A threonine residue in the seventh transmembrane domain of the ...

2 downloads 0 Views 515KB Size Report
Jan 28, 2016 - From the Garvan Institute of Medical Research, 384. Victoria St., Darlinghurst ... This research was supported by a grant from the Australian Na-.
THE JOURNAL OF BIOLCGICAL CHEMISTRY Vol. 269,No. 4, Issue of January 28,pp. 2373-2376, 1994 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U S A .

Communication

chains possessing seven hydrophobic transmembrane-spanning segments that couple to a n effector molecule through a trimeric G protein complex. Early biochemical and pharmacological studies had established the existence of at least two distinct adenosinereceptor subtypes (4, 5). In addition to their opposing effects on adenylate cyclase activity, these receptor subtypes could also be differentiated from one another by the potency order of a series of agonist and antagonistcompounds. With the adventof molecular cloning and subsequent pharma(Received forpublication, November 2, 1993, and in revised form, November 24, 1993) cological and functional analyses, four adenosine receptor subtypes havenow been identified from up to five different species Andrea Townsend-Nicholson$ and (6-22). The Al and A3 receptor subtypes inhibit adenylate cyPeter R. Schofield clase activity and are also able tocouple to other second mesFrom the Garvan Institute of Medical Research, 384 senger effector systems including phosphoinositol hydrolysis Victoria St., Darlinghurst, New South Wales 2010, and potassium channels (3,23), while the A,, and A2b receptor Sydney, Australia subtypes have only been shown to stimulate adenylatecyclase The Al adenosine receptor is a member of the seven- activity. transmembrane G protein-coupled, receptor superfam- Despite extensive pharmacological and functional characterily. This receptor binds the purine nucleoside adenosine ization of adenosine receptors, there has been little information with high affinity and inhibits the activity of adenylate published on structure-function studiesof adenosine receptors. cyclase. We haveusedsite-directedmutagenesisand The results fromchemical modification studies with groupfunctional expression studies to examine the role of the specific reagents have suggested that at least two histidine threonine residue, located at position 277 in transmem-residues (24, 25) as well as a carboxyl residue (25) arelocated brane domainVI1 of the humanAl receptor. Mutation of in the vicinity of the ligand binding site of the Ai adenosine Thr-277 to either serine or alanine resulted in the ex- receptor. Cysteine and arginine residues appear to be imporno change in tant in the coupling of this receptor to G proteins (25). The pression of receptors that had essentially binding affinity for theAl selective antagonist 8-cyclo- involvement of two histidine residues in the ligand binding pentyl-l,3-dipropylxanthine. Mutation of Thr-277 toserine resulted in modest (4.4-8.6-fold) but significant in- domain is also supported by a study of the rabbitA,, receptor creases in the observed Kivalues for three adenosine (26). All three studies reinforce the suggestion of two distinct agonists, namelyN-(l-methyl-2-phenethyl)adenosine (R- sites of interaction for agonists and antagonists.Site-directed mutagenesis of the bovine Al adenosine receptor has confirmed