Cj0596 is a periplasmic peptidyl prolyl cis-trans isomerase involved in Campylobacter jejuni motility, invasion, and colonization

Bacterial strains and routine culture conditions

Escherichia coli JM109 was used as the host strain for cloning experiments and E. coli BL21(DE3)pLysS [32] was used as the host strain for expression of the his-tagged Cj0596 protein. E. coli strains were cultured in Luria-Bertani (LB) broth or agar [33], supplemented with the following antibiotics as appropriate for selection of plasmids: ampicillin, 50 μg/ml; chloramphenicol, 30 μg/ml; streptomycin, 30 μg/ml.

Proteome analysis of C. jejunistrains

Proteomics experiments were performed on C. jejuni cells grown at 37°C and 42°C as described [34]. Briefly, cells were grown overnight at 37°C in Brucella broth, then diluted the following morning into two aliquots of fresh Brucella broth (OD₆₀₀ = 0.1), which were grown at 37°C and 42°C to mid-log phase (OD₆₀₀ = 0.1). Chloramphenicol (187 μg/ml) was added to stop protein synthesis [35], and the cells were harvested for proteome analysis as described [34]. Proteomics experiments were performed using Differential In-Gel Electrophoresis (DIGE) technology from GE Biosystems (Piscataway, NJ), Whole-cell protein lysates from the 37°C- and 42°C-grown C. jejuni (25 μg each) were labelled individually with Cy3 and Cy5 dyes according to the protocol supplied by the manufacturer (GE Biosystems), then mixed in equal mass and separated using two-dimensional (2D) SDS-PAGE.

For comparison of WT, mutant and revertant proteomes (see below), protein lysates (25 μg each) were labelled with Cy2 (wild-type 81–176), Cy3 (mutant 81–176cj0596), or Cy5 (revertant 81–176 cj0596⁺) dyes, mixed in equal mass, and separated by 2D SDS-PAGE.

Following separation, gels were scanned on a Typhoon fluorescent flatbed scanner (GE Biosystems), at the following wavelengths: Cy2, 488 nm excitation, 520 nm emission, Cy3, 532 nm excitation, 580 nm emission; Cy5, 633 nm excitation, 670 nm emission. Images were analyzed with Decyder Differential In-Gel Analysis (DIA) software (version 4.0, GE Biosystems) for identification of proteins with higher or lower expression in different samples. The identities of proteins of interest were determined using a matrix-assisted laser desorption ionization – time-of-flight/time-of-flight (MALDI-ToF/ToF) spectrometer (Applied Biosystems, Foster City, CA), using both tryptic fingerprint data and fragmentation-based MS/MS.

Purification of Cj0596 protein and antibody production

To confirm the proteomics-implicated temperature regulation of Cj0596, a western blot was performed on C. jejuni cells grown at 37°C or 42°C using anti-Cj0596 antibody (1:1,000) as the primary antibody and HRP conjugated goat anti-rabbit IgG (1:50,000) as the secondary antibody. A control western blot against Cj0355 (expression of which is unaffected by growth temperature (Fields and Thompson, unpublished results; [37]) was performed using anti-Cj0355 antibody (1:1,000) as the primary antibody and HRP conjugated goat anti-rabbit IgG (1:50,000) as the secondary antibody. The blots were developed using a DAB Substrate Kit (BD Biosciences). Densitometry measurements were conducted using ImageJ software [38].

Localization of the Cj0596 Protein

To determine the cellular location of Cj0596, C. jejuni cells grown at 37°C were separated into cytoplasmic, periplasmic, and inner membrane fractions [39], and outer membrane fractions as described [37]. Western blots were performed on C. jejuni cell fractions using anti-Cj0596 antibody (1:1,000) as the primary antibody, along with control blots using anti-Cj0355 (cytoplasmic control), anti-CetA (inner membrane control), and anti-MOMP (outer membrane control) polyclonal sera (all at 1:1,000) as the primary antibodies. HRP conjugated goat anti-rabbit IgG (1:50,000) was used as the secondary antibody for all blots, which were then developed using a DAB Substrate Kit (BD Biosciences).

