Portal de Periódicos da CAPES

The Fourth Transmembrane Segment Forms the Interface of the Dopamine D2 Receptor Homodimer

2003; Elsevier BV; Volume: 278; Issue: 7 Linguagem: Inglês

10.1074/jbc.c200679200

1083-351X

Wen Guo, Lei Shi, Jonathan A. Javitch,

Neurotransmitter Receptor Influence on Behavior

Considerable evidence suggests that G-protein-coupled receptors form homomeric and heteromeric dimersin vivo. Unraveling the structural mechanism for cross-talk between receptors in a dimeric complex must start with the identification of the presently unknown dimer interface. Here, by using cysteine cross-linking, we identify the fourth transmembrane segment (TM4) as a symmetrical dimer interface in the dopamine D2 receptor. Cross-linking is unaffected by ligand binding, and ligand binding and receptor activation are unaffected by cross-linking, suggesting that the receptor is a constitutive dimer. The accessibility of adjacent residues in TM4, however, is affected by ligand binding, implying that the interface has functional significance. Considerable evidence suggests that G-protein-coupled receptors form homomeric and heteromeric dimersin vivo. Unraveling the structural mechanism for cross-talk between receptors in a dimeric complex must start with the identification of the presently unknown dimer interface. Here, by using cysteine cross-linking, we identify the fourth transmembrane segment (TM4) as a symmetrical dimer interface in the dopamine D2 receptor. Cross-linking is unaffected by ligand binding, and ligand binding and receptor activation are unaffected by cross-linking, suggesting that the receptor is a constitutive dimer. The accessibility of adjacent residues in TM4, however, is affected by ligand binding, implying that the interface has functional significance. G-protein-coupled receptors (GPCRs) 1The abbreviations used are: GPCR, G-protein-coupled receptor; TM, transmembrane segment; D2R, dopamine D2 receptor; PBS, phosphate-buffered saline; NEM, N-ethylmaleimide; DM, dodecyl maltoside; BRET, bioluminescence resonance energy transfer; CuP, copper phenanthroline comprise a large superfamily of receptors that couple binding of a diverse group of ligands to activation of heterotrimeric G-proteins (1Gether U. Endocr. Rev. 2000; 21: 90-113Crossref PubMed Scopus (1002) Google Scholar). A number of class C GPCRs have been shown to form dimers in the plasma membrane, including the calcium-sensing receptor (2Bai M. Trivedi S. Brown E.M. J. Biol. Chem. 1998; 273: 23605-23610Abstract Full Text Full Text PDF PubMed Scopus (347) Google Scholar), the GABABreceptor (3White J.H. Wise A. Main M.J. Green A. Fraser N.J. Disney G.H. Barnes A.A. Emson P. Foord S.M. Marshall F.H. Nature. 1998; 396: 679-682Crossref PubMed Scopus (1023) Google Scholar, 4Jones K.A. Borowsky B. Tamm J.A. Craig D.A. Durkin M.M. Dai M. Yao W.J. Johnson M. Gunwaldsen C. Huang L.Y. Tang C. Shen Q. Salon J.A. Morse K. Laz T. Smith K.E. Nagarathnam D. Noble S.A. Branchek T.A. Gerald C. Nature. 1998; 396: 674-679Crossref PubMed Scopus (933) Google Scholar, 5Kaupmann K. Malitschek B. Schuler V. Heid J. Froestl W. Beck P. Mosbacher J. Bischoff S. Kulik A. Shigemoto R. Karschin A. Bettler B. Nature. 1998; 396: 683-687Crossref PubMed Scopus (1023) Google Scholar), and the metabotropic glutamate receptors (6Romano C. Yang W.L. O'Malley K.L. J. Biol. Chem. 1996; 271: 28612-28616Abstract Full Text Full Text PDF PubMed Scopus (452) Google Scholar). In the case of the GABAB receptor, heterodimerization is essential for proper trafficking to the cell surface (7Margeta-Mitrovic M. Jan Y.N. Jan L.Y. Neuron. 