943

Journal of Food Protection, Vol. 47, No. 12, Pages 943-949 (December 1984) Copyright®, International Association of Milk, Food, and Environmental Sanitarians

Campylobacter jejuni and Campylobacter coli Production of a Cytotonic Toxin Immunologically Similar to Cholera Toxin

BARBARA A. McCARDELL1*, JOSEPH M. MADDEN1 and EILEEN C. LEE2'3

Division of Microbiology, Food and Drug Administration, Washington, D.C. 20204, and Department of Biology, The Catholic University of America,Downloaded from http://meridian.allenpress.com/jfp/article-pdf/47/12/943/1655814/0362-028x-47_12_943.pdf by guest on 25 September 2021 Washington, D.C. 20064

(Received for publication September 6, 1983)

ABSTRACT monella typhimurium is related to CT (31), although its role in pathogenesis has not been determined. Production An enzyme-linked immunosorbent assay (ELISA) based on by some strains of Aeromonas species of a toxin which binding to cholera toxin (CT) antibody was used to screen cell- free supernatant fluids from 11 strains of Campylobacter jejuni can be partially neutralized by CT antiserum in rat loops and one strain of Campylobacter coli. Positive results for seven suggests some relationship to CT (21). of the eight clinical isolates as well as for one animal and one Although Campylobacter jejuni and Campylobacter food isolate suggested that these strains produced an extracellu­ coli have long been known as animal pathogens, only in lar factor immunologically similar to CT. An affinity column recent years have their importance and prevalence in (packed with Sepharose 4B conjugated to purified anti-CT IgG human disease been recognized (13,22). With the advent via cyanogen bromide) was used to separate the extracellular of improved methods (77), C. jejuni is now isolated from factor from cell-free supernatant fluids. Both unconcentrated human diarrheal stools more frequently in the United supernatant fluids and affinity-purified material caused rounding States than are Salmonella spp. and Shigella spp. (2,3). in a Y-l mouse adrenal cell assay, suggesting that the factor However, little is known about possible toxins or other was a cytotonic toxin. Rounding of Y-l cells caused by cell- free supernatant fluids, affinity-purified toxin or CT was neut­ factors responsible for the pathophysiology of Cam­ ralized by preincubation with CT or Campylobacter cytotonic pylobacter infections. This report describes the isolation toxin (CCT) antiserum. CCT and CT showed a reaction of par­ and partial characterization of a Campylobacter cytotonic tial identity by gel immunodiffusion, using IgG from CT anti­ toxin (CCT) produced by C. jejuni and C. coli, which serum. Sodium dodecyl sulfate polyacrylamide gel elec­ may be an enterotoxin. trophoresis (SDS PAGE) of purified CCT produced one band at 70,000 daltons. Cell-free concentrates were positive in the MATERIALS AND METHODS rabbit skin permeability test and caused fluid accumulation in rabbit ileal loops. However, cell-free supernatant fluids and Bacterial strains concentrates heated at 90°C for 15 min and tested by the suckl­ Table 1 lists the bacterial strains used in this study. All strains were ing mouse assay produced no fluid accumulation in the intes­ biotyped by the method of Hebert et al. (79). tines of mice. Culture conditions All C. jejuni and C. coli strains were stored at -80°C in casamino acid yeast extract (CYE) broth (23) containing 5 to 10% dimethylsul- Certain enteric bacteria, such as Vibrio cholerae and foxide. Thawed cultures were used to inoculate brucella broth (BB; a number of Escherichia coli strains, produce enteroto- Difco). Overnight BB cultures were used to inoculate BB, CYE broth, or CYE broth containing 0.3, 0.5, 1.0, 2.0 or 50 u,g lincomycin (LN)/ xins that stimulate secretion of fluid into the intestine and ml (1 ml BB culture per 100 ml CYE in a 500-ml flask). Final cell cause a watery diarrhea (9,15). These toxins, e.g., chol­ concentrations (in the absence of antibiotics) were between 5x 108 and era toxin (CT) and heat-labile E. coli toxin (LT), are im­ 8xl09 CFU/ml. Polymyxin (PO), at a concentration of 100 or 2,000 munologically related (5,20). Immunological cross-reac­ IU/ml, was added to some of the lincomycin-treated cultures for the tivity with CT has also been shown with toxins produced last 30 min of incubation. All Campylobacter spp. cultures were incu­ bated for 18 to 24 h with shaking (100 rpm) at 37°C in an atmosphere by other microorganisms. A toxin produced by Sal- of 85% nitrogen, 10% carbon dioxide and 5% oxygen. V. cholerae 569B was incubated under the same conditions, except that no special atmosphere was provided. 'Food and Drug Administration. Preparation of cell-free supernatant fluids and concentrates 2The Catholic University of America. Cell-free supernatant fluids were prepared by centrifugation at 10,000 ^Present address: Naval Medical Research Institute, Bethesda, MD x g for 30 min. An Amicon ultrafilter (Amicon Corp., Lexington, 20814. MA) with an XM50 membrane, which concentrates molecules larger

