Modeling the surface of Campylobacter fetus: protein surface layer stability and resistance to cationic antimicrobial peptides.

Campylobacter fetus is a Gram negative bacterium recognized for its virulence in animals and humans. This bacterium possesses a paracrystalline array of high molecular weight proteins known as surface-layer proteins covering its cell surface. A mathematical model has been made of the outer membrane of this bacterium, both with its surface-layer proteins (S+) and without (S-). Monte Carlo computer simulation was used to understand the stability of the surface-layer protein structure as a function of ionic concentration. The interactions of an electrically-charged antimicrobial agent, the cationic antimicrobial peptide protamine, with surface-layer proteins and with the lipopolysaccharides of the outer membrane were modeled and analyzed. We found that (1) divalent ions stabilize the surface-layer protein array by reducing the fluctuations perpendicular and parallel to the membrane plane thereby promoting adhesion to the LPS region. This was achieved via (2) divalent ions bridging the negatively-charged LPS Core. The effect of this bridging is to bring individual Core regions closer together so that the O-antigens can (3) increase their attractive van der Waals interactions and "collapse" to form a surface with reduced perpendicular fluctuations. These findings provide support for the proposal of Yang et al. [1]. (4) No evidence for a significant increase in Ca(2+) concentration in the region of the surface-layer protein subunits was observed in S+ simulations compared to S- simulations. (5) We predicted the trends of protamine MIC tests performed on C. fetus and these were in good agreement with our experimental results.

Identification of antigenic epitopes of the SapA protein of Campylobacter fetus using a phage display peptide library.

In this study, we immunized mice with prokaryotically expressed recombinant surface layer protein, SapA, of Campylobacter fetus, generated hybridomas secreting mouse monoclonal antibodies (mAb) targeting SapA, and purified the mAb A2D5 from mouse ascites using saturated ammonium sulfate solution. The mAb A2D5, coated onto ELISA plates, was used to screen the phage random 12-peptide library through three rounds of panning. Following panning, 15 phage clones were randomly chosen and tested for reactivity with mAb A2D5 by indirect ELISA. Single-stranded DNA from positive clones was sequenced and compared with the sequence of SapA to predict the key epitope. ELISA and/or Western blot analyses further validated that synthetic peptides and recombinant peptide mimotopes all interact with mAb A2D5. Nine of ten positive phage clones identified by screening were sequenced successfully. Seven clones shared the same sequence HYDRHNYHWWHT; one had the sequence LSKNLPLTALGN; and the final one had the sequence SGMKEPELRSYS. These three sequences shared high homology with SapA J05577 in the region GNEKDFVTKIYSIALGNTSDVDGINYW, in which the underlined amino acids may serve as key residues in the epitope. ELISA and/or Western blot analyses showed that mAb A2D5 not only interacted with the four synthetic peptide mimotopes, but also with 14 prokaryotically expressed recombinant peptide mimotopes. The mimotopes identified in this study will aid future studies into the pathological processes and immune mechanisms of the SapA protein of C. fetus.

Interbacterial macromolecular transfer by the Campylobacter fetus subsp. venerealis type IV secretion system.

We report here the first demonstration of intra- and interspecies conjugative plasmid DNA transfer for Campylobacter fetus. Gene regions carried by a Campylobacter coli plasmid were identified that are sufficient for conjugative mobilization to Escherichia coli and C. fetus recipients. A broader functional range is predicted. Efficient DNA transfer involves the virB9 and virD4 genes of the type IV bacterial secretion system encoded by a pathogenicity island of C. fetus subsp. venerealis. Complementation of these phenotypes from expression constructions based on the promoter of the C. fetus surface antigen protein (sap) locus was temperature dependent, and a temperature regulation of the sap promoter was subsequently confirmed under laboratory conditions. Gene transfer was sensitive to surface or entry exclusion functions in potential recipient cells carrying IncPα plasmid RP4 implying functional relatedness to C. fetus proteins. The virB/virD4 locus is also known to be involved in bacterial invasion and killing of cultured human cells in vitro. Whether specifically secreted effector proteins contribute to host colonization and infection activities is currently unknown. Two putative effector proteins carrying an FIC domain conserved in a few bacterial type III and type IV secreted proteins of pathogens were analyzed for secretion by the C. fetus or heterologous conjugative systems. No evidence for interbacterial translocation of the Fic proteins was found.

