Propertiesof the f3i- and /32-AdrenergicReceptor Subtypes Revealed by Molecular Cloning

CLIN.CHEM,35/5, 721-725 (1989) Propertiesof the f3i- and /32-AdrenergicReceptor Subtypes Revealed by Molecular Cloning ThomasFrielle,Marc G. Caron, and Robert J. Lefkowltz The f3- and f32-adrenergicreceptorsubtypesare biochemically and functionallysimilar,because both receptorsmediate the catecholamine-dependent activation of adenylate cyclasethroughthe GTP-bindingprotein,G. Pharmacologically, the two receptorscan be distinguishedon the basisof their relativeaffinitiesfor the agonistsepinephrineand norepinephrine as well as their affinitiesfor several selective antagonists.the primarystructuresof the humanf3- and /32- adrenergicreceptorshave recently been deduced from the cloning of their genes and (or) cdna5, revealing high sequence homology and a membrane topography of seven putativetransmembraneregionssimilarto that of rhodopsin. Chimenc 131/f32-adrenergic receptorcdnas have been constructed by site-directedmutagenesisand the chimericrna transcriptsexpressed in Xenopus Iaevis oocytes. The pharmacological properties of the expressed chimeric receptor proteinswere assessed by radioligandbindingutilizingsubtype-selectiveagonistsand antagonists.apparently,several of the putativetransmembraneregions contribute significantly to the determination of subtype selectivity, presumably by formation of a ligand-binding pocket, with determinants for agonistand antagonistbindingbeing distinguishable. The translation of extracellular signals into intracellular metabolic events is often dependent upon a system of three interacting components: (a) a receptor capable of specifically binding selected uganda, (b) an effector enzyme, and (c) a guanine nucleotide-binding regulatory protein that couples the receptor to the effector enzyme. The 13-adrenergic receptor-coupled adenylate cyclase (EC 4.6.1.1) system is one of the most thoroughly characterized transmembrane signaling systems. In this system the agonist-occupied /3-adrenergic receptor is capable of activating adenylate cyclase through its interaction with the guanine nucleotide-binding regulatory protein, O (1). Two pharmacologically distinct subtypes of the mainmalian /3-adrenergic receptor, termed f3 and 132, have been described. The definition of the two subtypes is based upon their relative affinities for the catecholamine agomsts epinephrine and norepinephrine (2). The /31-adrenergic receptor binds epinephrine and norepinephrine with approximately equal affinities, whereas the /32-adrenergic receptor binds epinephrine with about 3-fold greater affinity than it does norepinephrine. Subsequent to this initial pharmacological classification, various 13-adrenergic receptor agomsts and antagonists have been developed that distinguish between the two p-adrenergic receptor subtypes and that have allowed their quantification and characterization (3). The distinction between the /3i and /32 subtypes is demonstrated most clearly by the competition of a nonselective radiolabeled antagonist by a subtype-selective agonist or antagonist. The same selectivity can also be demonstrated by the Departments of Medicine, Biochemistry, and Cell Biology, Howard Hughes Medical Institute at Duke University Medical Center, Durham, NC 2771. Received (prereviewed) February 14, 1989; accepted February 15, 1989. ability of selective agomsts or antagonists to activate or inhibit stimulation of adenylate cyclase. The (3k- and 132-adrenergic receptors are functionally homologous, because they are both coupled to the stimulation of adenylate cyclase through activation of G. Biochemically, these two mammalian proteins possess identical molecular masses by sodium dodecyl sulfate/polyacrylamide gel electrophoresis, and peptide maps generated from photoaffinity-labeled receptors are virtually identical (4). Thus, the question has arisen whether the f3- and the (32-adrenergic receptors represent post-translationally modified products of a common gene or are products of different genes. Recent data, reviewed in this essay, establish that the two receptors are, in fact, products of different genes. It is becoming increasingly clear that, based on structural and functional similarities, the adrenergic receptors are members of a large family of -protein-coupled receptors. Recently cloned members of this family include human rhodopsin (5), the human color opsins (6), the a1- (7) and a2- adrenergic receptors (8), the hamster (9) and human /32- adrenergic receptors (1), the avian /3-adrenergic receptor (11), and the cerebral (12) and cardiac muscarinic cholinergic receptors (13), as well as the receptor for substance K (14) and two additional recently described muscarinic