Proc. Nadl. Acad. Sci. USA Vol. 91, pp. 8042-8046, August 1994 Neurobiology Agonists and antagonists bind to different domains of the cloned c opioid receptor HAEYOUNG KONG*, KAREN RAYNOR*, HIDEKI YANOt, JUN TAKEDAt, GRAEME I. BELLt, AND TERRY REISINE*t *Department of Pharmacology, University of Pennsylvania School of Medicine, Philadelphia, PA 19104; and tHoward Hughes Medical Institute and Departments of Biochemistry and Molecular Biology and Medicine, University of Chicago, Chicago, IL 60637 Communicated by L. B. Flexner, May 6, 1994 ABSTRACT Opium and its derivatives are potent analge- intracellular COOH termini are divergent. It seems reason- sics that can also induce severe side effects, including respira- able to assume that these divergent extracellular regions may tory depression and addiction. Opioids exert their diverse be responsible for the distinct ligand-binding profiles ofthe 8, physiological effects through specific membrane-bound recep- K, and 1i receptors. To test this hypothesis, we have ex- tors. Three major types of opioid receptors have been de- changed the extracellular NH2 termini of the mouse 8 and K scribed, termed 8, c, and ,t. The recent molecular cloning of receptors (3, 4) and examined the abilities of these chimeric these receptor types opens up the possibility to identify the receptors to bind 8- and K-selective agonists and antagonists, ligand-binding domains of these receptors. To identify the as well as to mediate inhibition of adenylyl cyclase activity. ligand-binding domains of the K and 8 receptors, we have expressed in COS-7 cells the cloned mouse 8 and K receptors METHODS and chimeric 8/K and K!8 receptors in which the NH2 termini have been exchanged. The opioid antagonist naloxone binds Generation of Chimeras. To exchange NH2 termini be- potently to wild-type K receptor but not to wild-type 8 receptor. tween the mouse 8 and K opioid receptors, a common The K!8 chimera bound [3H]naloxone with high affinity. In restriction site, Spe I, was generated at an equivalent position contrast, the K-specific agonist [3HJU-69,593 did not bind to the in the cDNAs in the region encoding the first trainsmembrane K8/ chimera. These fmdings indicate that selective agonists domain without altering the amino acid sequence of either and antagonists interact with different recognition sites in the receptor. Site-directed mutagenesis was done by using the K receptor and localize the antagonist-binding domain to the Altered Sites in vitro mutagenesis system (Promega) and NH2 terminus. Consistent with the results of radioligand- 27-mer oligonucleotides containing the Spe I site (8-receptor binding studies, the K!8 chimera did not mediate K-agonist oligonucleotide, CTGGGCAACGTACTAGTCATGTTTG- inhibition of cAMP formation. In contrast, the B/l chimera GC, and K-receptor oligonucleotide, GTGGGCAATTCAC- did mediate K-agonist inhibition of cAMP formation, but this TAGTCATGTTTGTC). After digestion with Spe I and the effect was not blocked by naloxone. Furthermore, a truncated appropriate 5' and/or 3' enzymes, the cDNA fragments Kreceptor lacking its NH2 terminus was able to mediate agonist encoding the NH2 and COOH termini of 8 and K were isolated inhibition of cAMP accumulation in a naloxone-insensitive from a 1.2% low-melting-point agarose gel. Fragments en- manner. This result further indicates that the NH2 terminus of coding the NH2 terminus of 8 receptor and the COOH the K receptor contains the selective antagonist-binding do- terminus of K receptor and vice versa were ligated together main. The ability to dissociate agonist- and antagonist-binding and cloned into the mammalian expression vector pCMV-6c. sites will facilitate the development of more specific K agonists, Truncated 8 and K receptors were generated by ligating the which could have analgesic properties devoid of side effects. fragments encoding the COOH termini directly into the expression vector. Translation of the receptor sequences in Opioids such as morphine are used clinically for the man- these constructs was predicted to begin at a conserved ATG agement of pain (1). However, the use of opioids has unde- just distal to the Spe I site. sirable side effects-including respiratory depression, de- Radioligand-Binding Assays. The chimeras K1 78/&70372 and creased gastrointestinal motility, sedation, nausea, and mood 8169/K79380 (Fig. 1) were generated and transfected into changes. Other major limitations include abuse potential, COS-7 cells in parallel with either wild-type K or 8 receptor tolerance, and physical dependence. Morphine and the en- by the calcium phosphate precipitation method as described dogenous opioid peptides, the enkephalins, endorphins, and (4, 6). For receptor-binding studies, COS-7 cells expressing dynorphins, exert their physiological effects through mem- the receptors were harvested 72 