Proc. Nati. Acad. Sci. USA Vol. 73, No. 5, pp. 1679-1683, May 1976
Cell Biology
Cholera toxin interactions with thyrotropin receptors on thyroid plasma membranes (adenylate cyclase/gangliosides/glycoprotein hormones/cooperativity/protein fluorescence)
BRIAN R. MULLIN*,; SALVATORE M. ALOJ*t, PETER H. FISHMAN*, GEORGE LEE*, LEONARD D. KOHN*, AND ROSCOE 0. BRADYT * Section on Biochemistry of Cell Regulation, Laboratory of Biochemical Pharmacology, National Institute of Arthritis, Metabolism, and Digestive Diseases, National Institutes of Health, Bethesda, Maryland 20014, USA; t Centro di Endocrinologia ed Oncologia Sperimentale del C.N.R., Naples, Italy; and t Developmental and Metabolic Neurology Branch, National Institute of Neurological and Communicative Disorders and Stroke, National Institutes of Health, Bethesda, Maryland 20014
Contrbuted by Roscoe 0. Brady, March 9, 1976
ABSTRACT Unlabeled cholera toxin inhibits [125I]thyrotropin binding to thyrotropin receptors on thyroid plasma membranes. Maximal inhibition by cholera toxin does not exceed 40%, whereas unlabeled thyrotropin completely inhibits [125I~thyrotropin binding to these same membranes. Kinetic analyses of the binding data are compatible with the view that the cholera toxin decreases the number of receptor sites available to thyrotropin and that the mechanism by which the cholera toxin inhibits [250Ithyrotropin binding to these receptor sites involves both competitive and noncompetitive elements. Cholera toxin can stimulate adenylate cyclase activity in thyroid plasma membranes. Its effect is not additive with that of thyrotropin, and cholera toxin can inhibit thyrotropin stimulation of the adenylate cyclase activity. NAD enhances cholera toxin stimulation of adenylate cyclase activity but has no enhancing effect on the stimulatory activity exhibited by thyrotropin. The GM1 ganglioside [galactosyl-N-acetylgalactosaminylin thyroid (N-acetylneuraminyl)galactosylglucosylceramideI plasma membranes can be tritiated by treating membranes with galactose oxidase, followed by reduction with 3-H-labeled sodium borohydride. Cholera toxin at a concentration which yields maximal inhibition of thyrotropin binding to thyroid plasma membrane receptors prevents the labeling of GMI. Fluorescence data indicate that the interaction between cholera toxin and GM1 results in a conformational change in the cholera toxin molecule. Analogous conformational alterations cannot be detected upon exposure of cholera toxin to 5-fold higher concentrations of N-acetylgalactosaminyl(N-acetylneuraminyl)-galactosylglucosylceramide (GMS) or N-acetylneuramin-
ylgalactosylglucosylceramide (GM3).
In a previous report (1) we suggested that thyrotropin (TSH) and cholera toxin have an analogous mode of interaction with receptors on thyroid plasma membranes. We hypothesized that in both the case of TSH and cholera toxin, the ,B or B subunit interacts with specific cell surface receptors having gangliosides or ganglioside-like oligosaccharides as part of their core structures. This binding induces a conformational change in the hormone or toxin which results in translocation of an "active" a or A subunit within the cell membrane and adenylate cyclase activation. The evidence which led to this hypothesis can be summarized as follows. Thyrotropin and cholera toxin both bind Abbreviations: TSH, thyrotropin; GM3, N-acetylneuraminylgalactosylglucosylceramide; GM2, N-acetylgalactosaminyl-(N-acetylneuraminyl)-galactosylglucosylceramide; GM,, galactosyl-N-acetylgalac-
to gangliosides in vitro (1-5). The binding of both molecules to the gangliosides is critically affected by the number and location of sialic acid residues on the carbohydrate portion of
the ganglioside molecule (1-5). The interaction of TSH and cholera toxin with gangliosides is associated with a change in their conformational states (1, 6), which in the case of cholera toxin is believed to ultimately result in a dissociation of the toxin molecule (6). Gangliosides which bind best to TSH in vitro, i.e., GD1b [galactosyl-N-acetylgalactosaminyl-(N-acetylneuraminyl-N-acetylneuraminyl)-galactosylglucosylceramide], GTI [N-acetylneuraminylgalactosyl-N -acetylgalactosaminyl-(Nacetylneuraminyl - N - acetylneuraminyl) - galactosylglucosylceramide], and GM1 [galactosyl-N-acetylgalactosaminyl(N-acetylneuraminyl)-galactosylglucosylceramide], are present in thyroid plasma membranes in higher quantities than have been previously found in extraneural tissue (1). A sequence analogy can be demonstrated in the B chain of cholera toxin and the (# subunit of TSH (1, 7), both of which are believed to carry the primary determinants for interactions with receptors on cell membranes (1-5, 8). Also, a sequence analogy exists between the a subunit of TSH and the cholera toxin A protein (7, 9), the subunit believed to be associated with adenylate cyclase activation (5, 6). Implicit in this hypothesis is the requirement that GM1 or a GM1-like structure in thyroid plasma membranes might under certain circumstances interact with either TSH or cholera toxin. Thus, cholera toxin should inhibit TSH binding to thyroid plasma membranes. Further, if TSH and cholera toxin do act through a common receptor containing a GM1 or a GMl-like structure, their ability to stimulate adenylate cyclase activity should not be additive and GM1 should cause cholera toxin to have a detectable conformational change analogous to that which we demonstrated for TSH (1). The present report demonstrates that these requirements are met. In addition, it details a labeling technique which shows that GM, is an integral part of the binding site for cholera toxin on thyroid plasma membranes. Last, the data indicate that the binding of cholera toxin to receptors on thyroid plasma membranes can influence the binding of TSH in a positive or negative fashion and that the inverse is equally valid, i.e., TSH can influence the binding of cholera toxin in an analogous fashion. MATERIALS AND METHODS TSH, [125I]TSH, and thyroid plasma membranes were bovine preparations previously described (11-13). Cholera toxin was obtained from Schwarz/Mann and evaluated by disc gel analysis (10) to insure purity. [1251]TSH binding to plasma membranes was assayed using a filtration technique (1, 13, 14).
tosaminyl-(N-acetyineuraminyl)-galactosylglucosylceramide; GDIa, aminyl)-galactosylglucosylceramide; GD1b, galactosyl-N-acetylgalactosaminyl-(N-acetyineuraminyl-N-acetyineuraminyl)-galactosylglucosylceramide; GT1, N-acetyineuraminylgalactosyl-N-acetylgalactosaminyl-(N-acetylneuraminyl-N-acetylneuraminyl)-galacto-
N-acetylneuraminylgalactosyl-N-acetylgalactosaminyl-(N- acetylneur-
sylglucosylceramide; C.T., cholera toxin. 1679
1680
Cell Biology: Mullin etal.
Proc. Natl. Acad. Sci. USA 73 (1976)
borohydride (specific activity 16 Ci/mmol), 3 mCi in a 200-AI volume of buffer, was added and the incubation at 23' was continued for 30 min with occasional shaking. Membranes were again sedimented by centrifugation at 10,000 X g and washed three times with cold 0.02 M Tris-acetate at pH 6.0. Gangliosz0 m z
0.001
10.0
1.0
0.1
0.01
CONCENTRATION OF UNLABELED TSH OR CHOLERA TOXIN (d0
6M)
FIG. 1. Inhibition of [125j]TSH binding to thyroid plasma membranes by unlabeled TSH or unlabeled cholera toxin. The assays used standard conditions previously described (1). The [125IJTSH concentration was approximately 0.5 nM.
