INECriON AND IMMUNITy, May 1975 p. 982-985 Copyright 0 1975 American Society for Microbiology
Vol. 11, No. 5 Printed in U.SA.
Inhibition of the Steroidogenic Effects of Cholera and Heat-Labile Escherichia coli Enterotoxins by GMl Ganglioside: Evidence for a Similar Receptor Site for the Two Toxins SAM T. DONTA* AND JOHN P. VINER
Department of Internal Medicine, University of Iowa and Veterans Administration Hospitals, Iowa City, Iowa 52242 Received for publication 2 December 1974
The effects of three different ganglioside preparations on cholera enterotoxin (CT) and heat-labile Escherichia coli enterotoxin (ECT)-induced steroidogenesis in Y1 and OS, adrenal tumor cells in tissue culture were examined. Only with GM1 ganglioside was any inhibition of the toxins' effects noted. Concentrations of the crude ECT preparation that gave similar morphogenic and steroidogenic effects as CT were inhibited by the same amount or less of GM1 as that required to inhibit the effects of CT. The results of competition experiments also demonstrated that previous incubation of GM1 with one toxin could inhibit the ganglioside's ability to inactivate the other toxin. These findings indicate that at least for Y1 and OS, adrenal tumor cells, GM1 may resemble or be the receptor for both CT and ECT.
Some strains of Escherichia coli produce an enterotoxin that is similar to cholera enterotoxin immunologically and in its mechanisms of action (6, 16). Recent work has established that the membrane receptor site for the purified cholera toxin closely resembles or includes a specific sialidase-resistant monosialosylganglioside referred to as GGnSLC (12) or GMl (1, 10). Although it might be expected from the similarities between the cholera and E. coli enterotoxins that the E. coli toxin would similarly bind and be inactivated by GM, ganglioside, the results of two separate investigations have not shown this to be the case (9, 14); furthermore, the results of studies employing the naturally occurring toxoid of cholera enterotoxin, choleragenoid, have also implied that the receptor sites for the two toxins are not closely related (9, 14). In contrast, however, studies with adrenal tumor cells in tissue culture have failed to demonstrate any inhibitory effects of choleragenoid on cholera enterotoxin-induced steroidogenesis (3, 13). Furthermore, in a preliminary report by us using GM1 ganglioside (7), and in a report by Kwan and Wishnow using mixed gangliosides (13), it was shown that the steroidogenic effects of both cholera and E. coli enterotoxins could be inhibited equally well by the gangliosides employed. This present report details the effects of three different ganglioside preparations on cholera enterotoxin- and E. coli
enterotoxin-induced morphologic and steroidogenic changes in two different clonal lines of adrenal tumor cells in monolayer tissue culture, and provides evidence that, at least for these cells, GM1 ganglioside may resemble or be part of the receptor site for both of the toxins. (A preliminary report of this work was presented at the Annual Meeting of the American Society of Clinical Investigation, Atlantic City, N.J., 5 May 1974.) MATERIALS AND METHODS Purified cholera enterotoxin was obtained from R. A. Finkelstein (8). Dialyzed, lyophilized culture filtrates of E. coli 078:H12 (strain 408-3) were obtained from R. B. Sack (16). Gangliosides SGGnSLC (GDIa), GGnSLC (GMl), and SGGnSSLC (GTl) were obtained from W. E. van Heyningen (12). Dilutions of the toxins and gangliosides were in a 0.05 tris(hydroxymethyl)aminomethane, 0.001 M ethylenediaminetetraacetate buffer system at pH 7.5. The tissue culture methods for the propagation and maintenance of the Y1 and OS, cells and measurements of the M4,3-ketosteroids are as those previously described (4, 6), with the exception that in some of the experiments the cells were grown in a Waymouth medium (2) supplemented with 10% fetal calf serum. For each experiment, equal numbers of cells ('106) were distributed onto plates (60 by 15 mm) (Falcon) and allowed to grow to near stationary-phase growth (4 to 6 days) before use in studies of steroidogenesis. The tissue culture medium was changed prior to the start of each experiment. For experiments in which ganglioside was used, the toxin was preincubated with
VOL. 11, 1975
INHIBITION OF ENTEROTOXINS BY GM, GANGLIOSIDE TABLE 1. Effects of gangliosides on Cholera enterotoxin (CT)- and E. coli enterotoxin (ECT)-induced steroidogenesis
ganglioside at 37 C for 30 min before a portion of this mixture was added to the tissue culture plates. The M4,3-ketosteroids present in the medium were assayed spectrophotometrically and the results were expressed as nanomoles of steroid per plate of cells per incubation time period (6). Some of the variation seen in the amounts of steroid produced between experiments is secondary to either the total incubation time or the density of cells per plate. In addition, some subclonal lines of Y1 and OS, cells have lower basal rates of steroidogenesis than do others.
