Downloaded by guest on September 28, 2021 Proc. Nat!. Acad. Sci. USA Vol. 89, pp. 12048-12052, December 1992 Neurobiology The 6opioid receptor: Isolation of a cDNA by expression cloning and pharmacological characterization (G protein-coupled receptor/NG 108-15 cells/transient expression/Tyr-D-Thr-Gly-Phe-Leu-Thr/analgesia) BRIGITTE L. KIEFFER*, KATIA BEFORT, CLAIRE GAVERIAUX-RUFF, AND CHRISTIAN G. HiRTHt Ecole Supdrieure de Biotechnologie, 11 rue Humann, 67085 Strasbourg Cedex, France Communicated by Pierre Chambon, August 31, 1992

ABSTRACT A random primed expression cDNA library dence that a cDNA encoding an opioid receptor has been was constructed from the RNA of NG 108-15 cells. Pools of cloned. plasmid DNA were transfected into COS cells, which were We report here the isolation ofa cDNA encoding a 8-opioid screened for their ability to bind 'H-labeled Tyr-D-Thr-Gly- receptorA We used a transient expression strategy where a Phe-Leu-Thr, a tritiated agonist for the -opioid receptor. A cDNA library derived from NG 108-15 cells was screened cDNA was isolated that encodes a 371-amino acid-residue using 3H-labeled Tyr-D-Thr-Gly-Phe-Leu-Thr (DTLET), a protein presenting all the structural characteristics ofreceptors tritiated agonist for the 8-opioid receptor. Our pharmacolog- that interact with guanine nucleotide-binding proteins. Notice- ical study of the cloned cDNA expressed in COS cells shows able features are (i) the high hydrophobicity of the encoded that it possesses all expected properties of a 8 receptor. A protein, (it) its low sequence similarity to both catecholamine comparative analysis of the deduced protein sequence with receptors and peptide-binding receptors, although it presents other members of the GPR family is also presented. the typical aspartate residue involved in catecholamine binding ofthe first group and the characteristic short third cytoplasmic MATERIALS AND METHODS loop of the second group. When expressed in COS cells, the receptor exhibits pharmacological properties similar to those of Library Construction and Screening. NG 108-15 cells were binding sites for 3H-labeled provided by B. Foucaud (URA 1836, Facultd de Pharmacie, the native receptor: high-affinity Strasbourg, France). Cells were harvested at 50% conflu- Tyr-D-Thr-Gly-Phe-Leu-Thr (Kd = 1.4 nM), stereospecific ency, RNA was prepared by LiCl/urea precipitation (6), and binding sites for the - enantiomers of levorphanol and nalox- polyadenylylated RNA was purified with an oligo(dT) col- one, and the selectivity prorde of a 8 receptor, as determined umn (Pharmacia). The cDNA library construction in the by competition experiments with a set of EL-, 8-, and K-oploid mammalian expression vector pCDM8 has been described ligands. (7). Plasmid DNA was prepared according to the alkaline lysis method (8) from pools of 3000 bacterial colonies. Opioid receptors have long been described as membrane One-tenth of each plasmid pool DNA was independently receptors of the nervous system mediating the analgesic transfected into COS-1 cells (American Type Culture Col- effects ofopium-derived alkaloids. Endogeneous ligands and lection CRL/1650) using the DEAE-dextran method. After their precursors have been characterized, and their role in 72 hr, transfected COS-1 cells were assayed for tritiated response to pain and stress has been widely studied (1). DTLET binding (61 Ci/mmol, Commissariat a l'Energie Pharmacological studies have shown the existence of three Atomique, Saclay, France; 1 Ci = 37 GBq). Cell monolayers subtypes of receptors, ,u (morphine), 8 (enkephalin), and K were washed twice with PBS/0.5% bovine serum albumin (dynorphin), the cellular inhibitory actions of which are and incubated 35 min at 370C with PBS/0.5% bovine serum linked to G protein activation (2). Opioid receptors are, albumin/1 nM [3H]DTLET. Dishes were chilled on ice for 5 therefore, believed to be part of the G protein-coupled min, and the cells were washed four times with ice-cold