Expression Cloning of Cdna Encoding a Seven-Helix Receptor from Human

Home , Eledoisin, Secretin receptor

Proc. Natl. Acad. Sci. USA Vol. 89, pp. 4124-4128, May 1992 Pharmacology Expression cloning of cDNA encoding a seven-helix receptor from human placenta with affinity for opioid ligands Guo-Xi XIE*t, ATSUSHI MIYAJIMA*t, AND AVRAM GOLDSTEIN§ *DNAX Research Institute of Molecular and Cellular Biology, Palo Alto, CA 94304; and §Stanford University, Stanford, CA 94305 Contributed by Avram Goldstein, December 27, 1991

ABSTRACT Here we report the expression cloning of eluted to DE-81 ion-exchange paper (Whatman) and eluted cDNA encoding a putative opioid receptor from a human with 1 M NaCl. After precipitation with ethanol, the cDNA placenta cDNA library. Placental opioid receptors are of the K was washed and unidirectionally inserted by T4 ligase into the type. As the dynorphin opioid peptides are K-selective, a BstXI-Not I sites ofpME18S. The resulting cDNA library had dynorphin ligand was used in an affinity-enrichment (panning) 1.4 x 106 independent colonies after transformation of DH5a procedure to select transiently transfected COS-7 cells express- competent E. coli by electroporation. ing K receptor binding sites. The cloned cDNA encodes a Affinity Enrichment (Panning). In preparation for expres- 440-residue protein of the seven-helix guanine nucleotide- sion cloning, we had developed a set of modified (13) and binding protein (G-protein)-coupled receptor family. Ligand chimeric (14) opioid peptides, the latter based on the structure binding reveals a stereospecific site with typical opioid prop- of Dyn-32. Dyn-32 is the 17-residue DynA linked at its C erties, which binds peptide and nonpeptide opioids with mod- terminus through Lys-Arg to DynB (15). We had also raised a erate affinity (Kd 100 nM) and which lacks the expected K monoclonal antibody (mAb), 17.M, that recognized the C-ter- selectivity. The deduced transmembrane domain is 93% iden- minal sequence of Dyn-32 (15), leaving the opioid-active tical to the homologous region of the human neuromedin K K-selective N-terminal domain free to interact with cell- (neurokinin B) receptor, but the N-terminal and C-terminal surface receptors. These reagents were used for panning. sequences have many dissimilarities. The expressed receptor COS-7 cells were transfected (16) as follows. Exponentially binds opioid ligands but not tachykinins; and under the same growing cells were plated (5 x 105 per 10-cm dish) in Dulbec- conditions, a cloned rat neuromedin K receptor binds tachy- co's modified Eagle's medium (DMEM) with 10o fetal calf kinins but not opioids. serum (FCS). After 24 hr at 37°C in a 5% C02/95% air incubator, cells were washed twice with Iscove's modified Saturable stereospecific binding of an opiate ligand was first Dulbecco's medium (IMDM), which contains L-glutamine and demonstrated by one ofus in 1971 (1), and since then multiple 25 mM Hepes (pH 7.4). Plasmid DNA [10 ,ug, purified by CsCl opioid receptors have been characterized (2-5), but cloning gradient centrifugation (11)], DEAE-dextran (Mr 500,000, 0.2 has not yet been achieved to our knowledge. The cloning of mg/ml; Pharmacia), and chloroquine (100 ,uM) were mixed in an opioid-binding protein, OBCAM (6), has been reported; 4 ml of IMDM and added to the dish, which was returned to but as it has no transmembrane domain, OBCAM seems not the incubator for 5 hr. The cells were washed with IMDM and to be a receptor. For expression cloning of cDNA encoding then with DMEM containing 5% FCS and were cultured in 10 the dynorphin (Dyn) (K opioid) receptor (7) by ligand binding, ml of DMEM with 10% FCS for 3 days in the incubator. we turned to transient transfection ofCOS-7 cells and affinity The panning procedure itself was modified from Seed and enrichment (panning), followed by successive dilutions of Aruffo (17). Petri dishes (60-mm diameter, FALCON 1007; plasmid pools. The cDNA library was derived from human Becton Dickinson) were coated with mAb 17.M (15) (30,g placenta, a rich source of K receptors (8). per dish in 3 ml of50mM Tris HCl buffer, pH 9.5) for 2 hr and AND METHODS then washed three times with 0.15 M NaCl. Phosphate- MATERIALS buffered saline [PBS; 3 ml containing 0.1% bovine serum Vector and Library Construction. We used the pME18S albumin (BSA)] was added, and the dishes were incubated at vector (9), a high-copy-number small-vector [3.4 kilobases 4°C overnight. Transfected cells (see above) were detached (kb)] plasmid with a strong promoter (10), suitable for con- with PBS (GIBCO, without Ca2' and Mg2+) containing 0.5 structing size-selected unidirectional cDNA libraries and for mM EDTA and 0.02% NaN3. After washing the cells twice mammalian expression. Total RNA was isolated from fresh with Krebs-Hepes buffer (KHB; 118 mM NaCI/4.8 mM human placenta by guanidinium isothiocyanate extraction KCI/2.5 mM CaCl2/1.2 mM MgCl2/25 mM Hepes, pH 7.4), followed by centrifugation in cesium chloride (11), and cells from two dishes were suspended in 1 ml of KHB poly(A)+ RNA was purified by using an oligo(dT)-cellulose containing 0.1% BSA and 100 nM DYN-32 and were incu- column (Pharmacia). Synthesis of cDNA (12) was with a bated 1.5 hr at room temperature and then 30 min on ice. Promega kit. First-strand cDNA was synthesized by avian After three washes to remove free Dyn-32, cells were resus- myeloblastosis virus reverse transcriptase with oligo(dT)-Not pended in 1 ml of PBS containing 0.5 mM EDTA and 0.1% I primer-adapter [oligo(dT)15 containing the Not I site on its 5' BSA and plated on the antibody-coated dishes. After 2 hr at end]. Second-strand cDNA was synthesized by using Esche- richia coli DNA polymerase I and RNase H. After treatment Abbreviations: Brem, bremazocine; Dyn, dynorphin; DynA and with T4 DNA polymerase to blunt the ends, the double- DynB, Dyns A and B; mAb, monoclonal antibody; hNKR and stranded cDNA was ligated with Bst XI linker (Invitrogen) and rNKR, human and rat neuromedin K receptors; NK, neuromedin K; T4 ligase at 14°C for 24 hr. After a treatment with Not I to SP, substance P; SK, substance K; G protein, guanine nucleotide- create sticky ends, the cDNA was fractionated on 1% agarose binding protein; DAGO, [n-Ala2, N-MePhe4, Gly-o15]enkephalin; gel by electrophoresis, and fractions > 1.5 kb were electro- DPDPE, [D-Pen2, D-Pen5]enkephalin in which Pen is penicillamine; G protein, guanine nucleotide-binding protein. tTo whom requests for the hK1R clone should be addressed at: The publication costs of this article were defrayed in part by page charge Mental Health Research Institute, University of Michigan, Ann payment. This article must therefore be hereby marked "advertisement" Arbor, MI 48109. in accordance with 18 U.S.C. §1734 solely to indicate this fact. tTo whom reprint requests should be addressed.

4124 Pharmacology: Xie et A Proc. Natl. Acad. Sci. USA 89 (1992) 4125 room temperature, the dishes were washed gently three times with 3-ml portions of PBS. Then cells remaining on the dishes x 5 were lysed, and plasmid DNA was recovered by the Hirt E method (18) and amplified in E. coli to obtain material for the 4 next cycle of panning. 3 Ligand Binding. Binding assays were performed with whole 2 COS-7 cells; membrane preparations gave similar results. As 0.s positive controls we included the H-187 line of small-cell lung cancer cells (kindly furnished by J. D. Minna, National Cancer 0 100 200 300 Institute), which express about 5 x 105 K receptors per cell. As Free [3H]Brem, nM negative controls we used mock-transfected (plasmid alone, no cDNA insert) and untransfected cells. 10 b Transfected cells were harvested in PBS (without Ca2l and 0. Mg2+) containing 0.5 mM EDTA and 0.01% NaN3, then 80 f washed three times with KHB, and resuspended in KHB (106 U, 60 cells per ml). Binding assays were performed with 106 cells in _I 2 ml of KHB containing radioligands and competitors. 40 [3H]Bremazocine ([3H]Brem; 37.0 Ci/mmol, NEN; 1 Ci = 37 20~~~~~~ GBq), a high-affinity opioid ligand with modest selectivity for 201 K sites, was used as radioligand. Assay tubes were incubated 2 hr at room temperature and then on ice for 5 min and were 0 10 100 1000 10000 centrifuged (250 x g at 20C for 5 min). Free radioligand Competitor, nM concentrations (>80%o of total) were measured in superna- FIG. 1. Binding isotherm and competition. Data are means tants. Cell pellets were washed twice with ice-cold KHB and SEM for multiple experiments, with triplicate determinations at each transferred to fresh tubes; radioactivity was determined in point in each experiment. (a) [3H]Brem binding isotherm. (b) Com- scintillation