Expression patterns and role of the CadF protein in Campylobacter jejuni and Campylobacter coli (2024)

,

Article history

Abstract

Introduction

Campylobacter jejuni and Campylobacter coli are major causes of gastrointestinal diseases worldwide (Altekruse et al., 1999; Akitoye et al., 2002). These pathogens colonize and invade the intestinal mucosa in vitro (Hu & Kopecko, 1999; Bacon et al., 2000; Biswas et al., 2000; Monteville et al., 2003; Nadeau et al., 2003; Konkel et al., 2004; Hu et al., 2005). Campylobacter jejuni synthesizes a set of proteins called Campylobacter invasion antigens (Cia proteins) that may contribute to the invasion of epithelial cells (Konkel et al., 1999a). Campylobacter jejuni also possesses a 37 kDa adhesin, termed CadF, that binds fibronectin and aids the adherence of C. jejuni to intestinal epithelial cells (Konkel et al., 1997, 1999b, 2005). CadF is a single-copy, highly conserved chromosomal gene of Campylobacter (Konkel et al., 1999b; Parkhill et al., 2000; Fouts et al., 2005; Hofreuter et al., 2006). Using an overlapping peptide library derived from CadF, maximal fibronectin-binding activity was localized within 4 amino acids (aa) (134–137 aa) consisting of the phenylalanine–arginine–leucine–serine motif (Konkel et al., 2005). Previous work based on immunoblot analysis of clinical isolates indicated that the CadF protein is highly conserved among C. jejuni strains from the US (Konkel et al., 1997, 1999b). Therefore, a variety of assays could be developed based on the detection of the cadF virulence gene and its product. In the present study, the CadF proteins of a large number of C. jejuni and C. coli strains of human and animal origin were compared, and the role of CadF in the attachment and internalization of INT-407 epithelial cells was determined.

Materials and methods

Campylobacter wild-type strains and growth conditions

Table 1 shows the collection of isolates used in this study. Bacteria were grown (48 h) on Campylobacter blood-free selective agar base with growth supplement at 37°C under microaerophilic conditions generated by CampyGen (Oxoid, Basingstoke, UK). Species identification was based on biochemical tests (catalase, oxidase, urease activity, hippurate and indoxyl acetate hydrolysis, and sensitivity to cephalothin and nalidixic acid) and a multiplex PCR assay (Cloak & Fratamico, 2002; Oyarzabal et al., 2005).

PCR and analysis of amplified products

The cadF gene and its flanking regions were amplified by PCR using the following primers. CadF1 Fwd: 5′-TTG CTC TAA AGG ATA ACC TAT GA-3′, CadF1 Rev: 5′-TAT GGA CGC CGC AAA GCA AG-3′, CadF2 Fwd: 5′-CCA CTC TTC TAT TAT CCG CTC TAC C-3′, and CadF2 Rev: 5′-GGT GCT GAT AAC AAT GTA AAA TTT G-3′. PCR conditions were as follows: denaturation (94°C, 2 min), six cycles of touchdown PCR (94°C for 30 s, 58°C for 45 s, decreasing 0.5°C per cycle, 72°C for 2 min), followed by 30 cycles of 94°C for 30 s, 55°C for 45 s, 72°C for 2 min and a final extension step at 72°C for 10 min. Amplified products were analyzed by agarose gel electrophoresis, cloned into pGEM-T-easy vector (Promega, Madison) and sequenced. Nucleotide sequence analysis and protein sequence alignments were performed using free software (http://searchlauncher.bcm.tmc.edu/seq-util/Options/sixframe.html; http://www.ebi.ac.uk/clustalw).