receptors (15). On the basis of hydropathicity analyses of the deduced amino acid sequences, it has been proposed (16) that the membrane organization of the receptors mimics the structure of bacterio-rhodopsin as determined by electron diffraction studies. This structure of seven a-helical membrane-spanning regions of 2 to 28 amino acid residues, each connected by intra- and extracellular hydrophilic loops, is highly conserved among all of these proteins (17). Like other members of this family, the human /32-adrenergic receptor possesses seven clusters of hydrophobic amino acids, likely representing a-helical membrane-spanning regions (18). Figure 1 compares the primary structures of human /31-adrenergic receptor with those of the avian /3- adrenergic receptor and the human /32-adrenergic receptor. The /32-adrenergic receptor consists of 477 amino acid residues, whereas the avian /3-adrenergic receptor and the human (32-adrenergic receptor have 483 and 413 residues, respectively. Sequence similarities between the human f3- adrenergic receptor and the avian /3-adrenergic receptor (Figure IA) and the human /3- and J32-adrenergic receptors (Figure 1B) are striking and highest in the presumed transmembrane regions. The percentage of amino acid homologies within the transmembrane regions and the overall sequence homologies, respectively, between the human /31-adrenergic receptor and the other two receptors are: avian /3-adrenergic receptor 84%, 69%, and human /34- adrenergic receptor 71%, 54%. Thus the /31-adrenergic receptor is remarkably similar to the avian /3-adrenergic receptor but is less closely related to the /34-adrenergic receptor. The extensive homology of the membrane-spanning regions among the three receptors suggests that these regions are subject to a selective pressure that acts to prevent divergence. These regions must therefore embody functions such as ligand binding and protein coupling common to all CLINICAL CHEMISTRY, Vol. 35, No. 5, 1989 721

1\ C E : ES Fig. 1. ComparIson of the primary structures for various/3-adrenergic receptorsubtypes The primary structure for the human p,-adreneic receptor is shown in both diagrams. A (), the residues highlighted represent residues identical with the avian -adrenergic receptor. B (bosom), the highlighted residues are those identicalwith the human /3-adrenergtc receptor. The receptore are represented as they may be organized inthe membrane, based on the structurededuced for bacterlo-rhodopsi, (26) the receptors and integral to the functioning of these receptors. In contrast to the extensive similarities within the membrane-spanning regions, the amino and carboxy termini of the receptors are quite divergent. The first two cytoplasmic loops are generally well conserved, whereas the extracellular loops are more divergent. The third cytoplasmic loop of the f31-adrenergic receptor diverges significantly from that of the avian receptor and includes an additional 17 amino acid residues. The third cytoplasmic loop of the /3kadrenergic receptor is also quite divergent from that of the /34-adrenergic receptor. Not only is it 27 amino acid residues longer than that of the f.32-adrenergic receptor, it also I contains 16 additional proline residues. Given the rotational constraints imposed by proline, the third cytoplasmic loop of the /34-adrenergic receptor could have a very different tertiary structure than that of the (32-adrenergic receptor. Recent mutagenesis studies (19,2) with the hamster /34-adrenergic receptor and the human f32-adrenergic receptor (21) have suggested two short segments of this third cytoplasmic loop, situated near transmembrane segments 5 and 6, respectively, as being important in coupling of the /32 receptor to G8. However, the variations in length and composition of this loop among receptors suggest that these domains may confer specificity for G protein coupling or some other function of the receptors. The data reviewed in this essay indicate that mammalian /3k- and /34-adrenergic receptors are products of distinct genes. This conclusively eliminates the possibility discussed previously, that differences in agonist and antagonist affinities for the two subtypes are due to post-translational modifications. As stated, the only apparent functional differences between the human /3- and fradrenergic receptors are the differences in epinephrine and norepinephrine affinities and in subtype-selective antagonist affinities. To determine which sequences within the receptors are responsible for such subtype specificity, chimeric receptor proteins containing sequences derived from both the /3- and the J34-adrenergic receptors have been constructed by using site-directed mutagenesis and expressed