Proc. Nat. Acad. Sci. USA Vol. 72, No. 2, PP. 590-594, February 1975 Morphine Receptors as Regulators of Adenylate Cyclase Activity (neuroblastoma X g&mm hybrid/prostaglandin E,/narmtic dependence) SHAIL K. SHARMA*$, MARSHALL NlRENBERG*, AND WERNER A. KLEEt * Laboratory of Biochemical Genetics, National Heart and Lung Institute, and the t Laboratory of General and Comparative Biochemistry, National Institute of Mental Health, Bethesda, Maryland 20014 Con~ribthd by Marshall Nirenberg, November 11, 1974 ABSTRACT Morphine inhibits adenylate cyclase (EC 4.6.1.1) activity of neuroblastoma X glioma hybrid cells. The inhibition is stereospecific and is reversed by the antagonist, naloxone. The relative affinities of narcotics for the opiate receptor agree well with their effectiveness as inhibitors of adenylate cyclase. Morphine-sensitive and -insensitive cell lines were found, and the degree of sensi- tivity was shown to be dependent upon the nbundnnce of narcotic receptors. Thus, morphine receptors are func- tionally coupled to adenylate cyclase. A molecular mech- anism for narcotic addiction and tolerance is proposed. A ncuroblastoma X glioma hybrid cell line with morphine receptors has been described (1). The receptors are stereo- specific with respect to narcotic binding, and receptor affinities for narcotics closely rcscmble those of rat brain (2-4) and correlate well with their pharmacologic potencies. Collier and Roy have reported that prostaglandin E de- pendent CAMP synthesis is inhibited by narcotics in rat brain homogenates (5, 6). In this communication the effects of morphine and other narcotics upon adenylate cyclase [EC 4.6.1.1; ATP pyrophosphate-lyase (cyclizing)] activity of neuroblastoma X glioma hybrids and parental cells are examined. We find that morphine and other narcotics inhibit adenylate cyclase activity and that the inhibition is dependent upon morphine receptors. MATERIALS AND METHODS JIalerials. [G-`H]Naloxone (26.3 Ci/mmol) and [G-`HI- CAMP (22.1 Ci/mmol) were from New England Nuclear; [(u-~*I']ATI (6-10 Ci/mmol) was from ICN; and [t'-"Cl- CAMP (0.295 Ci/mmol) was from Amersham/Searle. Mor- phinc sulfate was from Merck; naloxone.HCl was a gift of Endo Labs; etorphine was from Dr. R. Willette, NIDA; other narcotics were from Dr. Everette May; FGEr was from Dr. John Pike, Upjohn Co.; Ro20-1724 [4-(3-butosy-4- methosybenzyl)-2-imidazolidinone was from Hoffman La- Roche; IHMX (3-isobutyl-1-methylxanthine) was from rlldrich; Gpp(NH)p (5'-guanylylimidodiphosphate) was from ICN; cAAll', crcatine phosphate, and creatine kinase (155 U/mg of protein) were from Sigma. Cell Lines. Neuroblastoma X glioma hybrid NGlOS-15 was obtained5 by fusion of mouse neuroblastoma clone N18TG-2 Abbreviations: Ro29-1724,4-(3-butoxy-4-met.hoxybenzy1)-2-imid- azolidinone; IBMX, 3-isobutyl-1-methylxanthine; GPP- (NH)p, 5'-guanylylimidodiphosphate; PGE,, prostaglandin EI. t Fogarty International Fellow, on leave from the Department of Biochemistry, All India Institute of Medical Sciences, New I)elhi, India. 5 B. Hamprecht, T. Amano, and M. Nirenberg, in preparation. (7) and rat glioma clone C6BC-1 (8), derived from C6 (9). Cells were grown as described (1). Assay of cAJlP of In&l Cells. Confluent cells (3-4 mg of protein 60 min dish) were washed 3 times with 5 ml of medium A [Dulbecco's modified Eagle's medium with 25 mM Hepes (N-2-hydroxyethylpiperazine-N'2-ethanesulfonic acid) pH 7.4, instead of NaHCOa, adjusted to 340 mosmol/liter with 1.1 g of NaCl/liter] and incubated with medium A plus 0.5 mM Ro20-1724 and 0.5 mM 1BMX for 30 min at 37". Reactions were initiatecl by the addition of 30 ~1 of narcotic in water, 30 /II of PGE, in ethanol, or solvent. Ethanol (0.5% with Ro20-1724 present or 1% when Ro20-1724 and PGE, were present) had no effect upon CAMP formation. After incubation the medium was discarded and 3 ml of 5% trichloroacetic acid