Structure and Function of Thyrotropin Receptors (Hormone Mechanism/Cholera Toxin/Luteinizing Hormone/Human Chorionic Gonadotropin/Adenylate Cyclase) BRIAN R

Proc. Nat. Acad. Sci. USA Vol. 73, No. 3, pp. 842-846, March 1976 Cell Biology

Thyrotropin-ganglioside interactions and their relationship to the structure and function of thyrotropin receptors (hormone mechanism/cholera toxin/luteinizing hormone/human chorionic gonadotropin/adenylate cyclase) BRIAN R. MULLIN*, PETER H. FISHMANt, GEORGE LEE*, SALVATORE M. ALOJ*t, FRED D. LEDLEY*, ROGER J. WINAND§, 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; t Developmental and Metabolic Neurology Branch, National Institute of Neurological and Communicative Disorders and Stroke; f Centro di Endocrinologia ed Oncologia Sperimentale C.N.R., Naples, Italy; and § DNpartement de Clinique et de Semiologie Medicales, Institut de MWdecine, Universit6 de Liege, B4000 Liege, Belgium Contributed by Roscoe 0. Brady, December 18, 1975

ABSTRACT Gangliosides inhibit 1251-labeled thyrotropin since neuraminidase digestion eliminated the ability of both binding to the thyrotropin receptors on bovine thyroid plas- the purified receptor fragment and the crude solubilized re- ma membranes, on guinea pig retro-orbital tissue plasma membranes, and on human adipocyte membranes. This inhi- ceptor preparation to specifically bind TSH (1). bition by gangliosides is critically altered by the number and Recent studies have indicated that gangliosides, glyco- location of the sialic acid residues within the ganglioside sphingolipids that contain sialic acid, specifically interact structure, the efficacy of inhibition having the following with cholera toxin and that variations in the oligosaccharide order: GD1b > GTI > GM1 > GM2 = GM3 > GD1a. The inhibi- structure of the gangliosides influence this interaction (3-5). tion results from the interaction of thyrotropin and ganglios- The present report demonstrates that TSH also interacts ides, rather than the interaction of membrane and gangliosides. Fluorescence studies show that the inhibition is associ- with gangliosides and that the location and number of the ated with a distinct conformational change of the thyrotro- sialic acid residues on the ganglioside molecule is critical to pin molecule and that the progression from a "noninhibitory this interaction. The report further shows that there is a se- conformation" to an "inhibitory conformation" parallels ex- quence homology in the B component of cholera toxin and actly the order of effectiveness in inhibiting 125I-Yabeled thy- the ,B subunits of TSH, luteinizing hormone, and human rotropin binding. The ganglioside inhibition of 1251-labeled chorionic gonadotropin. It is suggested, therefore, that a thyrotropin binding appears to be hormonally specific in that it is not affected by a umin, glucagon, insulin, prolactin, fol- ganglioside-like structure is a basic component of glycopro- licle-stimulating hormone, growth hormone, or corticotropin. tein hormone receptors and that cholera toxin and the glyco- The possibility that a ganglioside or ganglioside-like struc- protein hormones might have a common mechanism by ture is a component of the thyrotropin receptor is suggested which their message is transmitted to the cell machinery. by the finding that gangliosides more complex than N-acetyl- neuraminylgalactosylglucosylceramide are present in bovine MATERIALS AND METHODS thyroid membranes in much higher quantities than have been previously found in extraneural tissue. The finding that TSH, 125I-labeled TSH, and thyroid plasma membranes the B component of cholera toxin, which also interacts with were bovine preparations prepared as described (6-9). 