NuPage 4 to 12% N,N-methylenebisacrylamide-Tris gel (Invitrogen) was used to analyze samples of aggregated protein corresponding to identical cell numbers, followed by Coomassie staining using Safe Stain (Invitrogen).

Proteome analysis by 2D gel electrophoresis.

Campylobacter cells were grown in BHI at 37°C under microaerobic conditions and harvested in late exponential phase (100 ml; OD₆₀₀ ≈ 0.4) by centrifugation (3,000 × g for 10 min). The cells were washed with Tris-buffered saline, pH 7.5, prior to lysis by glass bead beating (106 μm or finer) four times for 1 min each time in a lysis buffer containing 50 mM Tris, pH 7.5, 0.3% sodium dodecyl sulfate, 0.2 M dithiothreitol, 3.3 mM MgCl₂, 16.7 μg/ml RNase, and 1.67 U/ml DNase. Following beating of the cells four times for 1 min each time, the extract was kept on ice for 20 min before it was centrifuged at 18,500 × g for 20 min, and the supernatant was retained for further analysis. Protein concentrations were determined using the 2D Quant Kit (Amersham) according to the manufacturer's instructions. Proteomic analysis of the cell extracts, including two-dimensional (2D) electrophoresis, imaging, spot picking, digestion, and matrix-assisted laser desorption ionization (MALDI)-time of flight analysis was carried out as described by Holmes et al. (29) using 100 to 125 μg protein per immobilized pH gradient strip.

Motility assay.

C. jejuni strains were grown overnight on base II agar plates at 37°C under microaerophilic conditions. Bacterial cells were harvested using BHI, and the OD₆₀₀ was adjusted to 0.1. One microliter of the cell suspension was deposited in the center of each heart infusion broth (Difco) plate containing 0.25% agar. After 48 h of microaerophilic incubation at 37°C, the ability of the strain to move in the soft agar was evaluated.

Autoagglutination assay.

The abilities of the bacterial strains to autoagglutinate were determined essentially as described by Misawa and Blaser (41). Bacterial cells were harvested from base II agar plates using MilliQ water and washed once before the OD₆₀₀ was adjusted to 1 in phosphate-buffered saline (PBS) (10 mM; pH 7.2). Cell suspensions of 4 ml were incubated at 25°C for 24 h, and the OD₆₀₀ was measured by carefully removing 1 ml from the top phase. Bacterial cells that autoagglutinated gravitated to the bottom of the tube, resulting in a diminution of the OD₆₀₀ measurements. The OD₆₀₀ measured after 24 h was calculated as a percentage of the starting OD₆₀₀, and the values represented the mean counts ± standard deviations derived from three independent assays.

Adherence and invasion assay.

Adherence and invasion assays were performed with INT407 monolayers growing in Eagle's minimal essential medium (EMEM) supplemented with 10% fetal calf serum at 37°C in a humidified atmosphere containing 5% CO₂. Approximately 1 × 10⁸ bacterial cells in EMEM were added to a monolayer consisting of 5 × 10⁵ epithelial cells (multiplicity of infection, 200) and incubated for 2 h. For determination of adherence, the monolayers were washed three times with 10 mM of PBS (pH 7.2), and epithelial cells were lysed by adding 0.1% Triton X-100. The adhered bacteria were enumerated by a plate count. For determination of invasion, the monolayers were subsequently incubated with EMEM containing gentamicin (250 μg/ml) for 2 h at 37°C in 5% CO₂ to kill all extracellular bacteria. Then, the monolayers were washed three times with PBS, epithelial cells were lysed by adding 0.1% Triton X-100, and the internalized bacteria were enumerated by a plate count. The values obtained represented the mean counts ± standard deviations derived from four wells.

RESULTS

ClpP and Lon are required for growth under conditions that increase the level of nonnative proteins in C. jejuni.

