The Hydrophobic Tryptic Core of the @-Adrenergic Receptor ...

THE JOURNAL (c> 1987 by The American Society for Biochemistry and Molecular Biology,

OF BIOLOGICAL CHEMISTRY Inc.

Vol . 262, No. 34, issue of December 5, pp. 16655-16662,1987 Prrnted in CJ S. A.

The Hydrophobic Tryptic Core of the @-Adrenergic Receptor Retains G, Regulatory Activity in Response to Agonists and Thiols*

(Received for publication, April 27, 1987)

Ronald C. Rubenstein, Stephen K-F. Wong, and Elliott M. Ross From the Department of Pharmacology, Southwestern Graduate School of Biomedical Sciences, University of Texas Health Science Center, Dallas, Texas 75235

The function of structural domains of the B-adrenergic receptor were probed by studying the ability of tryptic fragments of the receptor to catalyze the binding of guanosine-5’-0-(3-thiotriphosphate (GTPyS) to the GTP-binding regulatory protein, G.. &Adrenergic receptor purified from turkey erythrocytes was treated with trypsin under nondenaturing conditions. Such treatment decreased &adrenergic ligand binding activity by only 15-25%. Active components of the limit digest were repurified by affinity chromatography on alprenolol-agarose and then reconstituted with purified G. into unilamellar phospholipid vesicles. After reconstitution, the proteolyzed receptor was able to catalyze agonist-stimulated binding of GTPyS to G. at a rate and extent equivalent to that of the nonproteolyzed receptor. The proteolyzed receptor was also partially activated upon reduction by dithiothreitol, as previously reported for the intact receptor (Pedersen, S. E., and Ross, E. M. (1985) J. Biol. Chern. 260, 14150-14157). The repurified, active tryptic digest contained two detectable peptides. One, of approximately 2 X lo4 Da, contained either four or five of the amino-terminal membrane-spanning domains plus the intervening hydrophilic loops but not the amino-terminal extracellular, glycosylated peptide. The second, of 9,000-10,000 Da, was composed essentially of the two carboxyl-terminal membrane-spanning domains and the intervening extracellular, hydrophilic loop. These data indicate that most of the large intracellular hydrophilic loop and the hydrophilic, carboxyl-terminal region of the receptor are not necessary for the agonist-stimulated regulation of G..

The @-adrenergic receptor is one member of a large class of cell surface receptors that control the activity of GTP-binding regulatory proteins (G proteins) involved in transmembrane signaling. The @-adrenergic receptor regulates G,,l which mediates the activity of adenylate cyclase (1). Other receptors that promote the accumulation of cyclic AMP also act upon

* This work was supported by National Institutes of Health Grant GM30355, Welch Foundation Grant 1-928, and National Institutes of Health Training Grants GM07062 and GM08014. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduer- tisernent” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The abbreviations used are: G., the GTP-binding regulatory protein that stimulates adenylate cyclase; GTPyS, guanosine-5’-0-(3- thiotriphosphate; SDS, sodium dodecyl sulfate; ICYP, iodocyanopin- dolol; TPCK, L-1-tosylamido-2-phenylethyl chloromethyl ketone; HPLC, high pressure liquid chromatography; CHAPS, 3-[(cholami- dopropyl)dimethylammonio]-1-propanesulfonate; Hepes, 4-(2-hy- droxyethy1)-1-piperazineethanesulfonic acid.

G , and receptors that mediate production of other intracellular signals are appropriately selective for different G proteins. Thus, rhodopsin interacts with transducin (G,) to cause activation of the retinal cyclic GMP phosphodiesterase (2). The muscarinic cholinergic receptor can activate both Gi and Go (3, 4), leading to phospholipid hydrolysis, inhibition of adenylate cyclase, and the activation of K’ channels ( 5 ) .

The recent sequencing of cDNAs for two P-adrenergic receptors (6, 7) and two muscarinic cholinergic receptors (8, 9) has also indicated that these receptors are homologous both to each other and to rhodopsin. The sequence homology among the G protein-coupled receptors is not uniform but rather is concentrated in seven strongly hydrophobic regions. In the case of rhodopsin, these hydrophobic regions each span the membrane bilayer, such that the glycosylated amino terminus is on the luminal face of the disc membrane and the carboxyl terminus is on the cytoplasmic face (10; see also Fig. 1). Sequence homology and the presence of N-linked glyco- sylation sites near the amino termini suggest that the p- adrenergic and muscarinic cholinergic receptors are similarly folded. The conservation of sequence in the membrane-spanning regions argues for their functional importance, but de- tails of their function are unknown.

The functions of the hydrophilic regions of the receptors have been only partially defined. The binding of G proteins presumably occurs on the hydrophilic, cytoplasmic face; the presence of certain homologous sequences in the first and second hydrophilic loops of the receptors make them likely sites for this interaction. The third intracellular, hydrophilic loop in these receptors displays relatively little conservation of sequence and its function is not well understood. Kuhn and Hargrave (11) and Litman et al. (12) found that a single tryptic cleavage in this loop of rhodopsin did not adversely affect light-stimulated control of transducin but that more extensive proteolysis here inhibited regulatory activity of the bleached rhodopsin molecule. In a conceptually similar study using site-directed mutagenesis, Dixon et al. (13) showed that deletion of a serine residue a t a potential cyclic AMP-dependent protein kinase substrate site did not decrease @-adrenergic regulation of adenylate cyclase. However, deletion of most of the large intracellular loop inhibited the receptor’s ability to activate G,. Because the mutant receptor retained @-adrenergic ligand binding activity, these authors speculated that this loop may be involved in binding to G,. Removal of all or most of the intracellular carboxyl terminal region of either rhodopsin or of the p-adrenergic receptor seems to have little effect on the regulatory function of these proteins (11-13).

In this study, we have used proteolysis of the @-adrenergic receptor to test the importance of three readily cleaved domains, the extracellular, amino-terminal region that contains bound carbohydrate, the large intracellular loop, and the carboxyl-terminal intracellular region. We find that removal

16655

16656 Regulatory Domains of the P-Adrenergic Receptor

of these three domains by proteolysis followed by affinity repurification yields two noncovalently bound fragments that can both bind to p-adrenergic ligands and stimulate guanine nucleotide binding to G, in response to p-adrenergic agonists or to reduction by thiols.