1251_ gangliosides, has a peptide sequence in common with the f, labeled TSH binding to plasma membranes was assayed by subunit of thyrotropin, suggests that thyrotropin and cholera the filtration techniques already described (8, 9) with the ex- toxin may be analogous in their mode of action on the membrane. ception that oxoid filters (Amersham/Searle Corp.) were re- placed by cellulose acetate filters (EHWP-02500, Millipore In previous studies that detailed the solubilization of the thy- Corp.). In addition to the agents tested for their ability to in- rotropin (TSH) receptor from bovine thyroid plasma mem- fluence binding, binding assays contained in a 130-Iu vol- branes, tryptic digestion was shown to yield a receptor frag- ume, 0.025 M Tris-acetate, pH 6.0, 0.6% bovine serum albu- ment that exhibited specific TSH binding and had proper- min, approximately 125,000 cpm (2 X 10-9 M) 125I-labeled ties similar to those exhibited by the TSH receptor prior to TSH, and 13-15 ,ug of membrane protein. The amount of solubilization (1, 2). This receptor fragment was purified by thyroid plasma membranes used was within the linear phase chromatography over TSH-Sepharose preparations (1, 2); it of binding when evaluated as a function of membrane pro- was shown to have a molecular weight of 25,000-30,000 by tein concentration. To insure that the binding and the inhi- gel electrophoresis in the presence of sodium dodecyl sulfate bition of binding measured in these assays were specific, (2), and to contain 30% carbohydrate and 10% sialic acid by control incubations containing 1.5 X 10-5 M unlabeled TSH, weight (1). The sialic acid was vital to receptor function gangliosides, or no membranes were included in each individual experiment. Gangliosides N-acetylneuraminylgalactosylglucosylcera- Abbreviations: TSH, thyrotropin; GM3, N-acetylneuraminylgalacto- mide (GM3), N-acetylgalactosaminyl-(N-acetylneuraminyl)- sylglucosylceramide; GM2, N-acetylgalactosaminyl-(N-acetylneura- galactosylglucosylceramide (GM2), and N-acetylneuraminyl- minyl)-galactosylglucosylceramide; GM1, galactosyl-N-acetylgalac- galactosyl-N-acetylgalactosaminyl-(N-acetylneuraminyl)- tosaminyl-(N-acetylneuraminyl)-galactosylglucosylceramide; GDla, galactosylglucosylceramide (GDLa) were obtained as pre- N-acetylneuraminylgalactosyl-N-acetylgalactosaminyl-(N-acetyl- viously described (10). Gangliosides galactosyl-N-acetylga- neuraminyl)-galactosylglucosylceramide; GDib, galactosyl-N-acetylgalactosaminyl-(N-acetylneuraminyl-N-acetylneuraminyl)-galactosaminyl-(N-acetylneuraminyl)-galactosylglucosylcera- lactosylglucosylceramide; GT1, N-acetylneuraminylgalactosyl-N-ac- mide (GM,), galactosyl-N-acetylgalactosaminyl-(N-ace- etylgalactosaminyl-(N-acetylneuraminyl-N-acetylneuraminyl)-gal- tylneuraminyl- N- acetylneuraminyl) - galactosylglucosylcer - actosylglucosylceramide. amide (GD1b), and N-acetylneuraminylgalactosyl-N-acetyl- 842 Downloaded by guest on October 1, 2021 Cell Biology: Mullin et al. Proc. Nat. Acad. Sci. USA 73 (1976) 843

Table 1. Effect of preincubation of membrane and ganglioside on inhibition of '25I-labeled TSH binding to thyroid plasma membranes* Inhibition Exp. Preincubation components (%) 1 None 95 2 Membranes + '25I-labeled TSH 87 3 Ganglioside + "2II-labeled TSH 97 O z 4 Ganglioside + membranes

z 60 without centrifugation before assay 93 5 Ganglioside + membranes followed by centrifugation before assay 0 A M 40- z * In the control experiment where no preincubation was performed (Exp. 1), all components (membranes, gangliosides, and 1251- labeled TSH) were added within 10 sec, mixed, and incubated for 0~~~~ ~ ~ ~ ~ ~ a total of 75 min prior to filtration. In Exps. 2, 3, and 4, the noted components were preincubated in assay buffer for 15 min before the missing component, ganglioside, membranes, and 125I- labeled TSH, respectively, were added; the binding assay then 20~~~~ proceeded for 60 min before filtration or a total of 75 min from