with 5 pmol of [i4C]cAM1' (3009 cpm) at 3" were added. The extract and 2 washes (each 1.5 ml of 5% trichloroacetic acid) were combined and centrifuged and the supernatant fluids applied to 0.8 X g-cm columns of AG 5OW-X4 resin, 200-400 mesh, H+ form (BioRad). The 3 ml of eluate following a 6-ml water wash was assayed for CAMP by the method of Gilman (10). Values reported arc for duplicate dishes and arc corrected to 100% recovery of c.4MP. Adenylate Cyclase Assay. Cells that had been washed three times were homogenized in 0.32 M sucrose, 10 mM Trisa HCL, 111I 7.4 (15 mg of protein per ml) with lo-15 strokes, by hand, of a ground-glass homogenizer. Enzyme activity was tle- termined by a modification of method C of Salomon el al. (11). Each tube contained: 45 mM Tris.HCl, pH 7.4; 5 mM MgCl,; 160 mM sucrose; 20 mM creatine phosphate; 10 U of creatine kinase; 1 mM CAMP; 0.5 m?ul Ro20-1724 (0.570 ethanol, final concentration); 1 mM [a-a21']hT1' (3 to 5 X 10' cpm) ; and 100-300 rg of homogenate protein in a final volume of 100 ~1. The reaction was terminated by addition of 50 ~1 of 157" trichloroacetic acid. For product characterization, the Y' product purified through the alumina column step (11) was subjected to Dowes-1 formate column chromat.ography. The 2 M formic acid eluate was lyophilizecl, and the radioac- tivity was characterized by pnper or thin-layer chromatog- raphy in: (A) isopropanol-NHIOI-I-O.1 M boric acid (7 : 1: 2) ; (B) 1 M ammonium acetate-95$& ethanol (3: 7) ; (C) isobutyric acitl-2 M NH,OH (2: 1); and (D) HIO-washed polyethylene- imine-cellulose thin-layer plates developed sequentially with HZ0 and 0.25 M LiCl. Grcatcr than 90% of the radioactivity rccovercd from the alumina columns was CAMP. Opiate Biding. The assay (1) was modified so that incuba- tion mixtures contained the components of the adenylate cyclase assay and 875 rg of homogenate protein; 5 X 1Om8 M [%]nalosone (4159 cpm); and unlabeled narcotic in a final volume of 206 ~1. Under these conditions, narcotic affinities Proc. Nat. Acad. Sci. USA 7% (1076) Morphine Receptors and Adenylate Gyclase 591 ,400 A CELLS Pea 0' I 2 4 6 8 IO MINUTES 0 MINUTES FIG. I. Inhibition by morphine of the rate of CAMP accumulation in intact neuroblastoma X glioma hybrid NG108-15 cells and of adenylste cyelase activity in homogenates. The effect of morphine on basal and PGEl-stimulated CAMP levels in intact cells is shown in part A. In B, the rate of adenylate cyclsee activity in NG108-15 homogenates is shown. Concentrations of PGEI and morphine were 10 PM. for the receptor are lower than those as measured (1) by the following factors: naloxone, 2; etorphine, 13; destrorphan, 67; levorphanol, 667; morphine, 200; and 3-allylprodinc, 100. RESULTS The effects of morphine upon basal and PGEi-stimulated CAMP levels in intact NGl08-15 hybrid cells and on adenylate cyclase activity in cell-free homogenates are shown in Fig. 1A and 13, respectively. The addition of PGE, to intact cells results in a 40-fold increase in intracellular CAMP levels. The addition of morphine markedly reduces cAMP levels, both in the presence and absence of PGE,. As shown in panel 13, the addition of PGEi to homogenates similarly results in a lo-fold incrcasc in adenylate cyclase activity (5- to la-fold increase in other experiments). Morphine inhibits basal and PGEidependent adenylate cyclase activity about 45yo (3@- 60% inhibition in other experiments). The rate of CAMP formation was linear for about 8 min; therefore, in all sub- sequent experiments reaction mixtures were incubated for 5 min. The effect of morphine upon adenylate cyclase activity was verified in the following ways: (i) the [arP]product of the re- action was shown to have the chromatographic mobility of authentic CAMP in five systems (see Jfateriala and Melhods) ; (ii) the concentration of substrate, ATP, did not change appreciably during