solution (Cytoscint, ICN Biochemicals) or in the petition by U-50488 (e) and DynA-(1-13) (o; triplicate determina- y counter. tions in one experiment). For competition studies [3H]Brem was used at about 2 x 105 cpm per tube (2.5 nM). Specific binding was defined as the (500 nM) (19). Enrichment required the presence of both reduction of bound radioligand by 4 ,uM U-50488 (19) (see Fig. antibody and Dyn-32. Thus, we selected for expression of a lb). The following unlabeled opioid ligands were tested as binding site with affinity for both peptide and nonpeptide competitors: DynA-(1-13), des-Tyr1-DynA-(1-13), [D-Ala2, opioid ligands. N-MePhe4, Gly-o15]enkephalin (DAGO), and [D-Pen2, The plasmids recovered after cycle 4 were used to trans- D-Pen5]enkephalin (DPDPE) in which Pen is penicillamine form E. coli. Discrete colonies were transferred for further (Peninsula Laboratories); U-50488, U-63939, and U-63940 growth to separate wells of six 96-well master plates. Con- (Upjohn); levorphanol and dextrorphan (Roche); and nalox- tents of all 96 wells of each plate were pooled. DNA prepared one and naltrexone (Endo Laboratories, New York). The from each of the six pools was assayed by transfecting COS-7 following tachykinin radioligands were used: [3H]eledoisin cells and measuring [3H]Brem binding with and without 500 (38.2 Ci/mmol, NEN); radioiodinated substance K, [2-(125I)- nM U-50488 as competitor. One positive pool was found, iodohistidy']SK (2,000 Ci/mmol, Amersham); and tritiated which was subdivided repeatedly until a single positive clone substance P [2-prolyl-3,4-3H2]-[Sar9, Met(02)11]SP in which ("hKlR") was obtained. Sar is sarcosine (32.5 Ci/mmol, NEN). Unlabeled neuromedin Ligand Binding. Untransfected or mock-transfected COS-7 K (NK), SP, and SK (Peninsula) were tested as competitors. cells, although they bind [3H]Brem nonspecifically, have no Specific binding was determined with 500 nM unlabeled NK, detectable specific sites, nor do they express hK1R mRNA, SP, or SK as appropriate. as determined by Northern blotting (not shown). Transfected Sequence Analysis. Both strands of cloned cDNA were cells bound specifically about 106 Brem molecules per cell at sequenced (20)¶ by using both double-stranded (pME18S- Competition, % cDNA) and single-stranded (Ml3mpl9, Boehringer Mann- -20 -10 0 10 20 30 40 50 60 70 80 heim) templates with Sequenase (United States Biochemical) Item n and Taq polymerase as required. For some G+C-rich re- Mock*U-50488 24 U-50488 (50) 4 gions, dITP and 7-deaza-dGTP were used. Sequences were 34 --4 confirmed in both directions by multiple redundant reactions U-50488 MockkDynA-(1-1 3) 4 WMEM and gel loadings. Especially in G+C-rich regions we found DynA-(1-13) 36 some consistent disagreements between 5'-to-3' and 3'-to-5' des-TyrI-DynA-(1-1 3) 17 sequencing gels, usually with a missing nucleotide in one Levorphanol 17 Dextrorphan 15 sequence or the other; in those cases we assumed that the -= extra nucleotide was correct. When single- and double-strand Nalox 11 readings disagreed, we accepted the former. Data base Naltrex 6 searching was by FASTDB (IntelliGenetics); DNA and de- U-63639 7 duced protein sequences were analyzed by the Wisconsin and U-63640 7 DAGO (2) 5 .-4 Intelligenetics programs. DAGO (20) 3 DAGO (200) RESULTS DPDPE (2) 4 Isolation of cDNA. At the fourth panning cycle, 80% of the DPDPE (20) 4 --I cells could be prevented from attaching by competition with DPDPE (200) 5 U-50488 the K-opioid receptor-selective arylacetamide ligand FIG. 2. Ligand competition profiles. Competitors were at 500 nM unless otherwise indicated in parentheses. Solid bars indicate reduc- $The sequence reported in this paper has been deposited in the tion of [3H]Brem binding by competitor, expressed as percent of GenBank data base (accession no. M84605). specific binding. n, Number of experiments. 4126 Pharmacology: Xie et al. Proc. Natl. Acad. Sci. USA 89 (1992) saturation with a Kd of 87 nM (Fig. la). Competition curves DynA-(1-13) did not compete at all. Levorphanol (a A-opioid for both U-50488 and DynA-(1-13) yielded IC50 values (here receptor-selective morphinan) competed, but its (+)- equivalent to K,) of about 100 nM (Fig. lb). The maximum enantiomer dextrorphan (21) did not. Two other p-selective reduction in binding by competition was approximately equal morphinans, naloxone and naltrexone, competed poorly. to the increment in binding due to transfection. Surprisingly, the peptides DAGO (19) (,u-selective) and Competition results obtained with various ligands are DPDPE (19) (&.selective) competed about as well as U-50488. shown in Fig. 2. U-50488 gave near-maximal competition, The inability of the site to distinguish A-selective from and DynA-(1-13) competed about as well, but des-Tyr1- K-selective ligands was also evident from the fact that