in Xenopus laeuis oocytes (22). Each chimeric receptor was assessed for subtype-specific agonist and antagonist binding characteristics, permitting delineation of those structural domains that determine the pharmacological differences between the /32-and /34-adrenergic receptor subtypes. Characteristic agonist competition profiles for the (.3k-and the /34-adrenergic receptors are shown in Figure 2, A and H, respectively. These were generated with oocytes injected with RNA for each of these two receptors. The /3-adrenergic agonist isoproterenol is the most potent competing ligand for either receptor subtype. Consistent with the well-established pharmacology of (3k- and p2-adrenergic receptors, epinephrine and norepinephrine are approximately equipotent in competing with the antagonist radioligand for binding to the (Jj-adrenergic receptor, whereas for the /34- adrenergic receptor, epinephrine is approximately 2 times more potent than norepinephrine. Replacement of the N-terminus and the first one, two, or three a-helices of the /34-adrenergic receptor with the corresponding regions of the /34-adrenergic receptor [chimeras, /32(I)//31(H-VU), f32(i,ll)//32(ffl-vll), and /34(I-ffl)//31(1V-VU)1 resulted in epinephrine and norepinephrine competition proffles resembling the proffles for the /31-adrenergic receptor (Figure 2, B, C, and D), i.e., epinephrine and norepinephrine are approximately equipotent. These results suggest that those amino acid residues that determine /3-adrenergic receptor subtype specificity reside within a-helices IV-VII and that the N-terminus and first three a-helices do not contribute to agonist subtype specificity. When helix W of the /34-adrenergic receptor is substituted for the corresponding helix of the /34-adrenergic receptor [/32(I-IV)//31( V-VU) (Figure 2E)}, the KD for norepinephrine increases by 1-fold when compared with the preceding /3- /32 chimera [f32(i-ffl)//31(w-vh)1 (Figure 2D). If one compares the KD norepinephrinelkd epinephrine ratio for the native receptors, the value is -1 for the /31-adrenergic receptor, -15 for the f32-adrenergic receptor, and -16 for f32(i-w/(31(v-vh) (Figure 2, A, H, and E, respectively). 722 CLINICALCHEMISTRY,Vol.35, No.5, 1989

L. \ L Fig. 2. Competition by agonists of lmiiabeied cyanopindololbinding to /3-, /32-adrenergicreceptor (AR) chimeras Binding of the nonselective antagonist I-labe4ed cyanopindolol at a final concentration of 75 pmovlwas assessed in the presence of various concentrations of isopioterenol (ISO, S), epinephnne (EP! ), and norepinephnne (NOR, A). The Ko for I-IabeIed cyanopindolol of each chimerawas determined and did not differ significantlyfromthose forthe wild-type receptors (5-1 pmol/l). The peptide sequence derived from the p-ar is indicated by an open line, that derived from the $2- AR by a solid fine. Panels A and H represent competition curves for the wild-type$1-arand /32-AR,respectively. PanelsB throughg representcurvesforthe following chimericreceptors, respectively: B, $2(l)/p1(II-Vll); C, $2(Hl)/$(llI-Vll); D, $2(Hllt)/$1(IV-Vll); E $2(Hv)/131(V-Vll);F, -V)/$(VI-VII); $2(l-Vl)/$1(Vll).Each competition curve shown is representative of two or three experiments from which geometric means were calculated for the agonist l( s. Inthe crude oocyte membrane preparations used for these studies, most of the agonist competition curveswereadequately modeled to a singlealfiorty state.when,as was found to be the case ina few instances, a two-site modelwas statisticallymore appropriate,the high-affinityk was reportedin the figure AGONIST (U) Thus, the relative affinities of norepinephrine and epinephrine binding to /32(I-IV)//31(V-VII) are essentially identical to those for the /32-adrenergic receptor. These data suggest that amino acid residues within a helix N contribute strongly to determining the agonist subtype specificity of the /3-adrenergic receptor. The sequential addition of (32- adrenergic receptor helices V, VI, and VU in chimeras /32(1- V)1f31(VI-V11),/34(I-VI)//31(Vfl), and the /34-adrenergic receptor, does not result in any further significant changes in the KD norepinephrine/kd epinephrine ratio. A decrease in the apparent potency of both epinephrine and norepinephrine was observed when the /34-adrenergic receptor was compared with the chimeras of Figure 2, E, F, and G. This may indicate a potential contribution of helix VU to the affinity of agonists. These data support the conclusion that residues in helices 1,11, ifi, V, and VU do not significantly determine (32- and (32-adrenergic receptor agonist subtype specificity. The subtype specificity of the chimeric receptors was also assessed by using the (32-adrenergic receptor-selective antagonist betaxolol