Structure and genotypic plasticity of the Campylobacter fetus sap locus.

The Campylobacter fetus surface layer proteins (SLPs), encoded by five to nine sapA homologues, are major virulence factors. To characterize the sapA homologues further, a 65.9 kb C. fetus genomic region encompassing the sap locus from wild-type strain 23D was completely sequenced and analysed; 44 predicted open reading frames (ORFs) were recognized. The 53.8 kb sap locus contained eight complete and one partial sapA homologues, varying from 2769 to 3879 bp, sharing conserved 553-2622 bp 5' regions, with partial sharing of 5' and 3' non-coding regions. All eight sapA homologues were expressed in Escherichia coli as antigenic proteins and reattached to the surface of SLP- strain 23B, indicating their conserved function. Analysis of the sap homologues indicated three phylogenetic groups. Promoter-specific polymerase chain reactions (PCRs) and sapA homologue-specific reverse transcription (RT)-PCRs showed that the unique sapA promoter can potentially express all eight sapA homologues. Reciprocal DNA recombination based on the 5' conserved regions can involve each of the eight sapA homologues, with frequencies from 10(-1) to 10(-3). Intragenic recombination between sapA7 and sapAp8, mediated by their conserved regions with a 10(-1)-10(-2) frequency, allows the formation of new sap homologues. As divergent SLP C-termini possess multiple antigenic sites, their reciprocal recombination behind the unique sap promoter leads to continuing antigenic variation.

Phenotypic and molecular characterization of bovine Campylobacter fetus strains isolated in Brazil.

The objective of the present study was to characterize the phenotypic and molecular aspects of Campylobacter fetus strains isolated from bovine herds with reproductive problems. Thirty-one Brazilian field isolates, together with one reference strain of each subspecies of C. fetus, were analyzed. The strains were submitted to phenotypic identification followed by subspecies characterization using the polymerase chain reaction (PCR) and numeric evaluation of restriction fragment length polymorphism (RFLP) separated by pulsed-field gel electrophoresis (PFGE). Phenotypically, 4 isolates (12.1%) were classified as C. fetus subsp. fetus, and 29 isolates (87.9%) were classified as C. fetus subsp. venerealis. However, according to molecular analysis, only 1 isolate (3.0%) was classified as C. fetus subsp. fetus (the reference strain), whereas 32 isolates (97.0%) were considered C. fetus subsp. venerealis. SalI digestion of C. fetus genomic DNA, obtained from the 33 strains, yielded 7-10 DNA fragments ranging in size from 40 to 373kb, with 12 distinct patterns. Furthermore, the numeric analysis by neighbor-joining of the DNA from the 33 strains resulted in a dendrogram in which 2 distinct groups were identified. It was concluded that phenotypic characterization of C. fetus subspecies might lead to erroneous classification of field isolates. Although RFLP-PFGE is a powerful and reliable technique to characterize C. fetus, it has the inconvenience of being time consuming and laborious. Whereas PCR, besides providing rapid results, was found to be reliable and convenient for the characterization of field isolates of C. fetus.

Comparative study using amplified fragment length polymorphism fingerprinting, PCR genotyping, and phenotyping to differentiate Campylobacter fetus strains isolated from animals.

A collection of Campylobacter fetus strains, including both C. fetus subsp. fetus and C. fetus subsp. venerealis, were phenotypically identified to the subspecies level and genotypically typed by PCR and amplified fragment length polymorphism (AFLP) analysis. Phenotypic subspecies determination methods were unreliable. Genotyping of the strains by PCR and AFLP showed a clear discrimination between the two subspecies.

Campylobacter fetus sap inversion occurs in the absence of RecA function.

