Search

Campylobacter fetus surface layer proteins (SLPs), encoded by sapA homologues, are important in virulence. In wild-type C. fetus strain 23D, all eight sapA homologues are located in the 54-kb sap island, and SLP expression reflects the position of a unique sapA promoter in relation to the sapA homologues. The extensive homologies in the sap island include both direct and inverted repeats, which allow DNA rearrangements, deletion, or duplication; these elements confer substantial potential for genomic plasticity. To better understand C. fetus sap island diversity and variation mechanisms, we investigated the organization and distribution of sapA homologues among 18 C. fetus strains of different subspecies, serotypes, and origins. For all type A strains, the boundaries of the sap island were relatively consistent. A 187-bp noncoding DNA insertion near the upstream boundary of the sap island was found in two of three reptile strains studied. The sapA homologue profiles were strain specific, and six new sapA homologues were recognized. Several homologues from reptile strains are remarkably conserved in relation to their corresponding mammalian homologues. In total, the observed differences suggest that the sap island has evolved differing genotypes that are plastic, perhaps enabling colonization of varied niches, in addition to antigenic variation.

The role of the surface (S)-layer proteins of Campylobacter fetus subsp. fetus has been investigated using an ovine model of abortion. Wild-type strain 23D induced abortion in up to 90% of pregnant ewes challenged subcutaneously. Isolates recovered from both dams and fetuses expressed S-layer proteins with variable molecular masses. The spontaneous S-layer-negative variant, strain 23B, neither colonized nor caused abortions in pregnant ewes. A series of isogenic sapA and recA mutants, derived from 23D, also were investigated in this model. A mutant (501 [sapA recA(+)]) caused abortion in one of five challenged animals and was recovered from the placenta of a second animal. Another mutant (502 [sapA recA]) with no S-layer protein expression caused no colonization or abortions in challenged animals but caused abortion when administered intraplacentally. Mutants 600(2) and 600(4), both recA, had fixed expression of 97- and 127-kDa S-layer proteins, respectively. Two of the six animals challenged with mutant 600(4) were colonized, but there were no abortions. As expected, all five strains recovered expressed a 127-kDa S-layer protein. In contrast, mutant 600(2) was recovered from the placentas of all five challenged animals and caused abortion in two. Unexpectedly, one of the 16 isolates expressed a 127-kDa rather than a 97-kDa S-layer protein. Thus, these studies indicate that S-layer proteins appear essential for colonization and/or translocation to the placenta but are not required to mediate fetal injury and that S-layer variation may occur in a recA strain.

Two variants of Campylobacter jejuni IN1 differing in the expression of flagella, IN1 (Fla+ Mot+, wild type) and IN1-NM (Fla- Mot-), were tested for their ability to establish infection in adult hamsters. Animals were challenged intracecally with 2 X 10(9) to 5 X 10(10) CFU and monitored for evidence of infection. None of the challenged animals developed illness. There was a significant difference, however, in the ability of IN1 to infect hamsters (35 of 43) compared with that of IN1-NM (1 of 42) (P less than 0.01). Additionally, eight animals challenged with IN1-NM excreted only the campylobacters of Fla+ phenotype, which were shown to be identical with the parental Fla+ Mot+ type. Both groups of animals developed serum immunoglobulin G antibodies to outer membrane proteins of IN1 as measured by an enzyme-linked immunosorbent assay; however, only animals that excreted Fla+ organisms developed antiflagellin antibodies. In vitro reversibility from Fla+ to Fla- occurred with a frequency of 9.2 X 10(-6) per cell per generation; however, reversion from Fla- to Fla+ could not be detected in vitro. Further characterization of IN1-NM showed that it produced a cytoplasmic 62K flagellin subunit protein, but this protein lacked six epitopes detected in IN1 and also differed in its two-dimensional gel electrophoresis pattern. The results of these studies support the concept that an intact flagellum is necessary for intestinal infection with C. jejuni and that this model may be useful for studying the immune response to C. jejuni.

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.