Control GDla (500 ng/ml)
6.8 ± 0.8 6.4 ± 1.6 6.4 ± 3.4
GT, (500 ng/ml)
GM, (50 ng/ml)
CT (2.5 ng/ml) CT + GDla CT + GT1 CT + GMl
26.0 ± 3.4 25.0 ± 1.6 25.2 ± 2.0
RESULTS 7.2 1.6 Of the three gangliosides tested, only the monosialosylganglioside GGnSLC (GM,) was ca17.8 X 1.4 ECT (5 Ag/ml) pable of inhibiting the morphogenic and steroi17.6 4 0.6 ECT + GDla dogenic effects of both the cholera and heat16.2 1.8 ECT + GT1 labile E. coli enterotoxins. Concentrations of ECT 6.0 1.2 GM, + the gangliosides GD,, and GT, 200 times that (on a Nanomoles of M4,3-ketosteroids per plate of cells a weight basis) of cholera toxin were without effect on either enterotoxin (Table 1). Although per 20 h of incubation period; values given are mean ± not detailed here, inhibition by GM, ganglioside standard deviation of triplicate determinations. of the morphogenic effects of both toxins paralleled that of their steroidogenic effects. Using equipotent concentrations of the two toxins, it could be shown that the toxins' effects were inhibited equally well by the various amounts of I4
GM, ganglioside employed (Fig. 1). Experi-
ments in which different concentrations of cholera enterotoxins were used revealed that approximately five to ten times as much ganglioside, on a weight basis (100 to 200 times on a molar basis), was necessary to fully inhibit the toxin's effects; 50% inhibition was achieved at a 1:1 weight ratio of ganglioside to toxin. Attempts to alter the effects of cholera enterotoxin by preincubating monolayer cultures of Y1 cells with any of the three gangliosides failed to demonstrate any such changes. Table 2 shows this lack of effect of GM, ganglioside-pretreated cells on submaximal steroidogenic concentrations of cholera enterotoxin. OS, adrenal cells differ from Y1 cells in that, whereas the latter respond to both adrenocorticotropin and cyclic 3',5'-adenosine monophosphate, OS, cells respond to cyclic 3',5'-adenosine monophosphate and not to adrenocorticotropin (15). We have previously demonstrated that OS, cells, like Y1 cells, are similarly induced by cholera and the heat-labile E. coli enterotoxins to undergo morphological changes (consisting of cell rounding) and increase their production of A4,3-ketosteroids (4). Again, as with Y1 cells, the effects of both enterotoxins could be inhibited equally well by GM, ganglioside. To determine if both toxins compete for the same binding site on GM1 ganglioside, an experi-
FIG. 1. Inhibition of steroidogenic effects of cholera enterotoxin (0) at 2.5 ng/ml, and E. coli enterotoxin (@) at 5 gg/ml, by G,,v ganglioside. Values shown are the means of five experiments.
ment was designed using an amount of ganglioside that would not result in complete inhibition of the toxins' effects. By preincubating this amount of ganglioside with one of the toxins and then further incubating this mixture with the other toxin, it should be possible to determine if the ganglioside's detoxification was specific for one or the other toxin. The results of such an experiment are shown in Table 3. It can be seen that, when either toxin was preincubated with GM, ganglioside, the ganglioside's ability to detoxify the effects of the other toxin was impeded. When the mixture of ganglioside and toxin was heated at 56 C for 30 min, after being preincubated together, to destroy unbound toxin, and then the second toxin was added to the mixture, again, the ganglioside's detoxifying ability was impaired by the previous incubation of ganglioside and toxin. (Heating alone did not
DONTA AND VINER
TABLE 2. Preincubation of Yl cells with GM1 ganglioside
cholera or heat-labile E. coli toxins to induce morphological or steroidogenic changes.