JOURNAL OF FOOD PROTECTION, VOL. 47, DECEMBER 1984 944 McCARDELL, MADDEN AND LEE

TABLE 1. Results of in vitro assays for the production of toxin by 12 strains of Campylobacter jejuni and one strain of Vibrio cholerae. Strain Source Donor Biotype ELISA" Y-lb CHOb C. jejuni CHI Clinical J. M. Lovett, 4 + 1:2 1:2 FDA, Cincinnati, OH CH2 Clinical Lovett, FDA, Cincinnati, OH 3 + 1:2 1:2 CH4 Clinical Lovett, FDA, Cincinnati, OH 3 + 1:16 1:8 CH5 Clinical Lovett, FDA, Cincinnati, OH 1 + 1:16 1:8 83 Food (raw milk) Lovett, FDA, Cincinnati, OH 3 - - - 88 Food (chicken) Lovett, FDA, Cincinnati, OH 1 + 1:2 1:2 55 Clinical G. Controni, 8 + 1:4 1:4 Children's Hospital

Center, Washington, DC Downloaded from http://meridian.allenpress.com/jfp/article-pdf/47/12/943/1655814/0362-028x-47_12_943.pdf by guest on 25 September 2021 82 Clinical B. Caldwell, 1 + 1:2 1:2 Naval Medical Research Institute, Bethesda, MD E3075 Clinical K. Wachsmuth, 4 Centers for Disease Control, Atlanta, GA 40(ATCC Clinical American Type Culture 1 + 1:4 1:2 29428) Collection, Rockville, MD 41 Animal (calf) N. J. Stern, USDA, 2 + 1:8 1:4 Beltsville, MD 65 Animal (calf) This study 3 - - - V. cholerae 569B Clinical W. Spira, Johns Hopkins 1:2048 1024 University, Baltimore, MD a+, positive; average O.D. 5=0.1 in four experiments; -, negative, average O.D. <0.1 in four experiments; ±, average O.D. 3=0.1, but negative (<0.1) in two experiments. Cell-free supernatant fluids of overnight cultures grown in CYE broth were tested by the ELISA method. bCell-free supernatant fluids were screened in the Y-l and CHO cell assays.

than 50,000 daltons. was used to concentrate cell-free supernatant was added to each well of a microtiter plate (Immunolon II, round bot­ fluids. Ammonium sulfate precipitates were prepared by the addition tom; Dynatech Labs, Inc., Alexandria, VA). After overnight incubation of 7 g ammonium sulfate/10 ml of concentrate (16). After overnight at 4°C, the plates were washed three times with 0.1 M phosphate-buf­ incubation at 4°C, the precipitates were centrifuged, resuspended in dis­ fered saline (PBS), pH 7.4, containing 0.1% bovine serum albumin tilled water and dialyzcd against phosphate-buffered saline (PBS), pH (Fraction V, Sigma) and 0.05% Tween 20 (PBS-Tween). The plates 7.4. were washed once with distilled water and stored at 4°C in a sealed plastic bag with dry silica gel until needed. Preparation of antisera Test solutions containing either CT (Sigma; 0.5 to 500 ng CT/ml), Antisera were obtained by subcutaneous injection of increasing doses uninoculated growth medium or cell-free supernatant fluids were added (4 to 50 jig) of affinity-purified CCT or purchased CT (Sigma Chemical in 50-ji.l amounts to precoated 96-well microtiter plates and incubated Co., St. Louis, MO) with Freund's incomplete adjuvant (Sigma) into at 35°C for 1 h. All solutions were tested in duplicate. After three New Zealand white rabbits over a period of several weeks. Burro anti- washes with PBS-Tween, 50 jil of whole burro anti-CT serum diluted CT (supplied by Dr. William Habig, FDA, Rockville, MD) contained to 1:2560 was added to each well. 6,950 antitoxin units/ml as determined by the rabbit incutaneous assay After incubation for 1 h at 35°C, the plate was again washed three method (7). times with PBS-Tween, and 50 jil of Protein A peroxidase conjugate (Zymed Laboratories, Burlingame, CA) was added to each well. After