Gangliosides GM3, GM2, GM1, GD1a, GD1b, and GT1 were obtained as previously described (1, 15, 16). Each ganglioside used in these experiments was at least 99% pure following rechromatography, visualization with resorcinol reagent, and densitometric analysis (16); gangliosides were quantitated from their sialic acid content using a micro-modification of the resorcinol method of Svennerholm (17). Gangliosides in bovine thyroid plasma membranes were labeled with tritium using the following procedure. Bovine thyroid plasma membranes (10 mg of membrane protein) were incubated in 0.01 M Tris-acetate at pH 6.0, for 30 min at 0° in the presence or absence of cholera toxin (9 ,M) and in a final volume of 3 ml. The membranes were sedimented by centrifugation at 10,000 X g at 0-4o, washed once with phosphatebuffered saline at pH 7.4, and resuspended to a final volume of 3 ml with phosphate-buffered saline at pH 7.4, which contained 40 units of galactose oxidase (Sigma Chemical Co.). The temperature was raised to 230 and the membranes were incubated for 2 hr with gentle shaking. The membranes were again centrifuged at 10,000 X g, washed with 0.05 M Tris-acetate at pH 7.8 which contained 0.05 M NaCl, and resuspended to a final volume of 3 ml in this same buffer. 3H-Labeled sodium
ides were extracted from the tritiated bovine thyroid membranes and purified by the method of Yu and Ledeen (18), separated by thin-layer chromatography, radioscanned, and visualized with resorcinol reagent, as previously described (1, 16). Fluorescence measurements of cholera toxin solutions were carried out as previously described for TSH solutions (1), with the following exceptions: cholera toxin concentrations are based on colorimetric protein analyses (19) using bovine serum albumin as the standard; N-acetyltryptophan-NH2 solutions were used to monitor the stability of the instrument. Adenylate cyclase assays were performed using the conditions of Rodbell and Krishna (20) as modified by Wolff and Jones (21). Cyclic AMP was measured using a binding protein assay (22). Thyroid plasma membranes for these experiments were prepared in 0.25 M sucrose, 1 mM EDTA, and 0.1% mercaptoethanol to enhance stability of the adenylate cyclase (21, 23). RESULTS The Influence of Cholera Toxin on [125IJSH Binding to Thyroid Plasma Membranes. Unlabeled cholera toxin, as well as unlabeled TSH, can prevent the binding of [125IJTSH to thyroid plasma membranes (Fig. 1). Although there is an initial enhancement of [125I]TSH binding at cholera toxin concentrations between 4 nM and 0.5 IsM, 40% inhibition occurs at a cholera toxin concentration of 2.5 MAM. The same concentration of unlabeled TSH yields only a slightly greater level of inhibition, 45% (Fig. 1). The addition of more unlabeled cholera toxin to the incubation medium causes no further inhibition of [125I]TSH binding, whereas the addition of more unlabeled TSH results in complete inhibition of binding. The 40% level of maximal inhibition correlates with a GM1 content of approximately 25-30% of the total gangliosides present in these membranes (1). An evaluation of the kinetics of this inhibition using a Scatchard analysis (24, 25) indicates that the interaction of cholera toxin with thyroid plasma membranes results in either an in-
z C
0
3
\
+ 1
X
.4
+
\
-_^
\
No Cholera Toxin
'Q
x
0.16fM Cholera Toxin
8
_
c /M oe ° o°:° 3.3 HL M Cholera Toxin °_
10
1
1
X 109 125I-TSH BOUND
1 X 10 -8
0.02
0.04
0.08
0.06
1 / 1251-TSH ADDED (cpm x
1
0-3)
FIG. 2. (A) Scatchard analysis (24,25) of [125IJTSH binding to thyroid plasma membranes in the absence (0) or presence of sufficient cholera toxin to cause the enhancement or inhibition of [1251]TSH binding noted in Fig. 1, 0.16 ,.M (A) or 3.3MgM (0), respectively. Assay conditions were those described [(1) and Fig. 1]. (B) Analysis of analogous data to those in Fig. 2A, but using Lineweaver-Burk plotting techniques (26). This experiment evaluates [125I]TSH binding to plasma membranes as a function of increasing amounts of labeled hormone in the absence (0) or presence of a sufficient amount of unlabeled cholera toxin to cause a 10% level of inhibition (o) or a maximal 40% level of inhibition (0), 0.78 MM and 3.3 MM, respectively (Fig. 1). The inset is a blowup of the extrapolated lines as they cross the origin; the data are plotted by a computer which calculated the best fit lines, using a linear least squares fitting procedure. "C.T." in the inset represents cholera toxin. Assay conditions were the same as Figs. 1 and 2A.
Proc. Natl. Acad. Sci. USA 73 (1976)
Cell Biology: Mullin et al. I
/ LLJ LU
z
Cholera / Toxin
C-) n LU -r
oGM 1
\~~~~
Adenylate cyclase stimulation (pmol of cAMP/13 min per mg of membrane protein)
0
\
~~~~~~ 0
+I
0 30
Table 1. Stimulation of adenylate cyclase activity in thyroid plasma membranes by cholera toxin, TSH, and fluoride, in the absence or presence of gangliosides
}GM2 + 0~~~ GM ~~1.2 ~
/ \
40
LU
1.4
12
50
~~~~~~L
GM1/
01
20 nmole
40 GAN(4 IGLIOSIDE
Effector
11
U-
LU
20
1/
-J
l--Cholera Toxin
\
LU
10i
-
Gangliosides + Gangliosides*
300
340
380
420
WAVELENGTH (nm)
FIG. 3. Fluorescence emission spectra of cholera toxin in the ) or absence of a 7 MM concentration of GM,. The cholera toxin concentration was 0.15 MM in this experiment and in the data depicted in the inset, where the GM1 effect on the emission of cholera toxin at 330 nm was compared to the effects of GM2 and GM3. Excitation was at 282 nm; the sample volume was 1.3 ml. A mixed ganglioside preparation containing 12% GM1, 16% GD1b, 47% GD1a, and 26% GT1 had the exact same effect as 7 MM GM,, suggesting that GD1b, GD1a, and Gri, like GM3 and GM2, have no significant effect on the spectra. In the inset, "C.T." stands for cholera toxin. presence ( ----
crease or decrease in the number of receptor sites available to bind TSH at concentrations of cholera toxin which enhance or inhibit binding, respectively (Fig. 2A). No significant alteration in the affinity of the sites for TSH is detectable in these analyses. A kinetic evaluation (Fig. 2B) which adapts Lineweaver-Burk plotting techniques (26) indicates that the mechanism by which the cholera toxin displaces [1251]TSH from the receptor sites appears to be "mixed," i.e., has both a competitive and noncompetitive element (27).