2000; 27: 97-106Abstract Full Text Full Text PDF PubMed Scopus (591) Google Scholar). Furthermore, the evidence supports a scenario in which the binding of GABA to the R1 subunit causes the R2 subunit to bind to and activate G-protein (8Havlickova M. Prezeau L. Duthey B. Bettler B. Pin J.P. Blahos J. Mol. Pharmacol. 2002; 62: 343-350Crossref PubMed Scopus (83) Google Scholar, 9Duthey B. Caudron S. Perroy J. Bettler B. Fagni L. Pin J.P. Prezeau L. J. Biol. Chem. 2002; 277: 3236-3241Abstract Full Text Full Text PDF PubMed Scopus (173) Google Scholar). In addition, mounting evidence supports the hypothesis that many class A rhodopsin-like receptors, including the dopamine D2 receptor (D2R) (10Armstrong D. Strange P.G. J. Biol. Chem. 2001; 276: 22621-22629Abstract Full Text Full Text PDF PubMed Scopus (119) Google Scholar, 11Lee S.P. O'Dowd B.F. Ng G.Y. Varghese G. Akil H. Mansour A. Nguyen T. George S.R. Mol. Pharmacol. 2000; 58: 120-128Crossref PubMed Scopus (145) Google Scholar, 12Ng G.Y. O'Dowd B.F. Lee S.P. Chung H.T. Brann M.R. Seeman P. George S.R. Biochem. Biophys. Res. Commun. 1996; 227: 200-204Crossref PubMed Scopus (248) Google Scholar, 13Zawarynski P. Tallerico T. Seeman P. Lee S.P. O'Dowd B.F. George S.R. FEBS Lett. 1998; 441: 383-386Crossref PubMed Scopus (120) Google Scholar), in membrane and in some cases in detergent, are dimeric as well (reviewed in Refs. 14Rios C.D. Jordan B.A. Gomes I. Devi L.A. Pharmacol. Ther. 2001; 92: 71-87Crossref PubMed Scopus (292) Google Scholar, 15George S.R. O'Dowd B.F. Lee S.P. Nat. Rev. Drug Discov. 2002; 1: 808-820Crossref PubMed Scopus (550) Google Scholar, 16Angers S. Salahpour A. Bouvier M. Annu. Rev. Pharmacol. Toxicol. 2002; 42: 409-435Crossref PubMed Scopus (518) Google Scholar). Unraveling the structural mechanism for cross-talk between receptors in a dimeric complex must start with the identification of the presently unknown dimer interface. Two hypotheses have been proposed for GPCR dimerization: domain swapping and contact dimerization (17Gouldson P.R. Higgs C. Smith R.E. Dean M.K. Gkoutos G.V. Reynolds C.A. Neuropsychopharmacology. 2000; 23: S60-S77Crossref PubMed Scopus (127) Google Scholar, 18Gouldson P.R. Snell C.R. Bywater R.P. Higgs C. Reynolds C.A. Protein Eng. 1998; 11: 1181-1193Crossref PubMed Scopus (115) Google Scholar), although recent experimental results seem inconsistent with domain swapping being the dominant form of dimerization (19Schulz A. Grosse R. Schultz G. Gudermann T. Schoneberg T. J. Biol. Chem. 2000; 275: 2381-2389Abstract Full Text Full Text PDF PubMed Scopus (88) Google Scholar, 20Hamdan F.F. Ward S.D. Siddiqui N.A. Bloodworth L.M. Wess J. Biochemistry. 2002; 41: 7647-7658Crossref PubMed Scopus (34) Google Scholar). Synthetic peptides of TM6 block dimerization and activation of the β2-adrenergic receptor (21Hebert T.E. Moffett S. Morello J.P. Loisel T.P. Bichet D.G. Barret C. Bouvier M. J. Biol. Chem. 1996; 271: 16384-16392Abstract Full Text Full Text PDF PubMed Scopus (685) Google Scholar), and it has been suggested that the GXXXG motif in TM6 may be involved in dimerization of this receptor, although this motif is not highly conserved in other family A GPCRs. Synthetic peptides of TM6 and of TM7 also blocked dimerization of the D2 receptor (D2R) (12Ng G.Y. O'Dowd B.F. Lee S.P. Chung H.T. Brann M.R. Seeman P. George S.R. Biochem. Biophys. Res. Commun. 