Eleven-day-old chicken embryos were used to compare the relative virulence of minimally passaged human isolates of Campylobacter jejuni and Campylobacter coli. Graded doses of bacteria were inoculated onto the chorioallantoic membrane, and 50% lethal doses were calculated at 72 h postinfection. Strains varied markedly in their ability to invade the chorioallantoic membrane and kill the embryos. The 50% lethal doses varied by about 6 logs for 25 strains of C. jejuni, and by 2 logs for 5 strains of C. coli. Although both outbred and inbred embryos were employed in the study, the latter were found to be more susceptible to infection with most strains. All isolates were screened for plasmid DNA, but there was no apparent relationship between plasmid content and virulence of strains for the embryos. Neither could virulence be associated with the production of siderophores by the strains. The ability of selected strains of C. jejuni to invade the liver of embryos was also studied. The number of campylobacters culturable from the liver was found to be inversely related to the 50% lethal dose of the strain. By inoculating 11-day-old embryos intravenously, it was possible to demonstrate that a strain of C. jejuni which was poorly virulent after chorioallantoic inoculation was relatively noninvasive. Invasiveness alone, however, could not fully account for the lethality of two highly virulent strains of C. jejuni administered by the intravenous route. Finally, there was no correlation between motility and virulence in this model system.

Adult athymic (nu/nu) and euthymic (+/nu) germfree BALB/c mice were orally challenged with pure cultures of Campylobacter jejuni (human clinical fecal strains) and a human blood isolate of Campylobacter fetus subsp. fetus. After a period of adaptation to the mouse intestinal tract, all three C. jejuni strains caused disease in gnotobiotic mice. Mouse-adapted, weakly cytotoxic C. jejuni 45100 consistently induced disease symptoms (transient diarrhea, cecal shrinkage, and acute inflammatory changes with eosinophilia in the lower intestinal mucosa) in nu/nu mice 7 to 9 days after oral challenges. Conversely, no overt disease or histopathology was evident in +/nu mice challenged with the same strain (45100). After periods of adaptation in the murine alimentary tract, the two C. jejuni strains, 24 and INN 73-83, with greater cytotoxin-producing capacities, decreased cecal size and caused minor mucosal inflammatory changes in both nu/nu and +/nu BALB/c mice 1 to 2 weeks after intestinal colonization. A transient splenomegaly was also evident at 1 to 2 weeks after germfree nu/nu mice were colonized with each of the three C. jejuni strains used in this study. Occult blood was observed in a small percentage (approximately 11%) of nu/nu and +/nu BALB/c mice that were colonized with C. jejuni strains 45100 and INN 73-83. C. fetus subsp. fetus 255 colonized the alimentary tract of gnotobiotic mice, but neither morbidity nor mortality was evident. The disease we observed in the gnotobiotic mice, along with the histological changes in the intestinal tract after oral challenges, resembles symptoms of campylobacteriosis in humans. The gnotobiotic BALB/c mouse model of Campylobacter disease provides a unique opportunity to detail basic aspects of the acute and chronic pathogenesis of and immunity to this recently recognized disease.

Campylobacter jejuni is able to colonize the human intestinal mucosa and cause disease. For this reason, it was important to investigate mechanisms by which C. jejuni adheres to epithelial cells and intestinal mucus gel. All strains of C. jejuni used were able to adhere to INT 407 epithelial cells and mucus, but high adherence to one substrate did not necessarily indicate comparable adherence to the other. The adherence of C. jejuni to cells was inhibited partially by treating the bacterial cells with proteases or glutaraldehyde or by adding a certain carbohydrate (fucose or mannose) to the medium. The flagellum of C. jejuni was identified as a potential adhesin by comparing adherence of flagellated and aflagellated variants. Shearing of the bacterial cells to remove the flagella reduced bacterial adhesion, whereas immobilization of the flagellum with KCN increased adhesion. Purified flagella showed specific, fucose-resistant binding to epithelial cells but not to intestinal mucus. The presence of a second, nonproteinaceous adhesin was suggested because no single treatment of the bacteria completely inhibited adhesion. Lipopolysaccharide (LPS) was identified as another C. jejuni adhesin. [3H]LPS specifically bound to epithelial cells, and this phenomenon was inhibited by periodate oxidation of the LPS or glutaraldehyde fixation of the epithelial cells. LPS, unlike flagella, was fucose sensitive and inhibited binding of whole bacterial cells to INT 407 cells. LPS was also able to bind to intestinal mucus gel. These data indicate that both flagella and LPS are important in adhesion to the mucosal surface.