and the (32-adrenergic receptor-selective antagonist ICI 118551 as competing ligands. In contrast to the results observed for agonists, it is apparent from Figure 3 that as /32-adrenergic receptor helices are sequentially substituted for the corresponding /31-adrenergic receptor helices, the potencies of betaxolol and of ICI 118551 gradually, change. However, several of the a-helices appear to contribute more significantly than others to determining antagonist potency. Thus the addition of /34-adrenergic receptor helices ifi, VI, and VII each result in six- to 1-fold changes in the KD betaxolollkd ICI ratio, when compared with each preceding chimera. Addition of helix I leads to no change in the ratio, whereas addition of helices II, N, and V lead to more modest (-twofold) changes in the ratio. The addition of helices III, N, and VU results in significant increases in the KD of betaxolol, whereas addition of helices II, V, and VI results in significant decreases in the KD of ICI 118551. This suggests that residues within all of these helices are involved in determining (32- vs /34-adrenergic receptor antagonist specificity. The high degree of sequence homology in the presumed transmembrane regions of the various adrenergic receptors as well as previous mutagenesis (19) and chimeric receptor work (23) all suggest that several of the transmembrane helices are involved in forming the ligand binding site of these receptors. The extracellular, hydrophilic loops are apparently not essential for ligand binding, because deletion of these residues does not alter ligand affinity (24). Therefore, even though our chimeric receptors contain both hydrophilic and hydrophobic sequences derived from both /32- and (32-adrenergic receptors, we have interpreted the data in terms of which receptor subtype contributes a particular membrane-spanning domain. Interestingly, when /32- and /32-adrenergic receptor sequences are compared (Figure 1), a-helices N, VI, and VU are the most divergent of the putative membrane-spanning domains, suggesting that those residues that determine the pharmacological differences between the subtypes may be located within these a-helices. Results of photoaffinity labeling techniques suggest that residues within several of the a- helices of the adrenergic receptors contribute to the ligand binding site. Labeled by such techniques are residues located within helix U of the /32-adrenergic receptor (25), and residues within helices 11-V and helix VU of the avian (32- adrenergic receptor (26). The use of chimeric receptors provides a potentially powerful approach for dissecting the structural basis of G- protein-coupled receptor function. On the basis of conclusions derived from photoaffinity labeling data and from CLINICAL CHEMISTRY, Vol. 35, No. 5, 1989 723

4 1 2 z C,.O S S.7.5.5.4.1.5 S S.7.5.5.4.1.1.1 S.7 5 4.5.3,5.5,5. 9 ANTAGONIST (U) Fig. 3. Competition by antagonistsof lssiiabeied cyanopindolol (AR)binding to /3-, (32-adrenergic receptor (AR) chimeras Bindingof the nonselective antagonist mnllebeled cyanopindolol (75 pmot/l)to Wild-typeand chimeric receptors was assessed in the presence of varying concentrations of the $,-AR subtype-selective antagonist betaxolol (BETA) #{149}) and the $2-AR subtype-selective antagonist ICI 118551 (Id, U). Wdd-type and chimeric receptor designations are as indicatedin the legend to Figure2. Each competitioncurve is representative of two or three experiments from which geometric means were calculated for the antagonisti( s chimeric receptor studies, a model can be proposed for the arrangement of the a-helical membrane-spanning domains within the plasma membrane. In such a configuration, those helices that most significantly contribute to /3-adrenergic receptor agonist or antagonist subtype specificity (N, VI, VU) are grouped together, suggestive of a ligand binding pocket. As a consequence of this arrangement, those additional helices that at least in part determine /3-adrenergic receptor antagonist subtype specificity (11, ifi, V) also could contribute to the formation of such a pocket. In summary, these results with chimeric Pi-, /32-adrenergic receptors suggest that subtype specificity is determined by most of the transmembrane spanning regions of the molecules. This further strengthens the emerging idea that binding of ligands by such receptors occurs in a pocket formed by the clustering of these membrane-spanning a- helices. While competitively occluding this binding pocket, antagonists nonetheless interact with structural determinants distinct from those with which agonists interact. These observations are consistent with the older notion of antagonists interacting with accessory binding sites on their receptors in addition to the sites with which the natural agonists interact. These studies further underscore the value of chimeric receptors for elucidating the structural basis of receptor functions. References 1. Gilman AG. G proteins: transducers of receptor-generated signals [Review]. Ann Rev Biochem 1987;56:615-5. 