1996; 227: 200-204Crossref PubMed Scopus (248) Google Scholar). Although these findings were specific to these particular synthetic peptides, the data do not necessarily establish TM6 and/or TM7 as the dimer interface, as the peptide-receptor interactions might modulate the ability of the receptor to form dimers at a different interface. In addition, if the receptors form higher order oligomers, multiple interfaces must exist. Computational studies have proposed a number of different potential interfaces (22Gouldson P.R. Dean M.K. Snell C.R. Bywater R.P. Gkoutos G. Reynolds C.A. Protein Eng. 2001; 14: 759-767Crossref PubMed Scopus (53) Google Scholar, 23Filizola M. Olmea O. Weinstein H. Protein Eng. 2002; 15 (in press)Crossref PubMed Scopus (78) Google Scholar, 24Filizola M. Weinstein H. Biopolymers (Peptide Science). 2002; 66: 317-325Crossref PubMed Scopus (80) Google Scholar), but these have not yet been experimentally verified. By using cysteine cross-linking, we have now explored the possibility that the D2R exists as a dimer or higher order oligomer in the plasma membrane. We show that the D2R is at least a homodimer, with the extracellular end of TM4 at a symmetrical dimer interface. Residues are numbered according to their positions in the human dopamine D2short receptor sequence. We also index residues relative to the most conserved residue in the TM in which it is located (32Visiers I. Ballesteros J.A. Weinstein H. Methods Enzymol. 2002; 343: 329-371Crossref PubMed Scopus (173) Google Scholar). The most conserved residue is assigned the position index “50,” e.g. Trp-1604.50, and therefore Val-1594.49 and Val-1614.51. Mutations were generated as described previously (30Javitch J.A. Shi L. Simpson M.M. Chen J. Chiappa V. Visiers I. Weinstein H. Ballesteros J.A. Biochemistry. 2000; 39: 12190-12199Crossref PubMed Scopus (63) Google Scholar) and were confirmed by DNA sequencing. Mutants are named as (wild-type residue)-(residue number)position index (mutant residue), where the residues are given in the single-letter code. The cDNA encoding the dopamine D2short receptor or the appropriate cysteine mutant was epitope-tagged at the amino terminus with the cleavable influenza-hemagglutinin signal sequence followed by the FLAG epitope (DYKDDDK) and inserted into the bicistronic expression vector pcin4 (30Javitch J.A. Shi L. Simpson M.M. Chen J. Chiappa V. Visiers I. Weinstein H. Ballesteros J.A. Biochemistry. 2000; 39: 12190-12199Crossref PubMed Scopus (63) Google Scholar), creating SFD2-pcin4. Cells were maintained, and stable transfections were performed as described previously (30Javitch J.A. Shi L. Simpson M.M. Chen J. Chiappa V. Visiers I. Weinstein H. Ballesteros J.A. Biochemistry. 2000; 39: 12190-12199Crossref PubMed Scopus (63) Google Scholar). For coimmunoprecipitation studies, an NH2-terminal signal sequence Myc epitope-tagged D2R was created and subcloned in the bicistronic vector pCIhyg (33Saunders C. Ferrer J.V. Shi L. Chen J. Merrill G. Lamb M.E. Leeb-Lundberg L.M. Carvelli L. Javitch J.A. Galli A. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 6850-6855Crossref PubMed Scopus (334) Google Scholar), creating SMD2-pcihyg. Double stable cells were created by transfecting stable cells expressing SFD2-pcin4 with the SMD2-pcihyg construct and selecting with 250 μg/ml hygromycin. EM4 cells, an adherent clonal line of HEK 293 cells, were maintained as described (33Saunders C. Ferrer J.V. Shi L. Chen J. Merrill G. Lamb M.E. Leeb-Lundberg L.M. Carvelli L. Javitch J.A. Galli A. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 6850-6855Crossref PubMed Scopus (334) Google Scholar) and were used for cAMP experiments (34Robbins A.K. Horlick R.A. BioTechniques. 1998; 25: 240-244Crossref PubMed Scopus (57) Google Scholar). HEK 293 cells or EM4 cells stably transfected with the appropriate D2R construct were washed and then reacted for 10 min at room temperature with the stated concentrations of copper sulfate and 1,10-phenanthroline (CuP) as described previously (35Hastrup H. Karlin A. Javitch J.A. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 10055-10060Crossref PubMed Scopus (174) Google Scholar). Phenanthroline was used at a 4-fold molar excess over copper. After removal of the CuP solution, the cells were washed twice with 4 ml of PBS and reacted for 20 min at room temperature with 10 mm N-ethylmaleimide (NEM) in PBS++ buffer (11 mmNa2HPO4, 154 mm NaCl, 2.7 mm KCl, 1 mm MgCl2, 0.1 mm CaCl2) to block free sulfhydryl groups. The cells were scraped into PBS++/PI buffer (PBS++ supplemented with 2 μg/ml Pefabloc, 2 μg/ml aprotinin, 1 μg/ml leupeptin, 1 μg/ml pepstatin A, and 10 mm NEM) and pelleted at 800 × g for 5 min at 4 °C. The pellet was suspended and incubated in 0.5% dodecyl maltoside (DM) for 60 min at 4 °C. The mixture was centrifuged at 20,000 × g for 30 min at 4 °C. Twenty μl of extract was mixed with an equal volume of 4× Laemmli sample buffer without reducing agent. HEK 293 cells, stably transfected with SFD2 or SMD2 or both were reacted with CuP and solubilized in DM as above. To 0.3 ml of DM extracts was added 1.5 μl anti-FLAG M1 mouse monoclonal antibody (Sigma), and the mixture was incubated for 1 h on ice. Thirty μl of prewashed rec-protein G-Sepharose (Zymed Laboratories Inc.) were added, and the mixture was incubated 1 h at 4 °C, washed four times in 1 ml of lysis buffer, and eluted at room temperature for 30 min in 50 μl of 2× Laemmli sample buffer without reducing agent. Samples were applied to 1.5-mm, 10-well 7.5% acrylamide gels prepared and run according to Laemmli (36Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (218087) Google Scholar). The bands were transferred to a polyvinylidene difluoride membrane (Millipore), which was blocked for 1 h at room temperature in 5% nonfat milk, 1% bovine serum albumin, 0.1% Tween 20 in TBS (50 mm Tris-HCl, 150 mm NaCl, pH 7.4). To detect FLAG-D2R, we typically incubated the blot with an anti-FLAG rabbit polyclonal antibody (Sigma) diluted 1:10,000 in blocking buffer for 1 h at room temperature. For Myc-D2R we incubated with an anti-Myc rabbit polyclonal (Santa Cruz Biotechnology) diluted 1:400 in blocking buffer for 1 h at room temperature. The membranes were washed three times for 10 min in TBS containing 0.1% Tween 20, incubated in horseradish peroxidase-conjugated anti-rabbit-antibody (Santa Cruz Biotechnology), diluted 1:15,000 in blocking buffer for 1 h at room temperature, washed three times for 10 min with TBS containing 0.1% Tween, and reacted with ECL-Plus reagent (AmershamBiosciences). Luminescence was detected and quantitated on a FluorChem 8000 (Alpha Innotech Corporation). Intact cells were harvested for binding and [3H]N-methylspiperone (PerkinElmer Life Sciences) binding was performed as described previously (30Javitch J.A. Shi L. Simpson M.M. Chen J. Chiappa V. Visiers I. Weinstein H. Ballesteros J.A. Biochemistry. 2000; 39: 12190-12199Crossref PubMed Scopus (63) Google Scholar). EM4 cells stably expressing the D2R mutants were used for assay of cAMP accumulation as described previously (37Liapakis G. Ballesteros J.A. Papachristou S. Chan W.C. Chen X. Javitch J.A. J. Biol. Chem. 2000; 275: 37779-37788Abstract Full Text Full Text PDF PubMed Scopus (169) Google Scholar) except that 10 μm forskolin was used to stimulate adenylyl cyclase, and increasing concentrations of dopamine were used to activate the D2R and inhibit the forskolin-stimulated cyclase. Our strategy to identify residues forming the dimer interface in the D2R was to use cysteine cross-linking. Thus, it was essential to develop a system that would allow non-cross-linked receptor to run as a monomer on non-reducing SDS-PAGE. In preliminary experiments using transient transfection of D2R, we found, consistent with previous reports (12Ng G.Y. O'Dowd B.F. Lee S.P. Chung H.T. Brann M.R. Seeman P. George S.R. Biochem. Biophys. Res. Commun. 