1141 CCTTTCATCCACGTCTCCAGCTACGACGAGCTGGAGCTCAAAGCCACCAGGCTCCACCCA P F I H V S S Y D E L E L K A T R L H P 369 : : : K : : : : : : : : : : T : : F : : 1201 ATGCGACAGAGCAGCCTATACACAGTGACAAGAATGGAGTCCATGAGCGTGGTATTCGAC a M R Q S S L Y T V T R M E S M S V V F D 389 GCGAGAACCCCGACTGACCGCGGCCACGGCGGCTCCCCGACCTGCCGCGTCCTGCGGGCG N : : : : M : : : : : : : : : T : : : : 61 GCGCTGGGCTCCGGGCACTCGGGCTGCGCCCCCATGGCCTCGCCCGCGGGAACCTGAGC M A S P A G N* L S 9 1261 TCCAACGATGGGGACAGTGCCAGGTCCAGTCACCAGAAGAGAGGGAC:GAUCCAGAGACGT1A V G A D : V : : T S N D G D S A R S S H Q K R G T T R D V 409 P : : A : T T : : : R K : : A : P : : P 121 GCGTGGCCGGGCTGGGGGTGGCCGCCGCCGGCCGCGCTGAGGAACCTGACCTCCTCCCCG A W P G W G W P P P A A L R+ N* L T S S P 29 1321 GGCTCCAATGTCTGCTCCCGCAGGAACTCCAAGTCCACCTCCACCACAGCCAGCTTCGTG S L A A : A A T G : V E T : : S : : : G S N V C S R R N S K S T S T T A S F V 429 S F : G F : : A : A S I 181 GCCCCGACCGCGTCCCCGTCCCCGGCCCCGTCGTGGACGCCCTCGCCGCGCCCCGGCCCC A P T A S Pt S P A P S W T P S P R P G P 49 1381 AGCTCCTCCCACATGTCGGTGGAAGAAGGCTCTTGATTTCTCTCTGGGGTCAAGGCCACT S A L G L : : : : : : . Q : W A N L S S S H M S V E E G S > 440 : : P Y T : : D : Y : > 241- -GCGCACCCGTTCCTGCAGCCGCCCTGGGCCGTGGCGCTCTGGTCGCTGGCCTACGGCGCC------A H P F L Q P P W A V A L W S L A Y G A 69 1441 GCAGGCACCCCTTCTCCTGTCACTGCTGCTGTCTCACACTCTCTGGAAGCTGAAGGACAG T N Q : V : S : R I : V 1501 TTTTTAGACAGCTACCCTTACAATAAGACAGATTGCACATAAATATAACAAAAATACTAC ------TM-1 ------1561 TAAGATATGAGCTCTCCCCCCCAAAAAAGAACAAAATGGGCTTTAAGAGTATGCCTTGAA 1621 AACTCTAAATTATTAATATGATACAAACAAAAATATAGATCCGAGAAATATTTATAAAGT GTGGTGGCCGTGGCGGTGCTCGGCAACCTCGTGGTGATCTGGATCGTGCTGGCCCACAAG 1681 GTCCAGTTTTGCTTATTTAAAAGTCACTGTGCACATTTGTGACACTGATATGGTAGTTTT V V A V A V L G N L V V I W I V L A H K 89 1741 TTCCCAAAATATTAAAGTTTAAAATTTAATACTGTCAGTGAAGAGAAGCCATGTTTTCCA I : I : : : : 1801 TTACAGAGCATAGAATGGAAAAGTTAAATGACTCATTTTCTTACAATAGTGATGGAAATT

,______1861 TAACCTCAAAAACTAACAATTAACGAAATCTCAAGAAAACCTATTTTGTACCATAACAAT 1921 TTTCAAAGACATTTAAATGAAAAGGAAACCTAAATCAAACCACTAGGCTTATCTAAATGC 361 CGCATGCGGACGGTCACCAACTCCTTCCTCGTGAACCTGGCCTTCGCCGACGCCGCCATG 1981 CTTTCTCTTATTTTTTTCTGAGAAAATGATTTCAAAGGAAAAAAATGTAGCTTTGATTGT R M R T V T N S F L V N L A F A D A A M 109 2041 TACATATTTTAAATGCCAAGTTAATATGTAGTTAAACTTAAGACCTTAAAAGGACAAACA :: : : : Y : : : : : : : S : : S : 2101 AAATTCCTATGATCCTCTATTTTTCAGAATTTTGTTCTAAGTAGGTAAGTTGTAAGACAT ------TM-2 ------2161 TAAATATACTTTCTGAGATGGAAGGAAAGAATCCCATTTGTCTCTGTAACTGGCTGCTAG 2221 CCTTTAGGcaGGAACCACCCACAGCCTCACGTAGCCATGAAGGTGGACAGGAACACCTCC 421 GCCGCGCTCAACGCGCTGGTCAACTTCATCTACGCGCTGCACGGAGAGTGGTACTTCGGC 2281 CAGCTCCAAGGCAGTTGTTTTTCCCCTGTACCCCAGCAAAAGTTCCAGACATGCAcTTTA A A L N A L V N F I Y A L H G E W Y F G 129 2341 TCAACCATATCGTGTCCTCCTCCTCCTTCATCAAAGAAGGAGTGTGGGCATGGGGGAAGG : : F : T : : : : : : : : : S : : : : : 2401 ATCAGAATGCGTCTTGTGAAAATCCTGAGAGGAAAAAGTTGTAAGAATTATGAAAGCAAA 2461 TATAGCI-GATGAAGTTAATATACATGTTGGAAAATCAGACAGGAAGTAGAAAGTTGAGTC 2521 AACTCTTTGAAAGATGTACCATAGTTTGGGTCACCCGTCAGgTGAGTGACAATATTACCC 481 GCCAACTACTGCCGCTTCCAGAACTTCTTCCCCATCACCGCCGTGTTCGCCAGCATCTAC i 581 TGCTGTTCCACACAGAGACCTGTACGCTCTGCATAGGTAACCCTTGTCCCTCCAGAAAGG A N Y C R F Q N F F P I T A V F A S I Y 149 264 1 ACGGGAAAGAGGCATTTGTTTTACTACAATAGTATATTTTTTGAGAACCATATTTGTGAG 2701 TGTTTTATGCCTCAATCTTGAAGCATGAACCTTTCCTTAAATTAGGAATACTGTCAATCC ------TM-3 ------2761 TGCTGAAGAAATCACAACCCTTCTGGAAATCTAAATGTGTTATATAAACTTCTGTAAAAT 2821 ATTGTTAGGTTTTGAAAACTGTCTAAAATAATTATCTCTAACATTTATTTCATTGCTATG 541 TCCATGACGGCCATCGCGGTGGACAGATACATGGCCATTATTGACCCCCTGAAGCCCAGG 2881 CCTTCCTTAGTGTCAGAACCAAATAACTTTTCAAAGATCAGCATAAAAGCAATTATCCAA S M T A I A V D R Y M A I I D P L K P R 169 2941 TGACAAGTGATGGTCTATTGTTACCCTGATATTAATCTCCCAATCCTGCTTTGGAGCCAA 3001 AGTCAGAAATATTTAGTTGTTAGTCTAAACAGCTTAACAACATGAGTTTGAGTTGAATTT 3061 CTTTTAATGACACCAATAAACACAAACAAGTAGATGGCACAATAAATTTGCAGACATATA 