receptor (GPR) family, a class of membrane-bound receptors PBS/0.5% bovine serum albumin. The cells were then solu- that exhibit a seven-transmembrane-spanning domain topol- bilized in 1% SDS and added to 7 ml of scintillation liquid for ogy and represent 80% of all known receptors (3). counting. Positive pools were fractionated by diluting an Several attempts to clone cDNAs encoding opioid recep- aliquot ofthe glycerol stock in selective medium, growing the tors have been reported. A cDNA encoding an opioid-binding bacteria on nitrocellulose membranes overlaid onto agar Pieces of membrane were then cut to isolate of protein (OBCAM) with g selectivity was isolated (4), but the plates. pools predicted protein lacks transmembrane domains, presumed colonies. Each of them was scraped into selective medium necessary for signal transduction. More recently, the isola- and grown to 0.5 OD6wn. unit. An aliquot was stored as a tion of another cDNA was reported, which was obtained by glycerol stock for further fractionation, whereas the rest was expression cloning (5). The deduced protein sequence dis- used for plasmid DNA preparation and COS cell transfection. plays seven putative transmembrane domains and is very Ligand Binding. [D-Ala2,D-Leu5ienkephalin (DADLE), cy- similar to the human neuromedin K receptor. However, the clic [D-penicillamine2,D-penicillamine5]enkephalin (DPDPE), affinity of opioid ligands for this receptor expressed in COS cells is two orders of magnitude below the expected value, Abbreviations: DADLE, [D-Ala2,D-Leu51enkephalin; DAGO, [D-Ala2,MePhe4,Gly-ol5lenkephalin; DTLET, Tyr-D-Thr-Gly-Phe- and no subtype selectivity can be shown. Finally, attempts to Leu-Thr; DPDPE, cyclic [D-penicillamine2,D-penicillamine5]en- clone opioid receptors via approaches relying on sequence kephalin; G protein, guanine nucleotide-binding regulatory protein; similarities in the GPR family, using the PCR technology, GPR, G protein-coupled receptor; Tm, transmembrane domain; U were unsuccessful. There is, therefore, no convincing evi- 50488, trans-(+)-3,4-dichloro-N-methyl-N-[2-(1-pyrrolidinyl)cyclo- hexyllbenzeneacetamide. *To whom reprint requests should be addressed. The publication costs of this article were defrayed in part by page charge tDeceased, May 29, 1992. payment. This article must therefore be hereby marked "advertisement" tThe sequence reported in this paper has been deposited in the in accordance with 18 U.S.C. §1734 solely to indicate this fact. GenBank data base (accession no. L06322). 12048 Neurobiology: Kieffer et al. Proc. Natl. Acad. Sci. USA 89 (1992) 12049 [D-Ala2,MePhe4,Gly-ol5]enkephalin (DAGO) and trans-(+)- vided twice, leading to a single clone, K56, which conferred 3,4-dichloro-N-methyl-N-[2-(1-pyrrolidinyl)cyclohexyl]ben- strong [3H]DTLET-binding capability to COS cells. The zeneacetamide (U 50488) were obtained from Sigma. (-)- signal was six times above the background value, naloxone Naloxone, (+)-naloxone, levorphanol, dextrorphan, brema- sensitive, and ligand-concentration dependent. zocin, and etonitazen were provided by B. Ilien (20). Binding Pharmacology ofthe Cloned cDNA Transiently Expressed in was assayed with membrane preparations. COS-1 cells were COS Cells. We first estimated the affinity of [3H]DTLET for transfected with plasmid K56 or with pCDM8 (mock), as K56-expressing COS cells. Fig. 1A shows that a high specific before, and were harvested after 72-hr expression time. NG binding is observed on K56-transfected COS cell membranes, 108-15 cells were collected at 50%o confluency. The cells (3-10 whereas it is negligible for mock-transfected cells. [3H]DT- x 107) were washed two times with PBS, pelleted, and frozen LET binding on NG 108-15 membranes, using the same at -80'C. The membranes were then prepared at 40C with a single buffer, 50 mM Tris HCI, pH 7.4/10 mM EDTA. The cell amount of membrane proteins in the assay, is shown for pellet was resuspended in 60 ml of buffer, treated with a