|Home | About | Journals | Submit | Contact Us | Français|
Campylobacter jejuni is one of the most frequent bacterial causes of food-borne gastrointestinal disease in developed countries. Previous work indicates that the binding of C. jejuni to human intestinal cells is crucial for host colonization and disease. Fibronectin (Fn), a major constituent of the extracellular matrix, is a ~250-kDa glycoprotein present at regions of cell-to-cell contact in the intestinal epithelium. Fn is composed of three types of repeating units: type I (~45 amino acids), type II (~60 amino acids), and type III (~90 amino acids). The deduced amino acid sequence of C. jejuni flpA (Cj1279c) contains at least three Fn type III domains. Based on the presence of the Fn type III domains, we hypothesized that FlpA contributes to the binding of C. jejuni to human INT 407 epithelial cells and Fn. We assessed the contribution of FlpA in C. jejuni binding to host cells by in vitro adherence assays with a C. jejuni wild-type strain and a C. jejuni flpA mutant and binding of purified FlpA protein to Fn by enzyme-linked immunosorbent assay (ELISA). Adherence assays revealed the binding of the C. jejuni flpA mutant to INT 407 epithelial cells was significantly reduced compared with that for a wild-type strain. In addition, rabbit polyclonal serum generated against FlpA blocked C. jejuni adherence to INT 407 cells in a concentration-dependent manner. Binding of FlpA to Fn was found to be dose dependent and saturable by ELISA, demonstrating the specificity of the interaction. Based on these data, we conclude that FlpA mediates C. jejuni attachment to host epithelial cells via Fn binding.
Members of the genus Campylobacter are gram-negative, asaccharolytic, motile bacteria, which grow optimally in the laboratory at temperatures between 37 and 42°C under microaerophilic conditions. Although members of Campylobacter spp. were initially recognized to cause disease in sheep and cattle, Campylobacter jejuni was not recognized as a human pathogen until much later (25). Infection of humans with C. jejuni is characterized by a rapid onset of fever, abdominal cramps, and diarrhea. C. jejuni is now recognized as one of the leading bacterial causes of gastroenteritis in the world. In spite of the incidence of campylobacteriosis, relatively few C. jejuni virulence genes have been characterized, and our understanding of the virulence properties of C. jejuni is limited compared with that of other enteric pathogens, including Salmonella, Shigella, and Yersinia spp.
The ability of C. jejuni to cause disease is a complex, multifactorial process. Virulence factors that contribute to the pathogenesis of C. jejuni are associated with motility, host (target) cell adherence, host cell invasion, protein secretion, alteration of host cell signaling pathways, induction of host cell death, evasion of host immune defenses, iron acquisition, and drug/detergent resistance (14, 18). The binding of C. jejuni to specific host cell ligands is hypothesized to play a fundamental role in host colonization and disease progression, since it prevents the organism's clearance from the intestine by peristalsis and fluid flow. Fauchere et al. (5) reported that C. jejuni isolates recovered from individuals with fever and diarrhea adhered to cultured cells in greater numbers than isolates recovered from asymptomatic individuals. While there is no evidence indicating that C. jejuni produces fimbriae that assist in host colonization (7), a number of constitutively synthesized proteins have been proposed to act as adhesins. Bacterial adhesins are surface-exposed macromolecules that facilitate an organism's binding to the host cell receptors. Known and putative C. jejuni adhesins include CadF, CapA, FlpA, and PorA (MOMP) (6).
An emerging theme among pathogenic microorganisms is their ability to utilize host cell molecules during the infectious process to facilitate their binding and entry into host cells (27). More specifically, many bacterial pathogens have been found to bind to fibronectin (Fn), which in turn modifies host cell signaling pathways to the pathogen's advantage. Fn exists as a dimer of nearly identical 250-kDa subunits that are linked by a pair of disulfide bonds near their C termini. Each Fn monomer is composed of three types of repeating units: type I (~45 amino acids), type II (~60 amino acids), and type III (~90 amino acids) (22). In total, each monomer contains 12 type I repeats, two type II repeats, and 15 to 17 type III repeats. Fn participates in many cellular interactions, including tissue repair, embryogenesis, blood clotting, and cell migration/adhesion. Plasma Fn, which is synthesized by hepatocytes, is soluble (22). In contrast, Fn involved in host cell-extracellular matrix (ECM) interaction, which is synthesized by chondrocytes, fibroblasts, endothelial cells, macrophages, and certain epithelial cells, is present in an insoluble form (22). Fn serves as an adhesion molecule that anchors cells to ECM components, including collagen and other proteoglycan substrates.