2. Lands AM, Arnold A. McAuIIff JP, Ludena FP, Brown PG. Differentiation of receptor systems by sympathomimetic amines. Nature (London) 1967;214:597-8. 3. Stiles GL, Lefkowitz RJ. Cardiac (3-adrenergic receptors [Review]. Annu Rev Med 1984;35:149-64. 4. Stiles GL, Strasser RH, Caron MG, Lefkowitz RJ. Mammalian /3- adrenergic receptors: structural differences in f3 and /3. subtypes revealed by peptide map. J Biol Chem 1983;258:1689-94. 5. Nathans J, Hogness DS. Isolation and nucleotide sequence of the gene encoding human rhodopsin. Proc Natl Acad Sci USA 1984;81:4851-5. 6. Nathans J, Thomas D, Hogness DS. Molecular genetics of human color vision: the genes encoding pigments. Science 1986;232:193-22. blue, green, and red 7. Cotecchia S, Schwinn DA, Randall RR, Lefkowitz RJ, Caron MG, Kobilka BK. Molecular cloning and expression of the cona for the hamster alpha1-adrenergic receptor. Proc NatI Acad Sci USA 1988;85:7159-63. 8 Kobilka BK, Matsui H, Kobilka PS, et al. Cloning, sequencing and expression of the gene coding for the human platelet a2- adrenergic receptor. Science 1987;238:65-6. 9. Dixon RAF, Kobilka BK, Strader 111, et al. Cloning of the gene and cdna for mammalian /3-adrenergic receptor and homology with rhodopein. Nature (London) 1986;321:75-9. 1. Kobilka BK, Dixon RAF, Frielle T, et al. cdna for the human /3-adrenergic receptor a protein with multiple membrane spanning domains and a chromosomal location shared with the PDGF receptor gene. Proc Natl Acad Sci USA 1987;84:46-5. 11. Yarden Y, Rodrigues H, Wong SKF, et al. The avian /3- adrenergic receptor: primary structure and membrane topology. Proc Nati Acad Sci USA 1986,;83:6795-9. 12. Kubo T, Fukuda K, Mikami A, et al. Cloning, sequencing and expression of cdna for muscarinic acetylcholine receptor. Nature (London) 1986;232:411-6. 13. Kubo T, Maeda A, Sugimoto K, et al. Primary structure of porcine cardiac muscarinic acetylcholine receptor deduced from the cdna sequence. FEBS Lett 1986;29:367-72. 14. Masu Y, Nakayama K, Tanmaki H, Harada Y, Kung M, Nakahishi S. cdnacloning of bovine substance-k receptor through oocyte expression system. Nature (London) 1987;239:836-8. 15. Peralta EG, Ashkenazi A, Winslow JW, Smith DH, Ramachandran J, Capon D,J. Distinct primary structures, ligand binding properties and tissue specific expression of four human muscarinic acetylcholine receptors. EMBO J 1987;6:3923-9. 16. Henderson R, Unwin PNT. Three-dimensional model of purple membrane obtained by electron microscopy. Nature (London) 1975;257:28-32. 17. Dohlman HG, Caron MG, Lefkowitz RJ. A family of receptors coupled to guanine nucleotide regulatory proteins [Review]. Biochemistry 1987;26:2657-64. 18. Frielle T, Collins S, Daniel KW, Caron MG, Lefkowitz RJ. 724 CLINICALCHEMISTRY,Vol.35, No.5, 1989

Cloning of the CDNA for the human -adrenergic receptor. Proc Natl Acad Sci USA 1987;84:792-4. 19. Dixon RAF, Sigal IS, Rands E, et al. Ligand binding to the /3- adrenergic receptor involves its rhodopsin like core. Nature (London) 1987;326:74-.-7. 2. Strader CD, Sigal IS, Blake AD, et al. The carboxyl terminus of the hamster /3-adrenergic receptor expressed in mouse L cells is not required for receptor sequestration. Cell 1987;49:855-63. 21. O Dowd BF, Hnatowich M, Regan JW, Leader WM, Caron MG, Lefkowitz RJ. Site-directed mutagenesia of the cytoplasmic domains of the human /32-adrenergic receptor localization of regions involved in G protein-receptor coupling. J Biol Chem 1988; 263:15985-92. 22. Frielle T, Daniel KW, Caron MG, Lefkowitz RJ. Structural basis of /3-adrenergic receptor subtype specificity studied with chimeric //32-adrenergic receptors. Proc Nati Acad Sci USA 1988 (in press). 23. Kobilka BK, Matsui H, Kobilka PS, et al. Cloning, sequencing and expression of the gene coding for the human platelet a2- adrenergic receptor. Science 1987;238:65-6. 24. Dixon RAF, Sigal IS, Candelore MR., et al. Structural features required for ligand binding to the /3-adrenergic receptor. EMBO J 1987;6:3269-75. 25. Dohlman HG, Caron MG, Strader CD, Amlaiky N, Lefkowitz RJ. Identification and sequence of a binding site peptide of the /32- adrenergic receptor. Biochemistry 1988;27:1813-7. 26. Wong SK-F, Slaughter C, Ruoho AR, Ross EM. The catecholamine binding site of the /3-adrenergic receptor is formed by juxtaposed membrane-spanning domains. J Biol Chem 1988;263:7925-8. CLINICALCHEMISTRY,Vol.35, No.5, 1989 725