comparison. Scatchard analysis of [3H]DTLET binding Dounce homogenizer, and centrifuged at 1100 x g for 10 min. The pellet was resuspended a second time in buffer (30 ml), A homogenized, and centrifuged again. Both supernatants were pooled and centrifuged at 110,000 x g for 15 min. The membrane pellet was resuspended in 5 ml ofbuffer, aliquoted, and stored at -800C. The protein concentration was deter- mined using the Bradford assay (9). Typical binding assays were done in 1-ml volumes by using membranes diluted in the same Tris/EDTA buffer (15-30 jg of protein per ml), com- petitors, and [3H]DTLET as the radiolabeled ligand. After a 30-min incubation at 370C, the membrane suspension was rapidly filtered on polyethylenimine-precoated GF/B What- man filters and washed three times with 3 ml ofcold buffer. For competition studies, [3H]DTLET was used at a 1 nM concen- tration. The Ki values were derived from the Cheng and Prussof equation: K, = lC50A(1 + L/Kd). Sequence Analysis. Oligonucleotides were synthesized on [3H]DTLET, nM an Applied Biosystems 394 DNA/RNA synthesizer. The B cDNA was sequenced in both directions by using the Sanger dideoxynucleotide chain-termination method (Sequenase kit, United States Biochemical). Data base searching was done on Swiss-Prot (release 21, Feb. 92) with BLASTP. The se- quence alignments and the dendrogram were done with the CLUSTAL multiple-alignment program (10). The hydrophobic- - ity analysis was determined by the method of Kyte and c)

Doolittle (11) with a window of 11 residues and the trans- *-. membrane-domain prediction according to Eisenberg (12) by using the PC GENE program (IntelliGenetics). PCR on Mouse Genomic DNA and Sequence Analysis. A set la of oligonucleotides was synthesized based on the cDNA log[competitor] sequence (position 1857-1877 for the forward primer and position 2154-2172 for the reverse primer). PCR amplifica- tion was done on a mouse genomic DNA preparation (13) V) using standard conditions (94°C 1 min, 500C 1.5 min, 72°C 2 min, 30 cycles). The PCR product was subcloned into EcoRI sites of pBluescript (Stratagene) and sequenced in both .0 directions. 0 r.~ RESULTS Isolation of cDNA K56 by Expression Cloning. We con- structed a random-primed cDNA library from NG 108-15 cells (14) expressing 5 x 104 B-opioid receptor molecules per -10 -9 -8 -7 -6 -5 cell, using the mammalian transient expression vector log[competitor] pCDM8 (15). The library contained 3.5 x 106 primary trans- FIG. 1. Pharmacological properties of K56-expressing COS cells. formants. A part of it was plated and partitioned into 100 For all figures a representative experiment is presented. (A) Affinity pools of 3000 different recombinant bacterial colonies. COS of[3H]DTLET for the expressed receptor. Membranes prepared from cells transfected with plasmid DNA from these pools were K56-(m) and pCDM8-(A) transfected COS cells or from NG108-15 (r) incubated with [3H]DTLET at a 1 nM concentration, which cells were incubated with increased concentration of [3H]DTLET ± is equivalent to its Kd value for the native receptor. The (-)-naloxone (10-6 M), the difference being defined as specific bound radioactivity was counted after cell lysis. The back- binding. Assays were done in triplicates. (Inset) Scatchard analysis: was 700 ± and a fraction a Kd is 1.4 nM for both expressed and native receptors, and B". values ground dpm 250, single produced are 4.2 and 1.2 pmol/mg, respectively. Three independent experi- signal that was consistently 10% above background. This ments yielded a Kd value of 1.45 ± 0.05 nM for [3H]DTLET binding fraction was subdivided, and the subfractions were tested in to K56-transfected COS cells. (B and C) Competition of [3H]DTLET COS cells again. Several subfractions produced a signal that binding with unlabeled competitors on K56-transfected COS cells. emerged 20% above background and disappeared in the Assays were done in duplicates, and specific binding without com- presence of naloxone. One of these subfractions was subdi- petitors was defined as 1N0. 