EXPERIMENTAL PROCEDURES

Materials-(-)-Propranolol, alprenolol, (+)-cyanopindolol, and Lubrol12A9 were gifts from Ayerst Laboratories, Hassle Pharmaceu- ticals, Dr. G. Engel of Sandoz Pharmaceuticals, and ICI, Ltd. TPCK- treated trypsin was purchased from Sigma or Millipore. Soy trypsin inhibitor, tosyl-lysyl-chloromethyl ketone, and diisopropylfluorophosphate were from Sigma. GTPyS from Boehringer Mannheim was purified as described previously (14). [35S]GTPyS, [1251]NaI, and lZ51-labeled goat Fab against rabbit IgG were from Du Pont-New England Nuclear. ['251]ICYP was prepared according to Engel et al. (15). [-p3']ATP, synthesized according to Johnson and Walseth (16), was a gift from J. Gregory and Dr. M. H. Cobb of this department. Phosphatidylethanolamine and phosphatidylserine were purchased from Avanti Polar Lipids. Molecular weight standards were purchased from Bio-Rad (high molecular weight) or Bethesda Research Labo- ratory (low molecular weight). Protein kinase C (Ca2+/phospholipid- dependent enzyme), partially purified from rat brain according to steps 1-3 of Kikkawa et al. (17), was a generous gift of Dr. M. H. Cobb. Alprenolol-agarose was synthesized according to Caron et al. (18). The sources of other reagents have been listed previously (14, 19).

Preparation and Assay of @-Adrenergic Receptor-P-Adrenergic receptor was purified from turkey erythrocytes as described previously (19). Soluble receptor was assayed in digitonin solution according to the binding of the antagonist ['251]ICYP using centrifugal gel filtration to separate free and bound ligand (20). Reconstituted receptor was assayed using filtration on glass fiber filters to remove unbound ligand (19, 20). Proteolytic fragments of the receptor were purified by HPLC on a 0.45 X 150 mm Synchrom C4 column eluted with a gradient of 0.1% trifluoroacetic acid to 50% 1-propanol in 0.06% trifluoroacetic acid (6).

Preparation and Assay of G.-G. was purified from rabbit liver by the method of Sternweis et al. (21). G. was assayed according to the binding of [35S]GTPyS in the presence of Lubrol 12A9 and 50 mM M F , as described by Asano et al. (14). Receptor-catalyzed [35S] GTPyS binding to GB in reconstituted phospholipid vesicles was assayed as described previously (14, 19).

Reconstitution of Receptor and G.-Purified P-adrenergic receptor, or receptor and G., were reconstituted into unilamellar phospholipid vesicles composed of phosphatidylethanolamine and phosphatidylserine (2:l) as described previously (19), except that CHAPS was sub- stituted for deoxycholate plus cholate (22). This procedure was not altered when proteolyzed receptor was used.

Polyacrylamide Gel Electrophoresis-Proteins and peptides were separated by electrophoresis on polyacrylamide gels in SDS. Uniform gels were prepared by the method of Laemmli (23). Gradient gels, 9- 27% acrylamide with a 0-3% glycerol gradient, were prepared and run as described by DeWald et al. (24), except that a stacking gel (23) was added to increase resolution. Other gradient gels, 10-21% acrylamide, were prepared according to the formula of Laemmli except that the concentration of Tris buffer in all solutions was doubled. Both gradient gels resolve and retain peptide standards of M , -3000 (glucagon, insulin subunits), as well as discrete peptic fragments of the receptor that migrate 3-4 mm below glucagon (extrapolated M, -2000). The glycerol-containing gels (24) resolve 2000-4000-Da peptides better than the high salt gels but give broader bands overall. Unless otherwise noted, samples for electrophoresis were prepared by heating at 100 "C for 2 min in 67 mM Tris-Pi (pH 6.7), 10% glycerol, 2% SDS. Gels were fixed overnight in 10% trichloroacetic acid, 10% sulfosalicylic acid, 50% methanol and silver stained (25) after washing in 50% methanol.

Immunologic Procedures-Immunogenic peptides (Fig. 1) were synthesized by Dr. John Burnier, of Genentech, Inc. and were coupled either to soy trypsin inhibitor or keyhole limpet hemocyanin using maleimidobenzoyl-N-hydroxysuccinimide ester (26). Procedures for immunizing rabbits and for Western blot analysis of peptides from polyacrylamide gels have been described previously (6).

Receptor Phosphorylution by Protein Kinase C-The 0-adrenergic receptor, reconstituted into phospholipid vesicles as described above,

u u Extracellular Space I

I92

FIG. 1. Proposed arrangement of the primary sequence of the ,%adrenergic receptor and location of antigenic sites. Ori- entation with respect to the plane of the bilayer is inferred by analogy to rhodopsin (6, 10). Solid bars indicate the sequences used for the synthesis of peptide antigens: Z, Hislm-Cyslg2; ZZ, Cy~~~~-Pro255; ZZZ, Ala23-Glu34; V, G1y250-Pro265; VZZ, A ~ n ~ ~ ~ - C y s ~ ; VZZZ, Cy~~~-Gln372. The residue numbers of some likely proteolysis sites are also shown.

was phosphorylated using protein kinase C.' Phosphorylation was carried out for 30 min at 30 "C in medium containing 20 mM Hepes (pH 8.0), 5 mM MgC1,0.5 mM EDTA, 1 mM CaC12, and 1-50 PM [y- 32P]ATP (100-1500 cpm/fmol). Phosphorylation was monitored by separating the labeled receptor on SDS-polyacrylamide electrophoresis gels and scintillation counting of the labeled band after its localization by autoradiography.