The genome of C. jejuni NCTC11168 (accession number NC_002163) encodes two proteins with high overall similarity to the ClpP and Lon proteases of other organisms. The 207 amino acids encoded by cj0192c show similarity to the ATP-dependent Clp protease proteolytic subunit (ClpP), while the 791 amino acids encoded by cj1073c show similarity to Lon proteases (also named ATP-dependent La proteases). To examine the roles of ClpP and Lon under conditions that increase the levels of nonnative proteins in C. jejuni, mutants were constructed in clpP and lon by replacing the central parts of the C. jejuni NCTC11168 clpP and lon genes with a cat gene (for details, see Materials and Methods). The clpP and lon mutants were obtained after selection for chloramphenicol-resistant bacteria at 37°C, demonstrating that ClpP and Lon are not essential for growth of C. jejuni NCTC11168 at this temperature. We isolated several independent mutants in each protease gene, and these mutants were phenotypically identical (data not shown). Furthermore, we constructed a clpP lon double mutant by introducing ΔclpP::aphA-3 encoding kanamycin resistance into the C. jejuni Δlon::cat mutant. A single transformant resistant to both chloramphenicol and kanamycin was isolated, and this mutant formed colonies of reduced size at 37°C compared to the wild-type strain and individual clpP and lon mutants, suggesting that loss of both proteases has an impact on growth even under nonstressful conditions.

In many bacteria, the Clp proteolytic complex and Lon are responsible for degradation of nonnative proteins; however, the contributions of the individual proteases in the protein quality control vary significantly among bacteria (17, 18, 37, 39, 45, 47, 57). To investigate the importance of ClpP and Lon for the growth of C. jejuni under conditions in which misfolded proteins accumulate, we compared the abilities of the wild type and protease mutants to form colonies in the presence of a tRNA homologue, puromycin. Puromycin increases the amount of misfolded proteins in the cell by competing with charged tRNAs for the ribosomal A site, thus leading to premature termination of protein synthesis (25). For this assay, we used the maximal concentration of puromycin (6 μg/ml) that wild-type C. jejuni cells can be exposed to without affecting their growth. The comparison revealed that the presence of puromycin severely reduced the colony size of the clpP lon double mutant, while no significant reduction in the ability to form colonies was observed for the individual clpP and lon single mutants compared to the C. jejuni NCTC11168 wild type (Fig. 1). Thus, the presence of both ClpP and Lon is required for optimal growth of C. jejuni under conditions in which the levels of misfolded proteins are increased.

FIG. 1.

Effects of puromycin on growth of C. jejuni clpP, lon, and clpP lon mutants on solid surfaces. Tenfold serial dilutions of C. jejuni NCTC11168 (wt), LB1277 (clpP), MTA21 (lon), and LB1313 (clpP lon) cultures at an OD₆₀₀ of 0.1 were spotted in volumes of 10 μl onto a base II agar plate containing 6 μg/ml puromycin. The plates were incubated for 5 days at 37°C under microaerobic conditions provided by CampyGen (Oxoid). As a control, the same dilutions were spotted onto base II agar plates. One representative of three experiments is presented.

Next, we compared the amounts of aggregated nonnative proteins in C. jejuni strains carrying single or double mutations of clpP and lon to that in the wild-type strain. For this purpose, the method developed by Tomoyasu and coworkers was employed, in which membrane proteins are solubilized and separated from aggregated proteins by repeated cycles of sonication and centrifugation steps in the presence of a detergent (58). Using this method, we could not detect an increased amount of aggregated protein in the clpP and lon mutants grown at 37°C compared to the wild type (data not shown). In contrast, aggregated proteins were reproducibly found to be more abundant in the clpP lon double mutant than in the wild type (Fig. 2). Thus, the levels of aggregated proteins increase when both proteases are absent, suggesting that both ClpP and Lon participate in the removal of nonnative proteins in C. jejuni.

FIG. 2.

Effects of the clpP lon mutations on the amounts of aggregated proteins. C. jejuni NCTC11168 (wt) and LB1313 (clpP lon) were grown at 37°C. Aggregated proteins were isolated according to the method of Tomoyasu et al. (58). Fractions corresponding to identical cell numbers were analyzed on a NuPage 4 to 12% N,N-methylenebisacrylamide-Tris gel (Invitrogen), followed by Coomassie staining using Safe Stain (Invitrogen).