12050 Neurobiology: Kieffer et al. Proc. NatL Acad. Sci. USA 89 (1992) shows a single class of binding sites with an apparent Kd of lated Mr of40,810. The native receptor is a glycoprotein, and 1.4 nM for both K56-transfected COS cell and NG 108-15 cell the difference between the predicted size and the 58,000-Da membranes. This value is similar to the described value (1.04 molecular mass of the purified NG 108-15 receptor, as nM) (16). Bm. values ranged from 3.9 to 6.4 pmol/mg of estimated by SDS/PAGE (21) is probably due to the attach- proteins, depending on the transfected membrane prepara- ment of carbohydrates on the two potential N-glycosylation tion. Considering that, on average, 10% only ofthe COS cells sites present at the N terminus of the predicted protein are efficiently transfected by the plasmid, the expression sequence. level of the cloned receptor could be estimated to 5 x 106 K56 cDNA encodes a protein that belongs to the class of molecules per transfected cell, which is 100 times higher than GPRs. The sequence is shown on Fig. 2A aligned (22) with the value determined for NG 108-15. Opioid receptors are four other representative members of the family: the %2- highly stereoselective. Levorphanol and (-)-naloxone are the neuromedin K and the known to bind to opioid receptors with high affinity, whereas adrenergic receptor, receptor, their (+) enantiomers, dextrorphan and (+)-naloxone, re- rhodopsin, as well as the N-formyl peptide receptor, with spectively, do not, irrespective ofthe receptor subtype under which it shares the best similarity (30%o identity, 10 gaps) study (5, 17). Competition experiments were done with among sequences in the Swiss-Prot Data Bank. The con- [3H]DTLET with the two pairs of ligands. Fig. 1B indicates served residues that are virtually found in all GPRs are that (+)-naloxone is not a competitor. A slight inhibition is present in K56 protein. The encoded protein also contains the seen with dextrorphan at a 1 gM concentration. This agrees potential sites for posttranslational modifications usually with previous data that report that dextrophan has a three- found in GPRs: in addition to putative N-glycosylation sites, orders-of-magnitude lower binding potency at opioid-binding three potential phosphorylation sites are present in the third sites, as compared with levorphanol. Both (-) enantiomers cytoplasmic loop, as well as three others in the C-terminal yielded Ki values that meet their reported affinities for a region. The localization of the seven transmembrane do- 8-opioid receptor (29.5 nM for naloxone and 20.9 nM for mains, as defined by homology with other receptors, agrees levorphanol, see Table 1). We further investigated the sub- well with the hydrophobicity analysis and meets the Eisen- type specificity of the cloned receptor by competition exper- berg prediction (Fig. 2B). The size of the protein is small iments between a set of,u, 6, and K agonists with [3H]DTLET. compared with other GPRs, which range from 324 (mas Fig. 1C shows that they all compete with an order ofpotency oncogene) to 744 (thyroid-stimulating hormone receptor) typical of the 6-opioid receptor (17, 18, 20): bremazocin = amino acid residues. The sequence contains short intra- and DADLE = DPDPE >> etonitazen > DAGO > U 50488 with extracellular loops, and the hydrophobic domains are, there- Ki values of 5.7, 6.2, 10.9, 1800, 5050, and 39,100 nM, fore, predominant. The regions that display the higher hy- respectively. We did the same experiment on NG 108-15 drophilic character concern short stretches of amino acid membranes and found the same profile with similar affinities. residues in the third cytoplasmic loop and in the C-terminal The Ki values are also comparable with values from the tail, regions which in GPRs presumably interact with the literature (see Table 1). In conclusion, the K56 cDNA- intracellular compartment of the cell and are important for