The Influence of Gangliosides on the Structure of Cholera Toxin. If GM1 is a structural component of the membrane which can interact with both TSH and cholera toxin, and if the mode of action of TSH and cholera toxin is similar, GM1 should induce in cholera toxin a conformational alteration which is both analogous to that imposed upon TSH and can be detectable by physical techniques such as fluorescence spectroscopy. The fluorescence emission spectrum of cholera toxin (Fig. 3) in 0.02 M Tris-acetate at pH 7 shows a maximum centered at 342 nm which is typical of proteins containing tryptophan (28). Upon addition of GM1 (7 ,uM), the peak emission is shifted to 330 nm. The 12-nm "blue shift" reflects a substantial decrease of the polarity of the indole chromophore environment which is likely to result from a conformational change of the protein. As the concentration of GM1 is increased, there is no further change in the emission maximum, although the relative intensity at 330 nm increases (Fig. 3 inset). The effect of GM1 on the conformation of cholera toxin appears to be specific; thus, the gangliosides GM2 and GM3 at equivalent or up to 5-fold higher concentrations have no effect on the emission maximum of the protein, which remains at 342 nm, and cause only a moderate increase of relative emission intensity at all wavelengths between 300 nm and 400 nm. Effect of Cholera Toxin on the Adenylate Cyclase Activity of Thyroid Plasma Membranes. In Table 1, cholera toxin is shown to stimulate adenylate cyclase activity in thyroid plasma membranes. The stimulation of adenylate cyclase by cholera toxin is significantly less than that exhibited by TSH, and at high concentrations of the two effectors there is no additive stimulation. The inhibitory effect of gangliosides on the stimulation
17 330 9 17 Not done
85 654
None Fluoride Cholera toxint
131 239 165
TSHt \
1681
Cholera toxin + TSHt * The gangliosides were a mixed preparation containing 12% GM1, 16% GDlb, 47% GDia, and 26% GT1; the concentration was approximately 0.9 mM. t The concentrations of TSH and cholera toxin 6.8 MM, respectively.
were
7.4 MM and
of the cyclase activity (Table 1) is predictable from the work already presented (above and ref. 1). Not anticipated is the inhibitory effect on the fluoride-stimulated and basal cyclase activities (Table 1). Gill has suggested that NAD is a necessary cofactor for the cholera toxin-induced stimulation of adenylate cyclase in plasma membrane preparations (29). In accord with his results, Fig. 4 shows that NAD significantly enhances the ability of cholera toxin to stimulate adenylate cyclase activity in thyroid plasma membranes, whereas NAD can actually inhibit the ability of TSH to stimulate adenylate cyclase activity in these same thyroid membranes. This finding suggests that despite the fact that both cholera toxin and TSH appear to have an analogous mode of interaction at a receptor level, they diverge in the molecular mechanism of adenylate cyclase stimulation. Effect of Cholera Toxin on the Labeling of Gangliosides in Thyroid Plasma Membranes. The terminal galactosyl residue in the carbohydrate moiety of gangliosides and glycoprotein membrane components is susceptible to galactose oxidase (30, 31). Thus, thyroid membranes incubated with galactose oxidase, and then exposed to sodium [3H]borohydride
0
0.
HE
600 -S TSH 500
0400
JR >-')
Cholera Toxin f
4Z
300 -
+ NAD
/
T~~SH W
200
a-
100
~
/
>
/
+
NAD
Cholera Toxin E
g
L-Y
0.0
----r~-
-,
7
7
7
0.1 10 1.0 CONCENTRATION OF TSH OR CHOLERA TOXIN (IM)
FIG. 4. The effect of NAD (1 mM) on the ability of cholera toxin and TSH to stimulate adenylate cyclase activity in thyroid plasma membranes. Membrane protein in these assays was approximately 176 Mg in a 50 Ml assay volume. The time chosen for incubation was 13 min. Assays were performed as described in Materials and Methods.
1682
Cell Biology: Mullin et al.
Proc. Nat!. Acad. Sci. USA 73 (1976)
THYROID PLASMA MEMBRANES
A
GM3
zi
0 cc I-
z 0
GM2
U
0
GM1
0
z 0
l-
G aria
STANDARD
TOP THYROID PLASMA MEMBRANES
THYROID PLASMA MEMBRANES
B
+
CHOLERA TOXIN
J It1t t ID~bGM2 GT1
GDlb
GM2
GM3 GM3
FIG. 5. (A) Tritium incorporation into the gangliosides of thyroid plasma membranes after exposure to galactose oxidase and subsequent reduction by 3H-labeled sodium borohydride. The gangliosides were subjected to thin-layer chromatography (picture), and the plates were scanned using a Vanguard radioscanning instrument (graph). Simultaneous chromatography of authentic GM1, GT1, GD1b, GM3, and GM2 allow the definition of the peaks of radioactivity. (B) Tritium incorporation into the gangliosides of thyroid plasma membranes preincubated with cholera toxin prior to galactose oxidase exposure and 3H-labeled sodium borohydride reduction. The concentration of cholera toxin during the preincubation period was sufficient to give maximal (40%) inhibition of [1251]TSH binding (Fig. 1). Total radioactivity applied to the plates in (A) and (B), respectively, was 176,000 and 104,000 counts per min. The radioscanning sensitivity in (A) and (B) is identical.
should have tritium incorporated into ganglioside residues if these are available to the enzyme. As seen in Fig. 5 (top), the major labeled ganglioside is GM1 when gangliosides are extracted from these membranes, chromatographed, and evaluated for their tritium content. In contrast, Fig. 5 (bottom) shows that the GM1 residues are not labeled with tritium if the thyroid membranes are preincubated with cholera toxin prior to galactose oxidase exposure and [3H]borohydride reduction. In addition, pretreatment with cholera toxin causes an enhancement of the labeling of other glycolipids contained in the membrane (Fig. 5, bottom), and there is no labeling of GM1 in the absence of galactose oxidase (data not shown).
IQ
1.0 2.0 UNLABELED TSH
3.0
WxO6M)
0.2 0.4 0.6 0.8 1.0 UNLABELED CHOLERA TOXIN (xO14M)
FIG. 6. (A) Inhibition of [125I]TSH binding to thyroid plasma membranes by unlabeled TSH (0) and by unlabeled cholera toxin (3) in the presence of sufficient unlabeled TSH, 0.5 MAM, to yield approximately 20% inhibition of [125I]TSH binding (arrow). Conditions are those of Fig. 1. (B) Inhibition of [125IJTSH binding to thyroid plasma membranes by unlabeled cholera toxin (3) and by unlabeled TSH (0) in the presence of sufficient cholera toxin to neither enhance nor inhibit [125I]TSH binding (arrow). Conditions are those of Figs. 6A and 1. The dotted line depicts the theoretical curve if unlabeled TSH were to inhibit [125I]TSH binding in a simple additive manner.