Phase variation of Campylobacter fetus surface layer proteins (SLPs) occurs by inversion of a 6.2-kb DNA segment containing the unique sap promoter, permitting expression of a single SLP-encoding gene. Previous work has shown that the C. fetus sap inversion system is RecA dependent. When we challenged a pregnant ewe with a recA mutant of wild-type C. fetus (strain 97-211) that expressed the 97-kDa SLP, 15 of the 16 ovine-passaged isolates expressed the 97-kDa protein. However, one strain (97-209) expressed a 127-kDa SLP, suggesting that chromosomal rearrangement may have occurred to enable SLP switching. Lack of RecA function in strains 97-211 and 97-209 was confirmed by their sensitivity to the DNA-damaging agent methyl methanesulfonate. Southern hybridization and PCR of these strains indicated that the aphA insertion into recA was stably present. However, Southern hybridizations demonstrated that in strain 97-209 inversion had occurred in the sap locus. PCR data confirmed inversion of the 6.2-kb DNA element and indicated that in these recA mutants the sap inversion frequency is reduced by 2 to 3 log(10) units compared to that in the wild type. Thus, although the major sap inversion pathway in C. fetus is RecA dependent, alternative lower-frequency, RecA-independent inversion mechanisms exist.

[Study on the physicochemical properties of Campylobacter jejuni enterotoxin].

Precipitate of Campylobacter jejuni cytotonic enterotoxin(CE) performed in an 80% saturated solution of ammonium sulfateit indicated that there were some little molecular proteins except the 68 kD main band on sodium dodecyl sulfate-polyacrylamide gel electrophoresis(SDS-PAGE), whereas the eluate from GM1 ganglioside affinity column chromatography exhibited only one 68 kD band on SDS-PAGE. The results suggest that CE mainly be consisted of 68 kD protein. The toxin is heat-labile, pH dependent and resistant to trypsin, It could be completely inactivated by heating at either 56 degrees C and 60 degrees C for 30 min or 100 degrees C for 15 min. The activity was maximum at pH 6.0 and was completely inactivate at pH 3.0 and pH 9.0, and rapidly reduced after storage over 3 d at 4 degrees C. The anti-LT serum could completely inhibited the activity of CE.

Campylobacter fetus surface layer proteins are transported by a type I secretion system.

The virulence of Campylobacter fetus, a bacterial pathogen of ungulates and humans, is mediated in part by the presence of a paracrystalline surface layer (S-layer) that confers serum resistance. The subunits of the S-layer are S-layer proteins (SLPs) that are secreted in the absence of an N-terminal signal sequence and attach to either type A or B C. fetus lipopolysaccharide in a serospecific manner. Antigenic variation of multiple SLPs (encoded by sapA homologs) of type A strain 23D occurs by inversion of a promoter-containing DNA element flanked by two sapA homologs. Cloning and sequencing of the entire 6.2-kb invertible region from C. fetus 23D revealed a probable 5.6-kb operon of four overlapping genes (sapCDEF, with sizes of 1,035, 1,752, 1,284, and 1,302 bp, respectively) transcribed in the opposite direction from sapA. The four genes also were present in the invertible region of type B strain 84-107 and were virtually identical to their counterparts in the type A strain. Although SapC had no database homologies, SapD, SapE, and SapF had predicted amino acid homologies with type I protein secretion systems (typified by Escherichia coli HlyBD/TolC or Erwinia chrysanthemi PrtDEF) that utilize C-terminal secretion signals to mediate the secretion of hemolysins, leukotoxins, or proteases from other bacterial species. Analysis of the C termini of four C. fetus SLPs revealed conserved structures that are potential secretion signals. A C. fetus sapD mutant neither produced nor secreted SLPs. E. coli expressing C. fetus sapA and sapCDEF secreted SapA, indicating that the sapCDEF genes are sufficient for SLP secretion. C. fetus SLPs therefore are transported to the cell surface by a type I secretion system.

Formation of cytotoxins by enteric Campylobacter in humans and animals.

Campylobacter (C.) jejuni from persons suffering from diarrhoea, from organs of poultry, C. jejuni and C. fetus ssp. fetus from the gastrointestinal tract of calves and adult cattle as well as a number of reference strains were examined for cytotoxin formation in a CHO-K1 cell culture test. During evaluation, three morphologically different pictures were observed. The first cytotoxin caused a formation of strikingly large, rounded or polymorphic and elongated cells which was associated with reduced growth. The progressive morphological changes corresponded to those described for the Cytolethal Distending Toxin (CLDT) and were assigned to it. The second cytotoxin produced a rounding of cells without a change in their size while at the same time, growth was reduced. In analogy to CLDT, this toxin was termed Cytolethal Rounding Toxin (CLRT). A third morphological picture consisted of cell changes characterized by enlarged polymorphic as well as by small rounded cells. These cell changes were considered as being distinct from the above mentioned ones and referred to as CLTD/CLRT effect. In none of the 39 Campylobacter strains isolated from humans and calves with diarrhoea, a noteworthy cytotonic activity could be detected that would indicate the presence of an enterotoxin.