encoded receptor expressed in COS cells displays binding signal transduction. characteristics similar to that of the native receptor. Species Origin of K56 Protein. The NG 108-15 cell line used Primary Structure. K56 contains a 2.2-kilobase (kb) insert to prepare the library is a hybrid cell line between mouse with an open reading frame of 1174 base pairs (bp) starting N18TG neuroblastoma and rat C6Bu-1 glioma cell lines, and from the 5' end. The first ATG at position 59 was assigned as the species origin of K56 DNA remained to be determined. the translational initiation codon for the protein. The nucle- Using mouse and rat genomic DNA as a template, we otides surrounding this ATG are in total agreement with the amplified by PCR a 350-nucleotide fragment corresponding to Kozack nucleotide sequence CCATGG (19). No stop codon positions 1857-2172 of the cDNA, a 3' noncoding region was found upstream of this ATG. The predicted protein presumed not conserved among different species. The PCR sequence consists of 371-amino acid residues, with a calcu- on mouse genomic DNA produced a single discrete band of was and We Table 1. of for K56 expressed in the expected size, which subcloned sequenced. Binding potency opioid ligands found a 100%6 match between this sequence and the cDNA COS cells sequence. We, therefore, conclude that K56 encodes a mouse pK; receptor. Data from Sequence Similarity Analysis Within the GPR Family. The literature sequence of K56 protein contains an Asp residue (position (ref.) 128) present in catecholamine GPRs only, but it also displays Subtype NG a short third cytoplasmic loop typical of GPRs that have Ligand preference COS-K56 108-15 (17) (18) (19) peptides as ligands. To determine the position of K56 recep- Bremazocin K = IL = 6 8.25 ± 0.12 8.75 8.8 9.14 tor in the GPR family, a dendrogram was established, based DADLE 8 = IL >>> K 8.22 ± 0.15 8.70 9.1 8.82 on sequence similarities of K56 protein with other members DPDPE 6>> ,I >> K 7.73 ± 0.48 8.30 8.9 8.57 of the family (Fig. 2C). Seventeen mouse and rat sequences Levorphanol u >> K = 8 7.68 ± 0.07 ND 7.8 - (preferably mouse when available except for the N-formyl (-)-Naloxone IL > 8 = K 7.53 ± 0.09 ND 7.6 peptide receptor, which is a human sequence), which dis- Etonitazen ,u >>> 8 5.75 6 - - <6 played the best homology with K56 protein, were selected DAGO >>» 6> K 5.21 ± 0.17 5.65 6.7 6.46 from the Swiss-Prot Data Base. As expected, the aminergic U 50488 K >>> /A > 8 4.35 ± 0.15 4.24 5.0 5.06 receptors and the peptidergic receptors appear as two distinct pKj values (- log Ki) are determined by competition of unlabeled groups, and rhodopsin appears as a single sequence. K56 and opioid ligands with [3HIDTLET on K56-transfected COS cells (mean the human N-formyl peptide receptor (23) form a group + SEM of three independent experiments) and NG 108-15 cells, as separate from the two main groups. In a second alignment described (see also Fig. 1 B and C). Values found in the literature that excluded all extra- and intracellular and considered [columns (17) and (18)] derive from competition experiments done on loops guinea pig brain membranes (except for etonitazen tested on rat) and the transmembrane domains only, K56 protein appeared as a show pKi values for the 6 subtype in this tissue. ND, not determined; group with the N-formyl peptide receptor again but closer to -, value not found in cited reference. The ligands, including [3H]DT- a group including the substance K, substance P, neuromedin LET, are agonists, except naloxone, which is an antagonist. K, and neuromedin B receptors (data not shown). Neurobiology: Kieffer et al. Proc. Natl. Acad. Sci. USA 89 (1992) 12051 A 1 # * * 50 ** +* 100 K56. MELVPSARAELQSSPLVNLSDAFPSAFPSAGANASGSPGARSASSLAL AIAI TALYSAVCAVGLLGNVLVMFGIV- RYTKLKTATN I YIFNLALADALATSTL I I I- N FORM P. METNSSLPTNISGGTPAVSAGYLFLD -I ITYLVFAVTFVLGVLGNGLVIWVAGF RMTHTVTT-- ISYLNLAVADFCFTSTL NEU K. MASVPRGENWTDGTVEVGTHTGNLSSALGVTEWLALQAGNFSSALGLPATTQAPSQVRANLTNQFVQPSWR IALWSLAYGLVVAVAVFGNLIVIWI IL- AHKRMRTVTN YFLVNLAFSDASVAAFN B2 ADR. MGPHGNDSDFLLAPNGSRAPDHDVTQERDEAW VVGMAILMSVIVLAIVFGNVLVITAIA- KFERLQTVTN YFIISLACADLVMGLAV RHODOP. MNGTEGPNFYVPFSNVTGVGRSPFEQPQYYLAEPWQ FSMLAAYMFLLIVLGFPINFLTLYVTV -HKKLRTPLN YILLNLAVADLFMVFGG