the onset of preincubation. In Exp. 5, after the ganglioside and membranes were preincubated for 15 min, the mixture was cen- trifuged at 12,000 x g for 15 min to sediment the membranes. The ~~~NANA membranes -- were then resuspended in buffer and the missing 20 40 60 component, 125I-labeled TSH, was added. The ganglioside prepa- GANGLIOSIDE ADDED (nmol) ration used in these experiments was a mixed preparation (from bovine brain) containing 47% GD1a, 25% GT1, 16% Dib, and 12% FIG. 1. Inhibition of to 125I-labeled TSH binding bovine thy- GM1; 70 nmol were added. In addition to serving as a control for roid plasma membranes by gangliosides or NANA (N-acetylneura- - Exp. 5, Exp. 2 shows that gangliosides can "chase" bound TSH off minic acid). Addition of 1 nmol of fetuin does not inhibit 1251-la- the membrane. All procedures were performed at 2-4°. beled TSH binding (data not shown); this amount of fetuin is equivalent to 13 nM of protein-bound NANA. RESULTS galactosaminyl-(N-acetylneuraminyl-N-acetylneuraminyl)- Gangliosides inhibit 125I-labeled TSH binding to thyroid galactosylglucosylceramide (GT1) were isolated from com- plasma membranes, and this inhibition is related to the car- mercial preparations (Supelco, Inc., Bellefonte, Pa.) by pre- bohydrate structure of the ganglioside (Fig. 1). The best in- parative thin-layer chromatography (11). After visualizing hibitor is GD1b, where two sialic acid residues are located on with iodine vapor, the individual gangliosides were scraped the internal galactose residue of the oligosaccharide portion from the plates, eluted with chloroform:methanol:water of the ganglioside structure. The addition of a sialic acid res- (10:10:3, vol/vol/vol), taken to dryness under a stream of ni- idue to the terminal galactose residue of the ganglioside trogen, and redissolved in 0.01 M Tris-acetate, pH 7.0. Each structure, as in GT1, reduces the inhibition only slightly; in- ganglioside used in these experiments was at least 99% pure hibition is nearly eliminated, however, if one of the internal after rechromatography, visualization with resorcinol re- sialic acid residues is simultaneously deleted, as in GD1a. agent, and densitometric analysis (11). Gangliosides were GM1, the most reactive ganglioside with cholera toxin (3-5), quantitated from their sialic acid content using a micromo- is intermediate in its ability to inhibit 125I-labeled TSH bind- dification of the resorcinol method of Svennerholm (12). ing by comparison to the most reactive and least reactive Gangliosides were extracted from bovine thyroid mem- ganglioside in these experiments, GDib and GDia, respective- branes, purified by the method of Yu and Ledeen (13), sepa- ly., rated by thin-layer chromatography, visualized with resor- The ganglioside inhibition of TSH binding to thyroid plas- cinol reagent, and quantitated by densitometry, as previous- ma membranes is the result of a specific interaction of gan- ly described (11). glioside with TSH rather than of an interaction of ganglios- Fluorescence measurements were carried out at 25° ides with plasma membranes. Evidence for the latter conclu- 0.100 with a Turner model No. 210 spectrofluorometer sion comes from experiments which show no inhibition of equipped with a thermostated cell holder. Small aliquots of binding when gangliosides are preincubated with mem- concentrated ganglioside solutions in 0.01 M Tris-acetate, branes, prior to exposure to 125I-labeled TSH (Table 1). Evi- pH 7.0, were added to TSH solutions which were being dence for the specific interaction of TSH and gangliosides magnetically stirred and which contained the same buffer. has been obtained in multiple separate experiments. In the Identical aliquots of buffer were added to TSH "control" so- first, preincubation of 125I-labeled TSH with gangliosides re- lutions. The