3121 CAACCAGCCAATGAATGTAACAATATCAAGAAGTAAATTAAAATTAATTCTAAAACAGTA 601 CTGTCTGCCACGGCCACCCGGATCGTCATTGGAAGCATCTGGATTCTGGCATTTCTACTT 3181 TAAGTGGTCTTTCCAGGGTTCCTAGAAATAACCTAATAAAATCTGTGAAACATGTGTGCA L S A T A T R I V I G S I W I L A F L L 189 3241 CTTTTTTAGATAAACAAATGTATCATAATTTAGAATCTAATTGTTTGAATGTTTTAACAT K : : : : : : : : : 3301 GTACGGGAGCTTGGTCTTCAAATTTCATATAGTCAGCCACTAACAAAGTATATCTGAAAT ------TM-4 ------3361 ACATACTCTTGACCTTCACATGCATTACGCAAATTCATGCTATGGCGTTTCTAAAGAAAA 3421 AATAGTAGCTTAATCTTGTTTTGTTCTGTTTGTTTGGAATTTTTTCTTTAGTAGATTTGT 661 GCATTTCCTCAGTGTCTGTATTCCAAAATCAAAGTCATGCCAGGCCGTACTCTTTGCTAC 3481 TGTTGCCTTGCTTACCGAGCATCACTCCTTCTAGTATGGCAGAAATACTGAGGTCCAGGT A F P Q C L Y S K I K V M P G R T L C Y 209 3541 CACATCTCTTAAATAGTTAAGAAAAACTGACATCATTTACTCAATAGTCATGACTTTTAA : : : : : : : : : T : : : : : : : : : F 3601 ACTAAGATTTATTATATATAATTTTCAAGTTCAAGAAATGTAAGCAATAACAGTAAAATG 3661 AATGAAAAAGGCTAAAGGTTAGCCCTTGTGTCTGAATTTCGAAGCTaaAAAGTATGAAAT 3721 GATGCCCATGCAGAGCCGCTTTAGTGGGCTCTCTGTGAGTAAATCTATGCCAGTGTTTTC 721 GTGCAGTGGCC 3781 ACATTTGCCAAGGCTTAGAAGCATTTGCCTCCAAATGCGCTCTACCCCAATACTAACGTC V Q W P E G S R Q H F T Y H M I V I V L 229 3841 CACGTCCATCTTCTTCATTATTTGCAGTCAAACACTACTCAGGACACTGAGCAGATAGGT : P K : : : : : I : I : 3901 ACAACATCTTAGGGTTTATTAAATTTAGATCAGCAGACAAAAATCCTAAACTATGTTGAG ------TM-s . 3961 AAAAATATGGGAAAAAAAAGCCTTGCCTTGTTTTAAATATTCTCCTTTTTGAAAGAACAT 4021 GCTAGTAAAACAAACAAACATTGAATTTCTATTATTTTGCACCTGGACAAAGTGACTGAA 781 GTGTACTGCTTTCCTTTGCTCATCATGGGCATCACCTACACCATAGTTGGAATCACGCTC 4081 GTGGCCTGCCGGGGAAAAGTTTAAAGCAAACGCgGCTTTGTGGAGTCCAGTCTAGCTTTT V Y C F P L L I M G I T Y T I V G I T L 249 4141 TTTTAGTGGTTCAGTATGTTGTTGCATGATTCCACCTCCCAGGTGACATTTCTGACCCAG : A : : : : : : : : : : : : : : : : : : 4201 AAGCCACATTTACGTTTTCAGGACGTAAATCTGAAAATCTCTTGCAAAAAGAAATCTGGC 4261 CAACTTCAAAGTTCCGCCGCCCTTAGAAGGCACACAAAGCACCAAGAAGCTTAGTACTAA 4321 ACCTAACAAACACAAAATAAATGTAAAAACCAACACTAGTTACCTCAGAATTTGGATTGG 841 TGGGGAGGGGAGATCCCAGGAGACACCTGCGACAAGTACCAGGAGCAGCTGAAGGCCAAG 4381 ATTTTGTTAATGCAGAATTTCCCCAGAAACCTGTAATCAGTGTCTGTTAAATTGCTCCAT W G G E I P G D T C D K Y Q E Q L K A K 269 4441 TACATACAAAGACAGGAGGATTAAAACAATTCAACTAACAGTAACAATCTGAGTTCCATT : : : : : : : : : : : : : H : : : : : : 4501 TTCCTTTGATGGTGTGCCAGAAGTTAAGGAAATCAAGCATAACATTGGCCATGAAGAAAA 4561 AAATTGTAACAATCTCACTGGAGGCCAAACAGGAATGGAGAATCACATTTAATGGAGCTG 901 CGGAAGGTTGTAAAAATGATGATCATCGTTGTGGTGACCTTTGCCATCTGCTGGCTGCCC 4621 TACAAAGTCATITATTGTGTGATTTAATATACATTACTGAAATCCTGCGAGCAAGAATT R K V V K M M I I V V V T F A I C W L P 289 4681 CATATATATAAAATTTGTAGGCAGTGCATAAAGTATTTTTCAAGTTGTGGAAATTATACT M : : : : 4741 GAGTATGCTAAAAATTCCATCTTCTGTATATGTGCCAGTATTTTGGAAAGTAAATCCA ------TM-6 - 4801 ATGTTTTTATAAATATATTAAAAATCATATGAAAAAT (A)22 TATCACATCTACTTCATCCTCACCGCCATCTATCAGCAGCTGAACAGGTGGAAATACATC Y H I Y F I L T A I Y Q Q L N R W K Y I 309 b : : : : : : : : : : : : : :

102 1 CAGCAGGTCTACCTGGCCAGCTTCTGGCTGGCCATGAGCTCGACCATGTACAACCCCATC Q Q V Y L A S F W L A M S S T M Y N P I 329 .------TM-7------l X 1081 ATCTACTGCTGTCTGAATAAGAGATTTCGTGCTGGCTTCAAGAGGGCCTTCCGCTGGTGC I Y C C L N K R F R A G F K R A F R W C 349 I 0 I o I : : : : : : : : : : : : : : I I m 4

FIG. 3. (a) Nucleotide and deduced amino acid sequence (in single-letter code) of hK1R and comparison with deduced amino acid sequence of hNKR (22). The sequence ofhNKR is shown under that of hK1R. Nucleotides are numbered on the left and amino acids on the right. Possible N-glycosylation sites (*) and stop codons (>) are marked. Alignment was by the Wisconsin GAP program, with gaps inserted to maximize matches. Colons indicate identities. The following sequence in hNKR precedes the start of hK1R: MATLPAAETWIDGGGG. Gaps: +, gap in hK1R after amino acid residue 23 with insert GWLQLLDQAG in hNKR; #, gap in hK1R after residue 35 with insert VA in hNKR; * * *, three-residue gap in hNKR opposite residues 41-43 of hK1R. Putative hydrophobic transmembrane domains TM-1 through TM-7 are according to ref. 22. (b) Hydropathy plot as described by Kyte and Doolittle (23), for hK1R showing hydrophobic segments upward and residue number on the x axis. Pharmacology: Me et al. Proc. Natl. Acad. Sci. USA 89 (1992) 4127