The bacterial proteins that bind to ECM components have been termed microbial surface components recognizing adhesive matrix molecules (MSCRAMMs) (23). The C. jejuni CadF protein is a member of the MSCRAMM family and one of the most extensively characterized C. jejuni virulence determinants (10-12, 15, 16, 19-21, 24, 28). CadF mediates the binding of C. jejuni to Fn, promotes bacterium-host cell interactions, and facilitates the organism's colonization of chickens (10, 11, 15, 16, 20, 21, 28). In addition to CadF, we recently reported that a mutation in Cj1279c resulted in a C. jejuni mutant that poorly colonized broiler chickens compared with a C. jejuni wild-type strain. The product encoded by the Cj1279c gene was termed Fibronectin-like protein A (FlpA) because the protein harbors Fn type III domains (6). The goal of this study was to characterize the binding properties of FlpA and to determine if this protein is a member of the MSCRAMM family. Here we provide experimental evidence that C. jejuni FlpA is surface exposed, promotes the bacterium's attachment to host epithelial cells, and has Fn binding activity. Assays were also performed to determine if CadF and FlpA act cooperatively to promote binding of C. jejuni to host cells and Fn. We submit that the identification of a second MSCRAMM in C. jejuni highlights the importance of Fn binding in host colonization and disease.
All Campylobacter jejuni strains were cultured on Mueller-Hinton agar plates supplemented with 5% bovine blood (MH-blood agar) under microaerobic conditions (5% O2, 10% CO2, 85% N2) at 37°C. Strains were passed to fresh plates every 24 to 48 h. The C. jejuni F38011 strain was recovered from an individual with campylobacteriosis. The C. jejuni F38011 flpA (tetracycline resistant [Tetr]) and cadF flpA (kanamycin resistant [Kanr] and Tetr) mutants were generated as outlined below. The C. jejuni F38011 cadF Kanr mutant was generated as outlined elsewhere (11). When appropriate, the growth media were supplemented with chloramphenicol (Chl) (8 μg/ml), Kan (50 μl/ml), Tet (2 μg/ml), or cefoperazone (30 μg/ml). Escherichia coli XL1-Blue MRF′ (Tetr) (Stratagene, Garden Grove, CA), E. coli BL21 (Novagen, Madison, WI), and E. coli LMG194 (streptomycin resistant and Tetr; Invitrogen, Carlsbad, CA) were grown aerobically at 37°C on Luria-Bertani (LB) agar plates or in LB broth. When necessary, growth media were supplemented with ampicillin (Amp) (100 μg/ml), Kan (50 μg/ml), Tet (12.5 μg/ml), or Chl (20 μg/ml).
The C. jejuni F38011 wild-type strain was grown to mid-exponential phase in MH broth, and total cellular RNA was extracted using the RiboPure bacterial kit (Ambion, Austin, TX) according to the manufacturer's instructions. Genomic DNA was degraded by treatment with 11 U of RQ1 RNase-free DNase at 37°C for 30 min. cDNA was synthesized from 500 ng of RNA using random hexamer primers and the ThermoScript reverse transcriptase (RT) PCR system (Invitrogen) according to the manufacturer's directions. As a negative control, RT-PCRs were performed without RT enzyme. Two separate RNA extractions and cDNA synthesis reactions were carried out on different days.
Table Table11 lists all primers used in this study. PCR was performed to determine which genes are cotranscribed with flpA using 1 μl of a 1:10 dilution of cDNA as a template in a total volume of 25 μl. As a positive control, the reactions were carried out using C. jejuni F38011 genomic DNA as a template. DNA fragments were amplified using Taq DNA polymerase (Invitrogen, Carlsbad, CA) with the following parameters: 94°C for 4 min, 1 cycle; 94°C for 45 s, 60°C for 30 s (−1°C per cycle), and 2 min at 70°C, 10 cycles; 94°C for 45 s, 50°C for 30 s, and 2 min at 70°C, 25 cycles. PCR products spanning the junctions between the genes Cj1280c, flpA, Cj1278c, Cj1277c, Cj1276c, and Cj1275c were amplified using the following primer pairs: MEK2386 and MEK2387, MEK2388 and MEK2389, MEK2412 and MEK2411, MEK2420 and MEK2421, and MEK2422 and MEK2423. The resulting PCR amplicons were analyzed by electrophoresis in a 1% agarose gel.
A mutation in the flpA gene of C. jejuni F38011 was generated by homologous recombination using a suicide vector harboring a disrupted copy of the flpA gene. The 5′ flanking region of the flpA gene was PCR amplified using HiFi Taq (Invitrogen) with the primers MEK1672 and MEK1671, containing BamHI and SstII restriction sites, and ligated into pCR2.1 (Invitrogen). The 3′ flanking region of the flpA gene was PCR amplified using primers MEK1673 and MEK1674, containing the SstII and BamHI restriction sites, and ligated into pCR2.1. The 3′ fragment was restricted with the SstII and BamHI restriction enzymes, gel purified, and ligated to the 5′ fragment in the pCR2.1 vector. The resultant vector was digested with SstII, and the tetO gene conferring Tet resistance was inserted. This vector was then digested with BamHI to liberate the fragment containing the 5′ and 3′ flpA flanking fragments with the tetO gene, which was subsequently ligated into the suicide vector pBSK (Stratagene, La Jolla, CA). The pBSK vector had previously been modified to include an aphA-3 gene cassette encoding Kanr. This vector was electroporated into the C. jejuni F38011 wild-type strain and C. jejuni F38011 cadF mutant, and colonies were picked that were Tetr. The C. jejuni flpA mutants were confirmed by PCR using flpA gene-specific primers.