It is well established that exposure to elevated temperatures leads to accumulation of misfolded proteins in the cell. To assess the significance of ClpP and Lon for growth of C. jejuni at elevated temperatures, we compared the abilities of the wild-type C. jejuni NCTC11168, the clpP and lon mutants, and the clpP lon double mutant to form colonies at different temperatures under microaerophilic growth conditions. At 37°C, we found no difference in growth between the clpP and lon mutants and wild-type C. jejuni, while the clpP lon mutant formed slightly smaller colonies (Fig. 3A). When the temperature was increased to 42°C, the clpP lon double mutant was unable to grow and the clpP mutant was severely impaired in forming colonies. The lon mutant grew like the wild type at 42°C, but it formed fewer colonies at 44°C than at 37°C, suggesting that a smaller fraction of the inoculum is able to grow at 44°C. Neither the clpP mutant nor the clpP lon double mutant could grow at 44°C. These results show that both ClpP and Lon are required for growth of C. jejuni at high temperatures. Interestingly, the ability to grow at 42°C, the optimal temperature for growth of C. jejuni and the temperature of the poultry gut, requires the activity of ClpP.

FIG. 3.

Effects of temperature on the growth of C. jejuni wild type (wt) and clpP, lon, and clpP lon mutants (A) and the wild type and clpP, clpX, and clpA mutants (B) on solid surfaces. Tenfold serial dilutions of C. jejuni NCTC11168 (wt), LB1277 (clpP), MTA21 (lon), LB1313 (clpP lon), MTA11 (clpA), and LB1263 (clpX) cultures at an OD₆₀₀ of 0.1 were spotted in volumes of 10 μl onto base II agar plates, which were incubated for 3 days at 37, 42, or 44°C under microaerophilic conditions provided by CampyGen (Oxoid). One representative of three experiments is presented.

ClpX, but not ClpA, is required for growth during heat stress.

The proteolytic activity of ClpP is dependent on association with Clp ATPases that mediate substrate recognition, unfolding, and subsequent delivery of substrate to the proteolytic chamber of ClpP (31, 32). The genome of C. jejuni NCTC11168 encodes two Clp ATPases designated ClpA and ClpX. To elucidate whether the increased sensitivity to heat of the clpP mutant is due to the absence of either ClpAP- or ClpXP-mediated functions, we constructed C. jejuni clpA and clpX mutants by deleting the central parts of the clpA and clpX genes (see Material and Methods). Growth at different temperatures was examined for the ATPase mutants and compared to the observed phenotypes of the clpP mutant. This comparison showed that at 42°C the clpX mutant, like the clpP mutant, grew poorly, and at 44°C, the clpX mutant was not able to form colonies (Fig. 3B). In contrast, the clpA mutant grew like the wild type at all temperatures tested. These identical phenotypes of the clpX and clpP mutants suggest that harmful proteins generated during heat stress are unfolded and delivered for degradation by ClpX to the proteolytic chamber of ClpP.

Inactivation of clpP or lon has a minor effect on the proteome of C. jejuni.

A proteomic approach was used to look for putative substrates of the Lon protease and the Clp proteolytic complex by comparing Sypro Ryby-stained 2D gels of protein extracts from exponentially grown clpP and lon mutants with that of the wild type. By analysis of the gels, we identified several protein species that were differentially expressed in the lon and clpP mutants. The proteins identified by MALDI (mass spectrometry) are listed in Tables 3 and 4. All differences listed were obtained in five or six independent replicas out of six.

TABLE 3.

Proteomic analysis of lon mutant cells

C. jejuni functional classification by Sanger Institute	cj no.	Protein identification	Induction compared to wild typea
Small-molecule metabolism	cj1508c	FdhD protein	1.39
	cj0538	Subunit C of 2-oxoglutarate:acceptor oxidoreductase	0.49
	cj1529c	Phosphoribosylformylglycinamidine cyclo-ligase methylenetetrahydrofolate	1.58
	cj0855	Dehydrogenase/methenyltetrahydro-folate cyclohydrolase	1.94
Signal transduction	cj1419c	Possible methyltransferase	3.40
	cj1491c	Putative two-component regulator	1.49
Synthesis and modification of	cj0551	Elongation factor P	1.37
macromolecules	cj0473	Putative transcription antitermination protein	1.41
Cell processes	cj0012c	Nonheme iron protein	0.78
	cj1663	Putative ABC transport system ATP-binding protein	1.65
	cj1221	GroEL 60-kDa chaperonin (cpn60)	1.84
	cj1220	GroES 10-kDa chaperonin (cpn10)	2.97
	cj1220	GroES 10-kDa chaperonin (cpn10)	2.20
	cj0759	Heat shock protein DnaK	1.71
	cj0758	GrpE	1.79
Unknown	cj1426c	Hypothetical protein	0.34
Miscellaneous	cj0414	Putative oxidoreductase subunit	1.45

^aAll changes in protein expression are considered significant, according to the t test (P < 0.05). Geometric average of five or six independent gels run together under the same conditions.