PPIase Assay

The PPIase activity of Cj0596 was determined in a coupled assay, which measures the ability of Cj0596 to convert the cis isomer of the oligopeptide substrate N-Suc-Ala-Ala-Pro-Phe-p-nitroanilide into the trans form which is cleavable by α-chymotrypsin [40–42]. Chymotrypsin (0.63 μg/ml) and varying concentrations of Cj0596 were combined in 50 mM Tris-HCl pH 7.8 and incubated at 4°C. The substrate (93.8 μg/ml) was added and the reaction was monitored at 10°C by the increase in absorbance at 390 nm (corresponding to the release of p-nitroanilide). The k_obs value for each PPIase concentration was found by plotting Ln [A₃₉₀(∞)-A₃₉₀(t)] vs. time (sec) and determining the slope. The catalytic efficiency (k_cat/K_m) of the PPIase activity was obtained by plotting k_obs vs. [PPIase] and determining the slope.

Analysis of cj0596 in additional Campylobacterstrains

The DNA sequences of the cj0596 locus (gene designation from strain NCTC 11168; the strain 81–176 designation is CJJ81176_0624) were determined from the Campylobacter strains listed in Table 1[22, 30, 43–45]. Primers cj0596-F1 and cj0596-R1 (Table 2) were designed based on the published NCTC 11168 genome sequence and used in PCR reactions with template genomic DNA prepared using a MasterPure DNA purification kit (Epicentre, Madison, WI) to amplify the region encompassing cj0596. Thermocycler parameters were 35 cycles of: 94°C for 30 seconds, 51°C for 30 seconds, and 72°C for 4 minutes. PCR products were purified using QIAquick PCR purification kits (Qiagen, Valencia, CA) and were then sequenced directly (both strands), using an ABI 3730xl sequencer (Applied Biosystems). VectorNTI (version 7, Invitrogen, Carlsbad, CA) was used to analyze DNA sequences. The protein sequences were analyzed for motifs using the ExPASy Prosite server http://au.expasy.org/prosite/[46, 47], and potential signal peptides were evaluated using SignalP 3.0 http://www.cbs.dtu.dk/services/SignalP/[48].

Isolation of a C. jejuni cj0596mutant

A cj0596 mutant of C. jejuni strain 81–176 was constructed using a streptomycin counterselection system similar to the heterologous H. pylori-C. jejuni method described by Dailidiene et al. [49] in which an rpsL gene from C. jejuni was used for counterselection in H. pylori, to decrease background gene conversion events. In the heterologous rpsL _HP system reported here, cj0596 was exactly replaced by the rpsL _HP (Str^S) gene from H. pylori strain 84–183 (Table 1; [50]) linked to a chloramphenicol acetyltransferase (cat) cassette (Cm^R). This strategy allows for selection of a mutant (Str^S/Cm^R) based on chloramphenicol resistance and then allows for selection of a revertant strain (Str^R/Cm^S) based on streptomycin resistance. First, a PCR-amplified cj0596 gene was amplified using primers cj0596-F1 and cj0596-R1 designed based on the published C. jejuni NCTC 11168 genome sequence (Table 2) and the resulting product was cloned into pCR II-TOPO creating pKR001 (Table 3). The plasmid pKR001 was subjected to inverse PCR (primers cj0596-inv1 and cj0596-inv2) to remove the cj0596 gene and add restriction sites. The inverse PCR product was self-ligated to form plasmid pKR002. The cat cassette was amplified from pRY111 [51] and Age I, Mfe I, and Nhe I sites were added using primers cat-F1 and cat-R1. The resulting PCR product was cloned into pCR II-TOPO to create plasmid pKR018. The rpsL _HP gene was amplified and Age I and Mfe I sites were added using primers rpsL _HP -F1 and rpsL_HP-R1 and the PCR product was cloned into pCR II-TOPO creating plasmid pKR019. Age I and Nhe I were used to excise the cat cassette from pKR018 and linearize pKR002. The cat cassette was ligated into pKR002 creating plasmid pKR020. Using Age I and Mfe I, the rpsL _HP gene was removed from pKR019 and ligated with linearized pKR020, creating plasmid pKR021, with the rpsL _HP /cat construct oriented in the same direction as the cj0596 and cj0597 genes. Care was taken to ensure that this replacement would not produce polar effects by preserving the cj0597 ribosome binding site and by leaving only 23 bp between the stop codon of cat and the start codon of cj0597. The mutagenized cj0596 allele was introduced into a spontaneous Strep^R derivative of C. jejuni 81–176 by electroporation. Several Cm^R/Strep^S transformants were verified as cj0596 mutants by PCR with primers cj0596-F1 and cj0596-R1 (Table 2) and DNA sequencing (data not shown), a representative of which was designated 81–176cj0596 (Table 1) and used for further analysis.