excitation and emission wave lengths were 278 sults in a change in the elution pattern of the 125I-labeled nm and 305 nm, respectively. The absorbance of the solu- TSH when it was subjected to gel filtration chromatography; tions used for fluorescence measurements never exceeded in contrast, preincubation of 125I-labeled TSH and ganglios- 0.1 cm-1 at the excitation wave length. TSH concentrations ides in the presence of a 1000-fold excess of unlabeled TSH are based on their absorbance at 276.5 nm using an EM = prevents the altered elution of the 125I-labeled TSH. Ultra- 25,700 liter/mole per cm (14). N-Acetyltyrosine-NH2 solu- centrifugation experiments confirm the interaction of TSH tions were used to monitor the stability of the instrument. and ganglioside and suggest that this interaction can result Downloaded by guest on October 1, 2021 844 Cell Biology: Mullin et al. Proc. Nat. Acad. Sci. USA 73 (1976)

1.5 tions inherent in relating sedimentation values to molecular mass (15). Fluorescence studies indicate that gangliosides induce a conformational change in the TSH molecule and that the 1.4 conformational change induced by the best inhibitor (GD1b) of the TSH-receptor interaction is distinct from that induced 0UJ by a minimal inhibitor (GD1a) of this interaction (Fig. 2). Lu Moreover, these studies show that the ability of a ganglioside uz z 0 to inhibit TSH binding to the receptor correlates directly 0 -J z with ability of the ganglioside to shift the TSH from a "non- Co I inhibitory conformation" to an "inhibitory conformation," i.e., the order of ganglioside inhibition of binding most ef- Co = 0 fective to least effective (GDlb > GT1 > GM1 > GM2 GM3 > is the same as the order of negative to positive fluo- U GDia) + LLCo z rescence changes induced by the gangliosides (Fig. 2). The ganglioside-TSH interaction appears to be hormonal- z Co w ly specific in that insulin, glucagon, prolactin, growth hormone, follicle-stimulating hormone, and corticotropin do not 0 D -J prevent the ganglioside inhibition of 1251-labeled TSH bind- lL ing to thyroid plasma membranes (Table 2). Binding studies

-J LL that have substituted human adipocyte and guinea pig retro- orbital tissue plasma membranes for the thyroid plasma membranes have thus far indicated that the inhibition by gangliosides is common to TSH receptors independent of the species or tissue from which these receptors were obtained (Table 3). The ganglioside content of thyroid plasma membranes has 50 100 been examined (Fig. 3). There are 5.6 nmol of lipid-bound GANG LIOSIDE CONCENTRATION (ymol/liter) sialic acid per mg of protein, of which 32% is GM3 and 68% FIG. 2. Fluorescence changes in the TSH molecule as mea- is higher ganglioside homologs. The higher homologs in- sured after interacting with specific gan'gliosides. The fluorescence clude gangliosides that migrate in the position of GM1, GD1a, studies measure the increase or decrease in quantum yield of the GD1b, and GT1; neuraminidase digestion can eliminate the tyrosine fluorescence of TSH when an interacting molecule alters last three of these with a corresponding increase in GM1. the environment of the phenolic chromophore (14). The altered This distribution of gangliosides is unusual for extraneural environment thus reflects a conformational change in the TSH molecule independent of whether this results from internalization tissues, where GM3 usually predominates (18, 19). of the tyrosine residues to a more "buried" location or from exter- nalization to a more "exposed" location. The TSH concentration DISCUSSION was 2 X 10-6 M. The present report demonstrates that TSH interacts with in the formation of multimolecular complexes. Thus, at ap- gangliosides to prevent binding to specific TSH receptors on proximately equivalent (l0-5 M) concentrations of GM1 and plasma membranes. The interaction is critically influenced TSH, the sedimentation value of nearly one-half of the TSH by the location of the sialic acid residues in the ganglioside present shifts from 2.5 S to 12.3 S, i.e., a significant portion structure and results in an altered TSH conformation. The of the TSH is associated with a complex having a molecular binding and fluorescence data show that the TSH-ganglios- weight of approximately 250,000, given the usual assump- ide interaction is distinct from that of cholera toxin and gan-