TABLE 4.

Proteomic analysis of clpP mutant cells

C. jejuni functional classification by Sanger Institute	cj no.	Protein identification	Induction compared to wild typea
Small-molecule methabolism	cj1536c	UTP-glucose-1-phosphate uridylyltransferase	0.24
	cj0835c	Aconitate hydratase	0.54
	cj0533	Succinyl-Coenzyme A synthetase beta chain	0.51
	cj1364c	Fumarate hydratase	1.40
	cj0537	Subunit B of 2-oxoglutarate:acceptor Oxidoreductase	0.67
	cj0105	ATP synthase F1 sector alpha subunit	1.78
	cj1096c	S-Adenosylmethionine synthetase	1.29
	cj0639c	Adenylate kinaseb	0.40
	cj0665c	Argininosuccinate synthase	0.58
	cj0332c	Nucleoside diphosphate kinaseb	8.69
	cj0298c	3-methyl-2-oxobutanoate hydroxymethyltransferase PanB	1.39
	cj0442	3-oxoacyl-[acyl-carrier-protein] synthase	0.78
Signal transduction	cj0355c	Two-component regulator	0.73
	cj0448c	Putative MCP-type signal transduction protein	0.70
Macromolecule metabolism	cj0477	50S ribosomal protein L7/L12	1.54
	cj0039c	GTP-binding protein TypA homolog	2.38
	cj0493	Elongation factor G	1.52
	cj1228c	Serine protease (protease DO) HtrA	2.62
	cj0129c	Outer membrane protein	1.86
	cj1131c	UDP-glucose 4-epimerase	0.73
	cj1339c	Flagellin FlaA	0.22
	cj0112	Periplasmic protein	1.31
Cell processes	cj1534c	Possible bacterioferritinb	1.32
	cj1534c	Possible bacterioferritinb	1.78
	cj0012c	Nonheme iron protein	0.36
	cj0759	Heat shock protein DnaK	1.67
	cj0193c	Trigger factorb	0.69
	cj0193c	Trigger factorb	3.67
	cj0696	Cell division protein FtsZb	3.00
	cj0696	Cell division protein FtsZb	0.28
	cj0285c	Chemotaxis protein CheVb	0.68
	cj0285c	Chemotaxis protein CheVb	2.85
	cj0779	Probable thiol peroxidase	1.55
Other	cj0930	Putative GTP-binding protein	1.76
	cj1214c	Hypothetical protein	0.80
	cj0706	Hypothetical proteinb	0.44
	cj0706	Hypothetical proteinb	6.80
	cj0069	Hypothetical protein	0.53
	cj1426c	Hypothetical protein	2.05
Miscellaneous	cj0559	Oxidoreductase	0.74
	cj0414	Putative oxidoreductase subunit	1.46

^aAll changes in protein expression are considered significant according to the t test (P < 0.05). Geometric average of five or six independent gels run together under the same conditions.

^bProtein that showed up more than once on the 2D gels, with the second spot positioned in a more acid area of the 2D gel than expected.

In the lon mutant, 14 protein species were increased and 3 protein species were reduced compared to the wild type (Table 3). Interestingly, the amounts of the major chaperones GroES, GroEL, DnaK, and GrpE were significantly increased in the lon mutant, suggesting a need for increased chaperone activity in the absence of Lon. The largest difference in protein level between the mutant and the wild type was observed for GroES (2.2- to 3.0-fold) and a putative methyltransferase (cj1419c) (3.4-fold).