PIA or S-PIA) and 1-(6-arnino-BH-purin-9-yl)-l-deoxy-Nthe importance of the histidine residues in transmembrane ethyl-P-tribofuranunamide) (NECA). However, mutation ofThr-277 to alanine resulted in no significant (TM)’ domains VI and VI1 (12). For example, mutation of the a histidine residue in TM VI results in a 3.8-fold decrease in changes in theKifor R-PIAor S-PIA but did result in reduce Ki for NECA. antagonist affinity. However, these mutations markedly highly significant 437-fold increase in the the efficiency of receptor expression, suggesting that other resiThis demonstrates that the hydroxyl moiety of -277 mediates agonist but not antagonist binding and, more dues may also play a role in ligand binding. specifically, that this residue forms a probable molecu- There have been numerous examples of the contribution of 5‘ substitution found inNECA. amino acid residues inTM domains VI and VI1 to ligand bindlar contact site with the ing (27-30). In this study, we examine thespecific contribution of the threonine residueat position 277, immediately adjacent TM VII, to the ligand binding domain of the The purinenucleoside adenosine is an important mediatorof to the histidine in human Al adenosine receptor. We demonstrate that this resiphysiological processes in almost all organ systems. In addition to therole of adenosine in cellular energy metabolism and the due forms a specific contact site for agonist butnot antagonist importance of adenosine nucleotides in the transmissionof ge- binding. netic information, adenosine functions as a neuromodulatory EXPERIMENTALPROCEDURES transmitter molecule. I t is in this latter capacity that adenosine, acting through specific cell-surface receptors, is able to Materials-Oligonucleotides were synthesized on an AB1 DNA syninfluence a diversity of processes that include the modification thesizer. AmpliTaq was from Perkin-Elmer Corp. Restriction and modiof cardiac rhythmicity and contractility (l), the inhibition of ficationenzymes were obtained from Promega and the mammalian exfrom Invitrogen. All cellculture reagentswere (21, and the potentiation of pression vector pRc/CMv excitatory neurotransmitter release obtained from either Cytosystems or CSL, and the Chinese hamster IgE-dependent mediator release(3). Adenosine receptors are members of the superfamily of G ovary (CHO.Kl)cells were from the American Qpe Culture Collection. was fromSigma.8-[dipropyl-2,3Adenosinedeaminase(EC3.5.4.4) protein-coupled receptors. As such, they aresingle polypeptide 3HlCyclopentyl-1,3-dipropylxanthine was from DuPontNEN,and 8cyclopentyl-l,3-dipropylxanthine (DPCPX), N-(l-methyl-2-phenethyl)* This research was supported by a grant from the Australian Na- adenosine (R-PIAor S-PIA), and1-(6-amin0-9H-purin-9-yl)-l-deoxy-Ntional Health and Medical Research Council. The costs of publication of this article were defrayed in part by the payment of page charges. This article must thereforebe hereby marked “aduertisement” in accordance T h e abbreviations used are: TM, transmembrane; DPCPX, 8-cyclowith 18 U.S.C. Section 1734 solely toindicate this fact. pentyl-l,3-dipropylxanthine;R-PIAorS-PIA, N-(l-methyl-2-phenethyl)$ Tu whom correspondence should be addressed. Tel.: 61-2-361-2050; adenosine; NECA, l-(6-amino-9H-purin-9-yl)-l-deoxy-N-ethyl-~-~-riboFax: 61-2-332-4876. furanuronamide.