U-63639 (IL-selective) and U-63640 (K-selective) (19) com- Untransfected cells bound radiolabeled eledoisin hardly at peted equally well. all, whereas cells transfected with rNKR (but not hK1R) Sequence Analysis. Fig. 3a shows the nucleotide and de- bound it very well (Table 1). Radiolabeled SP and SK ligands duced amino acid sequence of hK1R cDNA and also of bound relatively poorly (not shown). NK was the most potent human NK receptor (hNKR; see below). In hK1R the insert tachykinin tested as competitor (K, S nM), reducing binding of 4839 base pairs is followed by a poly(A) tract of 22 by 95% to the mock level. These results are in general nucleotides. A long open reading frame after the first ATG agreement with the report of Nakanishi's group (25). These encodes a protein of 440 residues (including the methionine) cells showed no specific binding of [3H]Brem, nor did the and calculated Mr of 49,422. The initial ATG (in CCCATGG) K-selective opioid ligand U-50488 compete with [3H]eledoi- meets the Kozak consensus criterion (ACCATGG) (24) only sin. In contrast, hKlR-transfected cells in the same experi- moderately well; as the insert has no upstream in-frame stop ments bound [3H]Brem specifically but did not bind any ofthe codon, additional 5' coding sequence cannot be excluded. tachykinin radioligands, nor did the tachykinins compete Fig. 3b shows the hydrophobicity analysis according to against [3H]Brem binding. Because of the considerable non- Kyte and Doolittle (23). No signal sequence is present. Seven specific binding of [3H]Brem, U-50488 reduced total binding hydrophobic segments are evident, each long enough to form by only 19o, but this represented 64% reduction of specific a membrane-spanning a-helix. These are placed in Fig. 3a binding or of the incremental binding caused by transfection. according to the assignments for the tachykinin receptors (22) Northern Hybridization. A 32P-labeled synthetic oligonu- (see below), but additional flanking hydrophobic residues cleotide probe, corresponding to base pairs 1-150 of hK1R, could have been included in all but transmembrane domains one of the regions of maximum dissimilarity to hNKR, TM-3 and TM-4. In the N-terminal domain are two potential hybridized in a sharp band of about 5.5 kb to mRNA from N-glycosylation sites, at amino acid residues 7 and 24. human placenta, brain, and other tissues (not shown). Cysteine residues occur in the first and second extracellular loops. In the C-terminal cytoplasmic tail, 27% ofthe residues DISCUSSION are serine or threonine, suggesting possible regulation by That the protein encoded by hK1R is an opioid receptor is protein kinases. supported by the following four considerations. (i) Opioid A search of the nucleic acid data bases revealed that the receptors bind both peptide and nonpeptide ligands with greatest similarity of hK1R was to receptors of the 7-helix stringent structural requirements. The pattern ofbinding here family and especially to the rat NK receptor (rNKR), with is classically opioid in that both opioid nitrogenous bases 79% identity of the deduced protein sequences (three gaps). (e.g., Brem, U-50488 and other arylacetamides, and levor- S. Nakanishi (Kyoto University) provided the rat clone phanol) and opioid peptides [e.g., DynA-(1-13), DAGO, and (rNKR-CDM8) (25), which was expressed in COS-7 cells. DPDPE] are bound. (ii) N-terminal tyrosine is a nearly Table 1. Ligand binding to hK1R and rNKR Competition, Competitor cpm, mean ± SEM % Radioligand n (500 nM) a b a b Transfection: rNKR [3H]Eled 4 Untransfected 44 ± 3 6 (5) Control 3688 ± 86 4341 ± 45 5 U-50488 4256 ± 77 2 6 (3) NK 213 ± 20 301 ± 30 95 94 6 SP 2798 ± 37 24 6 SK 3084 ± 43 17 [3H]Brem 28 Untransfected 3776 ± 270 8 (3) Control 3682 ± 392 5033 ± 96 8 (3) U-50488 3962 ± 430 4949 ± 357 -8 2 5 NK 3515 ± 280 4 7 SP 3882 ± 654 5 2 (3) SK 3757 ± 32 4992 ± 107 2 1 Transfection: hK1R [3H]Eled 4 Mock 44 ± 3 12 Control 44 ± 3 43 ± 4 3 U-50488 49 ± 2 8 NK 59 ± 7 8 SP 44 ± 4 8 SK 40 ± 3 [3H]Brem 28 Mock 3776 ± 270 17 Control 5388 ± 353 34 U-50488 4361 ± 216 19* 11 NK 5762 ± 514 -7 11 SP 5759 ± 542 -7 11 SK 5879 ± 509 -9 All data are total binding. "Control" is the binding of radioligand alone to transfected cells. Competition is the percent reduction of control binding. Columns a are independent transfections (number in column n) carried out over a period of weeks; columns b are independent transfections (second number in column n) carried out in a single day. To show clearly the differences in nonspecific binding between the two expressed receptors, raw counts are given here rather than specific binding data. [3H]Eled, [3H]eledoisin. *This is 64% of specific binding (compare Fig. 2). 