The Interaction of Receptor Components on Thyroid Plasma Membranes. The data show that low concentrations of cholera toxin enhance [125I]TSH binding to thyroid plasma membranes, and suggest the possibility that there is a cooperative relationship between the individual receptors. This cooperative relationship has been confirmed in Fig. 6 which compares the inhibition of [125I]TSH binding to the receptor by unlabeled cholera toxin alone; by unlabeled TSH alone; by unlabeled TSH plus an amount of unlabeled cholera toxin, 0.6 MuM, sufficient to yield no significant net inhibition or enhancement; and by unlabeled cholera toxin plus an amount of unlabeled TSH sufficient to cause only 20% inhibition. As can be seen (Fig. 6A), prior exposure of the membranes to a low concentration of TSH totally abolishes the enhancement effect of cholera toxin on [125I]TSH binding (Figs. 1 or 6B). Similarly (Fig. 6B), prior exposure of the membranes to 0.6 MM cholera toxin results in an increased ability of unlabeled TSH to inhibit binding. Thus, whereas 0.1 MAM TSH yields an inhibition of 15% when present alone, 0.1 MiM TSH results in a 45% inhibition in the presence of a concentration of cholera toxin which alone has no net effect on [l25I]TSH binding. DISCUSSION In our previous report, we showed that the efficacy by which gangliosides inhibited the interaction of TSH with thyroid plasma membrane receptors was GD1b > GT1 > GM1 and we suggested all three gangliosides could be elements or analogs of the TSH receptor, since these gangliosides coexist on thyroid plasma membranes. In the present report, we show that cholera toxin interacts with GM1 residues on thyroid plasma membranes, i.e., at 9 MiM cholera toxin, GM1 is completely protected from galactose oxidase attack (Fig. 5). In addition, we show that low concentrations of cholera toxin can enhance [125I]TSH binding to thyroid plasma membranes and can cause a dramatically increased ability at low concentrations of unlabeled TSH to inhibit [125I]TSH binding (Fig. 6B). These results could be accounted for by the following schema: TSH binds preferentially to receptor sites composed of GD1b or GT1 residues or
Cell Biology: Mullin et al. their glycoprotein analogs, whereas cholera toxin binds preferentially to GM1 or GM1-like receptors. Reactivity at low concentrations of cholera toxin with GM1 results in a perturbation of the membrane which makes the GD1b or GT1 sites more accessible to TSH in the medium. The toxin thus behaves as a positively cooperative ligand, i.e., it increases the availability of the receptors for the hormone. The implication of these data and this schema is that ganglioside or ganglioside-like receptors on thyroid plasma membranes do not interact with TSH in an independent fashion but rather in a cooperative mode. Further, the cooperativity may become negative under normal circumstances, i.e., faced with high concentrations of TSH interacting with the GM1 sites, the membrane perturbation can alter the environment of GD1b- or GTI-like sites already containing TSH and thereby enhance dissociation from these previously. occupied sites. Such a phenomenon could be the basis for the curvilinear Scatchard plots and "negative cooperativity" among TSH receptors which we previously described (13). By its interaction with the GM1 ganglioside, cholera toxin may be subverting a normal mechanism by which TSH interacts with the thyroid plasma membrane and transfers its message to the cell machinery. The present data, which show that cholera toxin undergoes a conformational transition upon interacting with GM1, are both in accord with this hypothesis and suggest that this conformational transition may well be a necessary prelude to the formation of an "active" A subunit of cholera toxin capable of interacting with components within the cell membrane. It is in this last respect that the adenylate cyclase data presented are pertinent. Cholera toxin stimulation of adenylate cyclase activity in thyroid plasma membranes requires NAD; TSH stimulation does not. A divergent molecular mechanism for cyclase activation must thus be anticipated, despite the analogies at the receptor level and at the level of effector penetration or translocation within the membrane. It is pertinent at this point to recall the molecular mechanism by which diphtheria toxin subverts the normal protein synthetic machinery of the cell, i.e., through the formation of an abnormal covalent linkage of NAD to elongation factor-2 (32). These results thus enhance the importance of the efforts by Gill (29) who has already initiated a search for a membrane component which accepts NAD in covalent linkage upon the interaction of cholera toxin with the membrane. A final point should be made concerning the data of Table 1. The ability of gangliosides to inhibit both the basal- and fluoride-stimulated adenylate cyclase activity has at this time no definitive explanation. These results are not caused by the gangliosides interfering with the cyclic AMP binding protein assay. A similar inhibition of basal adenylate cyclase activity by gangliosides has been observed in cultured mouse cells (33). Since exogenous gangliosides can be integrated into intact cells (34) and membranes (2), integration of gangliosides into the fluid phase of the membrane may perturb the membranebound adenylate cyclase activity.
Proc. Natl. Acad. Sd. USA 73 (1976)
16&3
1. Mullin, B. R., Fishman, P. H., Lee, G., Aloj, S. M., Ledley, F. D., Winand, R. J., Kohn, L. D. & Brady, R. 0. (1976) Proc. Nati. Acad. Sci. USA 73,842-846. 2. Cuatrecasas, P. (1973) Biochemistry 12,3547-3558. 3. Holmgren, J., LUnnroth, I. & Svennerholm, L. (1973) Infect. Immun. 8,208-214. 4. King, C. A. & van Heyningren, W. E. (1973) J. Infect. Dis. 127,
638-647. 5. Gill, D. M. & King, C. A. (1975) J. Biol. Chem. 250, 6424-6432. 6. Sahyoun, N. & Cuatrecasas, P. (19X6) Proc. Nati. Acad. Sci. USA
72,3438-3442. 7. Ledley, F. D., Mullin, B. R., Lee, G., Aloj, S. M., Fishman, P. H., Hunt, L. T., Dayhoff, M. 0. & Kohn, L. D. (1976) Biochem. Biophys. Res. Commun., in press. 8. Wolff, J., Winand, R. J. & Kohn, L. D. (1974) Proc. Natl. Acad. Sci. USA 71, 3460-3464. 9. Mendez, E., Lai, C. Y. & Wodnar-Filipowicz, A. (1975) Biochem. Biophys. Res. Commun. 67,1435-1443. 10. Laeinmli, U. K. (1970) Nature 227,680-682. 11. Winand, R. J. & Kohn, L. D. (1970) J. Biol. Chem. 245,967-975. 12. Kohn, L. D. & Winand, R. J. (1971) J. Biol. Chem. 246,65706575. 13. Tate, R. L., Schwartz, H. I., Holmes, J. M., Kohn, L. D. & Winand, R. J. (1975) J. Biol. Chem. 250, 6509-6515. 14. Amir, S. M., Carraway, T. F., Jr., Kohn, L. D. & Winand, R. J. (1972) J. Biol. Chem. 248, 4092-4100. 15. Fishman, P. H., McFarland, V. W., Mora, P. T. & Brady, R. 0. (1972) Biochem. Biophys. Res. Commun. 48,48-57. 16. Fishman, P. H., Brady, R. O., Bradley, R. M., Aaronson, S. A. & Todaro, G. J. (1974) Proc. Natt. Acad. Sci. USA 71, 298-301. 17. Svennerholm, L. (1957) Biochim. Biophys. Acta 24, 604-615. 18. Yu, R. K. & Ledeen, R. W. (1972) J. Lipid Res. 13,680-686. 19. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193,265-275. 20. Rodbell, M. & Krishna, G. (1974) in Methods in Enzymology, eds. Fleischer, S. & Packer, L. (Academic Press, New York), Vol. 31, part A, pp. 103-115. 21. Wolff, J. & Jones, A. B. (1971) J. Biol. Chem. 246,3939-3947. 22. Gilman, A. C. (1970) Proc. Natl. Acad. Sci. USA 67,305-310. 23. Winand, R. J. & Kohn, L. D. (1975) J. Biol. Chem. 250,65226526. 24. Scatchard, G. (1949) Ann. N.Y. Acad. Sci. 51, 660-674. 25. Klotz, I. M. & Hinston, D. L. (1971) Biochemistry 10,33065-309. 26. Lineweaver, H. & Burk, D. (1934) J. Am. Chem. Soc. 56,658666. 27. Webb, J. L. (1963) "General principles of inhibition," in Enzyme and Metabolic Inhibitors (Academic Press, New York), Vol. 1, p. 162. 28. Teale, F. W. J. (1960) Biochem. J. 76, 381-389. 29. Gill, D. M. (1975) Proc. Nati. Acad. Sci. USA 72,2064-2068. 30. Morell, A. G., Van Den Hamer, C. J. A., Scheinberg, I. H. & Ashwell, G. (1966) J. Biol. Chem. 241, 3745-3749. 31. Gahmberg, C. G. & Hakomori, S. (1973) J. Biol. Chem. 248, 4311-4317. 32. Robinson, E. A., Henriksen, 0. & Maxwell, E. S. (1974) J. Biol. Chem. 249,5088-5093. 33. Fishman, P. H., Moss, J. & Vaughan, M. (1976) J. Biol. Chem. 251, in press. 34. Moss, J., Fishman, P. H., Magianello, V. C., Vaughan, M. & Brady, R. 0. (1976) Proc. Natl. Acad. Sci. USA 73, in press.