Detection and characterization of two cytotoxins produced by Campylobacter jejuni strains.

Campylobacter jejuni strains are able to produce at least two different cytotoxins called "cytolethal distending toxin" (CLDT) and "cytolethal rounding toxin" (CLRT). In this study, we investigated the corresponding changes in CHO-K1 cells using the cell counter and analyzer system CASY 1. Determination of the cell volume after toxin treatment of the cells is a useful criterion for differentiation between the cytotoxic activities produced by Campylobacter strains. Incubation of the cells with crude CLDT resulted in a decrease in the cell count combined with a dramatic increase of the mean cell volume in comparison to the control culture. A decrease in the cell count was also seen as a response to CLRT preparations, while this toxin had no effect on the mean cell volume determined. It was shown that only CLDT caused histone-associated DNA fragments in the cytoplasm of CHO-K1 cells indicating an apoptotic pathway of cell death. In addition, the polymerase chain reaction (PCR) was employed to screen Campylobacter strains for the presence of the cdtB gene sequence, which was detectable in all strains investigated.

Biological characterization of Campylobacter fetus lipopolysaccharides.

Lipopolysaccharides (LPS) of three strains of Campylobacter fetus (subspp. fetus and venerealis, and serotypes A and B), a bacterium of veterinary importance but also a cause of various infections in humans, were assessed for their ability to induce mitogenicity, gelation of Limulus amebocyte lysate, lethal toxicity in mice, and pyrogenicity in rabbits. All C. fetus LPS exhibited activities lower than those of Salmonella typhimurium LPS. LPS of C. fetus subsp. fetus serotype A had the lowest activity in all assays. Since the majority of C. fetus subsp. fetus isolates from humans are serotype A, the lower biological activities of this LPS may aid the pathogenesis of such strains. The lower activities of C. fetus LPS compared with those of S. typhimurium LPS may reflect the presence of longer fatty acid chains in the lipid A of C. fetus LPS, whereas interstrain differences in C. fetus LPS bioactivities may be related to some property influenced by composition of the saccharide moiety.

Characterization of a post-translational modification of Campylobacter flagellin: identification of a sero-specific glycosyl moiety.

The flagellins of Campylobacter spp. differ antigenically. In variants of C. coli strain VC167, two antigenic flagellin types determined by sero-specific antibodies have been described (termed T1 and T2). Post-translational modification has been suggested to be responsible for T1 and T2 epitopes, and, using mild periodate treatment and biotin hydrazide labelling, flagellin from both VC167-T1 and T2 were shown to be glycosylated. Glycosylation was also shown to be present on other Campylobacter flagellins. The ability to label all Campylobacter flagellins examined with the lectin LFA demonstrated the presence of a terminal sialic acid moiety. Furthermore, mild periodate treatment of the flagellins of VC167 eliminated reactivity with T1 and T2 specific antibodies LAH1 and LAH2, respectively, and LFA could also compete with LAH1 and LAH2 antibodies for binding to their respective flagellins. These data implicate terminal sialic acid as part of the LAH strain-specific epitopes. However, using mutants in genes affecting LAH serorecognition of flagellin it was demonstrated that sialic acid alone is not the LAH epitope. Rather, the epitope(s) is complex, probably involving multiple glycosyl and/or amino acid residues.

Shift in S-layer protein expression responsible for antigenic variation in Campylobacter fetus.

