INFECTION AND IMMUNITY, Aug. 1990, p. 2414-2419 0019-9567/90/082414-06$02.00/0 Copyright © 1990, American Society for Microbiology
Vol. 58, No. 8
Partial Purification and Characterization of the Enterotoxin Produced by Campylobacter jejuni TOHRU DAIKOKU,t MARIKO KAWAGUCHI, KOZO TAKAMA, AND SATORU SUZUKI* Department of Food Science and Technology, Faculty of Fisheries, Hokkaido University, Hakodate, Hokkaido 041, Japan Received 18 October 1989/Accepted 16 April 1990 Campylobacterjejuni enterotoxin was partially purified from culture supernatant. The purified fraction after gel filtration indicated three bands at 68, 54, and 43 kilodaltons on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). This fraction enhanced the adenylate cyclase activity of HeLa cell membranes by 1.5-fold over that of the control. The study with anti-cholera toxin immunoglobulin G (IgG) and ganglioside affinity column chromatographies revealed that the eluent from the anti-cholera toxin IgG column chromatography exhibited a single band (68 kDa) on SDS-PAGE and native PAGE, whereas the eluent from ganglioside column chromatography exhibited two bands (68 and 54 kDa) on SDS-PAGE. These suggest that the 68-kDa polypeptide should have an immunological relationship with cholera toxin, and the 68- and 54-kDa polypeptides might be responsible for the recognition of ganglioside.
Campylobacterjejuni recently has been recognized as an important pathogen in human diarrheal disease. With the advent of improved methods (5), C. jejuni is now isolated from human diarrheal stools frequently worldwide. Symptoms of C. jejuni infection are generally those of gastrointestinal disorder, for example, watery diarrhea and a dysenterylike syndrome, or rarely those of an extraintestinal infection. Noteworthy characteristics of C. jejuni are as follows: the latent period is long (1 to 7 days) compared with those of other bacteria, and the diarrhea caused by this bacteria usually continues for 2 to 7 days. The organism spends from 2 weeks to 3 months in the host before being excreted without any treatment with antibiotics. The mechanism of the various symptoms of diarrhea caused by C. jejuni has been unclear despite the many studies that have been done on this bacteria (20). However, one important mechanism causing diarrhea is thought to be derived from the production of potent toxins by the bacteria. It is now known that C. jejuni produces at least two exotoxins: a heat-labile enterotoxin and a cytotoxin (8, 12). The former is immunologically related to cholera toxin (CT) and Escherichia coli heat-labile enterotoxin (LT), i.e., it can be neutralized with antiserum for CT and LT (10, 11). The latter has cytotoxic effect against cultured mammalian cells, including Vero and HeLa cells (8, 12, 21). These two exotoxins are considered to relate strongly to the bacterial pathogenesis of C. jejuni. However, the biochemical properties of the toxins are still unclear, although there are two reports on the purification of entero-
enterotoxin produced by C. jejuni as the first step to achieving the results described above. MATERIALS AND METHODS All experiments were done at 0 to 4°C unless otherwise
noted. Bacterial strain and growth conditions. The C. jejuni strain used in this study was isolated from the stool of a child in Hakodate Chuou Hospital. It was grown at 37°C in the Campy Pak system (BBL Microbiology Systems) for 48 h on a Skirrow plate without antibiotics. Harvested cells were inoculated into 600 ml of Casamino Acids-yeast extract broth (CYE) (Difco Laboratories) containing 1.0 Fg of ferric chloride per ml and incubated at 37°C in the presence of 10% CO2 for 24 h with shaking (17). This liquid culture was then inoculated into 12 liters of CYE broth and grown under the conditions mentioned above. These cultures were centrifuged at 10,000 x g for 30 min. Then, the supernatants were filtered through a 0.45-,m-pore size membrane filter