Human clinical strains of Campylobacter jejuni and Campylobacter fetus subsp. fetus colonized the gastrointestinal tracts of both athymic (nu/nu) and euthymic (+/nu) germfree mice (BALB/c). Viable Campylobacter spp. (10(9) to 10(10) CFU/g [dry weight] of cecum and colon contents) were isolated on day 3 after oral challenge, and similar large numbers of viable cells were evident at several intervals during a 10-month experiment. The stomachs and upper small intestines of nu/nu and +/nu mice that were monoassociated for 224 days with C. jejuni 45100 contained 3 to 4 logs fewer viable bacteria than did their ceca or colons. Athymic mice that were monoassociated for 224 days with C. fetus subsp. fetus had 2 to 3 logs more viable Campylobacter spp. in their upper gastrointestinal tracts than did their +/nu littermates. Large viable populations (approximately 10(9)/g of contents) of C. fetus subsp. fetus were in the ceca and colons of both nu/nu and +/nu mice. All C. jejuni strains used in this study chronically infected the mesenteric lymph nodes of both nu/nu and +/nu mice. C. jejuni strains 24 and INN 73-83, which were cytotoxic for Chinese hamster ovary cells in vitro, were also more frequently isolated from the livers, spleens, and kidneys of nu/nu mice than was the weak cytotoxin-producing strain 45100. Additionally, heat-labile-enterotoxin-producing C. jejuni INN 73-83 was recovered more frequently from the internal organs of monoassociated +/nu mice than were any other Campylobacter spp. tested. Natural gastrointestinal colonization of neonatal nu/nu and +/nu mice (born to Campylobacter-colonized mothers) with Campylobacter spp. appeared to be delayed until approximately 1 to 2 weeks after birth. Conventionalization of C. jejuni 45100-monoassociated BALB/c mice with a complex mouse fecal microflora eliminated viable C. jejuni from the mesenteric lymph nodes by day 14 and from the cecum by day 78. These findings show that the gnotobiotic BALB/c mouse is a new model for studying acute and chronic host-Campylobacter sp. interactions.

Three human isolates of Campylobacter jejuni were grown in a biphasic culture medium with and without the addition of a synthetic chelator to induce iron limitation. Cells grown in low-iron medium exhibited slower growth rates and altered cellular morphology. Increased numbers of longer, more filamentous forms were seen in Gram-stained smears. Three proteins, with apparent Mrs of 82,000, 76,000, and 74,000, were consistently present in the outer membrane of cells grown in low-iron medium. At least one of these proteins (76,000 to 74,000) was exposed on the cell surface. A bioassay was used to look for the production of siderophores by these and other strains of C. jejuni. Seven of 26 strains tested produced detectable amounts of siderophores. Growing strains at 42 degrees C failed to suppress siderophore synthesis or to alter the outer membrane protein profiles of iron-starved cells. The ability of three strains to utilize exogenously supplied siderophores for growth in low-iron medium was also examined. All three strains were able to utilize enterochelin and ferrichrome, but none utilized aerobactin, rhodotorulic acid, or desferrioxamine B. The effect of iron on the virulence of C. jejuni for 11-day-old chicken embryos inoculated via the chorioallantoic membrane was also determined.

We studied the antigenicity of a wild-type flagellate and motile (F+M+) Campylobacter jejuni strain (81116) and two daughter mutants, one flagellate and immotile (F+M-) and one aflagellate and immotile (F-M-). By sodium dodecyl sulfate-polyacrylamide gel electrophoresis of acid-extracted surface proteins, a 63-kilodalton (kDa) band identified from sheared flagella as the flagellar protein was present in the F+M+ and F+M- strains but not in the F-M- strain. No other differences in protein profile among the three strains were noted. By Western blotting, serum from rabbits immunized with either the F+M+ or F-M- strain detected a 63-kDa protein in the F+M+ and F+M- strains but not in the F-M- strain. That the F-M- antiserum recognized the 63-kDa band suggests that small amounts of this protein or a cross-reacting antigen is present on the F-M- strain. By counterimmunoelectrophoresis of the acid-extracted preparations with immune sera, all three strains were found to share three major antigens, but a fourth antigen with a net positive charge was present only in the F+M+ and F+M- strains. Antisera to five C. jejuni and two Campylobacter fetus strains recognized the 63-kDa protein of purified F+M+ flagella in Western blots, demonstrating a common antigen is present, but enzyme-linked immunosorbent assay results suggest that the sharing of this antigen among Campylobacter strains is variable.