* * 0* iso.150+ *200 K56. PF-QSA KYLMETWPFGELLCK AVLSIDYYNMFTSIFTLTMMSV DRYIAVCHPVKALDFRTPAKAKL - INICIWVLASGVGV-P IMVMA --VTNPGMVQWYACSSSPVQLVLDTV------III IV N FORM P. PFFMVR KAMGGHWPFGWFLCK FLFTIVDINLFGSVFLIALIAL DRCVCVLHPVWTQNHRTVSLAKK -VIIGPWVMALLLTL-PVII-- RVTTVPGKTGTVACTFNFSPWTNDPKERINVAVAML NEU K. TLINFI YGLHSEWYFGANYCR FQNFFPITAVFASIYSMTAIAV DRYMAIIDPLKPRLSATA--TK- IVIGSIWILAFLLAF-PQCLYS --- KIKVMPGRTLCYVQWPEGPKQHFT ------B2 ADR. VPFGAS HILMKMWNFGNFWCE FWTSIDVLCVTASIETLCVIAV DRYVAITSPFKYQSLLTKNKAR- VVI LMVWIVSGLTSFLP IQMHW --- YRATHKKAIDCYTEETCCDFFTNQA------RHODOP. FTTTLY TSLHGYFVFGPTGCN LEGFFATLGGEIALWSLVVLAI ERYVVVCKPM--SNFRFGENHAI MGVVFTWIMALACAA-PPLVGW --SRYIPEGMQCSCGIDYYTLKPEVNNES ------

* * * # 250# # * 0 300 K56. --- TKICVFLFAFVVPILIITVCYG LMLLRLRSVRLLSG------SKEKDRSLRRITRM VLVVVGAFVVCWAPIHIFVIVWTL VDINRRDPLVVA--- ALHLCI V VI N FORM P. TVR -GIIRFIIGFSAPMSIVAVSYG LIATKIHKQG------LIKSSRPLRV LSFVAAAFFLCWSPYQVVALIATV RIRELLQGMYKEIGI AVDVTS NEU K. ___ YHIIVIILVYCFPLLIMGVTYT IVGITLWGGEIPGDTC------DKYHEQLKAKRKVVKM MIIVVVTFAICWLPYHVYFILTAI --YQQLNRWKYIQQ- VYLASF B2 ADR. _ YAIASSIVSFYVPLCVMVFVYS RVFQVAKRQLQKIDKSEGRFHAQNLSQVEQDGRSGHGLRRSSKFCLKEHKALKT LGIIMGTFTLCWLPFFIVNIVHVI --RDNLIPKE ----- VYILLN RHODOP. FVIYMFVVHFTIPMIVIFFCYG QLVFTVKEAAAQQQE ------SATTQKAEKEVTRMr IIMVIFFLICWLPYASVAFYIFT QGSNFGPI ------FMTLPA

,, * # 350# * 371 K56. ALGYANSSLNPVLYAFL DENFKRCFRQLCRTPCGRQEPGSLRRPRQATTRERVTACTPSDGPGGGAAA - VII N FORM P. ALAFFNSCLNPMLYVFM GQDFRERLIHALPASLERALTEDSTQTSDTATNSTLPSAEVALQAK NEU K. WLAMSSTMYNPIIYCCL NKRFRAGFKRAFRWCPFIQVSSYDELELKTTRFHPTRQSSLYTVSRMESVTVLFDPNDGDPTKSSRKKRAVPRDPSANGCSHRGSKSASTTSSFISSPYTSVDEYS B2 ADR. WLGYVNSAFNPLIYC-- RSPDFRIAFQELLCLRRSSSKTYGNGYSSNSNGRTDYTGEPNTCQLGQEREQELLCEDPPGMEGFVNCQGTVPSLSVDSQGRNCSTNDSPL RHODOP. FFAKSSSIYNPVIYIML NKQFRNCMLTTLCCGKNPLGDDDASATASKTETSQVAPA

C Mouse rhodopsin B Rat neurotensin receptor 50 - Mouse thyrotropin-releasing hormone receptor 40 Rat neurokinin A (SK) receptor Rat neuromedin K receptor x 30 Rat substance P receptor la 20 Rat neuromedin B / bombesin receptor t. 10 Rat serotonin Ic receptor Rat dopamine D2 receptor o -10 Rat alpha 2 adrenergic receptor -20o Rat dopamine D3 receptor Mouse beta 2 adrenergic receptor -30 I II III IV V VI VII -40 Rat alpha adrenergic receptor Rat m4 muscarinic acetylcholine receptor -50 Mouse ml muscarinic acetylcholine receptor 1 70 140 210 280 350 Human N-formyl peptide receptor Residue number K56

FIG. 2. Sequence analysis of K56. (A) Deduced protein sequence aligned with other GPRs. N FORM P., human N-formyl peptide receptor; NEU K., rat neuromedin K receptor; B2 ADR., mouse P2-adrenergic receptor; RHODOP., mouse rhodopsin. The putative transmembrane domains for these receptors are boxed. *, Highly conserved amino acids among GPR family; #, potential phosphorylation sites; and *, consensus N-linked glycosylation sites. (B) Hydrophobicity analysis (11). Hydrophobic domains of the protein appear above the dotted line. Positions of putative transmembrane domains (12) are indicated by short lines, numbered I-VII, as in A. (C) Homology of K56 with other GPRs (10). This dendrogram represents the relative homology between the sequences. Horizontal distances are inversely proportional to percentage homology. DISCUSSION strate that the expressed receptor is stereoselective and binds preferably ligands. (ii) The 8-opioid receptor in NG 108-15 Our cloning strategy involved the use of a tritiated ligand (61 cells has been shown to inhibit adenylate cyclase activity Ci/mmol) for the detection of receptor-expressing COS cells. through the activation of a G protein (2) and is, therefore, All previously reported ligands used in this approach are expected to be a member of the