RESULTS

Comparison of the primary amino acid sequences of the /3- adrenergic receptor (6,7), rhodopsin (lo), and the muscarinic cholinergic receptor (8,9) suggests that the largest amount of sequence homology, both identity and conservation, occurs within the proposed membrane-spanning regions. We rea- soned that these regions are likely to be important for the integrity of the structure and/or function of the receptor. It also seemed likely that their hydrophobic nature might limit their accessibility to proteases. Fig. 2 shows the proteolysis products that were obtained when purified ,&adrenergic receptor from turkey erythrocytes was treated at room temperature with trypsin in digitonin solution. The untreated receptor displays two bands, a minor species at M , -52,000 and a major species at M , -42,000. The 42,000-kDa species is at least partially deglycosylated, probably by proteolysis of 15- 20 amino acids near the amino terminus (27, 28), and may also be proteolyzed at the carboxyl terminus (6). At increasing concentrations of trypsin, there was a gradual loss of the predominant M , -42,000 silver-stained band concomitant with the appearance of smaller species? An intermediate of M , -25,000 preceded the appearance of an M , -19,000 prod- uct. An M , -36,000 precursor of the M , -25,000 peptide is barely visible in the figure, as is a poorly stained M , -11,000 fragment. Treatment with high concentrations of trypsin did not result in complete proteolysis of the receptor; a limit tryptic digest resulted even at 7.5 mg/ml of trypsin (not shown). The limit digest contained a major silver-stained band with a mobility similar to that of soy trypsin inhibitor. This fragment, which usually behaves as an M , -19,000 peptide, will be referred to below as fragment A. The second component is a poorly stained M , -11,000 peptide, referred to as fragment B. The band at M , -27,000 in the right-most lanes is trypsin. This pattern of limit proteolysis resembles

M. H. Cobb and E. M. Ross, manuscript in preparation. We refer to the M , of peptides identified on polyacrylamide gels

based upon their electrophoretic mobility in SDS. In the case of proteolytic fragments A and B (see below), the correct molecular weight of each peptide can be more closely estimated based on the known sequence, as discussed in the text.

Regulatory Domains of the &Adrenergic Receptor 16657

Trypsin ( ng I m l Mr 0 7.5 25 75 250 750 2500 7500 -45

-31

- 21.5

- 14.4

400 406 107 105 408 401 97 77

ICYP Binding (%) FIG. 2. Tryptic proteolysis of the &adrenergic receptor.

Purified @-adrenergic receptor (45 nM in 20 mM NaHepes (pH 8.0). 1 mM EDTA, 0.2 mM ascorbate, and 0.2% digitonin) was incubated a t room temperature for 45 min with TPCK-treated trypsin at the concentrations shown. Part of each sample was diluted 600-fold with the same buffer containing 0.1 mg/ml soy trypsin inhibitor and assayed for [""IJICYP-binding activity as described under "Experi- mental Procedures." Concentrated diisopropylfluorophosphate (0.13 mg/ml final concentration) was added to the remainder. After 10 rnin at 0 "C, one-half volume of 3-fold concentrated SDS dissociation buffer containing 3% 2-mercaptoethanol was added, and the samples were incubated a t 37 "C for 30 min. Samples were analyzed on 12% polyacrylamide gels and silver stained as described. The band that appears a t -27,000 Da is trypsin, and the narrow band at M , -37,000 is a trypsin-resistant contaminant.

that displayed by rhodopsin (11,12). Surprisingly, receptors treated as described above retained

their ability to bind @-adrenergic ligands. In the limit digest, the capacity to bind the @-adrenergic antagonist [12511]ICYP was decreased by only 23% (Fig. 2). In other experiments, in which no fragments of M, >19,000 were found, the loss of binding capacity has been as little as 15% (not shown). The affinities for [l'sI]ICYP and for competing @-adrenergic ligands were also not substantially altered in the limit tryptic digest (not shown).

Similar patterns of proteolysis were observed when the receptor was treated with chymotrypsin, thermolysin, or sub- tilisin under nondenaturing conditions (not shown). Each protease produced limit digests that contained a prominent silver-stained band of M, -19,000-24,000 on SDS-polyacrylamide gels. These proteolyzed receptors also retained the ability to bind @-adrenergic ligands.

The trypsin-treated receptor also retained its ability to regulate G,. In preliminary experiments, a limit digest of the sort shown in Fig. 2 was reconstituted into phospholipid vesicles with purified G , using established techniques (19). In these vesicles, the addition of the @-adrenergic agonist isoproterenol stimulated the rate of GTPyS binding to G, to the same extent as observed for vesicles that contained untreated receptor (data not shown). This unexpected result prompted more complete characterization of the identities and functions of the proteolytic peptides that were derived from the receptor.

Composition of the Limit Tryptic Digest of the @-Adrenergic Receptor-The limit tryptic digest of the receptor contained two peptide components with M , >6000 that could be identified by silver staining (Fig. 2). The M, -19,000 fragment A

was isolated from the total digest by reverse-phase HPLC as described (6), identified by its mobility on SDS-polyacrylamide gels, and subjected to automated sequence analysis. The amino-terminal sequence of fragment A, which could be interpreted for 15-20 residues, begins a t GlnRo, indicating that tryptic cleavage occurred after Arg,, (6, confirmed in this study).

Because of uncertainty in assigning an exact molecular size to fragment A based only on its electrophoretic mobility on SDS gels, the location of its carboxyl terminus was approxi- mated according to its differential retention or loss of defined antigenic determinants. Antisera were raised against the syn- thetic peptides listed in the legend to Fig. 1 and used to probe Western blots that contained the untreated receptor, fragment A, and intermediate proteolytic fragments (Fig. 3). These immunoblots showed that both the untreated receptor (lanes I, 4, and 6) and a M, -25,000 proteolytic intermediate (lanes 3, 5 , and 7) contained peptides I, 11, and V. Fragment A, shown in lane 2 and as a slightly larger precursor in lane 3, contained the sequence only of peptide I, on the second extracellular loop. Antisera against peptides I1 or V, on the major intracellular loop, did not cross-react with fragment A in the samples shown in lanes 5 and 7, although fragment A was clearly apparent in these samples by silver staining (not shown). It is likely that the loss of immunoreactivity against peptide I1 and peptide V in fragment A reflects their proteolysis rather than an alteration of their tertiary structure. Both the proteolyzed receptor and its fragments were denatured before and during electrophoresis and the antisera were initially raised against short peptides.