Preparation of anti-CT antigen-binding fragment (F(ab)2) another 1-h incubation period at 35°C and three washes with PBS- The IgG fraction was removed from whole high-titered burro anti­ Tween, 100 jil of 2,2'-azino-di-(3-ethylbenzthiazoline sulfonic acid) serum to CT (7) on an affinity column packed with Protein A-Sepharose (ABTS) solution (Zymed) was added to each well. The piate containing CL-4B (Pharmacia, Inc., Uppsala, Sweden), The fragment crystallizable ABTS solution was incubated at 35°C until an optical density of 0.1 portion of the purified IgG was removed by treatment with 2 x recrys- at 410 nm was obtained (usually within 30 min) on a Dynatech Instru­ tallized pepsin (25) (Sigma; 3 mg pepsin per 100 mg IgG), followed ment Micro ELISA Minireadcr for wells containing 0.5 ng CT/ml. Test by double passage through an affinity column containing Protein A- solutions were considered positive if a reading of s=0.1 was obtained Sepharose CL-4B. The pepsin and the F(ab)2 fragments, which did not at 410 nm. CYE or BB controls gave a reading of 0. bind to the Protein A column, were separated on a Sephadex G75 col­ An ELISA which measured binding to gangliosides was performed umn (Pharmacia). All column eluants were concentrated by reverse os­ as described above, except that type III ganglioside (50 jig/ml; Sigma) motic dialysis in polyethylene glycol (average molecular weight or GM1 ganglioside (50 jig/ml, Supelco, Bellefonte, PA) was used as 20,000). the initial plate coating solution.

ELISA Tissue culture assays Optimal dilutions of reagents for the ELISA were determined by Tissue culture assays using Y-l mouse adrenal cells (12.30) or checkerboard titration by a modification of the method of Yolken et Chinese hamster ovary cells (CHO; 18,29) were used to determine al. (33). The F(ab)2 fragment, diluted in carbonate buffer (pH 9.6), cytotonic activity of cell-free supernatant fluids and concentrates. A

JOURNAL OF FOOD PROTECTION, VOL. 47, DECEMBER 1984 CAMPYLOBACTER TOXIN PRODUCTION 945

200-u.l portion of filter-sterilized supernatant fluid, CT or uninoculated iolate. The gels were incubated in a humidified atmosphere for 48 to growth medium, diluted in Eagle minimal essential medium with 72 h at room temperature. glutamine and supplemented with 14% fetal calf serum, 44 IU penicillin G/ml and 44 u,g streptomycin sulfate/ml (Flow Laboratories, McLean, Determination of CCT concentration in cell-free supernatant fluids VA) was added to the wells of a 96-well microtiter plate. A 50-u.l Decreasing volumes of cell-free supernatant fluid were applied to the portion of diluted Y-1 or CHO cells (ca. 103 to 104 cells) was added anti-CT affinity column until the smallest volume (25 ml) that eluted to each well. In a few experiments, test solutions were added to a 48- the maximum amount of CCT (average, 20 u.g) was determined. Protein to 72-h monolayer of cells in microtiter plates. After overnight incuba­ was measured with a Bio-Rad protein assay kit (Bio-Rad, Richmond, tion, the cells were examined with an inverted phase-contrast micro­ CA). This experiment was repeated 20 times and the average amount scope. Wells containing 3=50% rounded (Y-1) or elongated (CHO) cells of affinity purified CCT per ml of supernatant fluid applied to the col­ were considered positive. To demonstrate neutralization of cytotonic ac­ umn was calculated. tivity, cell-free supernatant fluid, affinity-purified CCT or CT was preincubated at 37°C for 1 h with an equal volume of diluted anti-CT Rabbit ileal loops or anti-CCT serum. Heat stability at 56°C for 30 min, 60°C for 10 Enterotoxigenicity was assayed with the permanently ligated rabbit min and 100°C for 10 min and trypsin sensitivity of cell-free