Reversion of the cj0596mutation

A revertant of C. jejuni 81–176cj0596 was isolated by replacing the mutated cj0596 allele in 81–176cj0596 with a wild-type cj0596 gene using streptomycin counterselection. C. jejuni strain 81–176cj0596⁺ was created by using electroporation to introduce pKR001 into 81–176cj0596 cells, selecting for colonies on plates containing streptomycin (30 μg/ml). Putative revertants were identified by screening Str^R colonies for sensitivity to chloramphenicol (30 μg/ml) to ensure loss of the rpsL _HP /cat cassette. Chromosomal DNA was isolated from these transformants and proper replacement of the rpsL _HP /cat cassette with wild-type cj0596 was confirmed by PCR using primers cj0596-F1 and cj0596-R1 (Table 2) and by DNA sequencing of the region.

Quantitative real-time reverse transcription PCR

cDNA was prepared from RNA samples of C. jejuni grown 81–176 and 81–176cj0596 using a GeneAmp RNA PCR kit (Applied Biosystems). An ICycler IQ real-time PCR detection system (Bio-Rad Laboratories, Hercules, CA) was used to run qRT-PCR with IQ Sybr Green Super Mix, and primers cj0597RT-F and cj0597RT-R (Table 2). Data were analyzed using Bio-Rad ICycler data analysis software. Control reactions used primers specific for 16S rDNA (16S-RT-F and 16S-RT-R, Table 2), which allows amplification of a non-regulated RNA [52]. Differences in transcript levels among samples were calculated from amplification profiles using the comparative threshold cycle (ΔΔCT) method, as previously described [53].

Growth experiments

The growth rates of C. jejuni wild-type 81–176, mutant 81–176cj0596, and revertant 81–176cj0596⁺ were assessed by growing cells overnight in MH broth, then diluting the following morning in MH Broth to OD₆₀₀ ~ 0.06 (the cj0596 mutant was additionally inoculated at OD₆₀₀ ~ 0.2 due to poor correlation between OD₆₀₀ and CFU for this strain; see Results) and shaking at 37°C under microaerobic conditions. Growth was monitored by measuring OD₆₀₀ and numbers of viable bacteria were determined by plating serial dilutions of the bacterial suspensions on MH agar and counting the resultant colonies.

Motility

The motility of C. jejuni 81–176, 81–176cj0596, and 81–176cj0596⁺ was determined as previously described [54]. Briefly, cells were grown overnight in MH broth, then diluted the following morning in MH broth to OD₆₀₀ = 0.1. A pipet tip dipped into the suspension was used to stab the center of a MH motility plate (0.4% agar). The plates were incubated at 37°C and the diameter of the motility zone was measured every 12 h.

Adherence/Invasion/Intracellular survival assay

A gentamicin protection assay [34, 55] was used to assess the ability of 81–176, 81–176cj0596, and 81–176cj0596⁺ to adhere to, invade, and survive within INT407 human intestinal epithelial cells. Briefly, bacteria were grown in biphasic [brain heart infusion (BHI)/1% yeast extract (YE)] cultures at 37°C under microaerobic conditions for ~20 h. Bacteria were harvested, resuspended in phosphate buffered saline (PBS), then added in triplicate to semi-confluent INT407 cell monolayers (~1 × 10⁵ cells/well) at a multiplicity of infection (MOI) of ~40:1 (bacteria:epithelial cells). The number of bacteria added was quantified by determination of CFU/mL. The cells were incubated for 3 h at 37°C under microaerobic conditions and were washed with Hanks' Balanced Salt Solution (HBSS), lysed with Triton X-100 and the number of adherent bacteria was quantified by viable counts. For determination of invasion, cells were incubated for 3 h with bacteria and then gentamicin was added to a final concentration of 250 μg/ml to kill any extracellular bacteria. After an additional 2 h of incubation, the cells were washed, lysed with Triton X-100 and intracellular bacteria were quantified by viable counts. The gentamicin and Triton X-100 MICs of the three strains were also determined.