These data indicate that fragment A includes a t least the four amino-terminal membrane spanning domains and the intervening loops. I t may also contain the fifth membrane- spanning region but does not extend significantly beyond CysZ4+ This maximum length would predict a true molecular

23.4 - "

1 2 3 4 5 6 7 FIG. 3. Reaction of peptide-directed antisera with proteo-

lyzed or untreated 8-adrenergic receptor. Purified receptor (200 ng) in digitonin-containing buffer was exposed to 63 ng/ml trypsin for 2 h a t 30 "C (lane 2), 1.6 pg/ml trypsin for 18 h a t room temperature ( l a n e 3), or 1.4 pg/ml trypsin for 2 h a t 30 ' C ([ones 5 and 7) . Reactions were terminated by the addition of diisopropylfluorophosphate to 0.1% (v/v). Samples in lanes 1, 4, and 6 were not proteolyzed. All samples were denatured and analyzed by electrophoresis on 10% polyacrylamide gels in the presence of SDS. Proteins were electrophoretically transferred to nitrocellulose and probed with antisera raised against peptide I, 11, or V (see Fig. 1, legend). The appearance of all bands was blocked by preincubating the antiserum with the antigenic peptide but not by preincubation with a heterolo- gous peptide (not shown, see Ref. 6). The samples in lanes 3.5, and 7 contained equal amounts of fragment A and the M , -25,000 peptide according to silver staining of duplicate lanes. The sample in lane 2 contained no visible M , -25,000 peptide. The sizes of the peptides were estimated according to the mobility of each peptide relative to standards.


weight of -24,000 for fragment A. Termination in the second extracellular loop (ArgZos) would predict a molecular weight of 19,500. Neither is inconsistent with the electrophoretic mobility within the uncertainty of that determination. There are multiple potential sites for tryptic cleavage just on the intracellular side of the fifth membrane span, and heteroge- neity at the carboxyl terminus of fragment A is consistant with its frequent appearance as a doublet or multiplet (see Figs. 2 and 3).

The other component of the limit digest of the receptor was the M , -11,000 fragment B. When fragment B was isolated by reverse-phase HPLC and subjected to sequence analysis, its amino terminus was found to be ValZm, within the major intracellular loop. Sequence data were interpretable for over 25-amino acid residues in three experiments. The apparent molecular size of fragment B suggested that it includes the two carboxyl-terminal membrane-spanning sequences and the intervening extracellular loop but a minimal amount of the intracellular carboxyl-terminal region. This assignment is also supported by the differential retention and loss of antigenic determinants. Fragment B prepared from radioiodi- nated receptor was specifically immunoprecipitated by anti- peptide VI1 serum, indicating that it contains most of the seventh membrane-spanning region. This minimum length was confirmed by automated Edman sequencing of the major CNBr peptide of fragment B.4 Readable sequence was obtained through AsP~~R, just beyond the carboxyl terminus of peptide VII. However, antisera against peptide VI11 that cross-reacted with the intact receptor did not react with fragment B.s Therefore, peptide B terminates just to the carboxyl-terminal side of the seventh membrane span and before Arg,,. These data are consistent with tryptic cleavage between Arg,,,-Arg,,, an area that is rich in basic residues.

To provide a more convenient monitor of the proteolysis of the receptor and to facilitate the detection of specific sites on tryptic fragments, the receptor was labeled with :'2P using protein kinase C and [y-:"P]ATP. The time course of tryptic proteolysis of the [:'2P]phosphorylated receptor is shown in Fig. 4, using an acrylamide gradient gel to improve the sepa- ration of small peptides. Initially there was a loss of the ?'P- labeled receptor bands a t M, -52,000 and 42,000. This loss paralleled the formation of the two R2P-labeled proteolytic intermediates that electrophorese at apparent M, -36,000 and 25,000. At later times, the conversion of these intermediates to fragment A coincided with the loss of all peptide- bound %'P from the gel, indicating that neither fragments A nor B was phosphorylated. These data suggest that at least one site of protein kinase C-catalyzed phosphorylation of the @-adrenergic receptor resides on the large intracellular loop, which is removed in the proteolytic cleavage of the apparent M, -25,000 intermediate to fragment A (Figs. 3 and 4). Phosphorylation in this region is probably on Serg7,, Thr277, or SerziR according to the specificity of protein kinase C (29- 31). The fragments containing the R2P label must have been cleaved to M, <2500, or they would have been detected in these or other experiments.

Preliminary experiments suggest that the receptor is also phosphorylated by protein kinase C between Arg,,, and the carboxyl terminus, most likely between Ser4,s and Ser4ai. However, the extent of phosphorylation remains uncertain and the site or sites have not been independently determined.

Autophosphorylated protein kinase C or "'P-labeled con- taminants do not give rise to labeled peptides that interfere with this analysis. When purified autophosphorylated protein

S. K-F. Wong, unpublished data. ' S. K-F. Wong and E. M. Ross, unpublished data.

- 0 1 2 4 6 91216203060 - 95 - 62

52 - 42- - 45 36 - 25 - 19-

-30

-18.5

11- -14.3

-65

"3

FIG. 4. Time-dependent tryptic proteolysis of phosphorylated 8-adrenergic receptor. ,%Adrenergic receptor was reconstituted into phospholipid vesicles and phosphorylated as described under "Experimental Procedures" using [y-''P]ATP at -100 cpm/ fmol. The vesicles (294 pl, -3 pmol) were solubilized by the addition of 6 pl of 5% digitonin and incubation at 0 "C for 1 h, after which additional unlabeled receptor (25 pmol in 25 pl) was added. Proteol- ysis was initiated by the addition of 2.83 pl of 200 pg/ml TPCK- treated trypsin to 280 pl of receptor. At the indicated times, shown in min above each lane, proteolysis was stopped by mixing 25-pl aliquots of reaction mixture with 0.25 volumes of 5-fold concentrated SDS dissociation buffer. The zero time sample was prepared by mixing 2.5 pl of 20 pg/ml TPCK-treated trypsin with 6.25 pl of 5- fold concentrated dissociation buffer, incubating 5 min at room temperature and then adding 22.5 pl of receptor. An equal amount of receptor (-) was similarly dissociated in the absence of trypsin. Samples were made 10 mM in dithiothreitol, heated for 30 min at 40 'C, made 25 mM in N-ethylmaleimide, and heated for 60 min at 40 "C. Samples were then analyzed by electrophoresis on a high salt gradient gel as described under "Experimental Procedures." The gel was silver stained and autoradiographed (17-h exposure; -80 "C). Markers on the [eft indicate peptides that electrophoresed with mo- bilities equivalent to M, -52,000,42,000,36,000,25,000, 19,000 (fragment A) and 11,000 (fragment R). Positions of molecular weight standards are shown on the right, with the position of the poorly separated subunits of insulin denoted -3.

kinase C was treated with trypsin, its proteolysis was rapid, and there was no overlap of either silver-stained or autoradi- ographic bands with those of the receptor (not shown).