The flpA open reading frame (ORF) with 0 bp of upstream sequence and 15 bp of downstream sequence was PCR amplified from C. jejuni F38011 genomic DNA using HiFi Taq and the primers MEK1681 and MEK1883, harboring the NdeI and KpnI restriction enzymes. The metK promoter sequence was amplified from C. jejuni NCTC 11168 using the primers MEK1687 and MEK1688, harboring the BamHI and NdeI restriction enzymes. The metK promoter-flpA gene product was cloned into the multiple cloning site of the pRY111 shuttle vector using BamHI and KpnI sites. The presence of the metK promoter-flpA in pRY111 was confirmed by DNA sequencing, and the resultant vector was electroporated into E. coli S17-1 λ-pir for conjugation into the C. jejuni F38011 flpA mutant. The conjugations were performed with overnight cultures of the C. jejuni F38011 flpA mutant grown in MH broth supplemented with Kan and E. coli S17-1 λ-pir harboring the pRY111 metK promoter-flpA construct grown in LB broth supplemented with Chl. The bacteria (the equivalent of 1 optical density at 540 nm [OD540] unit) were pelleted via centrifugation at 6,000 × g for 2 min, and the supernatant was discarded. The E. coli S17-1 λ-pir pellet was resuspended in 500 μl of MH broth and combined with the C. jejuni F38011 flpA mutant pellet. The cells were pelleted again, and the supernatant was discarded. The combined pellet was then spotted onto an MH-blood agar plate and incubated at 37°C in a microaerophilic environment for 14 h. The conjugation spot was then streaked onto MH-blood agar plates supplemented with Chl and cefoperazone and incubated for 48 h. Isolated transformants were selected, and the presence of the recombinant vector in the C. jejuni flpA mutant was confirmed by PCR. The complemented flpA mutant was designated the C. jejuni flpA (flpA+) complemented strain.
Recombinant histidine-tagged FlpA was generated using the pET expression system from Novagen. A fragment of the flpA gene was PCR amplified using the gene-specific primers MEK1679 and MEK1680, harboring the BamHI and XhoI restriction enzymes, and cloned into the pET24b (Kanr) vector using standard molecular biology techniques. The recombinant plasmid, flpA-pET24b, was introduced into E. coli BL21(DE3). The His-tagged FlpA protein was purified using Talon metal affinity resin (Clontech, Mountain View, CA) according to the manufacturer's directions.
To determine if FlpA facilitates the binding of E. coli to epithelial cells, we expressed the flpA gene in E. coli using the pBAD expression vector. The pBAD/Myc-His A vector (Ampr), referred to as pBADA from this point forward, was obtained from Invitrogen. A fragment of the flpA gene was PCR amplified using the gene-specific primers MEK1765 and MEK1766, harboring the NcoI and KpnI restriction enzymes, and cloned into the pBADA vector using standard molecular biology techniques. The recombinant plasmid, flpA-pBADA, was introduced into E. coli LMG194. Expression of the flpA gene in E. coli LMG194 was induced by the addition of l-arabinose as outlined by the supplier.
The ability of FlpA to bind Fn was determined by enzyme-linked immunosorbent assay (ELISA) using purified glutathione S-transferase (GST)-tagged FlpA. The flpA gene was PCR amplified using the gene-specific primers MEK1691 and MEK1692, harboring the BamHI and XhoI restriction enzymes, and ligated into the pGEX-5x-1 vector using standard cloning procedures. The FlpA-GST protein was purified using glutathione Sepharose 4B affinity resin (GE Healthcare/Amersham) according to the manufacturer's instructions. The cadF gene fragment was cloned into the pGEX-5x-1 vector using the primers MEK2522 and MEK2523. The GST-tagged CadF protein was purified as described for FlpA.
Female New Zealand White rabbits were subcutaneously and intramuscularly injected with 500 μg of purified His-tagged FlpA in TiterMax Gold (CyRx Corporation, Norcross, GA). Two booster injections, each containing 50 μg of protein in Freund's incomplete adjuvant (Sigma), were given at 4 and 6 weeks after the primary injection. Blood was collected prior to all immunizations and 7 days after the second booster injection. The serum was prepared using standard laboratory procedures and stored at −80°C. FlpA-specific antibody was generated in a New Zealand White rabbit using a protocol approved by the Institutional Animal Care and Use Committee (IACUC protocol no. 2433) at Washington State University.