incubation by direct measurement with the luciferase assay (12) ; and. (iii) the specific activity of ATP changed less than 5% during the incubation. The effect of naloxone, an antagonist that displaces mor- phine from the opiate receptor but has no narcotic activity, is shown in Table 1. Naloxone reverses the inhibitory effect of morphine upon adenylate cyclase activity both in the prcscnce and absence of PGE,, but naloxone alone has little or no effect upon basal or PGE,-stimulated adcnylate cyclase activity. The relation between molarity of PGEi and the specific activity of adenylate cyclase is shown in Fig. 2. Adenylate cyclase is stimulated by concentrations of PGEI greater than 10e8 M. A discontinuity in the concentration curve is observed at 5 x lo-' M PGEI. The effects of different concentrations of naloxone, in the presence and absence of morphine, on adenylate cyclase are shown in Fig. 3. The concentration of morphine was lo+ M, which inhibits adenylate cyclase activity maximally. Naloxone completely reverses the inhibitory effect of morphine on adenylate cyclase activity at 2 X 10ee5 M; naloxone at higher concentrations stimulates adenylate cyclase activity 15-20~o. Stimulation has not been observed in all experiments, and further experiments are needed to clarify this observation. In Fig. 4 the decrease in binding of [`Hlnaloxonc to the narcotic receptor, as a function of narcotic concentration, is compared with the effectiveness of the narcotics as inhibitors of adenylate cyclase activity. Naloxone, a pure antagonist, binds with high affinity to the opiate receptor but does not inhibit adenylate cyclsse activity (panel A). Etorphine, levorphanol, morphine, and 3-allylprodine, each a potent narcotic, inhibit adcnylate cyclasc at essentially the same TABLE 1. Reversal of morphine inhibition of adenylate cyclase by naloxone in cellfee preparations of NGI OS-16 Addition* None Morphine Naloxone Morphine + naloxone PGE, PGE, + morphine PGEl + naloxone PGEl + morphine + naloxone Membrane2 Homogenate+ fraction (pmol of CAMP formed/ min per mg of protein) 17 24 8 17 17 22 17 23 113 124 68 101 110 100 122 * 10 @I, each component. I Unfractionated homogenate. 1 The homogenate wss centrifuged at 1500 X 0 for 10 min and the supernatant fraction at 17,000 X g for 30 min. The 17,000 X Q pellet W&S washed and resuspended in the homogenizing me- dium (membrane fraction). 592 Biochemistry: Sharma el al. Proc. Nat. Acad. Sci. USA 72 (1976) FIG. 2. Activation of adenylate cyclase in homogenates of NGlOS-15 by PGE, as a function of PGEl concentration. concentrations as those needed to displace [SH]naloxone from the opiate receptor (panels 13, C, E, and F). In contrast to levorphanol, the analgesically inactive enantiomer dex- trorphan is without effect on adenylate cyclase activity and has a much lower affinity for the opiate receptor (panel D). Apparent dissociation constants for [narcotic. opiate re- ceptor] complexes are shown in Table 2, along with apparent K, values for adenylate cyclase calculated from the data of Fig. 4. A striking similarity is observed between the dissocia- tion constants and the concentrations of narcotics required for 50% inhibition of adenylate cyclase activity. Narcotic binding and enzyme inhibition curves shown in Fig. 4 are not superimposable, even though the narcotic concentrations for half-maximal inhibition of binding and enzyme activity agree well. With each of the four narcotics, enzyme activity changes over a relatively small range of narcotic concentration compared to narcotic binding. This relationship is seen more clearly in Fig. 5A and B, which shows Hill plots of narcotic binding and adenylate cyclase inhibition, respectively. Each slope for narcotic binding is approximately 1, whereas the maximum slopes for adenylate cyclase inhibi- tion are 2.3. These results suggest that the binding of nar- cotics to receptors is not a cooperative process, w!jereas the effects of narcotics on adenylate cyclase activity exhibit a cooperative character. The relation between the molarity of Gpp(NH)p, a rela.