and used as the starting material for toxin purification. Ultrafiltration. Ultrafiltration with a PT membrane (30,000Mr cut-off; Millipore Corp.) was used to concentrate a cellfree supernatant. Precipitation of proteins was performed in an 80% saturated solution of ammonium sulfate. The precipitate was centrifuged at 10,000 x g for 15 min after incubation at 4°C for 4 h. The precipitate was suspended in buffer A (50 mM Tris hydrochloride [Tris-HCl, pH 7.8], 1 mM EDTA, 10 mM 2-mercaptethanol) containing 1 mM phenylmethylsulfonyl fluoride. This was dialyzed against buffer A containing 1 mM benzamidine and 30% glycerol and stored at -200C. Gel filtration. The ammonium sulfate fraction was applied to a Sephadex G-100 (Pharmacia, Uppsala, Sweden) column (2.5 by 85 cm) equilibrated with buffer A containing 100 mM NaCl and 10% glycerol. Eluents were collected (1.5 ml per tube) at a flow rate of 30 ml/h. ELISA. C. jejuni enterotoxin could be neutralized by anti-CT serum and bound to ganglioside (16). Based on this property, the enterotoxin assay was performed by enzymelinked immunosorbent assay (ELISA) with a goat antiserum (LBL) for CT with gangliosides (SRL) in the solid phase.
toxin (11, 16). To reveal the expression mechanisms of the toxins and to clarify the mechanism of pathogenesis, it is necessary to obtain a lot of purified toxin. Moreover, examination of the biochemical properties of the purified toxin will give useful information for immunotherapy and chemotherapy for campylobacteriosis. In this study, we attempted to purify and characterize the * Corresponding author. t Present address: Nagoya University School of Medicine, Nagoya 466, Japan.
2414
ENTEROTOXIN PRODUCED BY C. JEJUNI
VOL. 58, 1990
Protein assay. Protein concentration was estimated by the method of Bradford (1). Bovine serum albumin (BSA) was used as the standard. PAGE. The purity of the toxin in each step of purification was examined by sodium dodecyl sulfate-polyacrylamide slab gel electrophoresis (SDS-PAGE), as described by Laemmli (14). Non-reductive PAGE (native PAGE) was carried out without SDS and heat treatment. Gels were stained with Coomassie brilliant blue. Molecular mass markers used were phosphorylase b (97.4 kilodaltons [kDa]), BSA (66 kDa), egg albumin (45 kDa), trypsinogen (24 kDa), and lactalbumin (14.2 kDa). Whether the sample had glycoprotein was determined by periodic acid-Schiff (PAS) stain of the gel (9). Treatment of HeLa cells with toxins. HeLa cells were grown in 35-mm (inner diameter) tissue culture dishes to confluence. Cells were incubated at 37°C for 24 h after the addition of an appropriate amount of the purified enterotoxin or CT (LBL). Preparation of HeLa cell membranes. Monolayers of HeLa cells were washed twice with phosphate-buffered saline (PBS), and cells were harvested by scraping them off into 0.2 ml of PBS and pelleted by centrifugation at 200 x g for 5 min. The cell membrane fraction was prepared by sonication (50 W for 15 s, four times) of the pellet in 1 ml of buffer B (50 mM Tris-HCl [pH 7.5], 0.25 M sucrose, 5 mM MgCl2). This mixture was centrifuged at 500 x g for 5 min, and then the supernatant was centrifuged at 25,000 x g for 10 min. The pellet was washed twice with buffer B and dispersed in the same buffer (protein concentration, ca. 4 mg/ml) (2). Assay for adenylate cyclase activity (assay 1). Adenylate cyclase activity was assayed by the method of Chang et al. (2) with some modifications. The total volume of the reaction mixture was 100 ,lI which contained 25 mM Tris-HCl (pH 7.5), 0.5 mM [a-32P]ATP (9 x 104 cpm/nmol) (New England Nuclear Corp.), 6 mM MgCl2, 1 mM EDTA, 20 mM creatine phosphate (Sigma Chemical Co.), 100 U of creatine phosphate kinase (Sigma) per ml; 1 mM 3',5'-cyclic AMP (cAMP) (Yamasa, Choshi, Japan), 20 ,uM GTP (Yamasa), 1 mM NAD (Oriental, Tokyo, Japan), and an appropriate amount of HeLa cell membranes. Cocktail without cell membrane was