For determination of intracellular survival, the cells were incubated for 3 h with bacteria, 2 h with gentamicin, and then the INT407 cells were washed and incubated for 4 h in minimal essential media containing 3% fetal bovine serum and gentamicin (10 μg/ml) as described by Candon et al. [56]. After the incubation period, cells were washed and lysed with Triton-X 100 and the number of bacteria that survived intracellularly was quantified by viable counts.

Mouse Colonization Experiments

The in vivo relevance of Cj0596 was investigated by testing the ability of 81–176, 81–176cj0596, and 81–176cj0596⁺ to colonize mice as described [34, 57, 58]. 10-week old female BALB/c-ByJ mice were given 500 μl of 5% sodium bicarbonate by oral gavage to neutralize stomach acid. The mice were then given a dose of 1 × 10⁹ CFU in 500 ml of BHI/1% YE broth by oral gavage. Because there was an observed discrepancy between OD₆₀₀ and CFU for the mutant (see Results), we first performed pilot experiments correlating OD₆₀₀ and CFU for all of the strains. After four repetitions, we found the mutant OD₆₀₀ that gave the same number of CFU as for the WT and revertant strain, and this is what we used for the mouse inocula. We also verified that each mouse received equal CFU by plating the inocula for viable counts at the time of inoculation. Colonization was determined by viable counts of bacteria in fecal pellets, which were enumerated on selective CVA media [59] starting 7 days after the challenge and continuing every 7 days thereafter until 35 days post-inoculation.

In parallel, experiments were carried out to determine the ability of cj0596 mutant bacteria to compete with wild-type bacteria in colonization. For competition experiments, wild-type and mutant bacteria were mixed in equal amounts (5 × 10⁸ CFU each) immediately prior to inoculation. Colonization was determined by enumerating bacteria on selective media with or without chloramphenicol (30 μg/ml). The number of bacteria counted on the plates containing chloramphenicol (viable mutant bacteria) was subtracted from the number of bacteria found on the plates without chloramphenicol (total of mutant and wild-type bacteria) to obtain the number of viable wild-type bacteria. Control experiments showed that the plating efficiency of the Cj0596 mutant was equivalent on media containing or lacking chloramphenicol. All vertebrate animal experiments were conducted in accordance with recommendations by the Office of Laboratory Animal Welfare, and were approved by the Medical College of Georgia Institutional Animal Care and Use Committee (MCG IACUC; protocol 04-03-379B, approved 3/18/2004).

Expression of cj0596is slightly higher at 37°C than at 42°C

In a search to identify C. jejuni genes with differential response to steady-state growth temperature (37°C vs. 42°C), several proteins were identified that were more highly expressed at 37°C than at 42°C. C. jejuni 81–176 was grown overnight at 37°C and then diluted into fresh media. The two cultures were grown in parallel at 37°C and 42°C to mid-log growth phase. Proteomics experiments were then performed on cultures of C. jejuni 81–176 grown at the two temperatures. One protein that was upregulated at 37°C had the approximate pI and molecular mass of the predicted Cj0596 protein (Figure 1). This protein was 1.8-fold more highly expressed at 37°C, a result that was consistent in five different proteomics experiments. The protein was excised from the polyacrylamide gel and subjected to MALDI-ToF/ToF mass spectrometry. This protein was identified with 100% confidence as Cj0596 (data not shown).

In an attempt to confirm the proteomics results, we performed western blots using anti-Cj0596 antibodies and C. jejuni 81–176 grown at 37°C and 42°C. While only semi-quantitative, in two separate experiments the western blots showed a more modest 1.3–1.6-fold greater expression of Cj0596 at 37°C (data not shown). Therefore, we conclude that while Cj0596 appears to be expressed at a somewhat higher level at 37°C compared to 42°C, the effect of growth temperature on Cj0596 expression is not dramatic. Because proteins homologous to Cj0596 are involved in virulence in other pathogenic bacteria, we nevertheless characterized the role of this protein in C. jejuni physiology and pathogenesis.