Twenty-one isolates were tested for their ability to adhere to and invade HEp-2 cells in vitro. Of the 21 organisms tested, 2 did not invade the HEp-2 cells, and 1 of these did not adhere to the epithelial cells. Campylobacter jejuni clinical isolates were more invasive than the nonclinical strains that were tested. When HEp-2 cells were treated with cytochalasin B, the invasiveness of C. jejuni was reduced, indicating active participation of the host cell in the uptake of these organisms. The number of intracellular C. jejuni isolates decreased when Campylobacter whole-cell lysates were absorbed onto HEp-2 cell monolayers. Experiments were also conducted to identify the functional sites of the antigens responsible for expression of Campylobacter invasion. Oxidation of lysates with sodium meta-periodate significantly affected its inhibitory capacity. This implies that the Campylobacter invasive ligand appears to be dependent upon an intact carbohydrate moiety.

To study the pathogenic mechanisms of Campylobacter jejuni infection, 36- to 72-h-old chickens were fed 10(3) to 10(6) live cells, using strains isolated from 40 patients with watery diarrhea and 6 with bloody mucoid diarrhea from whom no other known enteropathogen was detected. Chickens of Starbro strain were more likely to develop C. jejuni-induced diarrhea than were White Leghorn chickens. Diarrhea was defined on the basis of amounts of gut fluid in 288 chicks fed with live C. jejuni versus 183 saline-fed control as an accumulation greater than or equal to 0.4 ml of fluid in the guts (excluding ceca) of chickens. Twenty-five percent of the chickens developed diarrhea on day 2, 49% on day 4, and 81% on day 5. The intestines, including ceca, were distended with watery fluid. The majority of the strains, irrespective of whether they were isolated from watery or bloody mucoid enteritis patients, caused watery diarrhea in chickens, and a few caused mucoid diarrhea. No correlation was observed between the source of a strain and the outcome in the experimental model. Bloody diarrhea was never observed in chickens. The peak incidence of diarrhea on day 5 coincided with the mean of maximum fluid accumulation. The organisms multiplied by 3 to 4 logs in all parts of the intestine, with a steady increase in number until day 5. Systemic invasion occurred frequently: C. jejuni could be recovered from the spleen in 47% of the chickens on day 5, in 25% from the liver on day 6, and in 11% from heart blood on day 4. Histopathological examination of gut tissue of the chickens having watery diarrhea did not reveal any abnormality except slight submucosal edema. However, in chickens with mucoid diarrhea, the organisms were found to adhere to brush borders and penetrate into the epithelial cells with formation of a breach in continuity of the brush border lining. The electrolyte composition of the intestinal fluid from chickens infected with C. jejuni and from saline-fed controls did not show significant differences, except for depletion of K+ in the test group. The results obtained in this highly reproducible chicken diarrhea model indicate that (i) most chickens develop nonexudative watery diarrhea 2 to 5 days after oral feeding of 10(3) to 10(6) live cells of C. jejuni; (ii) the organism multiples in all parts of a chicken intestine, (iii) systemic invasion is common, and (iv) local invasion is sometimes observed.

Fourteen isolates of Campylobacter jejuni of different serotypes and one Campylobacter coli isolate, from various human and animal sources, were tested for potential pathogenic mechanisms. Enterotoxin production was not detected in the infant mouse test or by calf and piglet ligated intestinal loop studies. Isolates were not invasive by the Sereny test. All isolates associated with and penetrated HeLa cells, although both actions occurred generally in a minor way under the conditions of our study. The C. coli isolate showed extensive HeLa cell association, but three other C. coli isolated tested did not. None of the 15 isolates produced diarrhea or death in 3-day-old chickens inoculated orally and observed for 3 days, nor did they consistently produce diarrhea and death in 9- to 10-day-old infant mice over a 3-day period after oral inoculation. Diarrheal disease and mortality were not observed when 3-day-old gnotobiotic chickens were infected with one of five isolates and observed over a 2-week period.