preincubated at 30°C for 5 min. Then the membrane fraction of HeLa cells was added to the reaction mixture, and incubation was continued at 30°C for 30 min. The reaction was stopped by the addition of 100 ,ul of stop solution containing 2% SDS, 40 mM ATP, and 1.4 mM cAMP. This mixture was applied to a DEAE-cellulose column (5 by 45 mm) (bicarbonate form). Nucleotides were eluted by a linear gradient of 0 to 0.3 M triethylamine bicarbonate buffer (pH 8.3). Detection of [32PJcAMP converted from [a-32P]ATP. A 100-,u1 amount was taken from the fractionated sample, and the radioactivity was counted with an Aloka 673 liquid scintillation counter in a toluene-based scintillator. In parallel, cold nucleotides such as cAMP, AMP, ADP, and ATP were applied to the DEAE-cellulose column. Their elution profile was monitored by the A260 Radioimmunoassay of cAMP in HeLa cells (assay 2). To confirm the activation of adenylate cyclase in the toxintreated cells, cAMP in the cells was determined by radioimmunoassay. HeLa cells treated with enterotoxin or CT were
harvested and homogenized with 6% (wt/vol) trichloroacetic acid. Quantitative analysis of cAMP was performed with a [125 ]cAMP radioimmunoassay kit (New England Nuclear). Anti-CT IgG affinity column chromatography. A 20-mg
2415
Vo -4 -
0 00
r4 -J
Li
150 200 Elution vol. (ml) FIG. 1. Elution profile of C. jejuni enterotoxin on Sephadex G-100 gel filtration chromatography. Symbols: O-O, enterotoxin activity monitored by ELISA; - - - , A280. A 22-ml amount of crude toxin was applied to the column (2.5 by 85 cm). Markers, blue dextran 200 (VO [void volume]) and BSA monomer, are indicated by arrows.
amount of anti-CT goat immunoglobulin (IgG) (LBL), dissolved in 0.1 ml of NaHCO3 buffer (pH 8.5) containing 0.1 M NaCl, was reacted overnight with 1 ml of activated Affi-Gel 10 (Bio-Rad Laboratories) at 4°C with shaking (13). The remaining active sites of the resin were blocked with 0.1 M ethanolamine for 30 min at 4°C. Approximately 75% of the IgG coupled to the gel. The coupling rate was monitored by A280. The partially purified enterotoxin was applied to the column. After being washed with a 10- to 20-fold bed volume of buffer A, the column was eluted with 50 mM Tris-HCl (pH
10.5) by the stepwise method. Ganglioside affinity column chromatography. A 1-mg amount of gangliosides (SRL) dissolved in distilled water (1 mg/ml) and 30 mg of EDC [N-ethyl-N'-(3-dimethyl aminopropyl)carbodiimide hydrochloride] were added to 1 ml of AF-Amino Toyopearl (Toso, Tokyo, Japan), which was kindly provided by Y. Nagashima, Toso Co. After incubation overnight at 4°C with shaking, the column was washed with 10 ml of distilled water, 1 M NaCl, distilled water, and 50 mM Tris-HCl (pH 7.8) in that order. The column was then equilibrated with buffer A. The partially purified enterotoxin was applied to the column. After being washed with a 10-fold bed volume of buffer A, the column was eluted with 7 M urea in 50 mM Tris-HCl buffer (pH 7.8) by the stepwise method
(3). RESULTS Partial purification of the enterotoxin. Enterotoxin activity was measured in the cell-free supernatant, polymyxintreated supernatant, and sonicated supernatant fractions. The highest total activity was observed in the cell-free supernatant (3.84 x 10' ELISA units), followed by the
supernatant of polymyxin-treated cells (0.94 x 106 ELISA units), and the lowest activity was found in the supernatant obtained from sonicated cells (0.22 x 10' ELISA units). However, the opposite order was found in the amount of total protein. Therefore, the cell-free supernatant had the highest specific activity (41.3 x 103 ELISA units/mg). For this reason, the cell-free supernatant was used as the starting material for purification of the enterotoxin. Crude enterotoxin was concentrated twofold after ammonium sulfate precipitation. This concentrated sample was applied to a Sephadex G-100 column. A typical elution
2416
45k_:.iP,.