Affinity Chromatographic Repurification of the Trypsin- treated Receptor-For further studies of adrenergic ligand binding and the regulation of G. by the proteolyzed receptor, the active components of the limit tryptic digest were repurified by @-adrenergic affinity chromatography on alprenolol- agarose (18). This process should retain only those peptides that are directly involved in ligand binding or those that remain noncovalently associated with ligand-binding peptides. As a marker for the removal of the major cytoplasmic domains, receptors were labeled with '"P prior to digestion with trypsin. As shown in Fig. 5, A and B, the /3-adrenergic ligand-binding activity of trypsin-treated or untreated receptors were similarly adsorbed and eluted from alprenolol-agarose. However, there was a 60% loss of :'lP radioactivity in the peak fractions from the limit tryptic digest relative to the same fractions of the untreated receptor. Much of the remaining 40% was evidently eluted nonspecifically when the temperature of the column was raised (Fig. 5 legend). Similar results were obtained in three experiments.

When identical volumes of these peak fractions from alprenolol-agarose chromatography were analyzed by SDS-pol-

Regulatory Domains of the @-Adrenergic Receptor 16659

A

Elution Vo

"(85 -143

- 6 5

- 3 3

D Applied Eluted -~ Ag 32P A9 32P

-45

- 30

-185 -143

-65

- 3 3

FIG. 5. Alprenolol-agarose affinity chromatography of t ryps in- t rea ted and un t rea ted receptor. A small quantity of receptor (3 pmol) was '"P-labeled by protein kinase C-catalyzed phosphorylation in phospholipid vesicles to a specific activity of 375 cpm/fmol, as described under "Experimental Procedures." Labeled receptor was resolubilized by the addition of digitonin to 0.1% and incuhation a t 0 "C for 1 h. Excess n2P was removed by the centrifugal gel filtration method originally described for separating receptor from radioactive ligands (20), and 125 pmol of unlabeled receptor was added. This preparation was divided in half, trypsin was added to 1 aliquot (10 pg/ml final), and both were incubated for 1 h a t 25 "C. The untreated receptor (180 pl, 62.2 pmol of total [lZ"I]ICYP-binding activity, 8.1 X 10'' :'jP cpm) and the limit tryptic digest (180 pl, 51.7 pmol, 8.1 X 10' "P cpm) were applied to separate 0.2-ml columns of alprenolol-agarose in 20 mM NaHepes, pH 8.0, 1 mM EDTA, 0.1 M NaCI, 0.1% digitonin. The columns were washed with an equal volume of buffer and incubated for 90 min a t 4 "C. The columns were then washed with 15 volumes of buffer (5 ml/h, 4 "C), allowed to come to room temperature, and eluted with 10 volumes of 10 PM (-)alprenolo1 in the same buffer a t room temperature a t a flow rate of 1 ml/h. The oertical arrows indicate warming of the column and the addition of alprenolol. Aliquots of each 200-pl fraction were analyzed for [12'1] ICYP-binding activity and total '*P content and by SDS-polyacrylamide gel electrophoresis. Elution profiles for nonproteolyzed receptor and for trypsin-treated receptor are shown in panels A and B. Panels C and D show SDS-polyacrylamide gel electrophoresis of 2-pl aliquots of the receptor that were applied to the affinity columns and 2 0 4 aliquots of the two peak fractions that were eluted with room- temperature alprenolol-containing buffer. Untreated receptor is shown in panel C and proteolyzed receptor is shown in panel D. Samples were incubated for 1 h a t room temperature in SDS dissociation buffer. N-Ethylmaleimide was added to 5 mM, followed by a second 1-h incubation a t room temperature. The samples were then made 1% in 2-mercaptoethanol, placed in a boiling water bath for 2 min, and electrophoresed on a glycerol-containing acrylamide gradient gel as described under "Experimental Procedures." The gels were silver stained identically, and photographs (marked Ag) were identically exposed. Autoradiographs of the gels (marked " P ) were exposed a t -80 "C for 3 days for untreated samples (panel C ) or for 7 days for proteolyzed samples (panel D). Photographs of the autoradiographs were processed identically.

yacrylamide gel electrophoresis, the silver-stained bands of the native @-adrenergic receptor (Fig. 5 C ) and of fragment A (Fig. 5 0 ) were clearly visible. Fragment B was also visible on the original silver-stained gel. However, only the M , -42,000 band in the sample of untreated receptor retained detectable :'2P label (Fig. 5 C ) . No "'P was associated with either fragment A or B, even though the autoradiograph of the trypsin-treated preparation (Fig. 5 0 ) was exposed more than twice as long as was that of the intact receptor. Although pale, diffuse bands could be observed at this exposure, they did not corre- spond to the silver-stained bands or to previously identified tryptic peptides of the receptor. From these data, it appears that the major cytoplasmic loop (loop 5-6) of the proteolyzed receptor did not coelute with the ligand binding activity and fragments A and B.

Regulatory Activity of Trypsin-treated P-Adrenergic Recep- tor-The ability of trypsin-treated receptor to regulate the binding of guanine nucleotides to G, was assayed after the coreconstitution of both proteins into phospholipid vesicles. Aliquots of the limit tryptic digest of the receptor and of untreated receptor were "'P-labeled, affinity purified as described above, and reconstituted with purified G,. Both batches of vesicles contained similar amounts of both G, and of @-adrenergic ligand-binding activity. As shown in Fig. 6, the @-adrenergic agonist isoproterenol stimulated the binding of GTPyS to a similar extent and with a similar rate in the vesicles that contained the proteolyzed and repurified receptor as in the vesicles that contained untreated but repurified receptor. The rates of agonist-stimulated binding (inserts) were comparable to those reported previously (19) and multiple molecules of G. were stimulated by a single receptor in several experiments. This complete experiment has been re- peated twice with and twice without prior phosphorylation of the receptor with the same results (data not shown). The

0 1 2 3 4 5 8 1 2 . . ..