4128 Pharmacology: Xie et al. Proc. Natl. Acad. Sci. USA 89 (1992) universal requirement for recognition of opioid peptides by opioid peptides and compact opioid nonpeptides are recog- opioid receptors. Therefore, it is significant that whereas nized by the same binding site. DynA-(1-13) binds, its des-Tyr' derivative does not. (iii) A final caveat is in order. Two segments of the coding Stereoselectivity is a feature of all known opioid receptors. sequence (beginning at nucleotides 18 and 585) contain al- The site expressed here is stereoselective, as revealed by the ternative long open reading frames encoding 188 and 171 contrast between levorphanol and its enantiomer dextror- amino acids, respectively. Remote as the possibility may be phan. (iv) Binding and functional behavior ofopioid receptors in view of the multiple redundancies in our sequencing depend on coupling to a guanine nucleotide-binding protein strategy, it is nevertheless conceivable that one or two (G protein), probably G, or Go (26). It is significant, therefore, nucleotides were missed, shifting the reading frame and that the hydropathy plot here reveals a seven-helix resulting in a completely different deduced protein sequence. structure, Analogous cases are known (32, 33). However, the hydrop- typical of receptors coupled to G proteins. athy plot for other readingframes ofhK1R gives no indication As a K-receptor-selective peptide ligand was used for of panning, and as human placenta (source ofthe cDNA library) multiple transmembrane helixes. contains chiefly K receptors, we had expected to find typical We thank Drs. K. Arai and T. Yokota for helpful advice and K-receptor affinity (low nanomolar range) and also a prefer- guidance, Dr. K. Maruyama for providing pME18S, D. Robison for ence (by at least 1 order of magnitude) for K-selective over pu- synthesis of oligonucleotides, Dr. R. C. Thompson for assistance and 6-selective ligands. Thus, the observed affinity (2 orders with the rNKR clone, Dr. J. D. Minna for the H-187 cell line, Dr. P. of lower) and the poor selectivity of the expressed von Voigtlander for Upjohn arylacetamide ligands, and Dr. S. magnitude Nakanishi for providing the rat NKR clone and for information about receptor were unexpected. There are five possible explana- the amino acid sequence ofhNKR. Drs. H. Akil, D. B. Goldstein, F. tions. (i) As there is no stop codon upstream of the first ATG, Lee, L. Stryer, J. A. Waitz, and S. J. Watson kindly commented on a 5' coding sequence that makes a major contribution to earlier drafts of the manuscript. DNAX Research Institute is sup- affinity and selectivity could be missing. (ii) The correct G ported by Schering-Plough Corporation. protein could be missing from the COS-7 cells, as found recently with a secretin receptor, which also belongs to the 1. Goldstein, A., Lowney, L. I. & Pal, B. K. (1971) Proc. Nadl. Acad. Sci. USA 68, 1742-1747. seven-helix family (27). As the binding affinity in our exper- 2. Terenius, L. (1973) Acta Pharmacol. Toxicol. 32, 317-320. iments was unchanged in washed membrane preparations, a 3. Pert, C. B. & Snyder, S. H. (1973) Science 179, 1011-1014. high cytosolic GTP concentration (well known to reduce 4. Simon, E. J., Hiller, J. M. & Edelman, I. (1973) Proc. Natl. Acad. Sci. USA 70, 1947-1949. binding affinity) (26) could not be responsible. (iii) An ac- 5. Loh, H. H. & Smith, A. P. (1990) Annu. Rev. Pharmacol. Toxicol. 30, cessory protein other than a G-protein could be required for 123-147. high-affinity binding and selectivity. More than a single 6. Schofield, P. R., McFarland, K. C., Hayflick, J. S., Wilcox, J. N., Cho, polypeptide is sometimes T. M., Roy, S., Lee, N. M., Loh, H. H. & Seeburg, P. H. (1989) EMBO required for high-affinity binding J. 8, 489-495. [e.g., interleukin 2 receptor (28)], but this has not been found 7. Chavkin, C., James, I. F. & Goldstein, A. (1982) Science 215, 413-415. for any G-protein-coupled receptor. (iv) A posttranslational 8. Porthe, G., Frances, B., Verrier, B., Cros, J. & Meunier, J.-C. (1988) Life modification could be needed for high-affinity binding but not Sci. 43, 559-567. 