- tively stable analog of GTP, and adenylate cyclase activity, in the presence and absence of morphine, is shown in Fig. 6. TABLE 2. Comparison of narcotic aJ?inity for the opiate receptor and ability to inhibit adenylate cydase Narcotic K2 Kit Narcotic Adenylate receptor cyclase nM nM Etorphine 5 10 Levorphanol 200 200 Morphine 4,000 2,000 3-Allylprodine 10,000 50,000 Dextrorphan 10,000 - Naloxone 20 - * Apparent dissociation constant of the narcotic-receptor com- plex. t Narcotic concentrat.ion required for 50V0 of maximal in- hibition of adenylate cyclase. -, no inhibition. Fig. 3. The effect of naloxone upon adenylate cyclase activity of an NGlOS-15 homogenate in the presence and absence of 10 pM morphine. The GTP analog activates adenylate cyclase of NGlOS-15 homogenates, similar to the effects reported with other systems (13). Complete reversal of morphine inhibition of adenylate cyclase was observed in the presence of 1 X lo-' M 500 t '. MXTRWPHW b.. 400 --.: 0 F .I0 700 600 500 34LL"LPRclnNE 400 The effectiveness of various narcotics ss inhibitors of adenylate cyclase activity of NGlOS-15 homogenates is com- pared with their ability to displace [`Hlnaloxone from the opiate receptor. Symbols represent the following: 0, [aH]naloxone bound to receptor; A, [a*P]cA>IP formed per 5 min per tube (250 rrg of protein per tube). Proc. Nat. Acad. Sci. USA 72 (1976) Morphine Receptors and Adeny1at.c Cyclase 593 L- ---A 10.' lo-' 10" lo- 10-6 lci' N4RcoTc k4oLmTf FIN. 5. (A) kill plots of narcotic binding to receptor; and (B) inhibition of adenylate cyclsae activity by narcotics. The sym- bols represent the following: 0, etorphine; * , naloxone; 0, levor- phanol; A, morphine: and A, 3-allylprodine. In panel A the symbol n in the ordinate represents the maximum amount of [zH]- naloxone displaced at saturating concentrations of unlabeled narcotic and Y is the amount of [`Hlnaloxone displaced at non- saturating concentrations of narcotic. In panel B, n represents the maximum inhibition of adenylate cyclaae observed at satu- rating narcotic concentrations and Y is the inhibition observed at each nonsaturating concentration of narcotic. Gpp(NH)p. In the absence of morphine, <1 X 1O-6 M Gpp- (NH)p has little effect upon adenylate cyclase activity. Similar results were obtained with partially purified mem- brane preparations (not shown). The effects of morphine upon adenylate cyclase activity of intact neuroblastoma N18TG-2 and glioma @UU-1 cells, the parents of hybrid NGlOB-i5, are shown in Fig. 7A and B, respectively. Morphine reduces CAMP levels of neuro- blastoma N18TG-2 cells slightly but does not affect CAMP levels of glioma C6BU-1 cells. The effects of morphine and naloxone upon adenylate cyclase activity of homogenates prepared from parent neuroblastoma N18TG-2 and glioma C6BU-1 cells are shown in Table 3. Morphine inhibits adenylate cyclase activity of N18TG-2 slightly, and the inhibition is reversed by naloxone; however, these compounds do not affect adenylate cyclase activity of CGBU-I. Thus, the neuroblastoma parent, which has few morphine receptors, a is slightly sensitive to morphine whereas the glioma parent, which lacks these receptors, is insensitive to morphine. DISCUSSION shows no specific binding of either naloxone or di.hydromorphine. The results show that morphine reduces basal and PGE1- stimulated CAMP levels of intact neuroblastoma X glioma hybrid cells and inhibits adenylate cyclase activity of homoge- nates. The effects of morphine are blocked by the narcotic antagonist, naloxone, which competes with narcotics for sites on the opiate receptor. The' inhibition of adenylate cyclase activity by narcotics is specific for the