Analysis of the proteome in the clpP mutant revealed a large number of protein species that were significantly induced or reduced compared to the wild type. However, a number of these proteins were identified as the same protein located at two positions on the 2D gels. The clpP mutant had a second spot of similar molecular mass but located in a more acid region of the gel. A similar shift in the positions of protein spots has been observed in cells treated with actinonin, an antibiotic that inhibits deformylase activity (4). The formylated start methionine is negatively charged, and if the formyl group is not removed, this results in a more acidic protein spot. In C. jejuni NCTC11168, the gene coding for the deformylase (def) lies immediately downstream of clpP, and Northern blot analysis revealed that expression of the downstream def gene is reduced in the clpP mutant (data not shown). Furthermore, peptide mass data from the MALDI analysis for the clpP mutant predicted a tryptic peptide with an N-terminal formyl group among two of the negatively charged protein duplicates (data not shown), supporting the notion that the shift in protein position on the 2D gel is due to a reduced expression of def. Twenty-nine proteins were not present in duplicate spots on the 2D gel, and interestingly, no major changes in protein amounts (≤2.6-fold) were observed for the clpP mutant compared to the wild-type strain (Table 4). As in the lon mutant, the amount of the highly conserved major chaperone DnaK was also increased (1.7- to 1.8-fold) in the clpP mutant compared to the wild-type level, suggesting the need for chaperone activity in the absence of ClpP, as well.

In conclusion, a number of small changes in protein amounts were identified in the clpP or lon mutant cells compared to the wild type, but interestingly, no major change in protein amounts could be identified.

ClpP and Lon in C. jejuni are important for invasion of epithelial cells.

ClpP and Lon are required for virulence in a large number of pathogenic bacteria (18, 21, 33, 53-55), and we wanted to address the question of whether ClpP and Lon are important for virulence in C. jejuni. Because motility is an important virulence factor (26), we evaluated the motility of the clpP and lon mutants using a soft agar. We found that both mutants displayed reduced motility as visualized by a decreased zone of spreading on the soft agar compared to wild-type cells (Fig. 4). However, the reduced motility of the clpP mutant could not be found in either the clpA or the clpX mutant (data not shown). This may suggest that both ClpAP- and ClpXP-mediated proteolyses are important for motility and that the two complexes may functionally substitute for one another.

FIG. 4.

Effects of clpP and lon mutations on the ability of C. jejuni to move in a soft-agar medium based on heart infusion broth containing 0.25% agar. Plates were inoculated in the center with C. jejuni NCTC11168 (wt) or the LB1277 (clpP) or MTA21 (lon) mutant and incubated for 48 h at 37°C under microaerobic conditions using CampyGen (Oxoid). One representative of five experiments is presented.

Data have demonstrated that autoagglutination abilities correlate with the ability to invade human epithelial cells in several gram-negative bacterial pathogens, including C. jejuni (6, 41, 48). We investigated whether the Clp proteolytic complex and Lon influence the ability to autoagglutinate by measuring the bacterial precipitation that occurs during autoagglutination (41). After 24 h of static incubation at 25°C, the OD₆₀₀ of the wild-type C. jejuni strain was reduced to 9.5% ± 1.2% of the initial value, while the OD₆₀₀ of the C. jejuni clpP and lon mutants were less reduced, to 15.6% ± 0.7% and 14.3% ± 0.5%, respectively. Comparison of the phenotype of the clpP mutant with those of the ATPase mutants showed that the clpX mutant, like the clpP mutant, was attenuated in the ability to autoagglutinate, while the clpA mutant autoagglutinated to the same level as the wild type, as shown in Fig. 5. Here, the ability to autoagglutinate was visible to the naked eye, as the wild-type C. jejuni and the clpA mutant formed a distinct pellet at the bottom of the tube, leaving the suspension clearer than those seen for the clpP and clpX mutant. Thus, Lon, ClpX, and ClpP are required for wild-type autoagglutination in C. jejuni.

FIG. 5.

The abilities to autoagglutinate of the clpP, clpA, and clpX mutants after incubation for 24 h at 25°C in PBS. C. jejuni NCTC111168 (wt), LB1277 (clpP), MTA11 (clpA), and LB1263 (clpX) were suspended in PBS to an OD₆₀₀ of 1 and incubated at 25°C for 24 h.

Finally, cultured INT407 epithelial cells were used to compare the adherence and invasion capacities of the lon, clpP, clpA, and clpX mutants with the wild type. No significant difference in the ability to adhere to INT407 epithelial cells was found between mutants and wild-type cells of C. jejuni (Fig. 6A). However, the levels of invasion were reduced for all mutants (Fig. 6B). On average, the invasiveness of the mutants was 10-fold lower than that of the wild-type C. jejuni NCTC11168. Based on our results, we concluded that invasion of INT407 epithelial cells is dependent on Lon-, ClpAP-, and ClpXP-mediated proteolytic activities in C. jejuni.