9. Maruyama, K. & Takebe, Y. (1990) Med. Immunol. (Tokyo) 20, 27-32. carried out by COS-7 cells. It is noteworthy that there are two 10. Takebe, Y., Seikie, M., Fujisawa, J., Hoy, P., Yokota, K., Arai, K., potential glycosylation sites in the extracellular N-terminal Yoshida, M. & Arai, N. (1988) Mol. Cell. Biol. 8, 466-472. domain of the deduced sequence. (v) This opioid receptor is 11. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) in Molecular Cloning: actually of a novel type, which does not discriminate A Laboratory Manual (Cold Spring Harbor Lab., Cold Spring Harbor, among NY), 2nd Ed. the classical type-selective ligands. A nonselective opioid 12. Gubler, U. & Hoffman, B. J. (1983) Gene 25, 263-269. receptor has been described (29), but it has very high affinity 13. Goldstein, A., Nestor, J. J., Jr., Naidu, A. & Newman, S. R. (1988) Proc. for ju, K, and 8 ligands. Natl. Acad. Sci. USA 85, 7375-7379. 14. Xie, G., Miyajima, A., Yokota, T., Arai, K. & Goldstein, A. (1990) Proc. That a receptor with opioid binding properties should be a Natl. Acad. Sci. USA 87, 3180-3184. member of the tachykinin receptor family is interesting from 15. Barrett, R. W. & Goldstein, A. (1985) Neuropeptides 6, 113-120. an evolutionary standpoint. The similarity to hNKR also 16. Maruyama, K., Wang, H.-M., Miyajima, A., Takebe, Y. & Arai, N. poses the question of what residues are responsible for ligand (1991) in Methods in Nucleic Acids Research, eds. Karam, J. D., Chao, L. & Warr, G. W. (CRC, Boston) pp. 284-305. recognition. S. Nakanishi kindly informed us in advance of 17. Seed, B. & Aruffo, A. (1987) Proc. Nail. Acad. Sci. USA 84, 3365-3369. publication of the deduced amino acid sequence of hNKR 18. Hirt, B. (1967) J. Mol. Biol. 26, 365-369. (22). The comparison is shown in Fig. 3a. Overall, there is 19. Goldstein, A. & Naidu, A. (1989) Mol. Pharmacol. 36, 265-272. 81% sequence identity (with three gaps), but the distribution 20. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad. Sci. USA 74, 5463-5467. of identical and nonidentical residues is very uneven. 21. Goldstein, A. & Naidu, A. (1990) Proc. Natl. Acad. Sci. USA 87, Throughout the central transmembrane region the sequences 1629-1632. are 93% identical (no gaps), and the cytoplasmic C-terminal 22. Takahashi, K., Tanaka, A., Hara, M. & Nakanishi, S. (1992) Eur. J. tails are 74% identical (no gaps). The extracellular N-terminal Biochem., in press. 23. Kyte, J. & Doolittle, R. F. (1982) J. Mol. Biol. 157, 105-132. domains differ greatly (39% identical, three gaps). Therefore, 24. Kozak, M. (1986) Cell 44, 283-292. we suggest that the very different ligand binding selectivities 25. Shigemoto, R., Yokota, Y., Tsuchida, K. & Nakanishi, S. (1990) J. Biol. of NKR and hK1R are determined in large part by the Chem. 265, 623-628. extracellular N-terminal domains, which are substantially 26. Childers, S. R. (1991) Life Sci. 48, 1991-2003. 27. Ishihara, T., Nakamura, S., Kaziro, Y., Takahashi, T., Takahashi, K. & different in the two receptors. Furthermore, hNKR and Nagata, S. (1991) EMBO J. 10, 1635-1641. rNKR have three and four potential glycosylation sites, 28. Hatakeyama, M., Tsudo, M., Minamoto, S., Kono, T., Doi, T., Miyata, respectively, as compared with two in hK1R. This model is T., Miyasaka, M. & Taniguchi, T. (1989) Science 244, 551-556. consistent with thyrotropin and luteinizing hormone recep- 29. Pasternak, G. W. (1988) Adv. Exp. Med. Biol. 236, 81-93. 30. Atassi, M. Z., Manshouri, T. & Sakata, S. (1991) Proc. Nail. Acad. Sci. tors (30) as well as tachykinin receptors (22). That ligands of USA 88, 3613-3617. seven-helix receptors bind in a pocket in the membrane was 31. Dohlman, H. G., Thorner, J., Caron, M. G. & Lefkowitz, R. J. (1991) deduced from mutagenesis studies with small nonpeptide Annu. Rev. Biochem. 60, 653-688. ligands (31). For recognition of the extended pharmacophore 32. Yamamoto, F., Clausen, H., White, T., Marken, J. & Hakomori, S. (1990) Nature (London) 345, 229-233. of a peptide, the N-terminal extracellular domain of the 33. Watkins, S., Madison, J., Davis, E., Sakamoto, Y., Galliano, M., receptor may be required. However, this hypothesis would Minchiotti, L. & Putnam, F. W. (1991) Proc. Nail. Acad. Sci. USA 88, not resolve the long-standing problem of how extended 5959-5963.