Table 2. Ganglioside inhibition of "25I-labeled TSH binding to thyroid plasma membranes in the presence of unlabeled hormones cpm specifically Percent Addition bound inhibited '25I-labeled TSH + membranes 21,700 I1-labeled TSH + membranes + GM1 5,055 77 I'-labeled TSH + membranes + GM1 + unlabeled hormone Insulin (2 x 10-5 M) 5,295 76 Glucagon (3 x 10-5 M) 7,882 64 Prolactin (3 x 10-6 M) 7,240 67 Growth hormone (3 x 10-6 M) 6,941 68 Follicle-stimulating hormone (2 X 10-6 M) 5,487 75 Corticotropin (2 X 10-6 M) 5,291 76 The 125I-labeled TSH in these experiments was present at a concentration of approximately 2.0 x 10-9 M; GM1 (50 nmol) was the added ganglioside. Bovine insulin, bovine glucagon, and porcine corticotropin were obtained from Calbiochem, San Diego, Calif. Bovine prolactin, bovine growth hormone, and rat follicle-stimulating hormone were highest purity preparations available from the NIH Endocrinology Study Section, Bethesda, Md. 20014. The hormones tested above had no effect on '251-labeled TSH binding to membranes at the concentrations tested. The effect of unlabeled bovine TSH and bovine luteinizing hormone could not be accurately evaluated since levels of these hormones which reversed the inhibition also inhibited 1251-labeled TSH binding to the membranes. Downloaded by guest on October 1, 2021 Cell Biology: Mullin et al. Proc. Nat. Acad. Sci. USA 73 (1976) 845 Table 3. Ganglioside inhibition of 125I-labeled TSH .: binding to TSH receptors from different sources* _ily Inhibition by Tvl A> *R

... gangliosides >kSn...... :;:: * . ,. . .: :.. : .. : . Receptor origint (%) ..- Bovine thyroid 54 ...... Guinea pig retro-orbital tissue 57 Human adipocytes 75 * The ganglioside used in this experiment was a mixed preparation (see legend to Table 2). The ratio of gangliosides to membrane protein was the same as in Fig. 1. t The preparation of plasma membrane preparations containing bDla specific TSH receptors from these sources has been detailed in previous reports (2, 8, 9, 16, 17; B. R. Mullin, G. Lee, F. D. Ledley, R. J. Winand, and L. D. Kohn, manuscript submitted). The * guinea pig retro-orbital tissue used is the Harderian gland; the human adipocytes are subcutaneous in origin.

* ~~~~~~~~~~~~~~~~~~~~~~~.'. (cholera toxin) capable of interacting with the membrane. The finding that the TSH binding activity of the purified ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. 24,000 molecular weight fragment of the bovine thyroid STANDARD id ;THYROID TSH receptor is destroyed by neuraminidase digestion (1) is MEMBRANES especially significant in the context of the present data. In FIG. 3. Thin-layer chromatogram of the gangliosides extracted the same vein, the finding that such a small fragment of the from bovine thyroid plasma membranes. Thyroid plasma mem- TSH receptor continues to exhibit criteria for multiple TSH branes were prepared as described (8, 9). The gangliosides were ex- binding sites with negatively cooperative kinetics (1) is un- tracted using techniques applicable to crude non-neural tissues derstandable in the light of the ultracentrifuge data present- (13). Chromatograms were developed and standards prepared as ed in this report. Gangliosides appear to be able to form detailed in Materials and Methods. higher order aggregates which contain approximately