GPR family: the deduced radiolabeled with isotopes of high specific activity (>2000 protein sequence of K56 indicates unambiguously that this is Ci/mmol). Our experiment shows that a ligand with low the case. specific activity can be used successfully in expression Analysis of the K56 protein sequence, with respect to its cloning, provided that the library is partitioned into pools of similarity with other GPRs, brings insight to the understand- limited size. ing of the biochemistry and pharmacology of opioid recep- NG 108-15 cells express a single type of opioid receptor tors. that has long been characterized as the subtype (24). Two Opioid receptors are sensitive to reducing agents, and the lines of evidence indicate that we succeeded in isolating a occurrence of a disulfide bridge has been postulated as cDNA encoding this receptor: (i) the pharmacological profile essential for ligand binding (25). For rhodopsin, muscarinic, exhibited by the K56-expressing COS cells is convincing: the and 8-adrenergic receptors, the two conserved cysteine binding assay with [3H]DTLET, a prototypal 8 ligand shows residues in each of the two first extracellular loops have been the presence of high-affinity binding sites with Kd similar to shown critical for stabilizing the functional protein structure the native receptor, and competition experiments demon- and are presumed to do so by forming a disulfide bridge. A 12052 Neurobiology: Kieffer et al. Proc. Nadl. Acad. Sci. USA 89 (1992) S-S bond between Cys-121 and Cys-199 of K56 protein Eucaryotes Supdrieurs du Centre National de la Recherche Scien- might, therefore, be important for correct protein folding. tifique U/184-Laboratoire de Biologie Mol6culaire et de G6nie Structure/function studies of opioid ligands have shown Genetique de l'Institut National de la Santc et de la Recherche the importance ofa protonated. amine group for binding to the Medicale, Strasbourg, France (P. Chambon, Director). receptor with high affinity. The binding site of the receptor 1. Akil, H., Watson, S. J., Young, E., Lewis, M. E., Khacha- might, therefore, possess a critical negatively charged coun- turian, H. & Walker, J. M. (1984) Annu. Rev. Neurosci. 7, terpart. Catecholamine receptors display in their sequence a 223-255. conserved aspartate residue that has been shown necessary 2. Simonds, W. F. (1988) Endocr. Rev. 9, 200-212. for binding the positively charged amine group of their 3. Bockaert, J. (1991) Curr. Op. Neurobiol. 1, 32-42. ligands. By analogy, the Asp residue of K56 protein at 4. Schofield, P. R., McFarland, K. C., Hayflick, J. S., Wilcox, 128 be involved in the interaction of the J. N., Cho, T. M., Roy, S., Lee, N. M., Loh, H. H. & See- position might burg, P. H. (1989) EMBO J. 8, 489-495. 6-opioid receptor with its ligands. 5. Xie, G., Miyajima, A. & Goldstein, A. (1992) Proc. Nati. Acad. Compared with other membrane receptors, purified opioid Sci. USA 89, 4124-4128. receptors are extremely labile molecules (for review, see refs. 6. Auffray, C. & Rougeon, F. (1980) Eur. J. Biochem. 107, 26 and 27). Progress in solubilizing these receptors has long 303-314. been hampered by their high sensitivity to detergents. More- 7. Kieffer, B. L. (1991) Gene 109, 115-119. over, phospholipids added to detergent extracts have been 8. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular shown to stabilize the purified receptor conformation in an Cloning:A Laboratory Manual (Cold Spring Harbor Lab., Cold active state. These experiments suggest the occurrence of Spring Harbor, NY), 2nd Ed. the and its natural lipidic 9. Bradford, M. A. (1976) Annu. Rev. Biochem. 72, 248-254. critical interactions between protein 10. Higgins, D. G. & Sharp, P. M. (1988) Gene 73, 237-244. environment that modulate the protein conformation and, 11. Kyte, J. & Doolittle, R. F. (1982) J. Mol. Biol. 157, 105-132. thus, its binding capability. Our finding that the deduced 12. Eisenberg, D., Schwarz, E., Komaromy, M. & Wall, R. (1984) amino acid sequence of K56 is a highly hydrophobic protein, J. Mol. Biol. 179, 125-142. probably almost entirely buried in the membrane, agrees with 13. Blin, N. & Stafford, D. W. (1976) Nucleic Acids Res. 3, these observations. 