Cell Biology
Cholera toxin interactions with thyrotropin receptors on thyroid plasma membranes (adenylate cyclase/gangliosides/glycoprotein hormones/cooperativity/protein fluorescence)
BRIAN R. MULLIN*,; SALVATORE M. ALOJ*t, PETER H. FISHMAN*, GEORGE LEE*, LEONARD D. KOHN*, AND ROSCOE 0. BRADYT * Section on Biochemistry of Cell Regulation, Laboratory of Biochemical Pharmacology, National Institute of Arthritis, Metabolism, and Digestive Diseases, National Institutes of Health, Bethesda, Maryland 20014, USA; t Centro di Endocrinologia ed Oncologia Sperimentale del C.N.R., Naples, Italy; and t Developmental and Metabolic Neurology Branch, National Institute of Neurological and Communicative Disorders and Stroke, National Institutes of Health, Bethesda, Maryland 20014
Contrbuted by Roscoe 0. Brady, March 9, 1976
ABSTRACT Unlabeled cholera toxin inhibits [125I]thyrotropin binding to thyrotropin receptors on thyroid plasma membranes. Maximal inhibition by cholera toxin does not exceed 40%, whereas unlabeled thyrotropin completely inhibits [125I~thyrotropin binding to these same membranes. Kinetic analyses of the binding data are compatible with the view that the cholera toxin decreases the number of receptor sites available to thyrotropin and that the mechanism by which the cholera toxin inhibits [250Ithyrotropin binding to these receptor sites involves both competitive and noncompetitive elements. Cholera toxin can stimulate adenylate cyclase activity in thyroid plasma membranes. Its effect is not additive with that of thyrotropin, and cholera toxin can inhibit thyrotropin stimulation of the adenylate cyclase activity. NAD enhances cholera toxin stimulation of adenylate cyclase activity but has no enhancing effect on the stimulatory activity exhibited by thyrotropin. The GM1 ganglioside [galactosyl-N-acetylgalactosaminylin thyroid (N-acetylneuraminyl)galactosylglucosylceramideI plasma membranes can be tritiated by treating membranes with galactose oxidase, followed by reduction with 3-H-labeled sodium borohydride. Cholera toxin at a concentration which yields maximal inhibition of thyrotropin binding to thyroid plasma membrane receptors prevents the labeling of GMI. Fluorescence data indicate that the interaction between cholera toxin and GM1 results in a conformational change in the cholera toxin molecule. Analogous conformational alterations cannot be detected upon exposure of cholera toxin to 5-fold higher concentrations of N-acetylgalactosaminyl(N-acetylneuraminyl)-galactosylglucosylceramide (GMS) or N-acetylneuramin-
ylgalactosylglucosylceramide (GM3).
In a previous report (1) we suggested that thyrotropin (TSH) and cholera toxin have an analogous mode of interaction with receptors on thyroid plasma membranes. We hypothesized that in both the case of TSH and cholera toxin, the ,B or B subunit interacts with specific cell surface receptors having gangliosides or ganglioside-like oligosaccharides as part of their core structures. This binding induces a conformational change in the hormone or toxin which results in translocation of an "active" a or A subunit within the cell membrane and adenylate cyclase activation. The evidence which led to this hypothesis can be summarized as follows. Thyrotropin and cholera toxin both bind Abbreviations: TSH, thyrotropin; GM3, N-acetylneuraminylgalactosylglucosylceramide; GM2, N-acetylgalactosaminyl-(N-acetylneuraminyl)-galactosylglucosylceramide; GM,, galactosyl-N-acetylgalac-
to gangliosides in vitro (1-5). The binding of both molecules to the gangliosides is critically affected by the number and location of sialic acid residues on the carbohydrate portion of
the ganglioside molecule (1-5). The interaction of TSH and cholera toxin with gangliosides is associated with a change in their conformational states (1, 6), which in the case of cholera toxin is believed to ultimately result in a dissociation of the toxin molecule (6). Gangliosides which bind best to TSH in vitro, i.e., GD1b [galactosyl-N-acetylgalactosaminyl-(N-acetylneuraminyl-N-acetylneuraminyl)-galactosylglucosylceramide], GTI [N-acetylneuraminylgalactosyl-N -acetylgalactosaminyl-(Nacetylneuraminyl - N - acetylneuraminyl) - galactosylglucosylceramide], and GM1 [galactosyl-N-acetylgalactosaminyl(N-acetylneuraminyl)-galactosylglucosylceramide], are present in thyroid plasma membranes in higher quantities than have been previously found in extraneural tissue (1). A sequence analogy can be demonstrated in the B chain of cholera toxin and the (# subunit of TSH (1, 7), both of which are believed to carry the primary determinants for interactions with receptors on cell membranes (1-5, 8). Also, a sequence analogy exists between the a subunit of TSH and the cholera toxin A protein (7, 9), the subunit believed to be associated with adenylate cyclase activation (5, 6). Implicit in this hypothesis is the requirement that GM1 or a GM1-like structure in thyroid plasma membranes might under certain circumstances interact with either TSH or cholera toxin. Thus, cholera toxin should inhibit TSH binding to thyroid plasma membranes. Further, if TSH and cholera toxin do act through a common receptor containing a GM1 or a GMl-like structure, their ability to stimulate adenylate cyclase activity should not be additive and GM1 should cause cholera toxin to have a detectable conformational change analogous to that which we demonstrated for TSH (1). The present report demonstrates that these requirements are met. In addition, it details a labeling technique which shows that GM, is an integral part of the binding site for cholera toxin on thyroid plasma membranes. Last, the data indicate that the binding of cholera toxin to receptors on thyroid plasma membranes can influence the binding of TSH in a positive or negative fashion and that the inverse is equally valid, i.e., TSH can influence the binding of cholera toxin in an analogous fashion. MATERIALS AND METHODS TSH, [125I]TSH, and thyroid plasma membranes were bovine preparations previously described (11-13). Cholera toxin was obtained from Schwarz/Mann and evaluated by disc gel analysis (10) to insure purity. [1251]TSH binding to plasma membranes was assayed using a filtration technique (1, 13, 14).
tosaminyl-(N-acetyineuraminyl)-galactosylglucosylceramide; GDIa, aminyl)-galactosylglucosylceramide; GD1b, galactosyl-N-acetylgalactosaminyl-(N-acetyineuraminyl-N-acetyineuraminyl)-galactosylglucosylceramide; GT1, N-acetyineuraminylgalactosyl-N-acetylgalactosaminyl-(N-acetylneuraminyl-N-acetylneuraminyl)-galacto-
N-acetylneuraminylgalactosyl-N-acetylgalactosaminyl-(N- acetylneur-
sylglucosylceramide; C.T., cholera toxin. 1679
1680
Cell Biology: Mullin etal.
Proc. Natl. Acad. Sci. USA 73 (1976)
borohydride (specific activity 16 Ci/mmol), 3 mCi in a 200-AI volume of buffer, was added and the incubation at 23' was continued for 30 min with occasional shaking. Membranes were again sedimented by centrifugation at 10,000 X g and washed three times with cold 0.02 M Tris-acetate at pH 6.0. Gangliosz0 m z
0.001
10.0
1.0
0.1
0.01
CONCENTRATION OF UNLABELED TSH OR CHOLERA TOXIN (d0
6M)
FIG. 1. Inhibition of [125j]TSH binding to thyroid plasma membranes by unlabeled TSH or unlabeled cholera toxin. The assays used standard conditions previously described (1). The [125IJTSH concentration was approximately 0.5 nM.