Similarity of cj0596 sequences among Campylobacterspecies

Because Campylobacter genomes are quite diverse [60, 61], we characterized the conservation of the cj0596 gene in other Campylobacter strains. Using PCR primers designed from the C. jejuni NCTC 11168 genomic sequence and located in the cj0595 and cj0597 genes (Figure 2), we amplified a 2 kb segment encompassing the cj0596 locus from five additional C. jejuni strains and one C. coli strain. PCR products of the expected size were obtained from each strain, and were subsequently sequenced (total of 4000 bp sequence analyzed for each strain). A search of 17 additional Campylobacter genome sequences (Table 1) was also performed and showed that a cj0596 ortholog was found in every strain. The sequences of these orthologs were also included in the sequence comparison analysis. The nucleotide sequences between pairs of C. jejuni strains or C. coli D3088 were at least 98% identical. The corresponding sequences from C. coli RM2228 and other Campylobacter species were somewhat lower (84% to 60% identical). The predicted Cj0596 protein was also highly similar in all C. jejuni strains and C. coli D3088, with an amino acid sequence identity of at least 99%. As with the nucleotide sequences, the degree of identity of proteins from C. coli RM2228 and other non-jejuni Campylobacter strains was lower, with identities ranging from 87% to 45%. Together, these results indicate that cj0596 is highly conserved in C. jejuni (16 strains), C. coli (two strains), and one strain each of C. concisus, C. curvus, C. fetus, C. hominis, C. lari, and C. upsaliensis. We focused on Cj0596 from C. jejuni strain 81–176 (the strain 81–176 designation is CJJ81176_0624) for our subsequent work.

In silicoanalysis of Cj0596 protein features

In the NCTC 11168 genome, the predicted Cj0596 protein had a predicted molecular mass of 30.5 kDa and pI of 9.9 and was annotated as a major antigenic peptide PEB4\cell binding factor 2, similar to peptidyl prolyl cis-trans isomerases found in a variety of organisms [62]. Because some peptidyl-prolyl cis-trans isomerases are located in the periplasm, the SignalP algorithm [48, 63] was used to analyze the 81–176 Cj0596 protein for the presence of an N-terminal signal sequence. A signal sequence with a probable cleavage site between amino acids 21 and 22 of the preprotein (VNA↓AT) was predicted. Cj0596 was also analyzed for the presence of protein motifs [46] and was found to contain a PpiC-type peptidyl-prolyl cis-trans isomerase motif located between amino acids 131–228. Secondary structure predictions showed no transmembrane segments in the mature protein, suggesting that Cj0596 is likely to be a periplasmic protein.

Cj0596 is located in the periplasm of C. jejuni

The amino acid sequence of Cj0596 suggested that this protein is located in the periplasm. To test this experimentally, western blots were performed on cytoplasmic, inner membrane, periplasmic, and outer membrane fractions of C. jejuni 81–176 using anti-Cj0596, anti-Cj0355, anti-CetA, and anti-MOMP antibodies (Figure 3). As expected, anti-Cj0355 antibodies reacted with a ~25 kDa protein in the cytoplasmic fraction, anti-CetA antibodies reacted with a ~50 kDa protein in the inner membrane fraction, and anti-MOMP antibodies reacted with a ~45 kDa protein in the outer membrane. Anti-Cj0596 antibodies reacted with a ~30 kDa protein present primarily in the periplasmic fraction.

Cj0596 has PPIase Activity

Cj0596 has one rotamase domain and is similar to E. coli SurA, suggesting that it is a PPIase. The PPIase activity of purified Cj0596 was determined using a coupled assay in which the cleavage of the trans isomer of N-Suc-Ala-Ala-Pro-Phe-p-nitroanilide by α-chymotrypsin results in the release of p-nitroanilide, causing a colorimetric change over time. Conversion of the cis to the trans isomers of the substrate occurs spontaneously in solution, allowing chymotrypsin cleavage (Figure 4, squares). However, addition of Cj0596 accelerates this cis-trans conversion, indicative of PPIase activity (Figure 4, diamonds). By using varying concentrations of purified Cj0596 (data not shown) and plotting calculated k_obs vs. [PPIase], the PPIase activity (k_cat/k_m) was calculated to be 22.3 mM^-1sec^-1, an activity consistent with values published for other PPIases [64–66].