Time (Mid FIG. 6. Agonist-stimulated regulation of GTPrS binding to

G. by proteolyzed receptor. Samples of trypsin digested (0, .) and untreated (0 ,0) receptor were affinity purified as described in the legend to Fig. 5 and reconstituted with purified G. into phospholipid vesicles as described, but with 0.02% diisopropylfluorophosphate in all buffers. The binding of ['"S]GTPrS to the vesicles was assayed as described under "Experimental Procedures" in the presence of 5 p~ (-)-isoproterenol (0,O) or 0.1 p~ (-)-propranolol (., 0). Vesicles prepared with proteolyzed receptor contained 3.9 fmol of receptor and 75 fmol of G. per assay, and vesicles prepared with untreated receptor contained 6.7 fmol of receptor and 64 fmol of G,. Inset, increment in ["S]GTP-yS binding caused by isoproterenol (A, proteolyzed receptor; A, untreated receptor). Data shown are the mean of duplicate assays.

16660 Regulatory Domains of the &Adrenergic Receptor

Tune (Mid Tlme (Mln)

FIG. 7. Thiol-stimulated GTPrS binding to G. mediated by proteolyzed &adrenergic receptor. Receptor-G, vesicles containing either proteolyzed receptor or untreated receptor were prepared as described in the legend to Fig. 6. Vesicles were incubated for 90 min at 0 "C in 20 mM NaHepes (pH 8.0), 1 mM EDTA, 3 mM MgC12, 100 mM NaC1, and 0.1 mM ascorbate with either 0.1 PM propranolol m), propranolol plus 5 mM dithiothreitol (e), 5 p M isoproterenol (El), or isoproterenol plus dithiothreitol(0). Adrenergic ligands were added here to stabilize reduced receptor (30). Vesicles were then assayed for ["SIGTPyS binding after 20-fold dilution in assay medium that contained either 5 FM isoproterenol (0,n) or 0.1 p M propranolol (e, m). Vesicles prepared with untreated receptors contained 4.6 fmol of receptor and 105 fmol of G. per assay point, and vesicles prepared with proteolyzed receptor contained 2.4 fmol receptor and 84 fmol of G.. Data shown are the means of duplicate determinations.

experiment shown in Fig. 6 used the preparations of receptor shown in Fig. 5.

The @-adrenergic receptor is stimulated in its regulatory activity by reduction of one or more intramolecular disulfide bonds. The thiol-reduced receptor promotes the binding of GTPyS to G. in the absence of agonist; the reduced and agonist-liganded receptor is more effective than the agonist- liganded receptor alone (32). The trypsin-proteolyzed and affinity repurified receptor retained its positive response to reduction, as shown in Fig. 7 . Reduction by dithiothreitol potentiated the activation of the receptor by the agonist isoproterenol to approximately the same relative extent as was observed with the untreated receptor. These data imply that the tryptic core of the receptor retains the disulfide bond that mediates the receptor's regulatory activity. Tryptic fragment A contains at least one disulfide that may serve this function: However, the proteolyzed receptor was essentially unresponsive to treatment with thiol in the absence of agonist (Fig. 7 ) . Little stimulation of GTPyS binding by the reduced, antagonist-liganded receptor was observed in this or other experiments. Although reduction of the receptor does decrease it stability to thermal denaturation, the partial loss of respon- siveness to thiols in the proteolyzed receptor does not appear to be a result of denaturation. In separate experiments (not shown), the stability at 30 "C of the ICYP-binding capacity of the dithiothreitol-reduced receptor before and after tryptic cleavage was unchanged.

DISCUSSION

This study represents an initial attempt to map the functional domains of the @-adrenergic receptor in terms of its primary structure. In general such an undertaking would require either direct spectrophotometric or crystallographic data on the tertiary structure of the protein. In this case, however, assignment is facilitated because the receptor is structurally homologous and functionally similar to rhodopsin, which has been studied in great detail (reviewed in Ref. 10). The homology among mammalian and avian @-adrenergic receptors, rhodopsins from several sources, and two forms of the muscarinic cholinergic receptor is concentrated primarily

R. C. Rubenstein, S. K-F. Wong, and E. M. Ross, unpublished data.

in seven hydrophobic regions (see 6,8). In the case of rhodopsin, these sequences are known to span the membrane bilayer and are arranged approximately normal to the membrane in a quasi-globular structure (see 10). Specific amino acid residues can therefore be inferred to lie on the cytoplasmic or extracellular face of the receptor or within the hydrophobic core region. The sensitivity of the @-adrenergic receptor to proteolysis in regions that are predicted to extend into the aqueous medium is consistent with this assignment of tertiary structure.

Within this hypothetical tertiary structure, it is of interest to determine the location of the regulatory binding domains on the receptor's surface, for agonist, G protein, kinases or phosphatases, oxidizing or reducing agents, and cytoskeletal components. The receptor-G, interface presumably lies on the cytoplasmic side of the plasma membrane. If the receptor binds directly to the a subunit of G., which is relatively hydrophilic (33), the G, regulating domain will presumably be composed of one or more of the hydrophilic cytoplasmic loops or the carboxyl-terminal region. Such an interaction is consistent with a receptor's ability to discriminate among G proteins, which differ primarily in their a subunits (5). If, however, a receptor binds to the by subunits, which bind directly to phospholipid bilayers and are not water soluble (33), the the G, binding domain may include more hydrophobic parts of the intramembranous core of the protein. Inter- action of receptor and by is suggested, but not proven, by data showing that the Py subunits of transducin are required for the binding of transducin to bleached rhodopsin (34; see also 2).