1996; 227: 200-204Crossref PubMed Scopus (248) Google Scholar), that a substantial amount of receptor migrated as dimer or higher-order oligomer (data not shown). Stably expressed D2R produced substantially less of higher-order species on SDS-PAGE. Moreover, we found that mutation to Ser of Cys-3716.61 and Cys-3736.63 in the third extracellular loop, between TM6 and TM7, which leaves only two extracellular cysteines (the highly conserved disulfide-bonded Cys-1073.25 and Cys-182E2) further decreased oligomeric SDS-resistant species (data not shown). The resulting background construct (C1183.36S/C3716.61S/C3736.63S) ran almost exclusively as a heterogeneously glycosylated monomer of ∼65 kDa on non-reducing SDS-PAGE (Fig. 1 A). (As shown below, this background construct, subsequently referred to as D2R, was fully functional and bound multiple ligands normally.) Thus, if the D2R is oligomeric, the oligomer dissociates in SDS. In addition, the D2R is not an obligatory disulfide-linked dimer in the plasma membrane, unlike some class C receptors (2Bai M. Trivedi S. Brown E.M. J. Biol. Chem. 1998; 273: 23605-23610Abstract Full Text Full Text PDF PubMed Scopus (347) Google Scholar, 6Romano C. Yang W.L. O'Malley K.L. J. Biol. Chem. 1996; 271: 28612-28616Abstract Full Text Full Text PDF PubMed Scopus (452) Google Scholar). Using cells stably coexpressing two D2R constructs, one FLAG-tagged and the other Myc-tagged, we attempted to coimmunoprecipitate Myc-D2R with an anti-FLAG monoclonal antibody and protein G. After solubilization with DM we detected no or only trace coimmunoprecipitation of Myc-D2R under these conditions (Fig. 1 C), although the FLAG-D2R was immunoprecipitated as expected (data not shown). Thus, if the D2R is oligomeric in the membrane, the oligomer does not survive DM solubilization. As a control before introducing engineered cysteines into FLAG-D2R for disulfide cross-linking experiments we reacted FLAG-D2R in intact cells with copper phenanthroline (CuP), an oxidizing reagent that promotes the formation of disulfide bonds directly between cysteines (25Careaga C.L. Falke J.J. J. Mol. Biol. 1992; 226: 1219-1235Crossref PubMed Scopus (223) Google Scholar, 26Wu J. Kaback H.R. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 14498-14502Crossref PubMed Scopus (75) Google Scholar). Reaction with CuP produced a new band of ∼133 kDa (Fig. 1 A), approximately twice that of monomer. The fraction of total density that was present in the ∼133 kDa band was plotted against increasing CuP (Fig. 1, A and B), giving half-maximal cross-linking at 60 ± 10 μm CuP and maximal cross-linking of 80 ± 14% (n = 3). The cross-linked species size was consistent with it being a homodimer of D2R, but it was possible that it might represent D2R cross-linked to another protein of similar size. The partners in the cross-linked species were definitively identified by coimmunoprecipitation of Myc-D2R stably coexpressed with FLAG-D2R. After reaction with CuP, immunoprecipitation with anti-FLAG antibody produced an ∼133-kDa band that was recognized by anti-Myc antibody (Fig. 1 C). In contrast when FLAG-D2R and Myc-D2R were expressed separately and then cross-linked, precipitation with anti-FLAG antibody did not produce a ∼133-kDa species recognized by anti-Myc antibody (Fig. 1 C), demonstrating the specificity of the antibodies. These results establish that the ∼133-kDa band is a D2R homodimer that is disulfide cross-linked via one of the remaining endogenous cysteines. Mutation of Cys-1684.58 in the fourth transmembrane segment (TM4) to Ser (as well as to Ala or Phe, data not shown), but not mutation of Cys-561.54, Cys-1263.44, or Cys-3566.47, completely prevented CuP-induced cross-linking (Fig. 2 A), demonstrating that this Cys at the extracellular end of TM4 forms the disulfide cross-link at a symmetrical homodimer interface. To assess ligand binding effects on the dimerization state of the receptor, we treated cells expressing the background construct with the agonists, quinpirole or bromocriptine, or the antagonists, sulpiride or butaclamol. Neither acute nor 24-h treatment with these ligands significantly impacted cross-linking under conditions in which nearly all the receptor was cross-linked (Fig. 2 B). This suggests that the receptor is a constitutive dimer (or possibly a higher order oligomer) in the plasma membrane and that ligand interaction does not lead to dissociation of the dimer. To study the effects of cross-linking Cys-1684.58 on ligand binding, we oxidized with 1 mm/4 mm CuP to ensure that essentially all D2R was cross-linked (see Fig. 1). We observed no change in the K D or Bmax for the antagonist [3H]N-methylspiperone (Fig. 3 C), or in theK I for sulpiride or dopamine (Fig. 3, Aand B). We also assessed the effects of cross-linking Cys-1684.58 on dopamine-mediated inhibition of adenylyl cyclase by the D2R. In the absence of D2R, treatment with 1 mm/4 mm CuP significantly inhibited forskolin-stimulated adenylyl cyclase through a direct effect on adenylyl cyclase (data not shown). This resulted in an apparent decrease in the potency and efficacy of dopamine in both the Cys-1684.58 and C1684.58S constructs after treatment with CuP (Fig. 3 D). Nonetheless, after treatment with CuP, dopamine was equally efficacious at inhibiting cyclase in a Cys-1684.58 construct that was highly cross-linked and in a C1684.58S construct that did not cross-link. We observed no dopamine-mediated inhibition of cAMP levels in untransfected EM4 cells. This argues that the dopamine-mediated inhibition of cyclase occurred only via the stably expressed D2R. Because cross-linking requires that only one of the two cysteines involved is modified initially by the reagent, and the derivatized cysteine then reacts by collision with the second unmodified cysteine, the rate of collision must be much faster than the rate of initial modification. This is consistent with the cysteines being very close initially. The very high fraction of receptor that can be cross-linked, the apparent specificity of the cross-linking, based on the appearance of a single homodimer band, and the lack of cross-linking of Cys-561.54, which based on the bovine rhodopsin structure has a similar lipid accessibility as Cys-1684.58, all argue for the proximity of the TM4 cysteines in the native state. Thus, it is likely that in the membrane, untreated with CuP, D2R exists as a homodimer but that this dimer does not survive detergent solubilization. Cross-linking does not impair the ability of dopamine to inhibit cyclase via the D2R (Fig. 3), demonstrating that the receptor can bind dopamine and activate Gi with a disulfide between Cys-1684.58 in each “subunit” of the dimer. Thus, significant movement of the dimer interface at the Cys-1684.58 position is not necessary for function. Moreover, ligand recognition was unaltered by cross-linking, and the extent of cross-linking was unaltered by treatment with agonists or antagonists, suggesting that the receptor is a constitutive dimer. These results are consistent with the recent findings of Mercieret al. (27Mercier J.F. Salahpour A. Angers S. Breit A. Bouvier M. J. Biol. Chem. 2002; 277: 44925-44931Abstract Full Text Full Text PDF PubMed Scopus (424) Google Scholar). Using