DISCUSSION The results of these studies support the idea that, at least for Y1 and OS, adrenal tumor cells, the receptor sites for the cholera and heat-labile E. coli enterotoxins are very similar. These findings are not surprising in light of the similarities between the two toxins both immunologically and in their mechanisms of action; they are, however, in contrast to the findings of Pierce (14) and Holmgren (9). These two investigators, using rabbit ileal loops, found that much greater quantities (1,000 times or more) of GM, ganglioside were required to inhibit the secretory effects of a crude E. coli enterotoxin preparation than were required to inhibit cholera toxin's effects. Furthermore, the amount of ganglioside necessary for inhibition was 1/20 to 1/50 (on a weight basis) the amount of cholera enterotoxin, which represents approximately an equimolar ratio between ganglioside and toxin, and which findings provided further evidence for the specificity of the GM, ganglioside for the cholera toxin. Our findings with adrenal tumor cells in tissue culture demonstrate that relatively larger amounts of GM, are necessary to fully inhibit the morphogenic and steroidogenic effects of cholera enterotoxin on these cells (5:1 ratio of ganglioside-toxin on a weight basis), but that these same amounts will equally inhibit morphogenic and steroidogenic concentrations of crude heat-labile E. coli enterotoxin similar to those of the cholera toxin. The simplest explanation that can be offered to explain the discrepancies raised between our results and those of previous studies centers around the use of impure preparations of E. coli enterotoxins in all of these studies. As the ileal loop is 200 to 500 times less sensitive to the effects of the heat-labile E. coli toxin than are the adrenal tumor cells in tissue culture, larger quantities of the crude toxin preparation, and perhaps therefore greater quantities of impurities that could have interfered with the binding of toxin to ganglioside, had to be used for the intestinal studies. Experiments similar to those reported here (e.g., Table 3), in which ganglioside is preincubated with E. coli toxin to see if this mixture would interfere with the ganglioside's cholera detoxifying effects in ileal loop preparations, have not been done or reported, and would be helpful in confirming this hypothesis. It could also be that the receptor sites on Y1 and OS, adrenal cells in tissue culture are
GM, (1 gg/ml)
CT (0.2 ng/ml) CT (0.5 ng/ml)
GM,, then CT (0.2 ng/ml)
+ 4 4
2.7 3.3 3.2 0.7 4.5
18.5 4 5.4 GM,, then CT (0.5 ng/ml) a Cells were preincubated with ganglioside or buffer
for 24 h, then the medium was removed, cells were washed with fresh medium three times, and fresh medium was added to cells i cholera enterotoxin (CT) for a further 20-h incubation. Values are expressed as nanomoles of steroid produced per plate of cells per 20 h standard deviation (triplicate determinations).
TABLE 3. Detoxification of Cholera enterotoxins (C7) and E. coli enterotoxins (EC7) by GM, ganglioside Determinants
Control GM, (10 ng/ml)
CT (2.5 ng/ml) ECT (5 ug/ml)
GM, + CT
15.4 + 1.4 12.0 + 2.4 22.0 4 1.8
GM, + CT, heated' GM, + CT, then ECT GM, + CT, heated,' then ECT GM, + ECT GM, + ECT, heated' GM, + ECT, then CT GM, + ECT, heated,5 then CT
18.2 A 1.8 12.4 + 1.4 24.6 + 2.0 26.2 + 2.6
a Nanomoles of steroid per 24 h ± standard deviation (triplicate determinations). ° Mixture heated at 56 C for 30 min after initial preincubation at 37 C for 30 min.