Generation of cadF mutants and growth conditions

The cadF gene and its flanking regions from C. jejuni 81116 were amplified by PCR using the primers CadF3 Fwd: 5′-GAT AAA GCA TTC TAA ACA TT-3′ and CadF3 Rev: 5′-GAG CAC CCA CAC ACT GCA C-3′. The fragment was ligated into pGEM-T-easy vector and transformed into Escherichia coli JM110. An inactivated cadF of strain 81116 was obtained by insertion of the Aph-A3 kanamycin resistance cassette (1.5 kb) at the BclI site and introduced into the 81116 genome by hom*ologous recombination. The cadF mutant in C. jejuni F38011 was generated as described (Konkel et al., 1997). Disruption of the cadF gene in each strain was confirmed by PCR. The cadF mutant strains 81116ΔcadF and F38011ΔcadF were grown on Columbia agar with 5% blood and 20 µg mL⁻¹ kanamycin, and on Mueller–Hinton (MH) agar amended with 20 µg mL⁻¹ kanamycin, respectively.

Infection of INT-407 cells

Human embryonic intestinal epithelial cells (INT-407, ATCC-CCL-6) were grown in Eagle's minimum essential medium (MEM) containing l-glutamine and Earle's salts (Invitrogen), 100 U mL⁻¹ penicillin, 100 µg mL⁻¹ streptomycin and 10% fetal bovine serum (FBS, Invitrogen) in a humidified 5% CO₂ incubator. For infection assays, cells were grown (48 h) in 12-well tissue culture plates to reach ∼70% confluence. Then, the medium was replaced with MEM without antibiotics and bacteria were added at a multiplicity of infection (MOI) of 100. After 6 h of incubation, the cells were washed three times with 1 mL of medium and suspended, diluted and plated on MH agar plates to determine the total number of cell-associated bacteria (attached and intracellular), or incubated with gentamicin (250 µg mL⁻¹, 2 h) to kill all extracellular bacteria, and then disrupted with saponin (0.1%, 37°C, 15 min). Released intracellular bacteria were diluted and plated as described above. The level of total cell-associated and intracellular bacteria was determined by calculating the number of CFU. All experiments were performed in triplicate.

Generation of the polyclonal CadF antibodies

Polyclonal antiserum (α-CadF-1) was raised according to standard protocols (Biogenes, Berlin, Germany) by immunization of two rabbits with a conserved C. jejuni CadF-derived peptide (293–306 aa: QDNPRSSNDTKEGR) conjugated to Limulus polyphemus hemocyanin carrier protein. Immunoblot analysis verified that α-CadF-1 is specific for C. jejuni CadF and does not react with CadF from C. coli. The polyclonal rabbit antiserum α-CadF-2 was obtained by immunization with gel-purified CadF and reacts with both CadF from C. coli and C. jejuni (Konkel et al., 1997).

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblot analysis

Whole bacterial cells harvested from agar plates or infected INT-407 cells were lysed in SDS-PAGE buffer (2% SDS, 0.1 M dithiothreitol), boiled, separated on 12% polyacrylamide gels and either stained with Coomassie–Brilliant Blue or blotted onto polyvinylidene difluoride (PVDF) membranes (Immobilon-P; Millipore). The blots were incubated with the polyclonal antibodies or with a monoclonal α-GAPDH antibody (Santa Cruz Biotechnology, Santa Cruz) and, subsequently, with horseradish peroxidase-conjugated α-rabbit IgG or α-goat IgG (Dako, Hamburg, Germany). Immuno-reactive bands were visualized with ECL plus a Western Blotting Detection System (Amersham-Bioscience).

Statistical analysis

All data were analyzed using the Student t-test with sigmastat statistical software (version 2.0), with significance set at P≤0.01 (^*) and P≤0.001 (^**).

Results

Immunodetection of CadF in C. jejuni isolates

The 58 Campylobacter isolates used in this study were characterized as C. jejuni (40 strains) and C. coli (18 strains). The C. jejuni isolates included strains isolated from both humans and animals, while all the C. coli strains were recovered from animals (Table 1).

Using two CadF-specific antisera, a 37 kDa band (p37) and a less-prominent 32 kDa band (p32) were detected in C. jejuni strains by immunoblotting of total cell lysates. These bands corresponded to previously described CadF proteins (Konkel et al., 1997; Mamelli et al., 2006, 2007). While p37 was present in all C. jejuni isolates, five human isolates and one from a calf failed to exhibit the less-prominent p32 band (Table 1). A representative gel and immunoblot with α-CadF-1 of several C. jejuni isolates are shown in Fig. 1a. Equivalent amounts of proteins present were confirmed by Coomassie staining for all tested strains (Fig. 1b).