concen­ ileal loop (8). Fluid accumulation (ratio of volume to length of loop) trates to 1 mg trypsin/ml held at 26°C for 1 h were also tested by the was measured ca. 18 h after injection of 2 ml of cell-free concentrate. Y-1 mouse adrenal cell assay. Downloaded from http://meridian.allenpress.com/jfp/article-pdf/47/12/943/1655814/0362-028x-47_12_943.pdf by guest on 25 September 2021 Immunosorbent affinity column chromatography Rabbit skin permeability test The IgG fraction (isolated from whole burro antiserum to CT on a Skin permeability tests were performed in adult New Zealand white Protein A Sepharose CL-4B column as previously described) was rabbits (14). The area of increased permeability was measured 1 h after coupled to cyanogen bromide-activated Sepharose 4B (27). Cell-free intravenous injection of 2 ml of a 4% solution of Evans blue. supernatant fluid was slowly applied to the affinity column (0.9 x 30 cm Pharmacia C-column) equilibrated with 0.1 M PBS containing Suckling mouse assay for heat-stable toxin (ST) 0.003 M sodium azide (Sigma). Bound CCT was eluted with 0.1 M Cell-free supernatant fluids and concentrates (2X) were heated at Tris buffer, pH 10.5, in one peak measured at 280 nm. After elution, 90°C for 15 min and then injected into the stomachs of suckling mice. the peak was pooled, concentrated with an Amicon ultrafilter (XM50 Inoculated mice were held at room temperature for 3 h (4) or 37°C membrane), and dialyzed against 0.1 M PBS, pH 7.4. The column was for 1.5 h (70); the ratio of the combined weight of the intestines to regenerated by washing with 10 column volumes of 0.1 M sodium ace­ the combined weight of the carcasses was then calculated for each tate buffer (pH 4.5). Before application of another supernatant fluid, supernatant fluid or concentrate tested. Ratios of less than 0.083 were the column was re-equilibrated with 0.1 M PBS containing 0.003 M considered negative. sodium azide. RESULTS Molecular weight determination Molecular weight was determined by sodium dodecyl sulfate polyac- In vitro assays rylamide disc gel electrophoresis (SDS PAGE), as described by Weber Cell-free supernatant fluids from seven of eight C. je­ and Osborn (32), on 7.5% acrylamide gels with 0.01 M sodium phos­ phate buffer containing 0.2% SDS. Before electrophoresis, samples juni and C. coli clinical isolates and from one of two were boiled for 2 to 3 min in the presence or absence of beta-mercap- food isolates and one of two animal isolates were positive toethanol. Gels were stained with 0.5% Coomassie brilliant blue R250 in an antibody-based ELISA directed against CT (Table (Sigma) and then destained overnight. 1). The optical density (O.D.) of the cell-free supernatant fluids in the ELISA corresponded to that of 0.5 ng CT/ Agarose flat bed isoelectric focusing ml, which was the minimum detectable level of CT in Isoelectric focusing was carried out as described previously (28) in a 1% agarose gel (Pharmacia) with a pH interval of 3 to 10. A mixture this assay. Ammonium sulfate precipitation (70%) re­ of seven standards, encompassing an isoelectric point (pi) range of 3 moved all of the toxin as measured by ELISA and Y-1 to 10, was done on the same plate as affinity-purified CCT. assays (Table 2). Ammonium sulfate concentrates of 5 X (4.0 (xg CCT/ml), 20X (16 \x.g CCT/ml) and 65x (52.0 Ouchterlony immunodiffusion test (xg CCT/ml) gave average O.D. readings of 0.15, 0.26 Double immunodiffusion tests were carried out by the method of Ouchterlony (16,26), using a 1.2% agar gel in 0.8% sodium barbital and 0.59 in the ELISA, respectively, corresponding to buffer (pH 7.4) containing 0.85% sodium chloride and 0.0001% merth- ca. 0.5, 2.0 and 25 ng CT/ml, respectively. All Cam-