DAIKOKU ET AL.
A
M
M B
97.4k
132k 66k}
-
_
-68k 54k
45k
-43k
661 -
.*:.:.
INFECT. IMMUN.
TABLE 2. cAMP content in HeLa cells treated with C. jejuni enterotoxin or CT Toxin (Rg/ml)
cAMP assay ia of protein) (nmol/mg
2b cAMP assay cells) (pmol/106
None (control)
2.6
1.9 ± 0.355
4.0 NT
NTC 3.0 ± 0.044
5.5 NT
NT 3.3 ± 0.043
C. jejuni 18 21 CT 91 96
a Cells were incubated with no toxin or the indicated toxin for 24 h at 37°C. Assay conditions are described in the text. Values are for a single experiment. b Cells were incubated with no toxin or the indicated toxin for 24 h at 37°C. Assay conditions are described in the text. Values are means ± standard deviations for triplicate experiments. c NT, Not tested.
..)).. s
24k
...
ss
..
...
........
...
;..
.* !::@
._minmS~.
4
FIG. 2. SDS-PAGE and native PAGE of the fraction from gel filtration. Lanes: M, marker proteins (sizes in kilodaltons); A, SDS-PAGE (10% gel); B, native PAGE (10% gel). S, Band stuck on condensation gel. Arrowhead indicates tracking dye.
pattern of enterotoxin is shown in Fig. 1. The enterotoxin
activity, monitored by ELISA, was detected between fractions 140 and 155 (1-ml fractions) of eluent. Fractions 140 to 155 were combined, and the purity was checked by SDSPAGE and native PAGE (Fig. 2). In native PAGE, this sample stuck on the stacking gel. However, SDS-PAGE gave three main bands with molecular masses of 68, 54, and 43 kDa. Moreover, glycoproteins were not detected in the PAGE by the method of PAS staining (data not shown). These results suggest that the halotoxins might aggregate. The overall purification results are summarized in Table 1. The enterotoxin was purified 8.7-fold from the crude supernatant to the gel filtration step. The final fraction showed a specific ELISA titer of 35.82 x 104 ELISA units/mg. ELISA units were defined as described previously (4). Adenylate cyclase activity of the enterotoxin. The purified C. jejuni enterotoxin and CT were not toxic to HeLa cells, and no morphological changes were observed. Chang et al.