quantitative bioluminescence resonance energy transfer (BRET), they showed that more than 82% of β2-adrenergic receptor in the plasma membrane exists as a constitutive dimer (or higher order oligomer) over a broad range of expression levels. They also failed to observe a change in BRET upon addition of ligand and suggested that the dimer does not form or dissociate upon activation. Bovine rhodopsin was crystallized in detergent as a non-physiological dimer in which the cytoplasmic face of one molecule is in the same plane as the extracellular face of the second molecule (28Palczewski K. Kumasaka T. Hori T. Behnke C.A. Motoshima H. Fox B.A. Le Trong I. Teller D.C. Okada T. Stenkamp R.E. Yamamoto M. Miyano M. Science. 2000; 289: 739-745Crossref PubMed Scopus (5101) Google Scholar). It is possible, however, that the native oligomeric structure of rhodopsin is disrupted in detergent and is therefore not seen in the crystal structure. Indeed, our findings that the dopamine receptor is a dimer in the plasma membrane but cannot be immunoprecipitated from a DM extract argues that the D2 receptor native quaternary structure is not preserved in this detergent. Interestingly, squid rhodopsin in detergent forms two-dimensional crystals that show a TM4 dimer interface (29Davies A. Gowen B.E. Krebs A.M. Schertler G.F. Saibil H.R. J. Mol. Biol. 2001; 314: 455-463Crossref PubMed Scopus (66) Google Scholar) (Fig. 4 A). These results suggest that the TM4 dimer interface may be common to other GPCRs as well as the D2R. Using the substituted cysteine accessibility method in the D2R, we found that cysteines in TM4 substituted for the highly conserved Trp-1604.50 as well as for Phe-1644.54 and Leu-1714.61 were accessible to charged sulfhydryl reagents and that this reaction was protected by the antagonist sulpiride (30Javitch J.A. Shi L. Simpson M.M. Chen J. Chiappa V. Visiers I. Weinstein H. Ballesteros J.A. Biochemistry. 2000; 39: 12190-12199Crossref PubMed Scopus (63) Google Scholar). We originally inferred that these residues in TM4 faced the binding site crevice. Surprisingly, in the bovine rhodopsin structure, the aligned residues Trp-1614.50, Leu-1654.54, and Leu-1724.61 face outward and not into the binding site crevice (28Palczewski K. Kumasaka T. Hori T. Behnke C.A. Motoshima H. Fox B.A. Le Trong I. Teller D.C. Okada T. Stenkamp R.E. Yamamoto M. Miyano M. Science. 2000; 289: 739-745Crossref PubMed Scopus (5101) Google Scholar). Our observations regarding the anomalous conservation and accessibility of this “back face” of TM4 suggested to us the possibility that this surface might be at the interface between two D2R subunits (31Ballesteros J.A. Shi L. Javitch J.A. Mol. Pharmacol. 2001; 60: 1-19Crossref PubMed Scopus (407) Google Scholar). Our finding that the site of cross-linking in D2R is Cys-1684.58 at the extracellular end of TM4 directly adjacent to these accessible residues (Fig. 4 B) is consistent with this hypothesis that TM4 forms a symmetrical dimer interface. We are currently extending these studies to include other residues in TM4 in an attempt to map the entire interface and to assess whether conformational changes occur at this interface in different functional states of the receptor. Given that cysteines substituted for Trp-1604.50, Phe-1644.54, and Leu-1714.61 were protected from reaction by sulpiride, we think it likely that such conformational changes do occur and may mediate cross-talk between receptor subunits. We thank Myles Akabas, Juan A. Ballesteros, Marta Filizola, Arthur Karlin, and Harel Weinstein for helpful discussion and Gebhard F. X. Schertler and Helen R. Saibil for providing the photograph used to create Fig. 4 A.