TABLE 2. Cytotonic activity of C. jejuni CH5 cell-free supernatant fluids incubated at 37°C for 24 h. Medium Treatment Y-1 titer BB or CYEa None 1:16 CYEb None 1:2 CYE Sonicated0 1:2 BB or CYE 0.3 to 2.0 (jLg LN/mld 1:16 BB or CYE 0.3 to 2.0 |xg LN/ml + 100 IU PO/mle 1:32 BB or CYE 0.3 to 2.0 |xg LN/ml + 2,000 IU PO/ml 1:2 BB or CYE Ammonium sulfate prec 1:64 BB or CYE Supernatant fluid from a 0 "Incubated at 37 or 42°C. bIncubated for 48 to 72 h. cCells were sonicated (Bransen, Danbury, CT) on ice for 60 s. dLN, lincomycin. ePO, polymyxin.

JOURNAL OF FOOD PROTECTION, VOL. 47, DECEMBER 1984 946 McCARDELL, MADDEN AND LEE ter spp. supernatant fluids and concentrates were negative The eluate, concentrated by ultrafilter, was positive in the in an ELISA based on binding to gangliosides. ELISA and caused rounding of Y-1 cells, which could Ammonium sulfate precipitates (5 X) were positive in be neutralized by preincubation with anti-CT serum (rab­ tissue culture assays employing a monolayer of cells (Fig. bit). When individual fractions were assayed, the peak 1). However, unconcentrated cell-free supernatant fluids of biological activity corresponded to the affinity column were positive only in the tissue culture assays where cells peak. The amount of CCT produced by C. jejuni CH5, and test solution were added at the same time, usually estimated by measuring the amount of eluted protein from giving titers between 1:2 and 1:16 (Table 1). The sen­ a known volume of supernatant fluid, was 0.8 |xg CCT/ sitivity of this Y-1 assay for CCT was 50 ng/ml and for ml of supernatant fluid from 18- to 24-h cultures grown CT 1.25 ng/ml. Observation under an inverted phase-con­ in CYE or BB. Supernatant fluids from older cultures (48 trast microscope showed rounding of Y-1 mouse adrenal to 72 h) or sonicated cultures contained 0.1 to 0.4 |xg cells and elongation of CHO cells. Both effects were in­ CCT/ml. Mouse adrenal assays confirmed the decrease in distinguishable from those of CT (Fig. 1). Rounding of CCT concentration (Table 2). Y-1 mouse adrenal cells was not eliminated by heating Downloaded from http://meridian.allenpress.com/jfp/article-pdf/47/12/943/1655814/0362-028x-47_12_943.pdf by guest on 25 September 2021 the cell-free concentrates at 56°C for 30 min, 60CC for Effect of lincomycin and polymyxin on CCT production 10 min or 100°C for 10 min. However, all Y-1 activity and release was eliminated by preincubation with antiserum (rabbit) Since none of the Campylobacter spp. strains used in to CT or CCT for 1 h at 37°C (1:128 dilution of anti-CT this study grew in the presence of 50 (xg LN/ml [the or 1:512 dilution of anti-CCT serum neutralized a 1:2 di­ amount used to stimulate toxin production in E. coli and lution of C. jejuni CH5 supernatant fluid). Rounding of V. cholerae (24)], C. jejuni strain CH5 was grown in Y-1 cells was also eliminated by preincubation of CT the presence of 0.3, 0.5, 1.0 and 2.0 |xg LN/ml. Lin­ (100 ng/ml) with a 1:8 dilution of anti-CCT or a 1:1024 comycin alone had no effect on the amount of CCT in dilution of anti-CT serum. Preincubation with a 1:2 dilu­ the supernatant fluid; however, treatment with LN fol­ tion of normal serum had no effect on cell rounding lowed by 100 IU PO/ml usually doubled the amount of caused by either CT or CCT. The Y-1 activity of CCT CCT as measured by Y-1 assay (Table 2). When LN- was eliminated by incubation of supernatant fluid with grown cultures were treated with 2,000 IU PO/ml, the 1 mg trypsin/ml for 1 h. amount of CCT in the supernatant fluid decreased (Table 2).

Purification and measurement of CCT Estimation of molecular weight and isoelectric point CCT was eluted from an anti-CT affinity column in SDS PAGE of affinity-purified CCT produced one a single peak by 0.1 M Tris buffer, pH 10.5 (Fig. 2). band at 70,000 dalton (Fig. 3). No subunit structure was

Figure 1. Effect of CCT and CT on tissue culture cells: (a) untreated Y-1 mouse adrenal cell control; (b) cell rounding 18 h after addition of concentrated supernatant fluid (4.0 \Lg CCT/ml) from C. jejuni CHS; (c) cell rounding 18 h after addition of CT (50 ng/ml); (d) untreated CHO cell control; (e) cell elongation 18 h after addition of concentrated supernatant fluid (4.0 (xg CCT/ml) from C. jejuni CH5; and (f) cell elongation 18 h after addition of CT (50 ng/ml). X250.