2303-2314. The opioid receptors are heterogeneous. The origin of 14. Klee, W. A. & Nirenberg, M. (1974) Proc. Natl. Acad. Sci. different subtypes is currently a speculative question because USA 71, 3474-3477. no molecular characterization has been achieved. At least 15. Aruffo, A. & Seed, B. (1987) Proc. Natl. Acad. Sci. USA 84, in the 8573-8577. three classes of opioid receptors are present brain, 1z, 16. Bochet, P., Icard-Liepkalns, C., Pasquini, F., Garbay-Jau- 8, and K, which differ in their pharmacology (17) and ana- reguiberry, C., Beaudet, A., Roques, B. & Rossier, J. (1988) tomical distribution (28). However, biochemical character- Mol. Pharmacol. 34, 436-443. ization of opioid receptors from many groups reports a 17. Goldstein, A. & Naidu, A. (1989) Mol. Pharmacol. 36, 265-272. molecular mass of -60,000 Da for all three subtypes, sug- 18. Leslie, F. M. (1987) Pharmacol. Rev. 39, 197-249. gesting that they could be related molecules (27). Moreover, 19. Kozak, M. (1984) Nucleic Acids Res. 12, 857-872. the similarity between the three receptor subtypes is sup- 20. Ilien, B., Galzi, J. L., Mejean, A., Goeldner, M. & Hirth, C. ported by the isolation of (i) antiidiotypic monoclonal anti- (1988) Biochem. Pharmacol. 37, 3843-3851. bodies competing with both and 8 ligands but not competing 21. Simonds, W. F., Burke, T. R., Jr., Rice, K. C., Jacobson, A& A. E. & Klee, W. A. (1985) Biochemistry 82, 4974-4978. with K ligands (29, 30) and (ii) a monoclonal antibody raised 22. Probst, W. C., Snyder, L. A., Schuster, D. I., Brosius, J. & against the purified g receptor that interacts with both 1L and Seaffon, S. C. (1992) DNA Cell Biol. 11, 1-20. K receptors (31). Low-stringency hybridization experiments 23. Boulay, F., Tardif, M., Brouchon, L. & Vignais, P. (1990) with K56 cDNA as a probe may allow the isolation ofcDNAs Biochem. Biophys. Res. Commun. 168, 1103-1109. encoding the other opioid-receptor subtypes and shed light on 24. Chang, K. & Cuatrecasas, P. (1979) J. Biol. Chem. 254, the question of the opioid-receptor heterogeneity. 2610-2618. The availability of cDNAs encoding the opioid receptors 25. Gioannini, T. L., Liu, Y. F., Park, Y-H., Hiller, J. M. & will then permit detailed studies of the signal-transduction Simon, E. J. (1989) J. Mol. Recogn. 2, 44-48. mechanism and reveal the anatomical distribution of the 26. Demoliou-Mason, C. D. & Barnard, E. A. (1990) in Receptor these information on their Biochemistry:A PracticalApproach, ed. Hulme, E. C. (Oxford mRNAs of receptors, providing Univ. Press, New York), pp. 99-122. expression pattern in the nervous system. This information 27. Loh, H. H. & Smith, A. P. (1990) Annu. Rev. Pharmacol. should ultimately allow better understanding of the opioid Toxicol. 30, 123-147. system in nociception and analgesia and also the design of 28. Mansour, A., Khachaturian, H., Lewis, M. E., Akil, H. & more specific therapeutic drugs. Watson, S. (1987) J. Neurosci. 7, 2445-2464. 29. Gramsch, C., Sculz, R., Kosin, S. & Herz, A. (1988) J. Biol. We are grateful to E. Borrelli, R. Hen, and L. Maroteaux for Chem. 263, 5853-5859. helpful discussions and support; to F. Plewniak for help with the 30. Coscia, C. A., Szfics, M., Barg, J., Belcheva, M. M., Bem, computer alignments; to F. Ruffenach for the oligonucleotides; and W. T., Khoobehi, K., Donnigan, T. A., Juszczak, R., McHale, to M. Acker for the cell culture. We thank A. Menez for critical R. J., Hanlley, M. R. & Barnard, E. A. (1991) Mol. Brain Res. reading ofthe manuscript. This research was supported by the Ecole 9, 299-306. Superieure de Biotechnologie, Strasbourg, France (J. F. Lefevre, 31. Bero, L. A., Roy, S. & Lee, N. M. (1988) Mol. Pharmacol. 34, Director) and by the Laboratoire de Gdndtique Moleculaire des 614-620.