Campylobacter fetus strains possess regular paracrystalline surface layers (S-layers) composed of high-molecular-weight proteins and can change the size and crystalline structure of the predominant protein expressed. Polyclonal antisera demonstrate antigenic cross-reactivity among these proteins but suggest differences in epitopes. Monoclonal antibodies to the 97-kDa S-layer protein of Campylobacter fetus subsp. fetus strain 82-40LP showed three different reactivities. Monoclonal antibody 1D1 recognized 97-kDa S-layer proteins from all C. fetus strains studied; reactivity of monoclonal antibody 6E4 was similar except for epitopes in S-layer proteins from reptile strains and strains with type B lipopolysaccharide. Monoclonal antibody 2E11 only recognized epitopes on S-layer proteins from strains with type A lipopolysaccharide regardless of size. In vitro shift from a 97-kDa S-layer protein to a 127-kDa S-layer protein resulted in different reactivity, indicating that size change was accompanied by antigenic variation. To examine in vivo variation, heifers were genetically challenged with Campylobacter fetus subsp. venerealis strains and the S-layer proteins from sequential isolates were characterized. Analysis with monoclonal antibodies showed that antigenic reactivities of the S-layer proteins were varied, indicating that these proteins represent a system for antigenic variation.

Production of cytolethal distending toxin (CLDT) by Campylobacter fetus subsp. fetus isolated from calves.

Cytolethal distending toxin (CLDT) production by Campylobacter fetus subsp. fetus isolated from calves was examined using CHO cells. Twenty-five of the 26 strains tested were positive for CLDT with titer of ranging 1:8 to 1:2,048. CLDT positive strains were divided into low and high cytotoxin titer groups. Isolates from diarrhea cases tended to produce a significant amount of CLDT compared with isolates from liver. CLDT produced by C. fetus showed no effect on Y-1 cells and was heat-labile and trypsin-sensitive.

Pathogenesis of Campylobacter fetus infections: critical role of high-molecular-weight S-layer proteins in virulence.

Wild-type Campylobacter fetus strains possess high-molecular-weight S-layer proteins (S+) and are highly resistant to serum-mediated killing and phagocytosis. Spontaneous mutant strains lacking these proteins (S-) are serum and phagocytosis sensitive and have reduced virulence in a mouse model. Intact S+ cells were treated with pronase, which made them S- although genotypically S+ and had essentially no effect on other cellular proteins or on viability. Treatment with pronase, but not buffer alone, rendered these cells serum and phagocytosis sensitive and reduced mouse virulence to the level observed for the S- mutant cells. In related studies, purified S-layer proteins diminished neutrophil chemoluminescent responses to a heterologous particulate antigen. Finally, passive administration of antiserum to the 97-kDa S-layer protein partially protected mice against lethal challenge with the S+ strain. These studies define the contribution of the S-layer proteins to C. fetus virulence.

Reattachment of surface array proteins to Campylobacter fetus cells.

Campylobacter fetus strains may be of serotype A or B, a property associated with lipopolysaccharide (LPS) structure. Wild-type C. fetus strains contain surface array proteins (S-layer proteins) that may be extracted in water and that are critical for virulence. To explore the relationship of S-layer proteins to other surface components, we reattached S-layer proteins onto S- template cells generated by spontaneous mutation or by serial extractions of S+ cells with water. Reattachment occurred in the presence of divalent (Ba2+, Ca2+, Co2+, and Mg2+) but not monovalent (H+, NH4+, Na+, K+) or trivalent (Fe3+) cations. The 98-, 125-, 127-, and 149-kDa S-layer proteins isolated from strains containing type A LPS (type A S-layer protein) all reattached to S- template cells containing type A LPS (type A cells) but not to type B cells. The 98-kDa type B S-layer protein reattached to SAP- type B cells but not to type A cells. Recombinant 98-kDa type A S-layer protein and its truncated amino-terminal 65- and 50-kDa segments expressed in Escherichia coli retained the full and specific determinants for attachment. S-layer protein and purified homologous but not heterologous LPS in the presence of calcium produced insoluble complexes. By quantitative enzyme-linked immunosorbent assay, the S-layer protein copy number per C. fetus cell was determined to be approximately 10(5). In conclusion, C. fetus cells are encapsulated by a large number of S-layer protein molecules which may be specifically attached through the N-terminal half of the molecule to LPS in the presence of divalent cations.

Surface array proteins of Campylobacter fetus block lectin-mediated binding to type A lipopolysaccharide.