(2) reported that type II heat-labile enterotoxin of E. coli (LT-II), which is similar to CT, increased the cAMP amount by activating adenylate cyclase through the GTP-dependent ADP-ribosylation of specific membrane proteins. In this experiment, the enzymatic activity of enterotoxin was determined by the same method. Results are shown in Table 2. Enterotoxin was found to enhance adenylate cyclase activity in HeLa cells. Adenylate cyclase activity increased in membranes prepared from HeLa cells incubated with enterotoxin by 1.5-fold over that in the control under these conditions, i.e., 4.0 nmol of cAMP was converted per mg of membrane protein, whereas the control fraction converted 2.6 nmol of cAMP per mg. After treatment with CT, adenylate cyclase activity in the HeLa membranes was enhanced 2.0-fold over that in the control. We could not compare the specific activity of both toxins because the amounts used were not equal. However, the findings in this experiment suggest that the enterotoxin has a mechanism similar to that of CT, which causes induction of secretory diarrhea by stimulating adenylate cyclase activity in intestinal cells. Enhancement of adenylate cyclase activity in HeLa cells was confirmed by quantitation of cAMP (Table 2). As in the adenylate cyclase assay (assay 1), the results of assay 2 revealed that treatment with the enterotoxin or CT elevated the cAMP content in HeLa cells. Elution profiles of the enterotoxin in the affinity column chromatography. Elution profiles of enterotoxin from antiCT IgG affinity column chromatography are shown in Fig. 3. The enterotoxin bound to the column was eluted with 50 mM Tris-HCl buffer (pH 10.5). The elution pattern was monitored by ELISA. Fractions 80 to 95 were collected and subjected to SDS-PAGE and native PAGE. The material
TABLE 1. Summary of purification of enterotoxina Step
Cell-free supernatant Ultrafiltration 80% saturated
(NH4)2S04
Gel filtration
Total vol (ml)
protein
Total
Activity (ELISA unitsb)
Fold purification
Yield (%)
4.13 4.38 9.04
1.00 1.06 2.19
100 90 94
35.82
8.67
51
Specific (104
(mg)
Total (106 units)
12,000 1,800 22
93.0 79.0 39.8
3.84 3.46 3.60
16
5.5
1.97
units/mg)
a Fractions from affinity columns (see text) are not shown because the amount of protein in these fractions was too small for the assay. b ELISA units were calculated as ELISA titer (the amount of antigen giving a positive reaction in the ELISA [>0.1 optical density unit] at the highest dilution of each fraction) times the total volume of the fraction (4).
ENTEROTOXIN PRODUCED BY C. JEJUNI
VOL. 58, 1990
2417
p
0\
%
_ ff
O~ ~ ~ 0
.1
it-e
01-
50
100
F rac t i o n FIG. 3. Elution profile of C. jejuni enterotoxin on anti-CT IgG affinity column chromatography. Inset is SDS-PAGE of pooled fraction (p); tracking dye is indicated by the arrowhead. Elution (e) was performed with 50 mM Tris-HCl (pH 10.5).
p 97.4k
-
66k
-
_
0.6
-
68k
-
54k
L) 14.2k -
-
0.4 $DS-PA!
-I
w 0.2
e 0
20 10 Fract ion
30
FIG. 4. Elution profile of C. jejuni enterotoxin on ganglioside affinity column chromatography. Inset is SDS-PAGE of pooled fraction (p); tracking dye is indicated by the arrowhead. Elution (e) was performed with 7 M urea in buffer A.
2418
DAIKOKU ET AL.
INFECT. IMMUN.
gave a single band at 68 kDa on SDS-PAGE (Fig. 4) and also on native PAGE (data not shown). The purity of this fraction was determined by scanning of the stained gel with a Flying spot scanner (Shimadzu Co., Kyoto, Japan), and it was found to be homogeneous. This suggests that the immunological crossing site of the enterotoxin to CT should be present in the 68-kDa polypeptide. In the case of ganglioside affinity column chromatography, elution was performed with 7 M urea in 50 mM Tris-HCl (pH 7.8). The elution profile and the SDS-PAGE pattern of the eluent are shown in Fig. 4. Unlike the anti-CT IgG column, the material obtained from the ganglioside column gave two bands at 68 and 54 kDa in SDS-PAGE. This might have been caused by the different specificity of the subunits of the toxin for binding to the affinity resins. The homogeneity of this fraction monitored by gel scanning was greater than 73%. The fractions from the two affinity columns were not able to be examined for cAMP because not enough of the toxin was obtained for the cAMP assay in these steps.
band at 68 kDa was observed on SDS-PAGE in the fraction from the anti-CT IgG affinity column; however, two bands at 68 and 54 kDa were detected by ganglioside affinity column chromatography. These results suggest that the holotoxin should consist of at least two polypeptides of 68 and 54 kDa and that the large subunit should have antigenic sites similar to those of CT and LT. Both subunits might be responsible for binding the toxin to the ganglioside receptors on the surface of the cell. A vaccine for enterotoxigenic E. coli is currently being developed (11). One such candidate vaccine for E. coli consists of the heat-labile toxin cross-linked to the LT B subunit. C. jejuni enterotoxin is known to have a close immunological relationship with LT and CT. Therefore, precise knowledge of the structure of the toxin is needed in order to develop an effective vaccine for campylobacters.