FIG. 6.

Effects of lon, clpP, clpP lon, clpX, and clpA mutations on the ability of C. jejuni to adhere to and enter cultured intestine cells (INT407). Cells adhered to (A) and invaded (B) are expressed as percentages of the number of cells inoculated. The bars represent the mean values, while the error bars represent the standard deviations of four wells. One representative of three experiments is presented.

DISCUSSION

The physiological roles of ATP-dependent proteases in the epsilonproteobacterium C. jejuni has largely been unexplored. The genome of C. jejuni NCTC11168 encodes homologues of the ATP-dependent Lon and Clp proteases, and to examine the roles of these proteases under conditions in which the levels of nonnative proteins increase in C. jejuni, mutants were constructed and assessed for their abilities to grow under such conditions.

We have demonstrated that ClpP and Lon are required for growth at high temperature, a condition known to increase the levels of misfolded proteins (38). In contrast, the presence of puromycin, an antibiotic that promotes formation of misfolded protein, reduced only growth of a C. jejuni clpP lon double mutant and not that of the clpP or lon mutant. Thus, our results suggest that both proteases are involved in eliminating misfolded protein in C. jejuni. In accordance with this hypothesis, the amounts of misfolded protein aggregates were increased only when both proteases were absent, and the accumulation of aggregated protein was observed already at 37°C, demonstrating that ClpP and Lon are involved in removal of aggregation-prone proteins in C. jejuni under nonstress conditions. Protein aggregation was assessed by the method developed by Tomoyasu and coworkers, who showed that aggregated proteins accumulated in E. coli mutant cells lacking ClpP and Lon, as well as the major chaperone, DnaK (58). In contrast, depletion of just ClpP and Lon in C. jejuni was sufficient to increase the level of aggregated protein, indicating that the ATP-dependent proteases are critical for eliminating cellular protein species prone to aggregation. Notably, growth of C. jejuni is more dependent on protease function than that of E. coli, as growth of the C. jejuni clpP lon double mutant was compromised at 37°C while growth of a triple protease mutant of E. coli (clpPX lon hslVU) was not affected even at 42°C (58).

In E. coli, similar substrate binding domains have been identified in the chaperone domains of Clp and Lon, and they display distinct but overlapping binding preferences (51, 66). Also, in C. jejuni, similar substrate binding domains of ClpP and Lon may be present. We recognized no increase either in the amounts of aggregated proteins or in sensitivity to puromycin in the clpP or lon single mutants compared to wild-type cells, while in the clpP lon mutant, we observed a significant increase in the amounts of aggregated proteins and sensitivity to puromycin. We speculate that this is not simply due to the cumulative loss of ClpP and Lon, and we hypothesize that the substrate specificities of the two proteases must overlap with each other to some extent. Nevertheless, in C. jejuni, ClpP and Lon do not play equal roles in allowing growth at high temperatures, suggesting that the proteases may have different binding affinities for heat-damaged proteins. ClpP-mediated degradation of heat-damaged proteins appears to be more important for growth at high temperature than the proteolytic activity of Lon. Moreover, the proteolytic activity of ClpP is required for growth at 42°C, which is the optimum temperature for growth of C. jejuni, as well as the temperature of the intestinal tracts of birds, one of the primary carriers of C. jejuni. Growth of the clpX mutant was affected equally to the clpP mutant at 42°C, and we propose that ClpP is required as part of the ClpXP proteolytic complex for degradation of heat-damaged proteins. Thus, our results suggest that ClpP and ClpX may be suitable targets of new intervention strategies aimed at reducing C. jejuni in poultry production. Interestingly, a dominant role of ClpP for heat tolerance compared to Lon has previously been reported only in gram-positive bacteria and not in gram-negative bacteria, such as Campylobacter (17, 21, 37). However, the proteolytic complex performing this function in gram-positive bacteria is ClpCP and not ClpXP. In S. enterica serovar Typhimurium, a slight increase in heat sensitivity has been observed for a clpP mutant, but whether this is due to the absence of ClpAP- or ClpXP-mediated function has not been determined (57).