Creation of a non-polar cj0596mutation

To test the role of Cj0596 in C. jejuni physiology or pathogenesis, we created a non-polar cj0596 mutant. To facilitate mutant construction, we developed a modified streptomycin counter selection system based on a similar strategy used in H. pylori [49]. The rpsl _HP /cat cassette (Methods) was used to precisely replace cj0596, maintaining the ribosome binding site of the downstream cj0597 gene. Together with the lack of a transcriptional terminator on the rpsl _HP /cat cassette, this construct is expected to yield a non-polar mutation. The resulting mutant strain was designated 81–176cj0596. To confirm that mutation of the cj0596 gene did not alter the expression of the putative co-transcribed cj0597 gene, quantitative real-time RT-PCR was performed on RNA samples harvested from 81–176 and 81–176cj0596. These studies confirmed that mRNA levels downstream of the mutation in 81–176cj0596 were equivalent to those in 81–176, and that the mutation was non-polar (data not shown). However, to ensure that any phenotypes of the cj0596 mutant were specific for cj0596, we subsequently isolated a reversion of the mutation by replacing the rpsl _HP /cat cassette with a wild-type cj0596 gene. Analogous to the counterselection system of Dailidiene et al. [49], use of the H. pylori rpsL gene decreased background gene conversion events and facilitated the recovery of cells in which the mutagenized cj0596 allele was replaced with the wild-type cj0596 gene. This process was extremely efficient, as ~70% of the recovered streptomycin-resistant colonies contained the desired reversions of the cj0596 mutation.

Mutation of cj0596 causes a moderate growth defect of C. jejuniin MH broth

The role of Cj0596 in Campylobacter growth was studied by comparing the growth rates of C. jejuni 81–176, 81–176cj0596, and 81–176cj0596⁺ in MH broth (Figure 5). When inoculated at OD₆₀₀ ~ 0.06 and growth rate was measured by OD₆₀₀ (Figure 5A), the mutant initially appeared to have a significant growth defect illustrated both by a slower increase in OD₆₀₀ and a lower maximum OD₆₀₀. However, we also found that a similar starting OD₆₀₀ resulted in a significantly lower starting CFU for the mutant, indicating that the OD₆₀₀ did not accurately reflect the number of cj0596 mutant CFU. We then grew the mutant under two starting conditions: one had the same starting OD₆₀₀ as the wild-type and revertant (OD₆₀₀ ~ 0.06) and the other had approximately the same starting CFU (OD₆₀₀ ~ 0.2) as the wild-type and the revertant. The mutant inoculated at OD₆₀₀ ~ 0.2 showed a faster initial increase in OD₆₀₀, but reached a similar maximum OD₆₀₀ as the mutant inoculated at OD₆₀₀ ~ 0.06. However, when growth rate was monitored by plating for viable counts (Figure 5B), the mutant had an initial growth rate more similar to that of wild-type, although both mutant cultures yielded final viable counts lower than wild-type. Wild-type growth characteristics were restored in the revertant strain. Together, these data suggest that mutation of cj0596 resulted in a moderate growth defect, especially later in the growth curve, and that the cj0596 mutant had apparent changes in cell characteristics such that the OD₆₀₀ had poor concordance with CFU measurements.

Deletion of cj0596 increases the motility of C. jejuni

Because motility plays an important role in invasion of host intestinal cells and is required for animal colonization, the motility of C. jejuni 81–176, 81–176cj0596, and 81–176cj0596⁺ was compared at 37°C (Figure 6). The average diameter of the zone of motility for the wild-type was 39.3 mm ± 3.7 at 48 h. The mutant was significantly more motile with a zone diameter of 66.0 mm ± 2.4 (p < 0.0001). The revertant returned to wild-type motility levels with a zone diameter of 42.5 mm ± 3.0. A similar increase in motility was seen when the assay was performed at 42°C (data not shown). Thus, Cj0596 is involved in the expression of motility.