A Threonine Residuein the Seventh Transmembrane Domain of the Human AI Adenosine Receptor Mediates Specific Agonist Binding*

2373

2374

Adenosine AI Receptor Binding Agonist Site

ethyl-PL-ribofuranuronamide (NECA) were from Research Biochemicals Incorporated. Site-directedMutagenesis of the Human AI Adenosine Receptorand Stable Expression in CH0.Kl Cells-The cDNA encoding the humanAl adenosine receptor (10) was subcloned into the pRdCMV mammalian expression vector. Specific mutations of this cDNA were generated using theoligonucleotide-directed polymerase chain reaction mutagenesis method (31). Mutations are described by the single letter code for the amino acid in the wild type AI adenosine receptor, followedby the position number of that amino acid, followed by the amino acid being substituted at thatposition. As an example, T277S refers to the mutation of the amino acid a t position 277 from threonine to serine. All mutations were confirmed by double-stranded dideoxy DNA sequencing. Mutant receptor cDNA constructs were transfected (32) into CHO.K1 cells, and selection for stable transformants was camed out with the neomycin analogue G418 at a concentration of 800 ng/ml. Pools of surviving colonies were selected for radioligand binding assays. Radioligand Binding Assays-A stably transformed, clonal cellline that expresses the wild type Al adenosine receptor and stably transformed pools expressing mutated AI adenosine receptors were used in radioligand binding assays. Whole cellequilibrium binding assays were performed for a t least 30 min at room temperature (23 * 3 “C)in a Tris ions buffer (120 m~ NaCl, 5 nm KCl, 2 n m CaCl,, 10 l l l ~MgC12,and 50 nm Tris-HC1, pH 7.4) containing 2 IU/ml adenosine deaminase aRer a 30-min preincubation of the cells in Tris ions buffer containing 2 IU/ml adenosine deaminase. Appropriate numbers of cells wereused per point in the assay such that the disintegrationdmin bound never exceeded 10% of the total disintegrationdmidtube. Nonspecific binding was determined in the presence of l p~ DPCPX.DPCPX was dissolved in ethanol while R-PIA, S-PIA, and NECA were dissolved in dimethyl sulfoxide. The final concentration of ethanol and dimethyl sulfoxide in these assays was never greater than 0.1 and lo%, respectively. The assay was stopped with the addition of at least 10 volumes of ice-cold phosphate-buffered saline, and the samples were then rapidly filtered through 0.03% polyethyleneimine-treated glass fiber filters using a Brandel cell harvester. Filters were washed several times in ice-cold phosphate-bufferedsaline, and the disintegrationdmin bound weredetermined by liquid scintillation counting. For saturation binding assays, eight different concentrations of radioligand were used, while seven different concentrations of unlabeled ligand were used for the competition binding assays. The receptor affinities and B, values obtained for each cell line were similar throughout the study period. Data Analysis and Curve Fitting-Triplicate measurements were made of each data point. The Kd of[3HlDPCPXbinding for eachreceptor studied was estimated by fitting the untransformed data to a saturation binding isotherm using nonlinear regression (GraphPad InPlot, GraphPad Software) while the untransformed data from competition experiments were fitted to a single-site log scale competition curve to determine the ICso The ICso values were then converted toK, values using InPlot. Results are expressed as the mean * S.E. of n experiments. Statistical analyses were performed using a two-tailed, unpaired Student’s t test (StatView 11, Abacus Concepts Inc.), and a Bonferroni correction was applied to the obtained p values to correct for multiple comparisons. Dueto the limits of solubility ofR-PIA, S-PIA, and NECA, it was not possible to examine competition at concentrations of agonist greater than 1m ~ It.should be noted that the K, values obtained from the data analysis are approximate in instances where the percent specific binding at 1m~ of competing agonist was greater than 0. RESULTS AND DISCUSSION

In order to determine the contribution to ligand binding of the threonine residue at position 277 of the human A1 adenosine receptor, site-directed mutagenesis was used to change this residue to either a serine or an alanine residue. The change to alanine removes the hydroxyl group of the threonine residue while the change to serine retains the functional group but modifies the positioning of the hydroxyl moiety.Wild type and mutant cDNAs were stably expressed in CHO.Kl cells, and intact cells were used in equilibrium binding assays at room temperature. The ligands used in our assays were the A,-specific antagonist DPCPX (33) and the agonists R-PIA, S-PIA, and NECA. These three agonists have been of great utility in the classification of adenosine receptor subtypes (26). All four ligands have been used extensively in adenosine receptor assays in a variety of different tissues and species. Both R-PIA

FIG.1. Representative saturation bindingisotherms of the specific binding of [SH]DPCPXwith aScatchard transformation of the data (inset)to.A, the clonal celllime expressing the wild type human brainA, adenosine receptor(filled s q w r e ) ;B, the stably transformed pool expressing the mutant T277SAI adenosine receptor(filled circle);and C,the stablytransformedpool expressing the mutant T277AAI adenosine receptor(filled triangle).Binding assays were performedas outlined under “Experimental Procedures”using lo6, 2.5 x lo6,and 5 x lo6 celldpoint for A,B, and C, respectively.