DISCUSSION The enterotoxin produced by C. jejuni is similar to CT and LT in many ways, such as immunological properties and enzymatic activity. This enterotoxin is also known to be heat labile and to induce accumulation of fluid and electrolytes in both rat (19) and rabbit (15) ileal loops. The enterotoxin also causes cytotoxic changes in confluent-monolayer cell lines, such as elongation of CHO cells (6, 7) and rounding of Y-1 cells (15, 19). These findings are consistent with evidence that the mechanism of action of the toxin is mediated by the increase in cAMP in the cells (19). We purified the enterotoxin and examined the enzymatic activity and the structure of the toxin. We also examined the binding of the toxin to affinity columns. The crude toxin was purified 8.7-fold by three steps, including gel filtration chromatography. This purified toxin gave three bands on SDSPAGE, although it stuck on native PAGE. From this apparent discrepancy, it was hypothesized that the holotoxin might aggregate by itself. PAS staining showed that there were no detectable sugars in the toxin protein. Therefore, it is suggested that the holotoxins of the enterotoxin would aggregate each other and form oligomers as the native form. It is known that CT and LT ADP-ribosylate a guanine nucleotide-binding regulatory protein (G.) (16, 18), whereas pertussis toxin ADP-ribosylates an inhibitory guanine nucleotide-binding regulatory protein (Gi) (2). In this study, we examined the adenylate cyclase activity of HeLa cells treated with C. jejuni enterotoxin. Ruiz-Palacios et al. (19) reported on the production of cAMP in CHO cells treated with C. jejuni enterotoxin and observed that the cAMP concentration in the cell increased. The results of this study showed that the enterotoxin purified by gel chromatography increased adenylate cyclase activity in HeLa cells. From these findings, the enzymatic activity of enterotoxin should resemble that of CT. The subunit structure of the enterotoxin was reported as a single polypeptide by McCardell et al. (16). However, our experiment and the report by Klipstein and Engert (11) indicate that the holotoxin of the enterotoxin would be constituted of several subunits. Klipstein and Engert (11) found that the holotoxin treated with guanidine showed three protein peaks in gel filtration with Sephadex G-75, although their precise molecular sizes were not indicated. They also reported that the largest subunit among the three crossreacted with LT. In this study, it was found that a single
We thank T. Eguchi (Hakodate Chuou Hospital) for kindly supplying the bacterial strains used in this work, Y. Nishiyama (Nagoya University) and T. Suzutani (Asahikawa Medical College) for the gift of the cultured cell lines, and T. Yamaguchi (Research Laboratory, Yamasa Shoyu Co.) for kindly supplying the nucleotides used in this work. We thank T. Suzuki and Y. Ezura, Hokkaido University, for useful suggestions.