levorotatory, pharma- cologically active, stereoisomer. The relative affinities of narcotics for the opiate receptor agree well with their effective- ness as inhibitors of adenylate cyclase and with their pharma- cologic potency. Morphine-sensitive and -insensitive cell lines g Neuroblastoma clone N18TG-2 can be shown to have a rela- tively small number of morphine receptors when binding experi- ments are performed with ["Hjnaloxone in place of the [`HI- dihydromorphine used previously (1). Glioma clone CGBU-1 TABLE 3. Fffec,t of morphine on adenylate cyclase activity of neuroblastoma and gltima parents Homogenate Neuro- blastoma Glioma N18TG-2 CGBTJ-1 Addition* (pmol of CAMP formed/ min per mg of protein) None 6 20 Morphine 4 19 Naloxone 5 20 Morphine + naloxone 5 20 PGE, 75 24 PGE, + morphine 63 25 PGE, + nalox.one 72 24 PGEi + morphine + naloxone 70 23 * 10 PM, each component. were found, and the degree of sensitivity to morphine is dependent upoh the abundance of narcotic receptors. Thus, two kinds of adenylate cyclase complex were de- tected, one sensitive, the other insensitive, to narcotics. Sensitivity to narcotics is dependent upon functional inter- action of the [morphine.reccptor] complex with adenylate cyclase. The partial inhibition of adenylate cyclase observed in the presence of saturating concentrations of narcotics may be due, at least in part, to the presence of narcotic-sensitive and -insensitive adenylate cyclase molecules within cells and, possibly, heterogeneity in the cell population with respect to the number of narcotic receptors per cell. Interactions that couple morphine receptors with adenylate cyclase exhibit positive cooperativity ; however, the inter- action of narcotic with receptor is not a cooperative process. The Hill coefficient of 2.3 suggests that multiple interactions between [narcotic.receptor] and the adenylate cyclase com- plex may be required for inhibition. Alternate mechanisms for apparent cooperativity may be envisaged involving threshold effects and spare receptors. The cooperative activation of adenylate cyclase has been reported for thyroid stimulating hormone (14), but glucagon (15) and oxytociu (16) do not exhibit cooperativity. The available information suggests that one molecule of PLUS MORPHINE FIG. 6. Effect of Gpp(NH)p on adenylate cyclase activity of an NGlOS-15 homogenate, in the presence and absence of 10 anI mornhine. 594 Biochemistry: Sharma el al. Proc. Nut. Acad. Sci. USA 78 (1975) MlN"TES FIQ. 7. Cyclic AMP accumulation, as a function of time, in intact parental cells. Cells used were neuroblsstoma N18TC-2 in panel A and glioma COBU-1 in panel B. Concentrations of mor- phine and PGE, were 10 PM. adenylate cyclase can respond to different kinds of hormones, depending upon the species of receptors present, but that only one kind of hormone is translated at a time. Usually, hor- mones activate adenylate cyclase but enzyme inhibition mediated by some a- or prostaglandin receptors has been reported (17, 18). Morphine thus resembles an inhibitory hormone. In essence, three kinds of mechanism have been proposed for the regulation of adenylate cyclase activity by hormones and other effector molecules: (i) Hormones and GTP are allosteric effecters that activate the enzyme by reducing its affinity for free ATP, a competitive inhibitor of the substrate [ATP.lMg] (19, 20). (ii) Mg++ or Ca++ bind to other sites and thereby regulate enzyme activity (21, 22). (iii) Adenylate cyclase is inactivated by phosphorylation, catalyzed by a protein kinase, and activated by dephosphorylation, catalyzed by a phosphatase (23). Reactions that terminate reading of messages by adenylate cyclase regulate the duration of response of the enzyme to each message. GTP may be involved both in initiation and termination of reading of hormonal messages by adenylate cyclase-in ways that may be analogous to the roles of GTP