C. jejuni outer membrane proteins (OMPs) were extracted using N-lauroyl-sarcosine as described by de Melo and Pechère (3) with modifications. C. jejuni were grown in MH broth under microaerobic conditions overnight, pelleted by centrifugation, and suspended in 10 mM phosphate buffer (pH 7.4) containing 1 mM phenylmethylsulphonyl fluoride (Sigma, St. Louis, MO), 10 μg/ml DNase I (Sigma), and 10 μg/ml RNase A (Fermentas, Glen Burnie, MD). The bacterial cell suspensions were sonicated with a Branson sonifier cell disruptor (model 250; Branson Sonic Power Co., Danbury, CT) five times for 30 s each with a 30-s cooling period on ice between each pulse. Cell debris was removed by two successive centrifugations, each at 6,000 × g for 10 min. The crude membrane extracts were obtained by centrifugation at 100,000 × g at 4°C for 2 h. The resulting pellets were suspended in 10 mM Tris (pH 7.5), and the protein concentration of each sample was determined using the bicinchoninic acid assay as outlined in the manufacturer's instructions (Pierce, Rockford, IL). N-lauroyl-sarcosine (Sigma) was added to the crude extracts at a protein-to-detergent ratio of 1:4 (wt/wt). The samples were incubated at room temperature with gentle rocking for 30 min and centrifuged at 100,000 × g at 4°C for 2 h. The pellets were washed with 50 mM Tris (pH 7.5), suspended in the same buffer, and stored at −20°C. The protein concentration of the OMP extracts was determined by bicinchoninic acid assay.
Whole-cell lysates (the equivalent of 0.1 OD540 U) of the C. jejuni F38011 wild-type strain and mutants were solubilized in single-strength electrophoresis sample buffer and incubated at 95°C for 5 min. The proteins were separated in sodium dodecyl sulfate (SDS)-12.5% polyacrylamide gels using the discontinuous system described by Laemmli (17). Following electrophoresis, proteins were stained with Coomassie brilliant blue R-250 (CBB R-250; Bio-Rad Laboratories, Hercules, CA). For immunoblot analysis, proteins were electrophoretically transferred to a polyvinylidene difluoride membrane (Immobilon P; Millipore, Bedford, MA). Immunoblotting was performed by incubating the membrane overnight at 4°C with a 1:500 dilution of the anti-FlpA serum in phosphate-buffered saline (PBS)-Tween (20 mM sodium phosphate and 150 mM sodium chloride, pH 7.5, containing 0.01% [vol/vol] Tween 20) with 9% nonfat dry milk. After three washes with PBS-Tween, horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (whole molecule) diluted 1:5,000 in PBS-Tween was added as a secondary antibody, followed by incubation at room temperature for 1 h. Following two washes with PBS-Tween and a final wash with PBS, blots were developed using Kodak BioMax MR film and Western Lightning chemiluminescence (PerkinElmer, Boston, MA) according to the manufacturer's directions.
The C. jejuni F38011 wild-type strain and flpA mutant were harvested from MH-blood agar plates in PBS, and 20 μl of the bacterial suspension was air-dried on a glass microscope slide. The air-dried samples were quickly passed over a flame, and PBS was added to the surfaces of the slides. The bacteria were incubated for 45 min at 37°C in a humidified chamber with either a 1:20 dilution of a rabbit anti-C. jejuni whole-cell polyclonal serum (antiserum 1622) (13), rabbit anti-FlpA serum, or rabbit prebleed serum in PBS containing 0.75% bovine serum albumin (BSA). The slides were washed three times with PBS and then incubated for 45 min at 37°C in a humidified chamber with a 1:100 dilution of Cy2-conjugated AffiniPure goat anti-rabbit IgG(H+L) (Jackson ImmunoResearch, West Grove, PA). Following incubation, the samples were rinsed 10 times with PBS, placed on a glass slide with mounting medium containing 4′,6-diamidino-2-phenylindole (DAPI) (Vectashield, Vector Laboratories, Inc., Burlingame, CA), and visualized using a Nikon Eclipse TE2000 inverted epifluorescence microscope. DAPI, a fluorescent stain that binds to DNA, was used to visualize all bacteria. Images were captured using the imaging software MetaMorph version 5 and processed using the Adobe Photoshop 3.0.4 software program.
INT 407 human intestinal epithelial cells (ATCC CCL6; American Type Culture Collection, Manassas, VA) were maintained in minimal essential media ((Gibco, Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (HyClone Laboratories, Logan, UT) and 5% L-glutamine (1.8 mM). The cells were cultured at 37°C in a humidified, 5%-CO2 incubator and passaged every 48 to 72 h.
Each well of a 24-well tissue culture tray was seeded with INT 407 cells (1.5 × 105 cells/well) and incubated for 18 h at 37°C in a humidified, 5%-CO2 incubator. The cells were rinsed with the appropriate medium and inoculated with approximately 5 × 107 CFU of the various C. jejuni strains. Bacterium-host cell contact was promoted by centrifugation at 600 × g for 5 min. To determine the viable number of bacteria that adhered to the INT 407 cells, the trays were incubated for 30 min at 37°C in a humidified, 5%-CO2 incubator. Following this incubation period, the epithelial cells were rinsed three times with PBS to remove nonadherent bacteria. The epithelial cells were then lysed with a solution of 0.1% (vol/vol) Triton X-100 in PBS. The suspensions were serially diluted, and the number of viable, adherent bacteria was determined by counting the resultant colonies on MH-blood agar plates. The values reported represent the mean counts ± standard deviations derived from triplicate wells.