Deletion of cj0596 increases the ability of C. jejunito invade INT407 cells, but does not affect adherence or intracellular survival

The possibility that Cj0596 plays a role in interaction with host cells was studied by comparing the adherence and invasion abilities of C. jejuni 81–176, 81–176cj0596, and 81–176cj0596⁺ in an in vitro assay using INT407 intestinal epithelial cells (Figure 7). The mean percentages of the inoculum that adhered were 8.5 (± 1.4), 7.2 (± 0.7), and 4.7 (± 1.2) for the wild-type, mutant, and revertant, respectively, demonstrating that deletion of Cj0596 does not significantly affect the ability of C. jejuni to adhere to INT407 cells (p > 0.05; Figure 7A). In contrast, mutation of cj0596 had a significant effect on the invasion ability of C. jejuni. While the percentages of the wild-type and revertant inocula invading INT407 cells were 0.041 (± 0.007) and 0.027 (± 0.005), respectively, the cj0596 mutant showed a nearly 20-fold increase in invasion (0.76 ± 0.11, p < 0.001; Figure 7B). The gentamicin and Triton X-100 sensitivities of the three strains were tested to ensure that the invasion results were not due to altered killing of a strain, and no significant difference was found for either compound.

Compared to the number of bacteria that invaded the INT407 cells, the percentages of wild-type, mutant, or revertant that were able to survive intracellularly at timepoints from 2 to 16 hours post-invasion were not significantly different (data not shown).

81–176cj0596is defective in mouse colonization

To determine whether Cj0596 plays a role in mouse colonization, we used a BALB/c model that has been used previously to assess colonization differences between wild-type and mutant bacteria [34, 57, 67]. Female BALB/c-ByJ mice were given doses of C. jejuni 81–176, 81–176cj0596, and 81–176cj0596⁺ individually (1 × 10⁹ CFU each), as well as a mixture of wild-type and cj0596 mutant (5 × 10⁸ CFU each) in a competition experiment, and colonization was measured by determining viable counts of bacteria in fecal pellets at weekly intervals (Figure 8).

On days 7 and 14, viable bacteria were found in all seven mice receiving the wild-type, mutant, or the revertant (Figure 8A). Following the peak in colonization at 14 days, viable mutant bacteria were recovered from only four mice on day 21, and only three mice on days 28 and 35. At these latter three timepoints, the wild-type and revertant were recovered from all but one mouse. The mean colonization densities of the wild-type and revertant were 1.0 × 10⁶ and 8.4 × 10⁷ CFU/g, respectively, on day 7 and remained relatively consistent throughout the experiment. The mean colonization level of the mutant was significantly lower than wild-type and revertant on days 21 (1.51 × 10⁵ CFU/g; p < 0.05) and 28 (3.42 × 10⁶CFU/g; p < 0.05). When placed in competition with the wild-type, the mutant showed an inability to compete for colonization (Figure 8B). Wild-type bacteria were recovered from five mice on day 7, four mice on day 14, and then one mouse for the remainder of the experiment. Viable mutant bacteria were recovered from no mice on day 7 (p < 0.001), two mice on day 14 (p < 0.05; the peak in colonization, as observed in mice given the mutant alone), one mouse on day 21, and then were not recovered on days 28 and 35.

Deletion of cj0596 alters C. jejuniprotein expression

Because Cj0596 is thought to be a periplasmic chaperone, its loss could result in compensatory changes in the expression of other proteins. To determine the effect that deletion of Cj0596 had on the expression of other proteins, a comparison of total cell proteins from C. jejuni 81–176, 81–176cj0596, and 81–176cj0596⁺ was performed using 2D SDS-PAGE (Figure 9). Three proteins were found to be significantly upregulated in the mutant. They were identified as HtrA (2.5-fold), Cj0998 (2.1-fold), and FlaA (2.0-fold). As expected, the CAT protein was found only in the Cj0596 mutant. Conversely, the Cj0596 protein was found in wild-type and revertant strains, but was absent in the cj0596 mutant, as expected. Three proteins were found to be significantly downregulated in the mutant. These proteins were EF-Ts (2.9-fold), superoxide dismutase (SOD) (2.6-fold), and EF-Tu (two spots; 2.0-fold, 1.9-fold). All of the proteins that showed altered abundance in the mutant returned to near wild-type levels in the revertant.