eight TSH binding sites at a i0-5 M gariglioside concentration. glioside. Thus GD1b is the best inhibitor of TSH binding, The critical micelle concentration for mixed brain ganglios- whereas Gm1 is demonstrated to be the best inhibitor of chol- ides has been reported as lo-4 M (22), although there is indi- era toxin binding (3-5). GD1b has, in fact, been reported to rect evidence that this value may be significantly lower for be ineffective in blocking the cholera toxin action (20). solutions of individual gangliosides such as GM1 (23). The re- Despite their different specificities for gangliosides, an in- lationship of ganglioside interactions and micelle formation teresting relationship of the cholera toxin B protein and the to the formation of macromolecular complexes which are 13 subunit of TSH may exist. Analysis of the primary se- membrane analogs with a large number of binding sites for quence of these molecules, as well as of luteinizing hormone TSH and which can exhibit negative cooperativity is under and human chorionic gonadotropin, which are glycoprotein investigation. hormones similar in structure to TSH, shows a highly signifi- Our present hypothesis is that TSH and cholera toxin may cant structural homology among all the peptides in that a be viewed as strikingly analogous in their mode of interac- common sequence can be readily recognized (Fig. 4). tion with the membrane. The B component of cholera toxin The present study raises the possibility that gangliosides and the dl subunit of TSH have sequence homologies and de- are an important structural component of TSH receptors in terminants which dominate the binding of their respective all tissues or a structural analog of an "active site" important proteins. These determinants interact with a receptor that is for binding in these membranes. These studies further analogous in structure to a specific ganglioside with a unique suggest that by alterations in the location or number of sialic number and location of sialic acid residues. A specific con- acid residues, the ganglioside structure may be made more formational shift is induced, and a second subunit of the or less specific for a particular hormone or for a molecule TSH or cholera toxin molecule (the a subunit or A protein,

Posi tion Position CHOLERA TOXIN (partial fragment) 1 T P Q N I T D L CA E Y H N T Q I H T L N D K I F SY T ES L A GK R E M A I T F 41 TSH 0 CHAIN (bovine) 19 C L T I N TT VI C A G Y C M T R B V B G K L F L P K Y A L S Q D V C T YR D F M Y K 59 LH 0 CHAIN (bovine) 26 C I T F T T S I C A G Y C P II M K R V L P V I L PP M P ER V C TY H E L R FAS V 67 HCG B CHAIN (human) 26 C I T V N lTi T I I C A G _ C P T M T R V L QG V L P A L P Z L V C N Y R D V R F E S I 67 FIG. 4. Sequences of the amino-terminal 41 residues of B component of cholera toxin and of portions of the (3 subunits of TSH, luteinizing hormone (LH), and human chorionic gonadotropin (HCG) defined by positions from the amino-terminal residue of each molecule. The boxed-in residues show a high degree of sequence homology, as measured by a mutation data matrix (21). The analysis included an evalua- tion of the remainder of the sequence of the /3 subunits of the glycoprotein hormones as well as a comparison of cholera toxin to the a subunits of these glycoprotein hormones. In none of these cases did a significant homology exist beyond that predicted by random sequence con- struction. Similarly, insulin and glucagon exhibited no sequence homology. The residue symbols corresponding to the standard residue abbreviations are as follows (21): A, Ala; P, Pro; D, Asp; T, Thr; S, Ser; L, Leu; I, Ile; V, Val; E, Glu; G, Gly; H, His; C, Cys; N, Asn; M, Met; Q, Gln; Z, Glx; R, Arg; F, Phe; Y, Tyr. Downloaded by guest on October 1, 2021 846 Cell Biology: Mullin et al. Proc. Nat. Acad. Sci. USA 73 (1976) respectively) translocates within the membrane domain to 5. King, C. A. & van Heyningren, W. E. (1973) J. Infect. Dis. interact with adenylate cyclase and to significantly alter ion 127,638-647. transport. Such a mechanism is in accord with our previous 6. Winand, R. J. & Kohn, L. D. (1970) J. Biol. Chem. 245, 967- study (24) suggesting that the TSH (3subunit contains deter- 975. 