and S-PIA are modified at the N6 position of the purine moiety while NECA is modified at the5’ position of the ribose (see Fig. 3). In saturation binding assays, cells expressing the wild type human Al receptor cDNA bound [3H]DPCPXwith a Kd of 0.8 f 0.04 IIM and a B,, of 53,167 2 14,311 binding sitedcell (n = 3) (Fig. L4, Table I). The mutant AI T277S receptor yielded a K d of 1.12 0.13 IIM with a B,, of 510,954 2 24,910 binding sited cell (n= 5) while the Kdfor the A, T277Areceptorwas 0.4 2 0.04 n~ with a B,, of 97,223 2 16,442 binding sitedcell (n= 4) (Fig. 1, Table I). The only significant difference observedin antagonist binding from wildtype was with the AI T277A receptor ( p < 0.025). This difference in antagonist affinity, although significant, represents a modest 2-fold change from wildtype and should not be overinterpreted. The affinity for [3H]DPCPXdisplayed by the wild type A, receptor is consistent with values obtained in both membrane and whole cellbinding assays of A1 receptors from several different species (11,33,34). High levels of expression of the mutant A, receptors, particularly the T277S mutant Al receptor, were observed.Whether this is due to an increased transfection efficiency or reflects an increase in receptor expression for this specific mutant is not known. In competition binding assays, R-PIA displaced [3HlDPCPX from wild type A, receptors with a Kiof 210 2 57 IIM (B,, =

2375

Adenosine AI Receptor Agonist Binding Site

Tmm I Summary of binding affinities of the antagonist PHIDPCPX and the agonists R-PIA, S-PIA, and NECA at wild type and mutant A, adenosine receptors All affinities are expressed as the mean S.E.of at least three independent experiments. Also indicated for each compound tested on a mutant receptor is the -fold change from wild type in the affinity observed. Asterisks indicate significance at p e 0.025. A Bonferroni correctionhas been applied to the p values obtained to correct for multiple comparisons. DPCPX Receptor

Kd f S.E. nM

Wild type AIR T277S AI R T277AAl R a

0.8 * 0.04 1.1 f 1.4 0.13 0.4 f 0.04

s-PIA

R-PIA

(gF$l.

K,

S.E.

(from change wt)

Ki S.E.

Change (from wt)

-fold

nM

-fold

nM

-fold

1.0

210 'c 57 920 147 807 2 169

1.0 4.4* 3.8

1,094 'c 153 8,160 'c 1,485 7,490 'c 2,134

1.0 7.5* 6.9

2.0*

NECA Ki f S. E. n.w

366 * 86 3,160 * 472 160,000 * 40,140

Change (from wt) -fold

1.0 8.6* 437*

wt, wild type.

30,567 +: 1,217 sitedcell, n = 4) while theAl T277S receptor had a Kiof 920 2 147 IMI ( B , = 438,562 * 22,069 sitedcell, n = 41, and the Ki value obtained for the AI T277Areceptor was 807 2 169 IMI (B,, = 84,824 5,241 binding sitedcell, n = 4) (Fig. 2A, Table I). As with the affinity of [3H]DPCPX binding, interpretation of the statistically significant 4.4-fold change in the afA1 T277S finity of R-PIA between the wild type and the mutant receptor, when compared with the lack of significance of the 3.8-fold change in theaffinity of the A1 T277A receptor, should be interpreted with caution. S-PIA yielded results similar to those observed for R-PIA in that there was little difference between the statistically significant 7.5-fold decrease in S-PIA affinity seen between the mutantAI T277S receptor (Ki = 8,160 2 1,485 m, B,, = 429,011 2 24,748 sitedcell, n = 4) (Fig. 2 B , Table I) uersus the wild type AI receptor (Ki = 1,094 2 153 m, B,, = 26,680 2 1,381 sitedcell, n = 3) and the lack of significance of the 6.9-fold change in affinity from wildtype observed = 7,490 2 2,134 IMI,B,, with the mutantAlT277Areceptor (Ki = 101,633 +: 13,296 sitedcell, n = 4). The Kivalues obtained for agonist binding to the wild type AI receptor are similar to those obtained in other intact cell binding assays (35) and from membrane preparations of the human AI receptors expressed in CHO.Kl cells (11).The Kivalues obtained from intact cells correspond to the low affinity site observed in membrane binding assays (33) that is believed to be due to high intracellular levels of GTP (35). In contrast, the displacement of [3H]DPCPXby NECA gave significant and selective differences in affinity (Fig. 2C, Table I). The Kivalue for NECA at the wild type Al receptor was 366 2 86 m (Bma = 26,891 2 954 sitedcell, n = 4) while the Kiof the mutant AI T277S receptor was 3,1602 472 rn (B,, = 422,113 2 38,839 sitedcell, n = 41, and the estimated Kiof the mutantA1 T277A receptor was 160,000 2 40,140 m (B,, = 123,873 2 21,257 sitedcell, n = 3). There is a highly significant 437-fold ' change from wildtype observed with the T277A mutation. The magnitude of this change is indicative of the profound effectof this mutation upon ligand binding. The threonine residue does not appear to be critical for the binding of the NG-modified adenosine analogues (R-PIA and S-PIA) as itdid not matter whether the threonine residue was mutated to a serine residue or to an alanine residue; a similar decrease in binding affinity was observed. The hydroxyl moiety, however, appears crucial for the binding of the 5"modified analogue NECA. The change from threonine to serine engenders an 8.6-fold decrease in theaffinity of NECA, but theloss of the hydroxyl group results in a complete loss of NECA binding; it was, therefore, only possible to estimate an approximate Ki value for the AI T277A mutant receptor. The Al T277S mutant receptor has a rank order of potency that is the same as thatof the wild type Al receptor with R-PIA > NECA > S-PIA. The rank order of potency of the AI T277A mutant receptor is R-PIA > S-PIA >> NECA. A comparison of the -fold change from wildtype of the affinity ratios of R-PIA:S-