Enteroinvasive Salmonella, Shigella, and Escherichia coli strains were found to exert an effect which rendered Campylobacter jejuni capable of intracellular localization in epithelial cells in vitro. When monolayers of HEp-2 or A-549 cells were challenged with pure cultures of C. jejuni or Campylobacter coli, none of the eight strains tested invaded the cells. In contrast, four of these strains were able to localize intracellularly when the cells were challenged with a mixture of campylobacters and enteroinvasive Salmonella typhimurium, Shigella flexneri, Shigella boydii, Shigella sonnei, or E. coli strains. Invasiveness of campylobacters was also induced by one nonenteroinvasive strain of E. coli O124. Coinfection with other nonenteroinvasive E. coli strains did not induce invasiveness in C. jejuni. The degree of internalization induced by S. typhimurium was significantly higher than that induced by Shigella or E. coli strains. The invasive capacity of C. jejuni was found to differ considerably between strains. No evidence of an invasive potential was demonstrable for two C. coli strains or for two enterotoxigenic isolates of C. jejuni examined. C. jejuni was only able to localize intracellularly in cell cultures when the interaction occurred in a microaerobic atmosphere. None of the strains tested evoked keratoconjunctivitis in guinea pig eyes (Sereny test), regardless of the presence of coinfectants. The results indicate that a synergistic interaction that exists between C. jejuni and other enteropathogens facilitates invasion by C. jejuni.

Human isolates of the intestinal pathogen Campylobacter jejuni have been shown to colonize mucus on the outer surface and deep within the intestinal crypts of gnotobiotic or germfree mice. The cecal crypts are preferentially colonized. A model of mucus colonization by C. jejuni in the mouse cecum has been developed, using antibiotic- and magnesium sulfate-treated specific-pathogen-free animals. These spiral-shaped bacteria colonize the mucus in a similar manner to the normal spiral-shaped microbiota. No evidence of adhesion to the intestinal surface was found with a wide variety of microscopic techniques. The campylobacters were seen to be highly motile in living preparations of gut tissue and rapidly tracked along intestinal mucus. Just as many of the normal spiral-shaped bacteria of intestinal surfaces can achieve close association with the epithelium through mucus association and do not adhere to the surface, C. jejuni colonizes the intestinal mucosa via mucus colonization. Thus, a major determinant of pathogenicity in intestinal infection with C. jejuni is proposed to be an ability to colonize intestinal mucus. The possession of specific adhesins is unlikely to be a significant determinant of pathogenicity. Better understanding of the mechanism of mucus association and the properties of the bacterium that are responsible will provide a basis for the rational selection of preventative measures. The model of mucus association in adult antibiotic-treated mice provides an opportunity for colonization studies with variant organisms and immunization studies.

The pathogenic properties of 20 strains of Campylobacter jejuni isolated from persons with clearly defined clinical manifestations were determined. Cell-free broth filtrates were examined for (i) enterotoxin production by Chinese hamster tissue culture assay and an enzyme-linked immunosorbent assay (ELISA) employing GM1 ganglioside and affinity-purified antiserum to Escherichia coli heat-labile toxin, (ii) cytotoxin production by Vero and HeLa cell tissue culture lines, and (iii) their ability to cause fluid secretion in rat ligated ileal loops. Viable bacteria were examined for invasive properties by an ELISA with the immunoglobulin fraction of antiserum to Formalin-killed bacteria of an invasive strain, and by their effect on fluid secretion and morphology in rat ligated ileal loops. None of the eight isolates obtained from asymptomatic carriers had any detectable pathogenic properties. All six strains isolated from persons with bloody invasive-type diarrhea elaborated a cytotoxin; their viable bacteria had high titers in the ELISA for invasive properties and caused fluid secretion in ligated ileal loops, although consistent morphologic abnormalities and evidence of mucosal invasion, examined by immunofluorescence techniques, were not detected. All six strains isolated from persons with watery secretory-type diarrhea produced an enterotoxin, one elaborated a cytotoxin, and broth filtrates of all strains caused fluid secretion in ligated ileal loops; viable bacteria had low titers in the ELISA for invasive properties and evoked fluid secretion in ligated loops by means of enterotoxin production. These observations show (i) that a correlation exists between the pathogenic properties of the infective C. jejuni strain and gastrointestinal manifestations in the infected host, and (ii) that these pathogenic properties can be identified by in vitro assays, including ELISAs.

Campylobacter jejuni 81116 and A3249 undergo a bidirectional transition between flagellated (Fla+) and aflagellated (Fla-) phenotypes. When measured in culture medium, the Fla+----Fla- transition occurred at a rate of 3.1 X 10(-3) to 5.9 X 10(-3) per cell per generation, and the Fla- to Fla+ transition occurred at a rate of 4.0 X 10(-7) to 8.0 X 10(-7) per cell per generation. However, passage through a rabbit intestine markedly favored the Fla+ phenotype.