ACKNOWLEDGMENTS
LITERATURE CITED 1. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248-254. 2. Chang, P. P., J. Moss, E. M. Twiddy, and R. K. Holmes. 1987. Type II heat-labile enterotoxin of Escherichia coli activates adenylate cyclase in human fibroblasts by ADP ribosylation. Infect. Immun. 55:1854-1858. 3. Cuatrecasas, P., I. Parikh, and M. D. Holienbelg. 1973. Affinity chromatography and structural analysis of Vibrio cholerae enterotoxin-ganglioside-agarose and the biological effects of ganglioside-containing soluble polymers. Biochemistry 12:42534264. 4. Daikoku, T., S. Suzuki, S. Oka, and K. Takama. 1989. Profiles of enterotoxin and cytotoxin production in Campylobacter jejuni and C. coli. FEMS Microbiol. Lett. 58:33-36. 5. George, H. A., P. S. Hoffman, R. M. Smibert, and N. R. Trieg. 1978. Improved media for growth and aerotolerance of Campylobacter fetus. J. Clin. Microbiol. 8:36-41. 6. Goossens, H., J. P. Butzler, and Y. Takeda. 1985. Demonstration of cholera-like enterotoxin production by Campylobacterjejuni. FEMS Microbiol. Lett. 29:73-79. 7. Goosenns, H., E. Rummens, S. Cadranel, J. P. Butzler, and Y. Takeda. 1985. Cytotoxic activity on Chinese hamster ovary cells in culture filtrates of Campylobacterjejunilcoli. Lancet ii:511. 8. Johnson, W. M., and H. Lior. 1986. Cytotoxic and cytotonic factors produced by Campylobacterjejuni, Campylobacter coli,
and Campylobacter laridis. J. Clin. Microbiol. 24:275-281.
9. Keyser, J. W. 1964. Staining of serum glycoproteins after electrophoretic separation in acrylamide gels. Anal. Biochem. 9:249-252. 10. Klipstein, F. A., and R. F. Engert. 1984. Properties of crude Campylobacter jejuni heat-labile enterotoxin. Infect. Immun.
45:314-319.
11. Klipstein, F. A., and R. F. Engert. 1985. Immunological relationship of the B subunits of Campylobacter jejuni and Esche-
richia coli heat-labile enterotoxin. Infect. Immun. 48:629-633.
12. Klipstein, F. A., R. F. Engert, H. Short, and E. A. Schenk. 1985. Pathogenic properties of Campylobacter jejuni: assay and correlation with clinical manifestations. Infect. Immun. 50:43-49.
ENTEROTOXIN PRODUCED BY C. JEJUNI
VOL. 58, 1990
13. Krivan, H. C., and T. D. Wilkins. 1987. Purification of Clostridium difficile toxin A by affinity chromatography on immobilized thyroglobulin. Infect. Immun. 55:1873-1877. 14. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227:680-685. 15. Lecce, J. G. 1958. Some biochemical characteristics of Vibrio fetus and other related vibrios isolated from animals. J. Bacteriol. 76:312-316. 16. McCardell, B. A., J. M. Madden, and E. C. Lee. 1984. Campylobacterjejuni and Campylobacter coli production of a cytotonic toxin immunologically similar to cholera toxin. J. Food Prot. 47:943-949. 17. McCardell, B. A., J. M. Madden, and J. T. Stanfield. 1986. Effect of iron concentration on toxin production in Campylo-
18. 19.
20. 21.
2419
bacter jejuni and Campylobacter coli. Can. J. Microbiol. 32: 395-401. Moss, J., D. L. Burns, J. A. Hsia, E. L. Hewlett, R. L. Guerrant, and M. Vaughan. 1984. Cyclic nucleotides: mediators of bacterial toxin action in disease. Ann. Intern. Med. 101:653-666. Ruiz-Palacios, G. M., J. Torres, N. I. Escamilla, B. RuizPalacios, and J. Tamayo. 1983. Cholera-like enterotoxin produced by Campylobacter jejuni: characterization and clinical significance. Lancet ii:250-251. Walker, R. I., M. B. Caldweil, E. C. Lee, P. Guerry, T. J. Trust, and G. M. Ruiz-Palacios. 1986. Pathophysiology of Campylobacter enteritis. Microbiol. Rev. 50:81-94. Yeen, W. P., S. D. Puthucheary, and T. Pang. 1983. Demonstration of a cytotoxin from Campylobacterjejuni. J. Clin. Pathol. 36:1237-1240.