To verify the specificity of the α-CadF antibodies, two isogenic cadF mutants were produced in strains 81116 and F38011. These mutants lacked the p37 and p32 bands observed for the parent strain (Fig. 5a, arrows). As expected, whole-cell extracts of Campylobacter fetus, Helicobacter pylori or E. coli controls did not react with the CadF-specific antisera (data not shown).

Variability of CadF proteins among C. jejuni and C. coli isolates

Although the pattern of α-CadF-1 antibody reactivity was largely identical among the isolates, the number and intensities of the CadF protein species slightly varied among C. jejuni strains (Fig. 2a, arrows and asterisks), despite loading equivalent amounts of proteins (Fig. 2b). In some cases, intermediate CadF bands of ∼34 kDa were also observed (Fig. 2a, arrows).

Interestingly, in all C. coli isolates tested, CadF was slightly larger and had a weaker expression, as judged from Western blot analysis with α-CadF-1 antibody (Fig. 3a). All C. coli isolates exhibited a 39 kDa band (p39), while a lower migrating 34 kDa band (p34) was detected in 12 out of 18 C. coli strains (Fig. 3a, Table 1).

PCR amplification and sequencing of cadF genes

To elucidate the differences in CadF protein size and expression between C. jejuni and C. coli strains, sequence analyses on a set of cadF genes were performed. PCR analysis of the C. coli strains revealed a slightly larger cadF than that of C. jejuni 81116 (1320 vs. 1285 bp for cadF and some flanking sequence, respectively) (Fig. 3b, arrows). A different PCR with primers directed against the most conserved parts within the cadF gene yielded 930 bp for C. coli strains and 890 bp for C. jejuni 81116 (Fig. 3b, arrows). Insertion of a kanamycin resistance cassette in 81116ΔcadF mutant resulted in a 1.5 kb increase in product size in both PCRs, as expected (Fig. 3b, arrowheads).

Sequencing of the cadF coding region from three C. coli isolates consistently revealed an additional sequence (39 bp) at the indicated position compared with cadF of C. jejuni (Fig. 3c). Analysis of the cadF sequences from three C. jejuni and one C. coli available sequenced genomes (Parkhill et al., 2000; Fouts et al., 2005; Hofreuter et al., 2006) confirmed the findings of this study. Alignment of deduced amino acid sequences showed that the CadF protein from C. coli strains is 13 aa larger than those from C. jejuni (Fig. 3d), in agreement with the size differences seen in the Western blots (Fig. 3a).

Binding and invasion of INT-407 cells by differently CadF-expressing C. jejuni and C. coli strains

Possible differences in bacterial adhesion and invasion between the CadF-expressing C. jejuni and C. coli isolates were explored in infection assays with INT-407 cells. Quantification of cell-associated (Fig. 4a) and intracellular bacteria (Fig. 4b) by the gentamicin protection assay revealed that the C. jejuni isolates expressing p37 CadF exhibited significantly higher binding and invasion rates than C. coli strains expressing p39 CadF (P≤0.001). The C. coli isolates Han35 and Han153 exhibited the lowest values of cell-associated and intracellular bacteria. To determine the overall contribution of the CadF protein in the binding and invasion of C. jejuni to INT-407 cells, the interactions of C. jejuni 81116ΔcadF and F38011ΔcadF mutant strains with cells were examined. Immunoblot analysis with α-CadF-1 confirmed that the CadF protein was not synthesized by either cadF mutant strain (Fig. 5a). Significant reduction (∼50%) in adherence and invasion was observed for the C. jejuni 81116ΔcadF and F38011ΔcadF mutant strains (Fig. 5b and c). These findings demonstrate that CadF is an important pathogenicity factor.