To determine if antibodies against FlpA reduce the binding of C. jejuni to INT 407 cells, different dilutions of the FlpA-specific and prebleed sera were added to bacterial suspensions containing approximately 5 × 107 CFU. The bacterial suspensions were then incubated for 30 min at 37°C under microaerobic conditions (5% O2, 10% CO2, 85% N2). The binding assay was performed as outlined above.
Binding (adherence) assays were also performed with E. coli LMG194 harboring flpA-pBADA and E. coli LMG194 harboring pBADA without a DNA insert. For these assays, the bacteria were cultured overnight at 37°C in LB broth supplemented with Amp. The following morning, 5 ml of LB broth containing Amp was inoculated with 250 μl of the overnight culture and incubated with shaking for 90 min at 37°C. Expression of flpA was induced by the addition of l-arabinose; 0.0002% l-arabinose was added to all the cultures for 2 h. The amount of l-arabinose added to the bacterial cultures was determined based on preliminary experiments that examined relative FlpA protein levels versus time and bacterial viability. The adherence assays were performed as described above for the C. jejuni strains, with the exception that the INT 407 cells were inoculated with approximately 2 × 107 CFU of each E. coli isolate.
The wells of a 96-well plate were coated overnight at 4°C with either 1 μg of plasma Fn(Sigma, St. Louis, MO) or BSA-coated wells, which served as a negative control. The following day, the wells were rinsed with wash buffer (PBS containing 0.005% [vol/vol] Tween 20), blocked with 1% BSA for 1 h at 25°C, and rinsed once with wash buffer. To determine the Fn binding activity of each protein, twofold serial dilutions of FlpA-GST and CadF-GST proteins were made in PBS, added to the wells, and incubated for 90 min at 25°C. The CadF-GST protein was used as a positive control for Fn binding affinity. After the wells were washed three times with wash buffer, a 1:1,000 dilution of rabbit anti-GST antibody (Sigma) in incubation buffer was added to the wells and the plate was incubated at 25°C for 90 min. The wells were then washed three times with wash buffer, a 1:5,000 dilution of horseradish peroxidase-labeled goat anti-GST antibody (Sigma) diluted in PBS was added to the wells, and the plate was incubated at 25°C for 90 min. The wells were washed, and bound antibodies were detected by addition of TMB substrate solution (Thermo Scientific, Rockford, IL). Binding was quantitated by colorimetric detection at 492 nm.
The wells of 96-well flat-bottom plates (Costar, Corning, NY) were coated with a 1-mg/ml solution of Fn in 0.05 M Tris-buffered saline, pH 7.5 (Sigma), overnight at 4°C. For a control, wells were also coated with 1% BSA in PBS. The C. jejuni F38011 wild type and flpA, cadF, and cadF flpA mutants were harvested from overnight plate cultures and resuspended in PBS at an OD540 of 0.150 (approximately 108 CFU). Wells were rinsed with PBS, and 100 μl of the bacterial suspensions was added to each well and incubated at 37°C with 5% CO2 for 1 h. The wells were washed three times with PBS, and adherent bacteria were removed by the addition of 0.25% trypsin (Gibco, Invitrogen). To enumerate the number of adherent bacteria, serial dilutions of the trypsin suspension were plated on MH-blood agar.
Analysis of the flpA gene and predicted operon structure from four C. jejuni sequenced strains (i.e., NCTC 11168, RM1221, 81-176, and 811116) revealed conserved features (Fig. (Fig.1).1). The orders of the genes flanking flpA are identical in the four C. jejuni sequenced strains (i.e., Cj1280c, Cj1279c [flpA], Cj1278c, Cj1277c, and Cj1276c). Apart from in silico analysis of flpA, little is known regarding the operon structure in which this gene resides. To determine the number of genes in the operon in which flpA resides, PCR was initially performed using gene-specific primer experiments to determine the gene order in the C. jejuni F38011 clinical strain. The sizes of the PCR fragments were in agreement with those predicted from an NCTC 11168 genome analysis, suggesting that the C. jejuni F38011 strain likely has the same gene order (i.e., Cj1280c, Cj1279c [flpA], Cj1278c, Cj1277c, and Cj1276c) as the four C. jejuni strains indicated above. We then performed RT-PCR analysis with gene-specific primers to experimentally determine the number of genes in the flpA operon. In the C. jejuni F38011 strain, flpA is the second gene in an operon consisting of Cj1280c, Cj1279c (flpA), Cj1278c, and Cj1277c.