*

i '

e

s

100

L

-9 0

x

80 60

3 0

40 X

-g u

a0

I

L -9

-8

-7 -6 -5 [AGONIST] (log M)

-4

-3

-9

-8

-7 -6 -5 [AGONIST] (log M)

-4

-3

= 0

Y i Y

g g

100 100

L

0

x

9

80

60

3 0

40 X

a u

-g

20

I

0,

c 0 Y

80-

9

60-

v

3 0

2 o u 40-

1 ',

-9

-8

-7

-6

-5

-4

-3

[AGONIST] (log M)

F'IG. 2. Competition of the agonists R-PIA (A), S-PIA ( B ) ,and NECA ( C ) against [sH]DPCPX (1 ma) in CHO.Kl cella stably expressing eitherwild type or mutant AI adenosine receptors.For all curves, nonspecific binding was determined in the presence of 1 p~ DPCPX (approximately 10%of total binding for the wild type A1 adenosine receptor; approximately2 and 3% of total binding for the T277S and T277AA1adenosine receptors, respectively). It was not possible to extend the concentration of the agonists beyond 1 m~ due to problems of ligand solubility. Curves representative of at least three independent assays are shown for the wild type A, adenosine receptor (open square), the stably transformed pool expressing the mutant T277S Aladenosine receptor (open circle), and the stably transformed pool expressing the mutant T277AA1 adenosine receptor (open triungle).

PIA and R-PWNECA shows that at llgfold,only the R-PIA: NECA ratio of the Al T277A mutant receptor gives a different result; all other values show only a 1.7-2.0-fold change from wild type. It is interestingto note that the rankorder of potency of the Al T277A mutant receptor is similar to that of the bovine AI receptor (12, 13). It is on the basis of this altered order of potency that thebovine AI receptor was believed to represent a

Adenosine AI Receptor Agonist Binding Site

2376

not antagonist binding. Furthermore, it has been shown that while mutation of this residue has a moderate effect upon the binding of the W-modified agonists R-PIA and S-PIA(Fig. 3), it is thehydroxyl moiety of this amino acid that isnecessary for the binding of NECA. The identification of this element of the ligand binding domain of the Ai receptor will contribute to a better understandingof the manner in which adenosine receptor ligands bind to their receptors and effect physiological responses.