JOURNAL OF FOOD PROTECTION. VOL. 47, DECEMBER 1984 CAMPYLOBACTER TOXIN PRODUCTION 947

dpud

o

CO 0.25 < 200 o CO 92 c\j

68 Downloaded from http://meridian.allenpress.com/jfp/article-pdf/47/12/943/1655814/0362-028x-47_12_943.pdf by guest on 25 September 2021 20 40 60 80 100 120 140 160 180 200 220 Volume (ml) Figure 2. Purification of CCT by affinity chromatography on an anti-CT Sepharose column. At arrow, 0.1 M Tris buffer was added to elute CCT. observed even after boiling the CCT in the presence of a reducing agent. Under these same conditions, CT pro­ 25 duced two bands, one at 23,000 daltons and one at about 11,500 daltons, corresponding to the Al and B subunits. Agarose flat bed isoelectric focusing of purified CCT pro­ 18 duced one band which had a pi of 9.0. On the same gel, the isoelectric point of CT was 6.6. 12 Partial identity between CT and CCT B With double immunodiffusion, crude concentrates of cell-free supernatant fluid from C. jejuni CH5 gave a reaction of partial identity with purified CT, using protein A Sepharose purified IgG from anti-CT (Fig. 4). Im­ munodiffusion using IgG from anti-CCT gave a reaction which was indistinct and may or may not have been par­ tial identity (not shown). Figure 3. SDS polyacrylamide disc gel electrophoresis of CCT and CT. Tube 1, purified CCT, one band at 70,000 dalton; Tube 2, CT, subunit A,, 23,000 dalton and subunit B, 11,500 Biotyping of C. jejuni and C. coli strains dalton. Values to the left show molecular weight standards X Biotyping of the twelve Campylobacter spp. strains by W3. the method of Hebert et al. (79) indicated no correlation between biotype and production of CCT (Table 1). No other biotyping or serotyping system was used. noted when cell-free supernatant fluids and 2x concen­ trates were heated at 90°C for 15 min and tested by the In vivo assays suckling mouse assay. Two of the three in vivo assays used for testing en- DISCUSSION terotoxin activity yielded positive results (Table 3). In the permanently ligated rabbit ileal loop, five of six concen­ The most prominent symptom of infection with C. je­ trates of C. jejuni strain CH5, containing ca. 16 ^g CCT/ juni is diarrhea, which can be watery or mucoid and may ml, produced fluid accumulation ranging from 0.8 to 1.2 contain erythrocytes and leukocytes (7). Although the ml/cm as compared with 1.2 to 1.5 ml/cm for 250 ng pathogenic mechanisms have not been elucidated, at least CT/ml. In the rabbit skin permeability test, 0.1 ml of three reports have described possible toxins. Blankenship a concentrate of strain CH5 containing 8 ^jug CCT/ml pro­ et al. (1982, Abstr. Ann. Meet. Am. Soc. Microbiol., duced an area of increased permeability 22 mm in diame­ P3, p. 205) identified the presence of a cytotoxin in soni­ ter (detectable with 2 ml of 4% Evans blue) compared cated cultures of C. jejuni; Gubina et al. (77) reported with a zone of 17 mm produced by CT (0.1 ml of 2.5 a toxin that has a cytotonic effect on Y-l mouse adrenal |xg CT/ml). Purified CCT (0.1 ml of 2.5 |xg/ml) produc­ cells; and Ruiz-Palacios et al. (29) recently described an ed a zone 12 mm in diameter. No fluid accumulation was enterotoxin that causes fluid accumulation in rat loops

JOURNAL OF FOOD PROTECTION, VOL. 47, DECEMBER 1984 948 McCARDELL, MADDEN AND LEE

The immunological similarities between CCT and CT were demonstrated by the reaction of partial identity, using immunodiffusion; by binding of CCT to an anti-CT affinity column; and by neutralization of the Y-l activity of CT or CCT by either anti-CT or anti-CCT serum. Similarities in biological activity between CT and CCT were demonstrated by the morphological changes in tis­ sue culture assays, by the rabbit skin permeability test and by fluid accumulation in rabbit loops. Although only crude concentrates have been tested in rabbit loops, purified CCT was positive in tissue culture assays and in the rabbit skin permeability test. However, there are differences between CCT and CT.

Although both cause rounding of Y-l mouse adrenal Downloaded from http://meridian.allenpress.com/jfp/article-pdf/47/12/943/1655814/0362-028x-47_12_943.pdf by guest on 25 September 2021 cells, this assay was more sensitive for CT (1.25 ng/ml) than for CCT (50 ng/ml). The molecular weight of CCT, 70,000 daltons, was lower than both that of CT (84,000 daltons) and Escherichia coli LT (91,450 daltons) (6). When CT was heated in the presence of a reducing agent, SDS PAGE produced two bands corresponding to the Al and B subunits (75). Under these same conditions, CCT remained intact and no subunit structure was observed. Figure 4. Ouchterlony immunodiffusion of CCT and CT: Wells The isoelectric point of CCT (9.0) was higher than that A and C, 100 |xg CT/ml; wells B and D, C. jejuni CH5 concen­ of CT (6.6) or E. coli LT (8.0) (6). Since CCT does trate containing approximately 116 |xg CCT/ml; center well, not appear to bind to gangliosides, the structure of the anti-CT IgG. binding fragment (B subunit) may account for some of and is neutralized by CT antiserum. This paper describes the differences between CCT and CT. Further experi­ a cytotonic toxin (CCT) produced by C. jejuni and C. ments are required to determine the exact nature of these coli which is immunologically and biologically similar to differences. CT and which may, therefore, be an enterotoxin. CCT may well be the toxin described earlier by Gubina The best conditions for production of the CCT were et al. (17) and Ruiz-Palacios et al. (29), since all three overnight growth in CYE or BB at 37 or 42°C. Supernat­ toxins are cytotonic for CHO or Y-l cells. In addition, ant fluids from older cultures (48 to 72 h) and sonicated the toxin described by Ruiz-Palacios et al. (29) is neut­ cultures contained less toxin (Table 2). Lincomycin- ralized by anti-CT serum and when unconcentrated pro­ treated cultures showed no increase in CCT production. duces no fluid accumulation in rabbit loops. CCT has not The addition of 100 IU PO/ml for the last 30 min of been tested in rat loops, but only concentrated supernat­ incubation caused only a slight increase in CCT levels; ant fluids (^16 (xg CCT/ml, Table 3) were positive in larger amounts of PO (2,000 IU/ml) actually caused a rabbit loops. decrease. The decrease in CCT levels observed when If the CCT described here is an enterotoxin, as much cells were lysed or treated with large amounts of PO may of the evidence indicates, it may account for the watery be due to the release of proteases. diarrhea often observed in Campylobacter infections. The purification of CCT was attained in a single step There is little likelihood that an enterotoxin similar to CT by running supernatant fluid onto an anti-CT affinity col­ could be responsible for the dysentery-like symptoms also umn followed by elution with Tris buffer. Although am­ often observed during infections with Campylobacter spp. monium sulfate could precipitate the CCT, these concen­ Accordingly, Campylobacter spp. would have to possess trates were not used for purification because their prepa­ more than one pathogenic mechanism to account for all ration was too time-consuming. the symptoms observed, if one of those mechanisms is an enterotoxin similar to CT.