The flpA gene in C. jejuni NCTC 11168 is 1,236 nucleotides in size. The ORF from C. jejuni NCTC 11168 (Cj1279c) begins with two AUG codons in tandem, followed by an AAA codon (Lys residue), and is terminated by a UAG termination codon. One discrepancy in the annotation of the flpA ORF from C. jejuni NCTC 11168 versus other sequenced strains is that the ORF begins with a single AUG codon followed by an AAA codon in the ORF from C. jejuni strains RM1221, 81-176, and 81116. The proposed methionine start codon in all the C. jejuni sequenced strains is preceded by a typical Shine-Dalgarno sequence (AGGA). The ORF from C. jejuni NCTC 11168 encodes 411 amino acids that are predicted to synthesize a protein with a calculated mass of 46,124 Da (Table (Table2).2). In silico analysis of the deduced FlpA amino acid sequence further revealed that the protein shares greater than 99% identity at the amino acid level among the four C. jejuni strains (i.e., NCTC 11168, RM1221, 81-176, and 811116) (Table (Table2).2). Other than the 1 additional methionine at the amino terminus of FlpA from NCTC 11168, only 11 residues differed within the entire deduced amino acid sequence of the four strains. The nucleotide sequence of the flpA gene in the C. jejuni F38011 strain is identical to that of the C. jejuni NCTC 11168 strain except for a single silent nucleotide difference at base 882 (i.e., C in strain F38011 versus T in strain 11168). Examination of the C. jejuni NCTC 11168 FlpA protein predicted amino acid sequence identified an L-S-A-C motif at residues 18 to 21, which matches the prokaryotic lipoprotein signal consensus [LVI][ASTVI][GAS][C] (Fig. (Fig.2).2). This consensus sequence, found at the C-terminal end of a lipoprotein signal peptide, is referred to as the lipobox. The invariant Cys residue is lipid modified and presumably inserted into one leaflet of the lipid bilayer. The deduced FlpA amino acid sequence also harbors Fn type III domains.
A C. jejuni flpA mutant was generated as outlined in Materials and Methods and demonstrated to have growth rates similar to those of the C. jejuni wild-type strain (not shown). To determine the cellular location of FlpA, whole-cell lysates (WCL) and OMP extracts were prepared from a C. jejuni wild-type strain, flpA mutant, and flpA (flpA+) complemented strain and analyzed by SDS-PAGE coupled with immunoblot analysis using an FlpA-specific serum. A band with an Mr of 46,000 was readily observed in WCL extracts of the C. jejuni F38011 wild-type and flpA (flpA+) complemented strains but not in the isogenic flpA knockout (Fig. (Fig.2).2). Consistent with the notion that FlpA is a membrane-associated protein, as suggested by its amino-terminal leader, a 46-kDa immunoreactive band was also observed in the OMP extracts of the C. jejuni F38011 wild-type strain and the flpA (flpA+) complemented strain.
To determine if domains of the FlpA protein are surface exposed, C. jejuni were incubated with the FlpA-specific serum and indirect immunofluorescence microscopy was performed. All bacteria were incubated with a rabbit anti-C. jejuni whole-cell serum for a positive control. After the bacteria were incubated with either the rabbit anti-C. jejuni whole-cell or rabbit FlpA-specific serum, they were incubated with a Cy2-conjugated goat antirabbit secondary antibody and examined. The rabbit anti-C. jejuni whole-cell serum stained both the wild-type and flpA mutant bacteria (not shown). In contrast, the rabbit FlpA-specific serum stained only the wild-type bacteria (Fig. (Fig.3).3). Together, these results indicate that FlpA is a membrane-associated protein with surface-exposed domains.
Previous work in our laboratory demonstrated that FlpA plays a role in C. jejuni colonization of broiler chickens, since only 2 of 10 chickens inoculated with the C. jejuni flpA mutant were colonized (6). To build on this initial work, in vitro adherence assays were performed with human INT 407 cells and a C. jejuni wild-type strain, cadF mutant, flpA mutant, and flpA (flpA+) complemented strain (Fig. (Fig.4).4). The C. jejuni cadF mutant was included in these assays as a negative control (11). At a multiplicity of infection of 30:1, the C. jejuni flpA mutant showed a 62% reduction in adherence to INT 407 cells compared with that of the C. jejuni wild-type strain. In comparison, the C. jejuni cadF mutant showed a 72% reduction in adherence to INT 407 cells compared with that of the C. jejuni wild-type strain. The reduction in the binding of the C. jejuni flpA mutant was judged to be specific, since complementation of the mutant in trans with a wild-type copy of the gene driven by the metK promoter restored the organism's binding to the INT 407 cells. To alleviate the concern of a polar effect and to further demonstrate that the phenotype displayed by the C. jejuni flpA mutant was due to the presence of the FlpA protein, we tested if the binding of C. jejuni to INT 407 cells could be blocked with the FlpA-specific serum (Fig. (Fig.5).5). The FlpA-specific serum reduced the binding of C. jejuni to INT 407 cells in a dose-dependent fashion, reaching a maximum value of 77% inhibition at a 1:12.5 dilution of the serum. In contrast, a statistically significant difference was not observed in the binding of C. jejuni to INT 407 cells treated with the rabbit prebleed serum. Together, these findings demonstrate that FlpA mediates adherence to epithelial cells.