&w NH

I

NH

I

Acknowledgments-We acknowledge the technical support of Majon e Liu, Vikki Falls, and Peter Adams, and we are grateful to Lisa Selbie, Rob Vandenberg, and John Shine for valuable discussion. R-PIA

S-PIA

NECA

FIG.3. The structures of R-PIA, S-PIA,and NECA. Adenosine agonists are all closely related to the parent compound in structure. Both R-PIAand S-PIA are modified a t the I\rs position of the purine and are stereoisomers of one another; NECAis modified at the 5' position on the ribose moiety.

novel adenosine receptor subtype. With the cloning of the bovine Ai receptor, however,it was demonstrated that there was extremely high (>go%) conservation of amino acid residues with the Ai receptors from other species. Oneof the differences in the bovine sequenceis a t position 277 in the receptor protein, the same residue that was mutated in this study. Instead of a threonine at position 277, the bovine receptor has a serine residue. The Ai receptors from all other species characterized thus far allhave a threonine at position 277. As the Ai T277S mutant receptor fails to reproduce the bovine pharmacology,it would seem that thissingle amino acid is not solely responsible for the altered pharmacology and may act in conjunction with differences elsewhere in the bovine receptor sequence. It is possible that the bovine receptor pharmacology is due to a decreased afinity for NECA, perhaps in coincidence with an amelioration in the binding of R-PIA and S-PIA. Previous studies have identified the importance of histidine residues in the ligand binding domain of the Ai adenosine receptor. Oneof the observations made from the chemical modification studies was that the degree of protonation of the histidine residues did not exert a major influence upon the binding of the radioligand (25).It is difficult to reconcile the importance of the histidine residues with the seeming unimportance of their degree of protonation. Site-directed mutagenesis studies (12) were unable to give a much clearer idea of the role of the two histidine residues located within the transmembrane domains of the Ai receptor protein. Mutation of the histidine in TM VI gave rise to a receptor protein whose expression, as assessed by B,,, values from binding assays, was decreased by 74%.Mutation of the histidine in TM VII, adjacent to the threonine residue examined in the current study, gave rise to a receptor protein whose specific binding of both agonists and antagonists was less than 10% of wild type levels. Taking the results from the chemical modification and mutagenesis studies together, it ispossible to envisage the involvement of histidine residues in thelocal environment of ligand binding. HOWever, the histidine residues themselves maynot be actual points of contact with specific ligand constituents. It has been suggested previously that TM VI1 is implicated in the binding of antagonists (29, 30). This study has provided evidence that thethreonine residue at position 277 in TM VI1 of the humanAi adenosine receptor is required for agonist but

REFERENCES 1. Olsson, R., and Pearson, J. (1990) Physiol. Rev. 70, 761-845 2. Yoon, IC-W., and Rothman, S . (1991) J. Neurosci. 11, 1375-1380 3. Ftamkumar, V., Stiles, G., Beaven, M., and Ali, H. (1993) J. Biol. Chem. 268, 16887-13890 4. van Calker, D.,Muller,M., and Hamprecht, B. (1979) J. Neumhem. 39, 999-1005 5. Iandos, C., Cooper, D., and Wolff, J. (1980) P m . Natl. Acad. Sei. U. S. A. 77, 2551-2554