Gangliosides GM3, GM2, GM1, GD1a, GD1b, and GT1 were obtained as previously described (1, 15, 16). Each ganglioside used in these experiments was at least 99% pure following rechromatography, visualization with resorcinol reagent, and densitometric analysis (16); gangliosides were quantitated from their sialic acid content using a micro-modification of the resorcinol method of Svennerholm (17). Gangliosides in bovine thyroid plasma membranes were labeled with tritium using the following procedure. Bovine thyroid plasma membranes (10 mg of membrane protein) were incubated in 0.01 M Tris-acetate at pH 6.0, for 30 min at 0° in the presence or absence of cholera toxin (9 ,M) and in a final volume of 3 ml. The membranes were sedimented by centrifugation at 10,000 X g at 0-4o, washed once with phosphatebuffered saline at pH 7.4, and resuspended to a final volume of 3 ml with phosphate-buffered saline at pH 7.4, which contained 40 units of galactose oxidase (Sigma Chemical Co.). The temperature was raised to 230 and the membranes were incubated for 2 hr with gentle shaking. The membranes were again centrifuged at 10,000 X g, washed with 0.05 M Tris-acetate at pH 7.8 which contained 0.05 M NaCl, and resuspended to a final volume of 3 ml in this same buffer. 3H-Labeled sodium
ides were extracted from the tritiated bovine thyroid membranes and purified by the method of Yu and Ledeen (18), separated by thin-layer chromatography, radioscanned, and visualized with resorcinol reagent, as previously described (1, 16). Fluorescence measurements of cholera toxin solutions were carried out as previously described for TSH solutions (1), with the following exceptions: cholera toxin concentrations are based on colorimetric protein analyses (19) using bovine serum albumin as the standard; N-acetyltryptophan-NH2 solutions were used to monitor the stability of the instrument. Adenylate cyclase assays were performed using the conditions of Rodbell and Krishna (20) as modified by Wolff and Jones (21). Cyclic AMP was measured using a binding protein assay (22). Thyroid plasma membranes for these experiments were prepared in 0.25 M sucrose, 1 mM EDTA, and 0.1% mercaptoethanol to enhance stability of the adenylate cyclase (21, 23). RESULTS The Influence of Cholera Toxin on [125IJSH Binding to Thyroid Plasma Membranes. Unlabeled cholera toxin, as well as unlabeled TSH, can prevent the binding of [125IJTSH to thyroid plasma membranes (Fig. 1). Although there is an initial enhancement of [125I]TSH binding at cholera toxin concentrations between 4 nM and 0.5 IsM, 40% inhibition occurs at a cholera toxin concentration of 2.5 MAM. The same concentration of unlabeled TSH yields only a slightly greater level of inhibition, 45% (Fig. 1). The addition of more unlabeled cholera toxin to the incubation medium causes no further inhibition of [125I]TSH binding, whereas the addition of more unlabeled TSH results in complete inhibition of binding. The 40% level of maximal inhibition correlates with a GM1 content of approximately 25-30% of the total gangliosides present in these membranes (1). An evaluation of the kinetics of this inhibition using a Scatchard analysis (24, 25) indicates that the interaction of cholera toxin with thyroid plasma membranes results in either an in-
z C
0
3
\
+ 1
X
.4
+
\
-_^
\
No Cholera Toxin
'Q
x
0.16fM Cholera Toxin
8
_
c /M oe ° o°:° 3.3 HL M Cholera Toxin °_
10
1
1
X 109 125I-TSH BOUND
1 X 10 -8
0.02
0.04
0.08
0.06
1 / 1251-TSH ADDED (cpm x
1
0-3)
FIG. 2. (A) Scatchard analysis (24,25) of [125IJTSH binding to thyroid plasma membranes in the absence (0) or presence of sufficient cholera toxin to cause the enhancement or inhibition of [1251]TSH binding noted in Fig. 1, 0.16 ,.M (A) or 3.3MgM (0), respectively. Assay conditions were those described [(1) and Fig. 1]. (B) Analysis of analogous data to those in Fig. 2A, but using Lineweaver-Burk plotting techniques (26). This experiment evaluates [125I]TSH binding to plasma membranes as a function of increasing amounts of labeled hormone in the absence (0) or presence of a sufficient amount of unlabeled cholera toxin to cause a 10% level of inhibition (o) or a maximal 40% level of inhibition (0), 0.78 MM and 3.3 MM, respectively (Fig. 1). The inset is a blowup of the extrapolated lines as they cross the origin; the data are plotted by a computer which calculated the best fit lines, using a linear least squares fitting procedure. "C.T." in the inset represents cholera toxin. Assay conditions were the same as Figs. 1 and 2A.
Proc. Natl. Acad. Sci. USA 73 (1976)
Cell Biology: Mullin et al. I
/ LLJ LU
z
Cholera / Toxin
C-) n LU -r
oGM 1
\~~~~
Adenylate cyclase stimulation (pmol of cAMP/13 min per mg of membrane protein)
0
\
~~~~~~ 0
+I
0 30
Table 1. Stimulation of adenylate cyclase activity in thyroid plasma membranes by cholera toxin, TSH, and fluoride, in the absence or presence of gangliosides
}GM2 + 0~~~ GM ~~1.2 ~
/ \
40
LU
1.4
12
50
~~~~~~L
GM1/
01
20 nmole
40 GAN(4 IGLIOSIDE
Effector
11
U-
LU
20
1/
-J
l--Cholera Toxin
\
LU
10i
-
Gangliosides + Gangliosides*
300
340
380
420
WAVELENGTH (nm)
FIG. 3. Fluorescence emission spectra of cholera toxin in the ) or absence of a 7 MM concentration of GM,. The cholera toxin concentration was 0.15 MM in this experiment and in the data depicted in the inset, where the GM1 effect on the emission of cholera toxin at 330 nm was compared to the effects of GM2 and GM3. Excitation was at 282 nm; the sample volume was 1.3 ml. A mixed ganglioside preparation containing 12% GM1, 16% GD1b, 47% GD1a, and 26% GT1 had the exact same effect as 7 MM GM,, suggesting that GD1b, GD1a, and Gri, like GM3 and GM2, have no significant effect on the spectra. In the inset, "C.T." stands for cholera toxin. presence ( ----
crease or decrease in the number of receptor sites available to bind TSH at concentrations of cholera toxin which enhance or inhibit binding, respectively (Fig. 2A). No significant alteration in the affinity of the sites for TSH is detectable in these analyses. A kinetic evaluation (Fig. 2B) which adapts Lineweaver-Burk plotting techniques (26) indicates that the mechanism by which the cholera toxin displaces [1251]TSH from the receptor sites appears to be "mixed," i.e., has both a competitive and noncompetitive element (27).