The data presented here essentially rule out the hydrophilic carboxyl-terminal domain and most of the largest intracellular hydrophilic loop as being directly involved with the regulation of G,. That the carboxyl-terminal region is not required for the regulation of G, is consistent with previous studies showing that proteolytic removal of the small carboxyl-terminal domain of rhodopsin is not required for the light- dependent regulation of transducin (11, 12). It also agrees with a recent study showing that deletion of the cDNA that encodes this region of the @-adrenergic receptor does not compromise the ability of the truncated receptor to regulate G. (13). The finding that the proteolytic removal of approximately 5000 Da of the major intracellular loop also does not diminish the regulatory activity of the receptor was somewhat more surprising. Dixon et al. (13) recently reported that mutational deletion of 33 amino acids from this loop yielded a @-adrenergic receptor that retained ligand binding activity but had lost the ability to activate adenylate cyclase. Kuhn and Hargrave (11) showed that proteolytic cleavage of this loop in rhodopsin blocked the light-dependent binding of transducin to disc membranes and Litman et al. (12) showed that similar cleavage inhibited the activation of the retinal phosphodiesterase. A single proteolytic cleavage in this loop was not inhibitory. The disparity between our data and the data on rhodopsin may perhaps be explained by the fact that this loop in rhodopsin is significantly shorter than the homologous region of the p-adrenergic receptor, and constraints on its structure may therefore be more rigorous. Proteolysis of rhodopsin may also have been more extensive. Because the proteolytic cleavage of this loop in the present study is roughly equivalent to the deletion produced by Dixon and co-workers (13), it is more surprising that the proteolyzed receptor retained activity while the mutant receptor did not. The sim- plest interpretation of this discrepancy is that the genetic deletion imposes alterations in the tertiary structure of the mutant receptor; i.e. the loss of activity in the mutant reflects

Regulatory Domains of the P-Adrenergic Receptor 16661

a disseminated conformational change. Proteolysis of the native, wild-type receptor might leave a correctly folded con- formation otherwise unaltered. This interpretation agrees with more recent data (35) showing that individual mutational deletion of three smaller segments of the loop does not inac- tivate the receptor as long as the total deletion is kept small.

The most difficult aspect of critically analyzing the present data is determining which peptide fragments of the receptor may remain associated in the active, repurified limit tryptic digest. The amino termini of fragments A and B have been determined directly and their carboxyl termini have been reasonably well defined based on molecular size, sequence of internal CNBr peptides, and retention or loss of specific immunogenic determinants. They are not covalently bound to each other because they can be separated on SDS-polyacrylamide gels or by reverse phase HPLC without reduction of disulfides. The proteolytic digestion or the chromatographic removal of the bulk of the intracellular loop is strongly suggested by the loss of specific immunological determinants (Fig. 2, for example) and the loss of radioactivity after labeling with 32P (Figs. 4 and 5). The carboxyl terminal domain was also cleaved at least twice. Conversion of the 52,000-Da receptor to the 42,000-Da species removes at least the two most carboxyl-terminal CNBr fragments, indicating cleavage at or before Lys4** (6). Further cleavage with trypsin either destroys peptide VI11 or cleaves the remaining carboxyl-terminal peptide, which contains peptide VIII, to small enough fragments that they are undetectable on polyacrylamide gels. Based on experience with peptides produced by treatment with pepsin in dilute formic acid, peptides smaller than the glucagon or insulin standards can be resolved, fixed, and stained by the methods used here, suggesting that other products of tryptic cleavage are either very small (e20 amino acids) or not detectable by silver staining.

We have no data that argue whether the glycosylated amino-terminal peptide, which is cleaved from the receptor by trypsin after Arg,,, remains noncovalently associated with fragments A and B after affinity chromatography and reconstitution. The only available antigenic determinant in this region (peptide 111) is itself cleaved by trypsin rendering it undetectable. Relatively convincing data argue that amino- terminal glycopeptide, at least through Aml,, is removed when the M , -52,000 receptor is converted to the M , -42,000 form (27, 28, 36). Whether the amino-terminal region SerI6-ArgZ9 is involved in ligand binding remains uncertain.

These data, taken together, suggest that the seven membrane-spanning regions and the shorter connecting loops form a compact core structure that is active, quite stable, and resistant to proteases. If the large intracellular loop and the carboxyl-terminal region of G protein-linked receptors are not involved in G-protein binding, one wonders what their function may be. This question is intensified by the observa- tions that there is minimal homology among receptors in these two regions (6-9) and that the sequences in these regions are unusual. The carboxyl-terminal domain of rhodopsin is the site of phosphorylation by rhodopsin kinase, which leads indirectly to inactivation of rhodopsin (37-38, and references therein). A similar role has been proposed for this region of the P-adrenergic receptor and for potential sites of phosphorylation by cyclic AMP-dependent protein kinase that are found in the large intracellular loop and carboxyl-terminal region (7 , 39). Whether phosphorylation and desensitization are the principal functions of these regions and whether phosphorylation of these sites can be assigned to specific kinases remains to be seen.

Because of their asymmetric orientation in the bilayer, the

definition of specific structural domains, and the presence of several interaction sites, the G protein-coupled receptors are attractive models for correlating specific structures with specific protein-protein interactions involved with transmembrane signaling. The sort of positive results obtained here and in several of the mutants described by Dixon and co-workers (13) allows strong inferences about what structures are not required for specific functions. Inferences about the role of a specific domain based on the loss of a function is more difficult to substantiate. Because these receptors are quite specific in recognizing ligands and G proteins, the chemical and mutational alteration of their selectivity for ligands may yield a clearer mapping of their structure-function relationships.

Acknowledgments-We are grateful to Dr. John Burnier, of Genen- tech, Inc., for preparing antigenic peptides and to Dr. Clive Slaughter, of the Department of Biochemistry, University of Texas Health Science Center, for performing the peptide sequencing. We thank S. Rozmiarek for excellent technical assistance, B. Strifler for purifying P-adrenergic receptor, K. C. McFarland for raising some of the antisera, and J. Corder0 for preparation of the manuscript.