While a C. jejuni flpA mutant exhibited a reduction in binding to INT 407 cells compared with results for the wild-type strain and the FlpA-specific serum blocked adherence, it remained possible that other proteins could act indirectly to potentiate the adhesive property of FlpA. To determine if FlpA is sufficient to promote the binding of bacteria to epithelial cells, adherence assays were performed with E. coli expressing flpA. More specifically, we tested the binding properties of an E. coli LMG194 strain harboring the pBADA plasmid containing flpA (E. coli flpA-pBADA) and the E. coli LMG194 strain harboring pBADA without a DNA insert. Prior to these assays, experiments were performed to determine the minimal concentration of l-arabinose and time period sufficient to induce flpA expression. A 46-kDa band was readily visible in the whole-cell lysates of the E. coli flpA-pBADA strain that had been cultured in medium containing 0.0002% of l-arabinose for 2 h as judged by SDS-PAGE (Fig. (Fig.6).6). A statistically significant difference was observed in binding of the E. coli flpA-pBADA isolate to INT 407 cells ([1.1 ± 0.2] × 106) versus E. coli harboring an empty pBADA vector ([4.13 ± 1.1] × 105). This finding further demonstrates that FlpA is an adhesin.
Each Fn monomer has a molecular mass of 250 kDa and contains type I, II, and III repeat units. Sequence analysis of FlpA revealed the presence of at least three domains with similarity to the Fn type III domain (see Fig. Fig.11 and http://www.microbesonline.org/, VIMSS ID 47155). The Fn type III domain mediates Fn-Fn interactions (22). Based on the presence of the Fn type III domains, ELISAs were performed to determine whether FlpA has Fn binding activity. The C. jejuni CadF protein was included in these assays as a positive control because its Fn binding activity is well documented (10, 11). As a negative control, wells were coated with BSA. In addition, we assessed the binding of GST alone. Fn binding activity was evident with both the FlpA-GST and CadF-GST tagged proteins as judged by ELISA (Fig. (Fig.7).7). The specificity of these interactions was demonstrated in that the binding was both dose dependent and saturable at concentrations between 5 and 10 μg. However, under the conditions used, more CadF bound to Fn than FlpA, suggesting that the two proteins have different affinities for Fn. GST alone did not demonstrate significant Fn binding affinity; background absorbance values of 0.1 were obtained over a range of concentrations (not shown). In addition, all of the GST fusion proteins demonstrated only low-level nonspecific binding to BSA-coated wells. The reason for using the GST recombinant proteins was to alleviate the concern of using different antibodies to detect the bound proteins. The Fn binding activities of FlpA and CadF were also confirmed using FlpA- and CadF-specific antibodies (not shown). Based on these results, we concluded that FlpA has Fn binding activity.
Based on the data shown above, FlpA is an MSCRAMM family member. To determine if binding of FlpA and that of CadF to host cells and Fn are independent of each other, C. jejuni-host cell adherence and Fn binding assays were performed with a C. jejuni wild-type strain, C. jejuni cadF mutant, C. jejuni flpA mutant, and C. jejuni cadF flpA double mutant (Fig. (Fig.8).8). Each of the C. jejuni mutants (i.e., cadF, flpA, and cadF flpA mutants) demonstrated a statistically significant reduction in binding to INT 407 cells and Fn-coated wells compared with results for the wild-type strain (Fig. 8A and B). In addition, the C. jejuni cadF flpA double mutant exhibited a similar reduction in binding to INT 407 cells and Fn-coated wells compared with results for the individual C. jejuni cadF and flpA mutants. Collectively, these data indicate that FlpA and CadF are both needed to facilitate the maximal binding of C. jejuni to Fn and host cells.
Previous work indicates that C. jejuni adherence to gastrointestinal cells and extracellular matrix components is crucial for host colonization and subsequent disease. More specifically, a C. jejuni cadF mutant shows a significant reduction in adhesion to human INT 407 intestinal cells compared to a wild-type strain (21). Similar to the case with cadF, disruption of Cj1279c (flpA) results in a C. jejuni mutant impaired in its ability to bind to chicken LMH hepatocellular carcinoma epithelial cells and to efficiently colonize broiler chickens compared with results for a wild-type strain (6). The product encoded by the Cj1279c gene is termed FlpA, for Fibronectin-like protein A, based on the fact that the protein's deduced amino acid sequence harbors Fn type III domains. Here we conclude that FlpA is associated with outer membrane components as judged by SDS-PAGE coupled with immunoblot analysis using FlpA-specific serum and is surface exposed as judged by immunofluorescence microscopy. We also conclude that FlpA acts as an adhesin based on the following experimental findings: (i) The binding of the C. jejuni flpA mutant strain to INT 407 epithelial cells was significantly reduced compared with that of a wild-type strain; (ii) rabbit polyclonal serum generated against FlpA blocked C. jejuni adherence to INT 407 cells in a dose-dependent manner; and (iii) The expression of flpA in E. coli significantly increased the bacterium's binding to INT 407 cells compared that of with E. coli containing an empty vector. Finally, we submit that FlpA is a member of the MSCRAMM family because it binds to Fn in a dose-dependent and saturable fashion, as demonstrated by ELISA. Based on the sum of in vitro and in vivo assays, we conclude FlpA is a novel C. jejuni adhesin.