The Influence of Gangliosides on the Structure of Cholera Toxin. If GM1 is a structural component of the membrane which can interact with both TSH and cholera toxin, and if the mode of action of TSH and cholera toxin is similar, GM1 should induce in cholera toxin a conformational alteration which is both analogous to that imposed upon TSH and can be detectable by physical techniques such as fluorescence spectroscopy. The fluorescence emission spectrum of cholera toxin (Fig. 3) in 0.02 M Tris-acetate at pH 7 shows a maximum centered at 342 nm which is typical of proteins containing tryptophan (28). Upon addition of GM1 (7 ,uM), the peak emission is shifted to 330 nm. The 12-nm "blue shift" reflects a substantial decrease of the polarity of the indole chromophore environment which is likely to result from a conformational change of the protein. As the concentration of GM1 is increased, there is no further change in the emission maximum, although the relative intensity at 330 nm increases (Fig. 3 inset). The effect of GM1 on the conformation of cholera toxin appears to be specific; thus, the gangliosides GM2 and GM3 at equivalent or up to 5-fold higher concentrations have no effect on the emission maximum of the protein, which remains at 342 nm, and cause only a moderate increase of relative emission intensity at all wavelengths between 300 nm and 400 nm. Effect of Cholera Toxin on the Adenylate Cyclase Activity of Thyroid Plasma Membranes. In Table 1, cholera toxin is shown to stimulate adenylate cyclase activity in thyroid plasma membranes. The stimulation of adenylate cyclase by cholera toxin is significantly less than that exhibited by TSH, and at high concentrations of the two effectors there is no additive stimulation. The inhibitory effect of gangliosides on the stimulation
17 330 9 17 Not done
85 654
None Fluoride Cholera toxint
131 239 165
TSHt \
1681
Cholera toxin + TSHt * The gangliosides were a mixed preparation containing 12% GM1, 16% GDlb, 47% GDia, and 26% GT1; the concentration was approximately 0.9 mM. t The concentrations of TSH and cholera toxin 6.8 MM, respectively.
were
7.4 MM and
of the cyclase activity (Table 1) is predictable from the work already presented (above and ref. 1). Not anticipated is the inhibitory effect on the fluoride-stimulated and basal cyclase activities (Table 1). Gill has suggested that NAD is a necessary cofactor for the cholera toxin-induced stimulation of adenylate cyclase in plasma membrane preparations (29). In accord with his results, Fig. 4 shows that NAD significantly enhances the ability of cholera toxin to stimulate adenylate cyclase activity in thyroid plasma membranes, whereas NAD can actually inhibit the ability of TSH to stimulate adenylate cyclase activity in these same thyroid membranes. This finding suggests that despite the fact that both cholera toxin and TSH appear to have an analogous mode of interaction at a receptor level, they diverge in the molecular mechanism of adenylate cyclase stimulation. Effect of Cholera Toxin on the Labeling of Gangliosides in Thyroid Plasma Membranes. The terminal galactosyl residue in the carbohydrate moiety of gangliosides and glycoprotein membrane components is susceptible to galactose oxidase (30, 31). Thus, thyroid membranes incubated with galactose oxidase, and then exposed to sodium [3H]borohydride
0
0.
HE
600 -S TSH 500
0400
JR >-')
Cholera Toxin f
4Z
300 -
+ NAD
/
T~~SH W
200
a-
100
~
/
>
/
+
NAD
Cholera Toxin E
g
L-Y
0.0
----r~-
-,
7
7
7
0.1 10 1.0 CONCENTRATION OF TSH OR CHOLERA TOXIN (IM)
FIG. 4. The effect of NAD (1 mM) on the ability of cholera toxin and TSH to stimulate adenylate cyclase activity in thyroid plasma membranes. Membrane protein in these assays was approximately 176 Mg in a 50 Ml assay volume. The time chosen for incubation was 13 min. Assays were performed as described in Materials and Methods.
1682
Cell Biology: Mullin et al.
Proc. Nat!. Acad. Sci. USA 73 (1976)
THYROID PLASMA MEMBRANES
A
GM3
zi
0 cc I-
z 0
GM2
U
0
GM1
0
z 0
l-
G aria
STANDARD
TOP THYROID PLASMA MEMBRANES
THYROID PLASMA MEMBRANES
B
+
CHOLERA TOXIN
J It1t t ID~bGM2 GT1
GDlb
GM2
GM3 GM3
FIG. 5. (A) Tritium incorporation into the gangliosides of thyroid plasma membranes after exposure to galactose oxidase and subsequent reduction by 3H-labeled sodium borohydride. The gangliosides were subjected to thin-layer chromatography (picture), and the plates were scanned using a Vanguard radioscanning instrument (graph). Simultaneous chromatography of authentic GM1, GT1, GD1b, GM3, and GM2 allow the definition of the peaks of radioactivity. (B) Tritium incorporation into the gangliosides of thyroid plasma membranes preincubated with cholera toxin prior to galactose oxidase exposure and 3H-labeled sodium borohydride reduction. The concentration of cholera toxin during the preincubation period was sufficient to give maximal (40%) inhibition of [1251]TSH binding (Fig. 1). Total radioactivity applied to the plates in (A) and (B), respectively, was 176,000 and 104,000 counts per min. The radioscanning sensitivity in (A) and (B) is identical.
should have tritium incorporated into ganglioside residues if these are available to the enzyme. As seen in Fig. 5 (top), the major labeled ganglioside is GM1 when gangliosides are extracted from these membranes, chromatographed, and evaluated for their tritium content. In contrast, Fig. 5 (bottom) shows that the GM1 residues are not labeled with tritium if the thyroid membranes are preincubated with cholera toxin prior to galactose oxidase exposure and [3H]borohydride reduction. In addition, pretreatment with cholera toxin causes an enhancement of the labeling of other glycolipids contained in the membrane (Fig. 5, bottom), and there is no labeling of GM1 in the absence of galactose oxidase (data not shown).
IQ
1.0 2.0 UNLABELED TSH
3.0
WxO6M)
0.2 0.4 0.6 0.8 1.0 UNLABELED CHOLERA TOXIN (xO14M)
FIG. 6. (A) Inhibition of [125I]TSH binding to thyroid plasma membranes by unlabeled TSH (0) and by unlabeled cholera toxin (3) in the presence of sufficient unlabeled TSH, 0.5 MAM, to yield approximately 20% inhibition of [125I]TSH binding (arrow). Conditions are those of Fig. 1. (B) Inhibition of [125IJTSH binding to thyroid plasma membranes by unlabeled cholera toxin (3) and by unlabeled TSH (0) in the presence of sufficient cholera toxin to neither enhance nor inhibit [125I]TSH binding (arrow). Conditions are those of Figs. 6A and 1. The dotted line depicts the theoretical curve if unlabeled TSH were to inhibit [125I]TSH binding in a simple additive manner.