A proteomic approach was used to look for putative substrates of the Lon protease and the Clp proteolytic complex by comparing 2D gels of protein extracts from exponentially grown clpP and lon mutants with that of the wild type. It should be noted that this analysis does not discriminate between increased expression and decreased degradation. However, only minor differences in protein composition were identified between the different mutants and the wild type. Thus, specific proteins do not appear to accumulate in the absence of either ClpP or Lon. Apart from GroES, only a possible methyltransferase (cj1419c) was increased more than twofold in the lon mutant, while one of the very few proteins that were increased more than twofold in the clpP mutant was the periplasmic protease HtrA. Moreover, in both the clpP and lon mutants, the level of the chaperone DnaK was increased (1.7-fold). DnaK is a highly conserved chaperone involved in refolding of misfolded proteins and in resolving protein aggregates in E. coli (3). In C. jejuni, DnaK and HtrA have previously been found to be induced upon a temperature upshift, which is believed to increase the levels of misfolded proteins (38, 52), and HtrA was suggested to be involved in degradation of misfolded protein in the periplasm (13). Thus, if the HtrA and DnaK protein levels are used as indicators of the levels of misfolded proteins in the cell, our results suggest that nonnative protein accumulates in the absence of either ClpP or Lon in C. jejuni and emphasize the important roles of both ClpP and Lon in the removal of misfolded protein.

Proteins having two species differing only in charge were excluded from further consideration in this analysis because we attributed this to a polar effect on def gene expression in the clpP mutant (see Results for an explanation). In E. coli, the def gene is suggested to be essential (40). However, in C. jejuni, the reduced level of def expression did not affect the general viability. Furthermore, both formylated and deformylated versions of proteins were present on our 2D gels, and thus, deformylase activity must be present in the clpP mutant.

The amounts of several proteins were slightly decreased in the clpP mutant, which we did not observe in the absence of Lon. Interestingly, the amounts of FlaA and UTP-glucose-1-phosphate uridylyltransferase were reduced more than threefold in the clpP mutant; the latter is known to be essential for capsule biosynthesis in Streptococcus pneumoniae (11). Both flagellar structural proteins, such as FlaA, and capsular composition may affect autoagglutination of bacteria (8, 41), and our data actually showed that the abilities of the clpP and clpX mutants to autoagglutinate were reduced. However, many factors influence the autoagglutination abilities of cells, but especially, several proteins involved in flagellar structure have been reported to affect autoagglutination of C. jejuni (26, 41).

Increasing evidence indicates that protease activity affects virulence in both gram-negative and gram-positive pathogenic bacteria (12, 19, 34, 53, 55). A correlation between invasion properties and pathogenicity has been observed for C. jejuni (35), and furthermore, invasive strains seem to have increased abilities to autoagglutinate (26). Both invasiveness and the ability to autoagglutinate were tested for the mutants and compared to those of the wild type. Our data showed that the activities of ClpP and Lon affected both the ability to autoagglutinate and the invasiveness of C. jejuni. The ability of C. jejuni to invade, but not adhere to, human INT407 epithelial cells was reduced for the single and double protease mutants, and both ClpA and ClpX seem to be required for optimal invasiveness in C. jejuni. Interestingly, no additive effect was observed for the absence of both ClpP and Lon, despite the growth defect of the clpP lon mutant. To date, there is little published information about proteins that play a role in Campylobacter invasion of epithelial cells, but our findings implicate ClpP and Lon proteases as having roles. Recent studies have shown that in other pathogenic proteobacteria the activities of proteases both decrease and increase invasion of intestinal cells. In S. enterica serovar Typhimurium, Lon decreases invasiveness by controlling the expression of a type III secretion system essential for invasion (53, 55), while in E. coli, ClpXP positively controls the expression of the locus of enterocyte effacement of enterohemorrhagic E. coli (34). Thus, the roles of the ATP-dependent proteases differ among bacteria, even for species within the same subdivision of proteobacteria.

In conclusion, we found that ATP-dependent proteases influence heat tolerance and virulence-associated phenotypes, such as autoagglutination and cell invasiveness, of C. jejuni. Notably, we observed a significant increase in the levels of aggregated proteins in the clpP lon double mutant, indicating that ClpP and Lon work in concert to protect the cell against the accumulation of misfolded proteins in the cytoplasm of C. jejuni. In addition, the growth of the clpP lon mutant was reduced even at 37°C, and thus, we predict that ATP-dependent proteolytic activity is essential for the growth of C. jejuni in the gastrointestinal tract of poultry, one of the most important reservoirs of pathogenic C. jejuni.