in aminoacyl-tRNh binding to ribosomes and translocation reactions in protein synthesis. The GTP analog, Gpp(NH)p, activates adenylate cyclase, and morphine reduces the threshold for activation lo-fold. Thus morphine and Gpp(NH)p act in a concerted fashion, which suggests that both are physically coupled to the adeny- late cyclase complex. The K, for glucagon stimulation of adenylate cyclase is reduced lo-fold by GTP (24). The possi- bility that GTP is required for morphine action, as it is for hormone action, should be considered. Collier and Roy have presented cogent arguments support- ing the thesis that PGE-stimulated adenylate cyclase is the primary site of morphine action (5, 6). Our results support their view, although we find that morphine inhibits adenylate cyclase both in the presence and absence of added PGEi. The inhibition of adenylate cyclase should have many secon- dary consequences which might account for the varied pharma- cologic effects of morphine. We wish to propose the following hypothesis concerning the molecular basis for narcotic addiction and tolerance. Inhibition of adenylate cyclase by morphine reduces intra- cellular levels of CAMP. This may lead to a compensatory shift in enzyme synthesis, degradation, or activity which restores the normal level of CAMP. The cell then is dependent upon narcotic because the level of &Ml' is normal in the presence of the drug and abnormally high upon withdrawal. Our working hypothesis is that the number of adenylate cyclase molecules increases as cells become addicted to nar- cotics. This also leads to tolerance since, at a given narcotic concentration, the amount of uninhibited enzyme is greater in addicted than in normal cells. We thank Doyle Mullinex for growing cells and Richard Streaty for his help. 1. 2. 3. 4. 2. 6. 7. 8. 9. 10. Il. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. Klee, W. A. & Nirenberg, M. (1974) Proc. Nat. Acad. Sci USA 71, 3474-3477. Pert, C. B. C Snyder, S. H. (1973) Science 179, 1011-1014. Terenius, L. (1973) Acta Pharmacoi. l'ozicol. 32, 317-320. Simon, E. J., Hiller, J. M. & Edelman, I. (1973) Proc. Nat. Acad. Sci. USA 70, 1947-1949. Collier, H. 0. J. & Roy, A. C. (1974) Mature 248, 24-27. Collier, H. 0. J. I% Roy, A. C. (1974) Prostuglandins 7, 361- 376. Minna, J., Glazer, U. & Nirenberg, M. (1972) ~Vature New Biol. 235, 223-231. Amano, T., Hamprecht, B. & Kemper, W. (1974) Ezp. Cell Res. 85, 399-408. Benda, P., Lightbody, J., Sato, G., Levine, L. & Sweet., W. (1968) Science 161, 370-371. Cilmah, A. G. (1970) Proc. Nut. Acad. Sci. USA 67, 305- 312. Salomon, Y., Londos, C. & Rodbell, M. (1974) Anal. Uio- them. 58, 5417348. Klofat, W., Picciolo, G., Chappelle, 1~. W. & Freese, E. (1969) J. Biol. C/tern. 244, 3270-3276. Londos, C., Salomon, Y., Lin, M. C., Harwoocl, J. P., Schramm, M., Wolff, J. & Bodbell, M. (1974) Proc. Nat. Acad. Sci. USA 71, 30873090. Pochet, It., Boeynaems, J. M. & Dumont, J. 15. (1974) Biochem. Biophys. Res. Commun. 58, 446-453. Pohl, 6. L., Birnbaumer, L. & Rodbell, M. (1971) J. Biol. Chem. 246, 1849-1856. Bockaert,, J., Roy, C. & Jard, 6. (1972) J. Biol. Chem. 247, 7073-7081. Robison, G. A., Butcher, 12. W. & Sutherland, E. W. (1971) Cyclic AMP (Academic Press, N. Y.), pp. 145-231. Kantor, H. S., Tao, P. & Kiefer, II. C. (1974) Proc. Nat. Acad. Sci. USA 71, 1317-1321. deHaen, C. (1974) J. BioE. Chem. 249, 2766-2762. Rendell, M., Lin, M., Salomon, Y., Rodbell, M. & Berman, M. (1974) Abstracts, Second International Conference on Cyclic AMP (University of British Columbia, Vancouver, B. C.), p. 38. Birnbaumer, L., Pohl, S. L. & Rodbell, M. (1969) J. Biol. Chem. 244, 3468-3476. Severson. L). L.. Drummond. G. I. & Sulakhe P. V. (1972) J. Bid. khem. i47, 2949-29%. Constantopoulos, A. & Najjar, V. A. (1973) Biochcm. Bio- phys. Rcs. Commun. 53, 794-799. Rodbell, M., Lin, M. C. & Salomon, Y. (1974) J. Biol. Chem. 249, ;39-65.