7. Kohn, L. D. & Winand, R. J. (1971) J. Biol. Chem. 246, minants for TSH-receptor binding, whereas the a- subunit 6570-6575. does not. It is further in accord with data (24) that indicate 8. Amir, S. M., Carraway, T. F., Jr., Kohn, L. D. & Winand, R. J. that a specific conformation of the TSH molecule is required (1972)J. Biol. Chem. 248, 4092-4100. for adenylate cyclase stimulation and that this conformation, 9. Tate, R. L., Schwartz, H. I., Holmes, J. M., Kohn, L. D. & despite a common and interchangeable a subunit, is distinct Winand, R. J. (1975).J. Biol. Chem. 250,6509-6515. from the conformation of luteinizing hormone after it inter- 10. Fishman, P. H., McFarland, V. W., Mora, P. T. & Brady, R. acts with the TSH receptor. 0. (1972) Biochem. Biophys. Res. Commun. 48,48-57. The sequence homologies between luteinizing hormone, 11. Fishman, P. H., Brady, R. O., Bradley, R. M., Aaronson, S. A. human chorionic gonadotropin, TSH, and cholera toxin (Fig. & Todaro, G. J. (1974) Proc. Nat. Acad. Sci. USA 71,298-301. 12. Svennerholm, L. (1957) Biochim. Biophys. Acta 24,604-615. 4) raise the possibility that luteinizing hormone and human 13. Yu, R. K. & Ledeen, R. W. (1972) J. Lipid Res. 13,680-686. chorionic gonadotropin also have a similar mechanism of re- 14. Ingham, K. C., Aloj, S. M. & Edelhoch, H. (1973) Arch. Bio- ceptor interaction but that each of these agents recognizes chem. Biophys. 195,497-504. carbohydrate sequences distinct from those recognized by 15. Martin, R. G. & Ames, B. N. (1961) J. Biol. Chem. 236, TSH or cholera toxin. Each target organ might thus have a 1372-1379. receptor with a specific carbohydrate sequence on a gan- 16. Winand, R. J. & Kohn, L. D. (1972) Proc. Nat. Acad. Sci. USA glioside-like structure. The interaction of the appropriate 69, 1711-1716. hormone with this specific oligosaccharide would result in a 17. Bolonkin, D., Tate, R. L., Luber; J. H., Kohn, L. D. & Win- unique conformational shift such that the a subunit would and, R. J. (1975) J. Biol. Chem. 250, 6516-6526. be placed in the position favored for adenylate cyclase acti- 18. Ledeen, R. W. & Yu, R. K. (1973) in Lysosomes and Storage vation in that particular target tissue. Interaction with the Diseases, eds. Hers, H. G. & van Hoof, F. (Academic Press, wrong hormone would result in a different conformation, an New York), p. 105. unfavorable position, and no second message. The predic- 19. Fishman, P. H. (1974) Chem. Phys. Lipids 13,305-326. in are examined. 20. Craig, S. W. & Cuatrecasas, P. (1975) Proc. Nat. Acad. Sci. tions inherent this hypothesis being USA 72, 3844-3848. 1. Tate, R. L., Holmes, J. M., Kohn, L. D. & Winand, R: J. (1975) 21. Dayhoff, M. 0. (1972) in Atlas of Protein Sequence and J. Biol. Chem. 250,6527-6533. Structure (The National Biomedical Research Foundation, 2. Tate, R. L., Winand, R. J. & Kohn, L. D. (1976) Proceedings Georgetown University Medical Center, Washington, D.C.), of the 7th International Thyroid Conference, June 9-13, Vol. 5, pp. 101-110. 1975 (Excerpta Medica, Amsterdam, The Netherlands), in 22. Gammack, D. B. (1963) Biochem. J. 88,373-383. press. 23. Kanfer, J. N. & Spielvogel, C. (1973) J. Neurochem. 20, 3. Cuatrecasas, P. (1973) Biochemistry 12,3547-3558. 1483-1485. 4. Holmgren, J., Lonnroth, I. & Svennerholm, L. (1973) Infect. 24. Wolff, J., Winand, R. J. & Kohn, L. D. (1974) Proc. Nat. Immun. 8. 208-214. Acad. Sci. USA 71, 3460-3464. Downloaded by guest on October 1, 2021