TABLE 3. Results of rabbit ileal loop and rabbit skin permeability tests. Concentration of CCT or CT tested Ave. fluid accumulation Ave. zone size (ml/cm) (mm) Unconcentrated supernatant fluid (0.8 ji,g CCT/ml) 0 0 20 x ammonium sulfate precipitate concentrate (16 |xg CCT/ml) 1.0 NDa 0.25 ^g CT/ml 1.4 ND 10 x ammonium sulfate precipitate concentrate (8 ^g CCT/ml) ND 22 Purified CCT (2.5 n-g/ml) ND 12 2.5 ^g CT/ml ND 17 aND, not determined.

JOURNAL OF FOOD PROTECTION, VOL. 47, DECEMBER 1984 CAMPYLOBACTER TOXIN PRODUCTION 949

REFERENCES 188-195. In D. G. Newel (ed.) Campylobacter: epidemiology, pathogenesis, and biochemistry. MTP Press, Hingham, MA. 18. Guerrant, R. L., L. L. Brunton, T. C. Schanitman, L. L. Rebun, 1. Blaser, M. J., I. D. Berkowitz, F. M. LaForce, J. Cravens, L. and A. C. Gilman. 1974. Cyclic adenosine monophosphate and alt­ B. Reller, and W-L. L. Wang. 1979. Campylobacter enteritis: clin­ eration of Chinese hamster ovary cell morphology: a rapid, sensi­ ical and epidemiologic features. Ann. Intern. Med. 91:179-185. tive in vitro assay for the enterotoxins of Vibrio cholerae and Es­ 2. Blaser, M. J., L. B. Reller, N. W. Leuchtefeld, and W-L. L. cherichia coli. Infect. Immun. 10:320-327. Wang. 1982. Campylobacter enteritis in Denver. West. J. Med. 19. Hebert, G. A., D. G. Hollis, R. E. Weaver, M. A. Lambert, M. 136:287-290. J. Blaser, and C. W. Moss. 1982. 30 Years of Campylobacters: 3. Blaser, M. J., J. G. Wells, R. A. Pollard, and J. R. Allen. 1983. biochemical characteristics and biotyping proposal for Campylobac­ Campylobacter enteritis in the United States. Ann. Intern. Med. ter jejuni. J. Clin. Microbiol. 15:1065-1073. 98:360-365. 20. Holmgren, J., I. Soderlind, and T. Wadston. 1973. Cross-reactivity 4. Boyce, J. M., D. J. Evans, Jr., D. G. Evans, and H. L. Dupont. between heat-labile enterotoxin of Vibrio cholerae and Escherichia 1979. Production of heat-stable, methanol-soluble enterotoxin by coli in neutralization tests in rabbit ileum and skin. Acta Pathol. Yersinia enterocolitica. Infect. Immun. 25:532-537. Microbiol. Scand.(B) 81:757-762. 5. Clements, J. D., and R. A. Finkelstein. 1978. Immunological, 21. James, C, M. Dibley, V. Burke, J. Robinson, and M. Gracey.