While the primary focus of this research was to demonstrate the adhesive properties of FlpA, multiple observations indicate that FlpA is associated with the C. jejuni outer membrane. We visualized a 46-kDa band in OMP extracts prepared from C. jejuni F38011 using an FlpA-specific serum. In addition, a 46-kDa band was apparent in the OMP extracts prepared from the C. jejuni flpA (flpA+) complemented strain. Noteworthy is that the FlpA protein (i.e., CJJ81176_1295) was detected by liquid chromatography/matrix-assisted laser desorption ionization-time of flight/time of flight mass spectrometry in C. jejuni 81-176 OMP extracts previously (2, 24). We also found that FlpA is exposed on the surface of the bacterium as judged by immunofluorescence microscopy using the FlpA-specific antibodies. Consistent with the notion that the domains of FlpA are surface exposed, the FlpA-specific antibodies used for the immunofluorescence assays reduced the adherence of C. jejuni to INT 407 cells in a dose-dependent manner.
Inspection of the amino terminus of FlpA indicated the presence of a lipoprotein signal consensus sequence. Although a few experimental methods are available to conclusively demonstrate that a protein is lipid modified, presumptive evidence for the identification of a lipoprotein is evident from inspection of its deduced amino acid sequence. The amino-terminal signal sequence of a lipoprotein is characterized by a tripartite structure of positively charged residues at the amino terminus, a hydrophobic core region, and the lipobox with the invariant Cys residue at the carboxy terminus of the signal. The deduced FlpA amino acid sequence contains each of these key features. The presence of the prokaryotic lipoprotein signal consensus sequence strongly suggests that FlpA is a lipoprotein.
Adherence assays were performed to determine the contribution of FlpA in the binding of C. jejuni to human INT 407 epithelial cells. A C. jejuni flpA mutant showed a 62% reduction in adherence to INT 407 cells compared with the C. jejuni wild-type strain. In comparison, the C. jejuni cadF mutant showed a 72% reduction in adherence to INT 407 cells. Given that both proteins demonstrate Fn binding activity, we tested if a C. jejuni cadF flpA mutant would exhibit a greater reduction in binding to INT 407 cells than either the C. jejuni flpA mutant or C. jejuni cadF mutant. Here we found that the reduction in binding of the C. jejuni cadF flpA double mutant was indistinguishable from a C. jejuni cadF or flpA mutant alone. Subsequently, we tested if purified FlpA and CadF compete for binding to Fn-coated wells as determined by ELISA. However, we were unable to identify conditions under which the two proteins compete for Fn binding. Studies are currently under way to localize the FlpA Fn binding domain and further elucidate the specific binding characteristics of CadF and FlpA. Based on these data, we favor a model whereby CadF binds to one portion of Fn and FlpA binds to another portion with both interactions being required for intimate host cell and Fn attachment. Studies are in progress to further define the CadF-Fn and FlpA-Fn interactions. Regardless of the specifics of these interactions, it is noteworthy that C. jejuni possesses at least two Fn binding proteins (i.e., MSCRAMMs). Similar to the case with C. jejuni, Streptococcus dysgalactiae harbors more than one MSCRAMM (9).
Fn binding plays an important role in the pathogenesis of a number of medically relevant bacteria, including Staphylococcus aureus (FnBPA), Streptococcus pyogenes (protein F1; PrtF1/SfbI), Salmonella enterica serotype Typhimurium (ShdA), Borrelia burgdorferi (BBK32), and Treponema pallidum (TP0136) (1). Here we have provided experimental evidence demonstrating that FlpA promotes the attachment of C. jejuni to host epithelial cells and has Fn binding activity. The identification and characterization of FlpA, along with CadF, highlights the potential importance of C. jejuni binding to Fn for host colonization and disease. Specifically, it is well established that Fn binding to the α5β1 integrins can stimulate integrin clustering, which in turn initiates a cascade of events leading to rearrangements in the cytoskeleton, including actin-based membrane protrusions (8). Numerous examples in the literature demonstrate the manipulation of host cell behavior by a pathogen binding to an ECM component (4, 26). Studies are currently in progress to further dissect the roles of CadF and FlpA binding to Fn in C. jejuni-host cell adherence and invasion.
We thank Matt Gay for assistance with this study, Jason Neal-McKinney for performing operon analysis using RT-PCR, and Darcy Duong for preparation of the recombinant CadF and FlpA proteins. We also thank Daelynn Buelow, Jeffrey Christensen, and Jason Neal-McKinney (School of Molecular Biosciences, Washington State University) for critical review of the manuscript.
This work was supported by funds awarded to M.E.K. from the USDA National Research Initiative's Food Safety 32.0 program (2006-35201-16553 and 2006-35201-17305).
Published ahead of print on 30 October 2009.