Campylobacter fetus strains with type A lipopolysaccharide (LPS) and a surface array protein layer (S+) have been found to be pathogenic in humans and animals. Spontaneous laboratory mutants that lack surface array proteins (S-) are sensitive to the bactericidal activity of normal human serum. The ability of lectins to determine the presence of the S-layer and differentiate LPS type was assessed. We screened 14 lectins and found 3 (wheat germ agglutinin, Bandeiraea simplicifolia II, and Helix pomatia agglutinin) that agglutinated S- C. fetus strains with type A LPS but not S- strains with type B or type C LPS or S+ strains. However, the S+ type A strains were agglutinated after sequential water extraction, heat, or pronase treatment, all of which remove the S-layer, whereas there was no effect on the control strains. Specific carbohydrates for each lectin and purified LPS from a type A C. fetus strain specifically inhibited agglutination of an S- type A strain. In a direct enzyme-linked lectin assay, binding to the S- type A LPS strain was significantly greater than binding to the S+ strain (P = 0.01) or to a Campylobacter jejuni strain (P = 0.008). Consequently, these results indicate that the three lectins bind to the O side chains of C. fetus type A LPS but that the presence of the S-layer on intact cells blocks binding.

Identification of Campylobacter jejuni surface proteins that bind to Eucaryotic cells in vitro.

To understand the role of Campylobacter jejuni surface proteins in the interaction of C. jejuni with cultured mammalian cell lines in vitro, we developed a ligand-binding assay. This procedure allowed us to antigenically identify C. jejuni outer membrane proteins (OMPs) that attach to intact host cell membranes. OMPs isolated from an invasive strain and a less invasive strain were antigenically indistinguishable. However, we found that proteins with molecular masses of 28 and 32 kilodaltons (kDa) from just the invasive strain bound to HEp-2 cell monolayers. Binding of the 32-kDa OMP was cell line specific and correlated directly with the ability of the invasive C. jejuni strain to penetrate. Such a correlation was probably also true for the 28-kDa OMP. We also investigated the binding of glycine acid extracts with cell line HEp-2. We identified four proteins with apparent molecular masses of 28, 32, 36, and 42 kDa in the invasive strain extracts that bound to HEp-2 cells. In contrast, only the 36-kDa protein from the less invasive strain bound to HEp-2 cells. Our data suggest that binding of these surface exposed proteins may play a key role in C. jejuni-host cell interactions and ultimate invasion.

Isolation, characterization, and host-cell-binding properties of a cytotoxin from Campylobacter jejuni.

A 68,000-molecular-weight protein was isolated by polyacrylamide gel electrophoresis from the organism-free filtrate of a fully virulent clinical strain of Campylobacter jejuni. The eluted protein was heat labile, was inactivated at either pH 3.0 or 9.0, was sensitive to trypsin, and was lethal for fertile chicken eggs. It also had toxic effects on chicken embryo fibroblast, Chinese hamster ovary (CHO), and intestinal 407 (Int407) cells. A monoclonal antibody (CETPMAb4) raised to this eluted toxic protein (ETP) from C. jejuni abolished these toxic activities. Homology between C. jejuni ETP and Vibrio cholerae toxin was not observed in that specific antisera to each did not block their respective toxic activities. In enzyme-linked immunosorbent assays, ETP, unlike chlorea enterotoxin, did not bind to GM1 ganglioside. Furthermore, the C. jejuni toxin had cytotoxinlike properties and induced rounding of CHO cells. Binding of ETP to Int407 and primary chicken embryo fibroblast cells was maximal after 2 h as assessed by enzyme-linked immunosorbent assay, and this toxin adherence to host cell membranes was significantly reduced by prior treatment of the cells with proteolytic enzymes, neuraminidase, or glutaraldehyde but not by treatment with beta-galactosidase, lipase, Nonidet P-40, or sodium metaperiodate. In competitive binding assays, sugars, lectins, or GM1 ganglioside did not adversely influence uptake of ETP by these cells. These results suggest that the ETP produced by C. jejuni is a cytotoxin which binds to Int407 cells via a protein- or glycoproteinlike receptor on cell membranes and possesses properties dissimilar to those of V. cholerae toxin.