The Interaction of Receptor Components on Thyroid Plasma Membranes. The data show that low concentrations of cholera toxin enhance [125I]TSH binding to thyroid plasma membranes, and suggest the possibility that there is a cooperative relationship between the individual receptors. This cooperative relationship has been confirmed in Fig. 6 which compares the inhibition of [125I]TSH binding to the receptor by unlabeled cholera toxin alone; by unlabeled TSH alone; by unlabeled TSH plus an amount of unlabeled cholera toxin, 0.6 MuM, sufficient to yield no significant net inhibition or enhancement; and by unlabeled cholera toxin plus an amount of unlabeled TSH sufficient to cause only 20% inhibition. As can be seen (Fig. 6A), prior exposure of the membranes to a low concentration of TSH totally abolishes the enhancement effect of cholera toxin on [125I]TSH binding (Figs. 1 or 6B). Similarly (Fig. 6B), prior exposure of the membranes to 0.6 MM cholera toxin results in an increased ability of unlabeled TSH to inhibit binding. Thus, whereas 0.1 MAM TSH yields an inhibition of 15% when present alone, 0.1 MiM TSH results in a 45% inhibition in the presence of a concentration of cholera toxin which alone has no net effect on [l25I]TSH binding. DISCUSSION In our previous report, we showed that the efficacy by which gangliosides inhibited the interaction of TSH with thyroid plasma membrane receptors was GD1b > GT1 > GM1 and we suggested all three gangliosides could be elements or analogs of the TSH receptor, since these gangliosides coexist on thyroid plasma membranes. In the present report, we show that cholera toxin interacts with GM1 residues on thyroid plasma membranes, i.e., at 9 MiM cholera toxin, GM1 is completely protected from galactose oxidase attack (Fig. 5). In addition, we show that low concentrations of cholera toxin can enhance [125I]TSH binding to thyroid plasma membranes and can cause a dramatically increased ability at low concentrations of unlabeled TSH to inhibit [125I]TSH binding (Fig. 6B). These results could be accounted for by the following schema: TSH binds preferentially to receptor sites composed of GD1b or GT1 residues or
Cell Biology: Mullin et al. their glycoprotein analogs, whereas cholera toxin binds preferentially to GM1 or GM1-like receptors. Reactivity at low concentrations of cholera toxin with GM1 results in a perturbation of the membrane which makes the GD1b or GT1 sites more accessible to TSH in the medium. The toxin thus behaves as a positively cooperative ligand, i.e., it increases the availability of the receptors for the hormone. The implication of these data and this schema is that ganglioside or ganglioside-like receptors on thyroid plasma membranes do not interact with TSH in an independent fashion but rather in a cooperative mode. Further, the cooperativity may become negative under normal circumstances, i.e., faced with high concentrations of TSH interacting with the GM1 sites, the membrane perturbation can alter the environment of GD1b- or GTI-like sites already containing TSH and thereby enhance dissociation from these previously. occupied sites. Such a phenomenon could be the basis for the curvilinear Scatchard plots and "negative cooperativity" among TSH receptors which we previously described (13). By its interaction with the GM1 ganglioside, cholera toxin may be subverting a normal mechanism by which TSH interacts with the thyroid plasma membrane and transfers its message to the cell machinery. The present data, which show that cholera toxin undergoes a conformational transition upon interacting with GM1, are both in accord with this hypothesis and suggest that this conformational transition may well be a necessary prelude to the formation of an "active" A subunit of cholera toxin capable of interacting with components within the cell membrane. It is in this last respect that the adenylate cyclase data presented are pertinent. Cholera toxin stimulation of adenylate cyclase activity in thyroid plasma membranes requires NAD; TSH stimulation does not. A divergent molecular mechanism for cyclase activation must thus be anticipated, despite the analogies at the receptor level and at the level of effector penetration or translocation within the membrane. It is pertinent at this point to recall the molecular mechanism by which diphtheria toxin subverts the normal protein synthetic machinery of the cell, i.e., through the formation of an abnormal covalent linkage of NAD to elongation factor-2 (32). These results thus enhance the importance of the efforts by Gill (29) who has already initiated a search for a membrane component which accepts NAD in covalent linkage upon the interaction of cholera toxin with the membrane. A final point should be made concerning the data of Table 1. The ability of gangliosides to inhibit both the basal- and fluoride-stimulated adenylate cyclase activity has at this time no definitive explanation. These results are not caused by the gangliosides interfering with the cyclic AMP binding protein assay. A similar inhibition of basal adenylate cyclase activity by gangliosides has been observed in cultured mouse cells (33). Since exogenous gangliosides can be integrated into intact cells (34) and membranes (2), integration of gangliosides into the fluid phase of the membrane may perturb the membranebound adenylate cyclase activity.
Proc. Natl. Acad. Sd. USA 73 (1976)
16&3
1. Mullin, B. R., Fishman, P. H., Lee, G., Aloj, S. M., Ledley, F. D., Winand, R. J., Kohn, L. D. & Brady, R. 0. (1976) Proc. Nati. Acad. Sci. USA 73,842-846. 2. Cuatrecasas, P. (1973) Biochemistry 12,3547-3558. 3. Holmgren, J., LUnnroth, I. & Svennerholm, L. (1973) Infect. Immun. 8,208-214. 4. King, C. A. & van Heyningren, W. E. (1973) J. Infect. Dis. 127,
638-647. 5. Gill, D. M. & King, C. A. (1975) J. Biol. Chem. 250, 6424-6432. 6. Sahyoun, N. & Cuatrecasas, P. (19X6) Proc. Nati. Acad. Sci. USA
72,3438-3442. 7. Ledley, F. D., Mullin, B. R., Lee, G., Aloj, S. M., Fishman, P. H., Hunt, L. T., Dayhoff, M. 0. & Kohn, L. D. (1976) Biochem. Biophys. Res. Commun., in press. 8. Wolff, J., Winand, R. J. & Kohn, L. D. (1974) Proc. Natl. Acad. Sci. USA 71, 3460-3464. 9. Mendez, E., Lai, C. Y. & Wodnar-Filipowicz, A. (1975) Biochem. Biophys. Res. Commun. 67,1435-1443. 10. Laeinmli, U. K. (1970) Nature 227,680-682. 11. Winand, R. J. & Kohn, L. D. (1970) J. Biol. Chem. 245,967-975. 12. Kohn, L. D. & Winand, R. J. (1971) J. Biol. Chem. 246,65706575. 13. Tate, R. L., Schwartz, H. I., Holmes, J. M., Kohn, L. D. & Winand, R. J. (1975) J. Biol. Chem. 250, 6509-6515. 14. Amir, S. M., Carraway, T. F., Jr., Kohn, L. D. & Winand, R. J. (1972) J. Biol. Chem. 248, 4092-4100. 15. Fishman, P. H., McFarland, V. W., Mora, P. T. & Brady, R. 0. (1972) Biochem. Biophys. Res. Commun. 48,48-57. 16. Fishman, P. H., Brady, R. O., Bradley, R. M., Aaronson, S. A. & Todaro, G. J. (1974) Proc. Natt. Acad. Sci. USA 71, 298-301. 17. Svennerholm, L. (1957) Biochim. Biophys. Acta 24, 604-615. 18. Yu, R. K. & Ledeen, R. W. (1972) J. Lipid Res. 13,680-686. 19. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193,265-275. 20. Rodbell, M. & Krishna, G. (1974) in Methods in Enzymology, eds. Fleischer, S. & Packer, L. (Academic Press, New York), Vol. 31, part A, pp. 103-115. 21. Wolff, J. & Jones, A. B. (1971) J. Biol. Chem. 246,3939-3947. 22. Gilman, A. C. (1970) Proc. Natl. Acad. Sci. USA 67,305-310. 23. Winand, R. J. & Kohn, L. D. (1975) J. Biol. Chem. 250,65226526. 24. Scatchard, G. (1949) Ann. N.Y. Acad. Sci. 51, 660-674. 25. Klotz, I. M. & Hinston, D. L. (1971) Biochemistry 10,33065-309. 26. Lineweaver, H. & Burk, D. (1934) J. Am. Chem. Soc. 56,658666. 27. Webb, J. L. (1963) "General principles of inhibition," in Enzyme and Metabolic Inhibitors (Academic Press, New York), Vol. 1, p. 162. 28. Teale, F. W. J. (1960) Biochem. J. 76, 381-389. 29. Gill, D. M. (1975) Proc. Nati. Acad. Sci. USA 72,2064-2068. 30. Morell, A. G., Van Den Hamer, C. J. A., Scheinberg, I. H. & Ashwell, G. (1966) J. Biol. Chem. 241, 3745-3749. 31. Gahmberg, C. G. & Hakomori, S. (1973) J. Biol. Chem. 248, 4311-4317. 32. Robinson, E. A., Henriksen, 0. & Maxwell, E. S. (1974) J. Biol. Chem. 249,5088-5093. 33. Fishman, P. H., Moss, J. & Vaughan, M. (1976) J. Biol. Chem. 251, in press. 34. Moss, J., Fishman, P. H., Magianello, V. C., Vaughan, M. & Brady, R. 0. (1976) Proc. Natl. Acad. Sci. USA 73, in press.