6. Libert, F., Parmentier, M., Lefort,A,, Dinsart, C., Van Sande, J., Maenhaut, C., Simons, M.J., Dumont, J., and Vassart, G . (1989) Science 24,569-572 7. Libert, F., Schiffiann, S . , Lefort, A., Parmentier, M., Gerard, C., Dumont, J., Vanderhaeghen, J.-J., and Vassart, G. (1991) EMBO J. 10,1677-1682 8. Reppert, S., Weaver, D., Stehle, J., and Rivkees, S . (1991) Mol. Endocrinol. 5, 1037-1048 9. Mahan, L., McVittie, L., Smyk-Randall, E., Nakata, H., Momma, E , Jr., Gerfen, C., and Sibley, D. (1991) Mol. Pharmacol. 40, 1-7 10. Townsend-Nicholson, A,, and Shine, J. (1992) Mol. Brain Res. 16, 365-370 11. Libert, F., Van Sande, J., Lefort, A., Czernilofsky, A., Dumont, J., Vassart, G., Ensinger, H., and Mendla, K (1992) Biochem. Biophys. Res. Commun. 187, 91S926 12. Olah, M., Ren, H., Jacobson, K., and Stiles, G. (1992) J. Bwl. Chem. 267, 1076410770 13. Tucker, A,, Linden, J., Robeva, A,, DAngelo, D., and Lynch, K.(1992) FEBS Lett. 297, 107-111 14. Bhattacharya, S., Dewitt, D., Burnatowska-Hledin, M., Smith, W., and Spielman, W. (1993) Gene (Amst. 128,285-288 15. Maenhaut, C., Van Sande, J., Libert, F., Abramowicz, M., Parmentier, M., Vanderhaeghen, J.-J.,Dumont, J., Vassart, G., and Schiffmann, S . (1990) Biochem. Biophys. Res. Commun. 3, 1169-1178 16. Furlong, T.,Pierce, K., Selbie, L., and Shine, J. (1992) Mol. Brain Res. 15, 62-66 17. Fink, J., Weaver, D., Rivkees, S . , Peterfreund, R., Pollack, A., Adler, E., and Reppert, S . (1992) Mol. Brain Res. 14, 186-195 18. Chern, Y., King, K., Lai, H.-L., and Lai, H.-T. (1992) Biochem. Biophys. Res. Commun. 186,304-309 19. Pierce, K., Furlong, T., Selbie, L., and Shine, J. (1992) Biochem. Biophys. Res. Commun. 187,813-93 20. Stehle, J., Rivkees, S.,Lee, J., Weaver, D., Deeds, J., and Reppert, S . (1992) Mol. Endocrinol. 6, 384-393 21. Meyerhof, W., Muller-Brechlin, R., and Richter, D. (1991) FEBS Lett. 284, 155-160 (1992) Proc. 22. Zhou, Q.-Y., Li,C., Olah, M., Johnson, R., Stiles, G., and Civelli, 0. Natl. A c a d . Sci. U. S. A. 89,7432-7436 23. Linden, J. (1991) F M E B J. 6,266%2676 24. Klotz, K-N., Lahse, M., and Schwabe, U. (1988) J. B i d . Chem. 263, 1752217526 25. Gamtaen, A,, IJzerman, A,, Beukers, M., and Soudijn, W. (1990) Biochern. Pharmacol. 40,835-842 26. Jacobson, K, Stiles, G., and Ji, X.-D. (1992) Mol. Pharmacol. 42, 123-133 27. Suryanarayana, S., Daunt, D., Von Zastrow, M., and Kobilka, B. (1991) J. Biol. Chem. 266, 15488-15492 28. Wess, J., Maggio,R., Palmer, J., and Vogel, Z. (1992) J. Biol. Chern. 267, 19313-19319 " . ~. ~ 29. Oksenberg, D., Marsters, S.,ODowd, B., Jin, H., Havlik, S., Peroutka, S., and Ashkenazi, A. (1992) Nature 360,161-163 Pharmacol. 41, 30. Guan, X., Peroutka, S., and Kobilka, B. (1992) Mol. 695-698 31. Ho, S . , Hunt, H., Horton, R., Pullen, J., and Pease, L. (1989) Gene (Amst.) 77, 51-59 32. Chen, C., and Okayama, H. (1987) Mol. Cell. Biol. 7,2745-2751 33. Bruns, R., Fergus, J., Badger, E., Bristol, J., Santay, L., Hartman, J.. Hays, S., and Huang, C. (1987) Naunyn-Schmiedebergk Arch. Pharmacol. SS6,5%63 34. Ukena, D., Jacobson, K., Padgett, W., Ayala, C.,Shamim, M., Kirk, K.,Olsson, R., and Daly, J. (1986) FEBS Lett. 209, 122-128 35. Gerwins, P., Nordstedt, C., and Fredholm, B. (1990) Mol. Pharmacol. 98, 66N66 ~~~