destroy the ganglioside's ability to inactivate the toxins.) The heat-stable E. coli enterotoxins are not active on Y1 or OS adrenal cells (5), which implies that they have membrane receptor sites that are different from those of cholera and the heat-labile E. coli toxins and/or that their mechanisms of action are different than those of the other two toxins. Experiments in which a crude preparation of heat-stable toxin was preincubated with GM, ganglioside demonstrated no inhibition of the ganglioside's detoxifying effects; similarly, preincubation of Y1 cells with heat-stable toxin did not affect the ability of
VOL. 11, 1975
INHIBITION OF ENTEROTOXINS BY GM, GANGLIOSIDE
We are grateful for the superb technical assistance prodifferent from or more plentiful than those in small bowel intestinal mucosa. Certain tumor vided by Shelley Kreiter. cells in tissue culture contain primarily GM1 LITERATURE CITED ganglioside in their membranes (11), and it might be that this type of ganglioside is also 1. Cuatrecasas, P. 1973. Gangliosides and membrane receptors for cholera toxin. Biochemistry 12:3558-3566. plentiful in Y1 and OS, adrenal tumor cells. In 2. Donta, S. T. 1973. The growth of functional rat glial cells contrast to reports of increased responsiveness in a serumless medium. Exp. Cell Res. 82:119-124. 3. Donta, S. T. 1974. Comparison of the steroidogenic effects to cholera enterotoxin of epididymal fat cells of cholera enterotoxin and ACTH. Am. J. Physiol. preincubated with GM1 ganglioside (1), Y1 adre227:109-113. nal cells show no different responses to cholera 4. Donta, S. T. 1974. Differentiation between the steroidotoxin whether or not preincubated with gangliogenic effects of cholera enterotoxin and ACTH through side. use of a mutant adrenal cell line. J. Infect. Dis. 129:728-731. These results may be interpreted as suggestS. T., H. W. Moon, and S. C. Whipp. 1974. ing that the cultured adrenal cells are naturally 5. Donta, Detection of enterotoxigenic E. coli with the use of already saturated with ganglioside-type recepadrenal cells in tissue culture by cholera enterotoxin. tors and might also help explain the increased Science 183:334-336. sensitivity of adrenal tumor cells to cholera 6. Donta, S. T., and D. M. Smith. 1974. Stimulation of steroidogenesis in tissue culture by enterotoxigenic toxin over all other systems thus far examined. Escherichia coli and its neutralization by specific The possibility that adrenal tumor cells in antiserum. Infect. Immun. 9:500-505. tissue culture may contain different types of 7. Donta, S. T., and J. P. Viner. 1974. Inhibition of the steroidogenic effects of cholera and heat-labile E. coli receptor sites than are present on intestinal enterotoxins by gangliosidq. J. Clin. Invest. 53:20a. mucosa and perhaps fat cells draws support 8. Finkelstein, R. A., and J. J. LoSpalluto. 1970. Production from the observations that choleragenoid was of highly purified choleragen and choleragenoid. J. incapable of blocking the steroidogenic effects Infect. Dis. 121:S63-S72. of cholera enterotoxin (3), and from preliminary 9. Holmgren, J. 1973. Comparison of the tissue receptors for Vibrio cholerae and Escherichia coli enterotoxins by observations that the 25,000-molecular weight means of gangliosides and natural cholera toxoid. piece unique to the toxin (17) is capable of Infect. Immun. 8:851-859. morphogenic and steroidogenic activities, al- 10. Holmgren, J., I. Lonnroth, and L. Svennerholm. 1973. Tissue receptor for cholera exotoxin: postulated structhough the concentrations required to produce ture from studies with GMI ganglioside and related these effects are 10,000 to 100,000 times greater glycolipids. Infect. Immun. 8:208-214. than that of the native toxin (unpublished 11. Keenan, T. W., and D. J. Morre. 1973. Mammary observations). carcinoma: enzymatic block in disialoganglioside biosynthesis. Science 182:935-938. Although the results of our experiments in C. A., and W. E. van Heyningen. 1973. Deactivawhich the toxins competed with each other for 12. King, tion of cholera toxin by a sialidase-resistant monosialothe same binding site on GM1 ganglioside (Table sylganglioside. J. Infect. Dis. 127:639-47. 3) do not conclusively prove that the binding 13. Kwan, C. N., and R. M. Wishnow. 1974. Escherichia coli enterotoxin-induced steroidogenesis in cultured adresites for the cholera and heat-labile enterotoxins nal tumor cells. Infect. Immun. 10:146-151. are identical, they strongly suggest that this is N. F. 1973. Differential inhibitory effects of the case. Further experiments using radioac- 14. Pierce, cholera toxoids and ganglioside on the enterotoxins of tively labeled toxins and more purified preparaVibrio cholerae and Escherichia coli. J. Exp. Med. 137:1009-1023. tions of E. coli enterotoxin would help to resolve 15. Schimmer, B. P. 1969. Phenotypically variant adrenal the questions raised by our findings. tumor cell cultures with biochemical lesions in the ACKNOWLEDGMENTS These studies were supported by grants from the Veterans Administration Research and Education Service and Public Health Service grant AI 11416 from the National Institute of Allergy and Infectious Diseases. S. T. Donta is a Clinical Investigator of the Veterans Administration.
ACTH-stimulated steroidogenic pathway. J. Cell Physiol. 74:115-122. 16. Smith, N. W., and R. B. Sack. 1973. Immunologic cross-reactions of enterotoxin from Escherichia coli and Vibrio cholerae. J. Infect. Dis. 127:164-170. 17. van Heyningen, S. 1974. Cholera toxin: interaction of subunits with ganglioside GM1. Science 183:655-656.