cross-reactivity between heat-labile enterotoxin(s) of Escherichia 1982. Immunological cross-reactivity of enterotoxins of Aeromonas Downloaded from http://meridian.allenpress.com/jfp/article-pdf/47/12/943/1655814/0362-028x-47_12_943.pdf by guest on 25 September 2021 coli and subunits of Vibrio cholerae enterotoxin. Infect. Immun. hydrophila and cholera toxin. Clin. Exp. Immunol. 47:34-42. 21:1036-1039. 22. King, E. O. 1957. Human infections with Vibrio fetus and closely 6. Clements, J. D., R. J. Yancey, and R. A. Finkelstein. 1979. Pure related vibrios. J. Infect. Dis. 101:119-128. heat-labile enterotoxin (LT) from Escherichia coli is similar to 23. Kusama, H., and J. P. Graig. 1970. Production of biologically ac­ choleragen immunologically, structurally, and in specific activity. tive substances by two strains of Vibrio cholerae. Infect. Immun. Proc. 15th Joint Cholera Res. Conf. U.S.-Japan Coop, Med. Sci. 1:80-87. Prog. pp. 27-41. 24. Levner, M., F. P. Wiener, and B. A. Rubin. 1977. Induction of 7. Dafni, Z., and J. B. Robbin. 1976. Purification of heat-labile en­ Escherichia coli and Vibrio cholerae enterotoxins by an inhibitor terotoxin from Escherichia coli 078:H11 by affinity chromatography of protein synthesis. Infect. Immun. 15:132-137. with antisera to Vibrio cholerae toxin. J. Infect. Dis. 25. Nisonoff, A., F. C. Wissler, L. N. Lipman, and D. L. Woernley. 133(Suppl.):S138-S141. 1960. Separation of univalent fragments from the bi-valent antibody 8. De, S. N., and D. N. Chatterjee. 1953. An experimental study by reduction of disulfide bonds. Arch. Biochem. Biophys. 89:230- of the mechanism of action of Vibrio cholerae on the intestinal 244. mucus membrane. J. Pathol. Bacteriol. 66:559-562. 26. Ouchterlony, O., and L. A. Nilsson. 1978. Immunodiffusion and 9. De, S. N., K. Bhattcharya, and J. K. Sarkar. 1956. A study of immunoelectrophoresis. In D. M. Weir (ed.) Handbook of experi­ the pathogenicity of strains of Bacterium coli from acute and mental immunology. Blackwell Scientific Publication, Oxford, Eng­ chronic enteritis. J. Pathol. Bacteriol. 71:201-209. land. 10. Dean, A. G., Y. C. Ching, R. G. Williams, and L. B. Harden. 27. Pharmacia Fine Chemicals. 1979. Affinity chromatography: princi­ 1972. Test for Escherichia coli enterotoxin using infant mice: appli­ ples and methods. Pharmacia Fine Chemicals, Uppsala, Sweden. cation in a study of diarrhea in children in Honolulu. J. Infect. 28. Pharmacia Fine Chemicals. 1980. Agarose IEF. Pharmacia Fine Dis. 125:407-411. Chemicals, Uppsala, Sweden. 11. Dekeyser, P., M. Goussuin-Detrain, J. P. Butzler, and J. Sternon. 29. Ruiz-Palacios, G. M., J. Torres, N. I. Torres, E. Escamilla, B. 1972. Acute enteritis due to related vibrio: first positive stool cul­ Ruiz-Palacios, and J. Tamayo. 1983. Cholera-like enterotoxin pro­ tures. J. Infect. Dis. 125:390-392. duced by Campylobacter jejuni: characterization and clinical sig­ 12. Donta, S. T., D. M. Smith, H. W. Moon, and S. C. Whipp. 1974. nificance. Lancet ii:250-251. Detection of heat-labile Escherichia coli enterotoxin with the use 30. Sack, D. A., and R. B. Sack. 1975. Test for enterotoxigenic Es­ of adrenal cells in tissue culture. Science 183:334-336. cherichia coli enterotoxin Y-l adrenal cells in miniculture. Infect. 13. Doyle, M. P. 1981. Campylobacter fetus subsp. jejuni: an old Immun. 11:334-336. pathogen of new concern. J. Food Prot. 44:480-488. 31. Sandefur, P. D., J. W. Peterson. 1977. Neutralization of Sal­ 14. Evans, D. J., Jr., D. G. Evans, and S. L. Gorbach. 1973. Produc­ monella toxin-induced elongation of Chinese hamster ovary cells tion of vascular permeability factor by enterotoxigenic Escherichia by cholera antitoxin. Infect. Immun. 15:988-992. coli isolated from man. Infect. Immun. 8:725-730. 32. Weber, K., and M. Osborn. 1969. The reliability of molecular 15. Finkelstein, R. A. 1973. Cholera. Crit. Rev. Microbiol. 2:553-623. weight determination by dodecyl sulfate polyacrylamide gel elec­ 16. Food and Drug Administration. 1978. Bacteriological analytical trophoresis. J. Biol. Chem. 244:4406-4412. manual, 5th ed. Association of Official Analytical Chemists, Ar­ 33. Yolken, R. H., H. B. Greenberg. M. H. Merson, R. B. Sack, lington, VA. and A. Z. Kapikian. 1977. Enzyme-linked immunosorbent assay 17. Gubina, M., J. Zajc-Satler, A. Z. Dragas, Z. Zeleznik, and J. for detection of Escherichia coli heat-labile enterotoxin. J. Clin. Mehle. 1982. Enterotoxin activity of Campylobacter species, pp. Microbiol. 6:439-444.

JOURNAL OF FOOD PROTECTION, VOL. 47, DECEMBER 1984