REFERENCES 1. Smigel, M. D., Ross, E. M., and Gilman, A. G. (1984) in Cell

Membranes: Methods and Reviews (Elson, E. L., Frazier, W. A., and Glaser, L., eds). Vol. 2, pp. 247-294, Plenum Publishing Corp., New York

2. Stryer, L. (1986) Annu. Rev. Neurosci. 9, 87-119 3. Florio, V. A., and Sternweis, P. C. (1985) J. Bwl. Chem. 260 ,

4. Haga, K., Haga, T., and Ichiyama, A. (1986) J. Biol. Chem. 261 ,

5. Stryer, L., and Bourne, H. R. (1986) Annu. Rev. Cell Biol. 2,391- 419

6. Yarden, Y., Rodriguez, H., Wong, S. K-F., Brandt, D. R., May, D. C., Burnier, J., Harkins, R. N., Chen, E. Y., Ramachandran, J., Ullrich, A., and Ross, E. M. (1986) Proc. Natl. Acad. Sci. U.

7. Dixon, R. A. F., Kobilka, B. K., Strader, D. J., Benovic, J. L., Dohlman, H. G., Frielle, T., Bolanowski, M. A., Bennett, C. D., Rands, E., Diehl, R. E., Mumford, R. A., Slater, E. E., Sigal, I. S., Caron, M. G., Lefiowitz, R. J., and Strader, C. D. (1986) Nature 321 , 75-79

8. Kubo, T., Fukuda, K., Mikami, A., Maeda, A., Takahashi, H., Mishina, M., Haga, T., Haga, K., Ichiyama, A., Kangawa, K., Kojima, M., Matsuo, H., Hirose, T., and Numa, S. (1986) Nature 323,411-416

9. Kubo, T., Maeda, A., Sugimoto, K., Akiba, I., Mikami, A., Taka- hashi, H., Haga, T., Haga, K., Ichiyama, A., Kangawa, K., Matsuo, H., Hirose, T., and Numa, S. (1986) FEBS Lett. 209 ,

10. Findlay, J. B. C., and Pappin, D. J. C. (1986) Biochem. J. 238 ,

11. Kiihn, H., and Hargrave, P. A. (1981) Biochemistry 20, 2410-

12. Litman, B. J., Aton, B., and Harley, J. B. (1982) Vision Res. 2 2 ,

13. Dixon, R. A. F., Sigal, I. S., Rands, E., Register, R. B., Candelore, M. R., Blake, A. D., and Strader, C. D. (1987) Nature 326,73- 77

14. Asano, T., Pedersen, S. E., Scott, C. W., and Ross, E. M. (1984) Biochemistry 23,5460-5467

15. Engel, G., Hoyer, D., Berthold, R., and Wagner, H. (1981) Nau- nyn-Schmiedebergs Arch. Pharmacol. 3 1 7 , 277-285

16. Johnson, R. H., and Walseth, T. F. (1979) Adu. Cyclic Nucleotide Res. 10 , 135-168

17. Kikkawa, U., Minakuchi, R., Takai, Y., and Nishizuka, Y. (1983) Methods Enzyml . 99, 288-298

18. Caron, M. G., Srinivasan, Y., Pitha, J., Kociolek, K., and Lefiow- itz, R. J. (1979) J. Bwl. Chem. 254, 2923-2927

19. Brandt, D. R., and Ross, E. M. (1986) J. Biol. Chem. 261,1656- 1664

20. Fleming, J. W., and Ross, E. M. (1980) J. Cyclic Nucleotide Res.

21. Sternweis, P. C., Northup, J. K., Smigel, M. D., and Gilman, A.

3477-3483

10133-10140

S. A. 83,6795-6799

367-372

625-642

2417

1439-1442

6,407-419


G. (1981) J. Biol. Chem. 256,11517-11526

J. Biol. Chem. 260, 15829-15833 22. May, D. C., Ross, E. M., Gilman, A. G., and Smigel, M. D. (1985)

23. Laemmli, U. K. (1970) Nature 227, 680-685 24. DeWald, D. B., Adams, L. D., and Pearson, J. D. (1986) Anal.

25. Wray, W., Boulikas, T., Wray, V. P., and Hancock, R. (1981) Anal. Biochem. 118,197-203

26. Green, N., Alexander, H., Olson, A., Alexander, S., Shinnick, T. M., Sutcliffe, J. G., and Lerner, R. A. (1982) Cell 28, 477-487

27. Jiirss, R., Hekman, M., and Helmreich, E. J. M. (1985) Biochem-

28. Boege, F., Jiirss, R., Cooney, D., Hekman, M., Keenan, A. K.,

29. Turner, R. S., Kemp, B. E., Su, H., and Kuo, J. F. (1985) J. Biol.

30. Gould, K. L., Woodgett, J. R., Cooper, J. A., Buss, J. E., Shallo-

Biochem. 154, 502-508

istry 24,3349-3354

and Helmreich, E. J. M. (1987) Biochemistry 26, 2418-2425

Chem. 260, 11503-11507

way, D., and Hunter, T. (1985) Cell 42,849-857

31. House, C., Wettenhall, R. E. H., and Kemp, B. E. (1987) J. Biol.

32. Pedersen, S. E., and Ross, E. M. (1985) J . Biol. Chem. 260,

33. Sternweis, P. C. (1986) J. Biol. Chem. 261, 631-637 34. Fung, B. K.-K. (1983) J. Biol. Chem. 258,10495-10502 35. Strader, C. D., Sigal, I. S., Blake, A. D., Cheung, A. H., Register,

R. B., Rands, E., Zemcik, B. A., Candelore, M. R., and Dixon, R. A. F. (1987) Cell 49, 855-863

36. Cervantes-Olivier, P., Durieu-Trautmann, O., Delavier- Klutchko, D., and Strosberg, A. D. (1985) Biochemistry 24,

37. Sitaramayya, A., and Liebman, P. A. (1983) J. Biol. Chem. 258,

38. Miller, J. L., Fox, D. A., and Litman, B. J. (1986) Biochemistry

39. Benovic, J. L., Mayor Jr., F., Somers, R. L., Caron, M. G., and

Chem. 262, 772-777

14150-14157

3765-3770

12106-12109

25,4983-4988

Lefkowitz, R. J. (1986) Nature 321,869-872

The Hydrophobic Tryptic Core of the @-Adrenergic Receptor ...

Documents

Transcript of The Hydrophobic Tryptic Core of the @-Adrenergic Receptor ...