Discussion

The pathogenicity of several Campylobacter species is dependent on their ability to attach and invade the human intestine. One of the adhesion factors that C. jejuni uses to attach, and eventually to invade mammalian cells, is CadF, a protein that binds to fibronectin − a component of the extracellular matrix (Konkel et al., 1997). The importance of CadF for the binding of C. jejuni to epithelial cells has been demonstrated in vitro (Konkel et al., 1997) and in vivo (Ziprin et al., 1999). The primary aim of this study was to determine the genetic and functional diversity of CadF protein among a large number of C. jejuni and C. coli isolates.

Western blotting analyses with two highly specific α-CadF antibodies showed a prominent 37 kDa CadF protein (p37) as well as a less-prominent 32 kDa (p32) protein for all tested C. jejuni isolates. Both bands were absent in two isogenic cadF knockout mutants. The results, which are consistent and extend earlier observations (Konkel et al., 1997, 1999b), also revealed that the number and intensity of CadF bands varied among C. jejuni strains. While p37 was detected in all C. jejuni isolates of human and animal origin, the less-prominent p32 band was found only in 62% of the C. jejuni isolates of human origin and in 96% of the C. jejuni isolates of animal origin. Heat modifiablity is a well-known feature of membrane proteins (Nakamura & Mizushima, 1976; Bolla et al., 1995), including CadF (Konkel et al., 1997, 1999b; Mamelli et al., 2006, 2007). Therefore, the migration of CadF as two protein species is likely caused by their heat-modifiable conformational state, where p32 is the incompletely denaturated and partially folded form of CadF.

In contrast to earlier reports, where the CadF protein was found to be conserved in size and antigenicity among C. jejuni and C. coli isolates from US (Konkel et al., 1999b), it was observed that all C. coli isolates tested possessed a larger CadF (p39 and p34) than C. jejuni. Sequence analysis of three C. coli isolates confirmed this difference between species and indicated that C. coli carried a stretch of 13 aa in the middle region of the protein. Interestingly, the latter insertion sequence was not found in one C. coli isolate from the US, which instead contained another insertion sequence of 7 aa (Konkel et al., 1999b). However, whether the differences in amino acid sequence or a lower expression level accounted for the apparent weaker immunoreactivity of the C. coli CadF with the polyclonal antisera of this study remains to be determined. Nevertheless, data of this study strongly suggest that the differences in molecular size and differences in nucleotide sequence between the C. jejuni and C. coli isolates may be a suitable diagnostic marker to discriminate between these species in food and clinical specimen.

The possible biological significance of the variation in CadF was investigated by comparing a subset of C. jejuni and C. coli strains for their ability to infect INT-407 intestinal epithelial cells, which serves as an in vitro model system for C. jejuni and C. coli attachment and invasion (Hu & Kopecko, 1999; Bacon et al., 2000; Biswas et al., 2000; Monteville et al., 2003; Nadeau et al., 2003; Konkel et al., 2004; Hu et al., 2005). Interestingly, C. jejuni strains adhered and invaded INT-407 cells at significantly greater levels than C. coli strains. This effect was at least in part caused by CadF as the 81116ΔcadF and F38011ΔcadF mutants showed reduced adhesion, which is consistent with previous studies showing a reduced adherence to INT-407 cells of a C. jejuni cadF mutant (Konkel et al., 1997; Monteville et al., 2003). These results may indicate that C. coli CadF is less functional than its C. jejuni counterpart.

Acknowledgements

The work of S.B. is supported through NBL-3 project (Magdeburger Forschungsverbund PFG4) and Priority Program SPP1150 of the Deutsche Forschungsgemeinschaft (Ba1671/3-2). M.K.-G. is supported by an NBL3-fellowship, 01ZZ0407. A grant from the USDA National Research Initiative Food Safety 32.0 Program (2006-35201-17305) supports the CadF project in the Konkel lab.

References

Author notes

Editor: Bruce Ward

© 2007 Federation of European Microbiological Societies

Issue Section:

Research Letters

Email alerts

Citing articles via

More from Oxford Academic