hr after transfection in 50 brane-bound receptors expressed in the central and periph- mM Tris-HCl, pH 7.8/1 mM EGTA/5 mM MgCl2/leupeptin eral nervous systems and in target tissues. at 10 pg/ml/pepstatin at 10 a/ml/bacitracin at 200 pg/ The three major types ofopioid receptors, 8, K, and ,u, have ml/aprotinin at 0.5 ,ug/ml (buffer 1) and centrifuged at 24,000 recently been cloned and functionally characterized (2-5). x g for 7 min at 4°C. The pellet was homogenized in buffer They belong to the Asp-Arg-Tyr (DRY)-containing subfamily 1 using a Polytron. The homogenate was centrifuged at 48,000 of seven transmembrane-spanning receptors. There is =60% x g for 20 min at 4°C, and the pellet was resuspended in buffer amino acid identity among the sequences of the 8, K, and , 1 and used in the radioligand-binding assay. Cell membranes opioid receptors. The sequences of the putative membrane- (10-20 ,ug of protein) were incubated with [3H]U-69,593 (2 spanning segments (TM I-VII) and the three intracellular nM, specific activity 47.4 Ci/mmol; 1 Ci = 37 GBq), loops connecting these segments are highly conserved, [3H]naloxone (6 nM, specific activity 72.1 Ci/mmol), [3H][D- whereas the sequences of the extracellular NH2-termini Pen2, D-Pen5]enkephalin, where Pen is penicillamine ([3H]- segments, second and third extracellular loops, and the DPDPE; 2 nM, specific activity 34.3 Ci/mmol), or [3H]nal- The publication costs of this article were defrayed in part by page charge Abbreviations: DPDPE, [D-Pen2, D-Pen5]enkephalin, where Pen is payment. This article must therefore be hereby marked "advertisement" penicillamine; DSLET, [D-Ser2, D-Leu5]enkephalin-Thr. in accordance with 18 U.S.C. §1734 solely to indicate this fact. tTo whom reprint requests should be addressed. 8042 Downloaded by guest on September 26, 2021 Neurobiology: Kong et al. Proc. Natl. Acad. Sci. USA 91 (1994) 8043 wild-type 8 wild-type K NH2 NH2 I| mli(extr~elular W~- ;g5r-. e _o ntracellular HOOC - K1-78/&70-372 81-69/K79-380 ( I II II FIG. 1. Schematic of wild-type and chimeric 8 and K opioid receptors. trindole (1 nM, specific activity 31.2 Ci/mmol) in a final and the cAMP was quantified by using an RIA kit from volume of200 A.l for 40 min at 250C in the presence or absence DuPont/NEN, as described (4). of competing agents. All radioligands were obtained from DuPont/NEN. Nonspecific binding was defined as radioac- RESULTS tivity remaining bound in the presence of 1 uM naltrindole or The radioligand-binding properties of the chimeras Kl-78/t70- naloxone for 8- and K-selective ligands, respectively. The 372 and 81-9/K7938o and the mouse 8 and K receptors were binding reaction was terminated by the addition ofice-cold 50 examined. As shown (4, 6), the wild-type K receptor could be mM TrisHCl, pH 7.8 and rapid filtration over Whatman labeled with the K-selective agonist [3H]U-69,593 and the GF/B glass fiber filters pretreated with 0.5% polyethylene- antagonist [3H]naloxone, whereas the wild-type 8 receptor imine and 0.1% bovine serum albumin. The filters were was labeled with the 8-selective agonist [3H]DPDPE and washed with 12 ml of ice-cold buffer and soaked overnight in antagonist [3H]naltrindole. These K-selective and 8-selective scintillation fluid. The bound radioactivity was determined ligands have minimal cross-reactivity (4, 6). Although nalox- by using a scintillation counter. IC50 values were analyzed as one and dynorphin A also have high affinity for the ,u receptor described (6) and obtained using the curve-fitting program (6), in the context of these studies they are referred to as FITCOMP on the National Institutes of Health-based Prophet K-selective because they have very low affinities for wild- system. Saturation analysis of [3H]naloxone binding was type 8 receptor (6). Only subtype-selective ligands were used used to determine Kd values and densities of receptors in our studies because we anticipated that at least one of the expressed in the COS-7 cells, and these Kd values were used chimeric receptors may contain a mixture of K and 8 ligand- to convert the IC50 values to Ki values as described (6) using binding sites. Only through the use of subtype-selective the Cheng-Prusoff equation. Binding to the Kk78/870-372 chi- ligands were we able to discern these different sites on the mera (Fig. 2A) was expressed as a percentage of binding to same chimeric receptor. wild-type 8 for [3H]DPDPE and [3H]naltrindole or wild-type The two chimeric opioid receptors had specific agonist- K for [3H]U-69,593 and [3H]naloxone. and antagonist-binding properties. The K178/870 372 receptor cAMP Accumulation Assays. cAMP accumulation in COS-7 bound the antagonist [3H]naloxone (which poorly labels the cells expressing the wild-type or mutant receptors was mea- wild-type 8 receptor) and the 8-selective agonist and antag- sured as described (7).