Proc. Natl. Acad. Sci. USA Vol. 74, No. 12, pp. 5482-5486, December 1977


Two distinct adenosine-sensitive sites on adenylate cyclase (purine moiety/ribose moiety/Leydig membranes/liver membranes/platelets)

CONSTANTINE LONDOS AND J. WOLFF National Institute of Arthritis, Metabolism and Digestive Diseases, National Institutes of Health, Bethesda, Maryland 20014

Communicated by Elizabeth F. Neufeld, October 5, 1977

ABSTRACT The effects of adenosine and adenosine analogs on adenylate cyclases from several tissues have been examined. Two adenosine-reactive sites have been identified: (i) the "R" site, occupancy of which usually leads to activation of cyclase and which requires integrity of the ribose ring for activity, and (ii) the "F" site, which mediates inhibition and requires integrity of the purine ring for activity. Biphasic effects of adenosine are explained by the presence of both sites on a single adenylate cyclase. Comparison of these data with those in the literature indicates that adenosine-reactive "P" and "R" sites are present generally.

Diverse physiological effects of adenosine have been known for a long time. Effects of adenosine on vasomotor and other autonomic functions have led to the concept of purinergic nerves (1). Moreover, adenosine is released from tissues in response to various stimuli (2, 3). A possible unifying mechanism for the effects of adenosine was suggested-by the studies of Sattin and Rall (4) who showed that adenosine led to augmentation of cyclic AMP (cAMP) levels in brain slices. Subsequently, isolated neuronal and other cells have been shown to respond similarly to adenosine (5-11). In a number of tissues it has been possible to demonstrate direct stimulation of adenylate cyclase by adenosine in isolated membrane preparations (12-17). Paradoxically, adenosine has also been shown to inhibit cAMP accumulation and adenylate cyclase activity in certain tissues (2, 18-23) as well as to cause biphasic stimulatory and inhibitory effects in others (discussed below). Such findings suggest that the stimulatory and inhibitory effects of adenosine may be exerted through different sites. In the present report we demonstrate that this is indeed the case because the two effects of adenosine differ in their chemical specificity, sensitivity to incubation conditions, and distribution in various cells. MATERIALS AND METHODS The sources for materials used in the assay of adenylate cyclase activity have been reported (24). Formicin, 2-methyladenosine, 2-aminoadenosine, 9-f-D-arabinofuranosyl adenine, 7-deazaadenosine, 8-azaadenosine, and 8-bromoadenosine were gifts from the International Chemical and Nuclear Corp. 9-3-LRibofuranosyl adenine, 2'5'-dideoxyadenosine, and 9-f-Dxylofuranosyl adenine were provided by the Drug Research and Development Branch of the National Cancer Institute; N6phenylisopropyladenosine was a gift from John Fain, Brown University. N6-Methyladenosine, 2'-deoxyadenosine, and 5'deoxyadenosine were purchased from Pabst Laboratories. Cordycepin, 9-f-D-ribofuranosyl purine, and 2',3'-isopropylidene adenosine were purchased from the Sigma Chemical

Preparation of Membranes. Rat liver plasma membranes were prepared as described (25). Membranes from I-10 Ley-dig tumor cells, from Y-1 mouse adrenal tumors, and from beef thyroid were prepared as described (17, 26). Human platelets were obtained from the National Institutes of Health Blood Bank; platelet membranes were prepared by freezing and thawing, as described by Haslam and Lynham (12). Adenylate Cyclase Assays. Adenylate cyclase activity was determined by the production of [P~cAMP formed from [a32P]ATP; in all cases [P]cAMP was isolated and purified by the method of Salomon et al. (24). Liver membranes were incubated under conditions shown previously to enhance the inhibitory effects of adenosine (23). The assay medium contained 10 ,uM ATP, 2 MCi of [a-32P]ATP, 4 mM MgCl2, 1 mM MnC12, 50 ,uM cAMP, 1 mM dithiothreitol, 2 mM creatine phosphate, creatine phosphokinase at 10 units/ml, 10 ,uM 5'-guanylylimidodiphosphate, 1 AM glucagon, and 25 mM Tris-HCI, pH 7.5. The reaction was initiated with approximately 10,ug of liver membrane protein to give a total volume of 0.1 ml and was stopped after 4 min at 300. Leydig cell and Y-1 adrenal membranes were incubated for 10 min at 370 in medium containing 25 mM Tris-HCI (pH 7.6), 1.5 mM MgCl2, 0.1% crystalline bovine serum albumin, creatine phosphokinase at 30 units/ml, 6.7 mM creatine phosphate, 30 ,uM 5'-guanylylimidodiphosphate, 2 ,ACi of [a-32P]ATP, and 1.0 mM ATP for 1-10 and 0.1 mM for Y-1 membranes. In the case of adrenal membranes, 0.1 ,AM corticotropin (1-24) was present. Beef thyroid membranes were assayed similarly except that 2 mM MgCl2 and 0.1 mM ATP were used. Platelet adenylate cyclase activity was assayed for 10 min in 0.1 ml of medium containing 1 mM MgCl2, 50,M ATP, 2 ,uCi of [a-32P]ATP, 10 MM GTP, 2 mM creatine phosphate, and 25 mM TrisIHCl, pH 7.5. Typically, 25 ,g of platelet membrane protein was added to start the reaction. ATP concentrations used with the different membranes were adjusted to optimize detection of the adenosine effects and to minimize interference from breakdown of the substrate to adenosine. RESULTS AND DISCUSSION Adenosine and its analogs have opposing effects on the adenylate cyclase activity of membranes from liver and Leydig cells: the liver enzyme is inhibited, whereas the Leydig cell enzyme is activated. Table 1 shows the half-maximal concentrations required to elicit these effects. Except for 5'-deoxyadenosine, all of the compounds listed as activators produced maximal activity equivalent to that seen with adenosine; that is, all were full agonists. None of the analogs tested increased activity of the hepatic enzyme and none was inhibitory toward the Leydig cell enzyme. On the basis of ability to react with different adenosine analogs, the site through which adenosine inhibits the hepatic system differs markedly from the site through which the nucleoside activates the Leydig cell enzyme.

Co. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact.

Abbreviation: cAMP, cyclic AMP.


Biochemistry: Londos and Wolff Table 1. Effects of adenosine and adenosine analogs on adenylate cyclases from liver and I-10 Leydig cells Leydig cell, Liver, Test compound Ka(app), AM Ki(app)q AM 7 24 Adenosine Class A 30 NI 2-Methyladenosine 35 NI 2-Aminoadenosine 40 NI N6-Methyladenosine 40 NI N6-Phenylisopropyladenosine 35 NI 9-fl-D-Ribofuranosylpurine 200 NI Formicin Class B NA 0.7 2',5'-Dideoxyadenosine 700 70 2'-Deoxyadenosine NA 200 2',3'-Isopropylidene adenosine NA 25 9-fl-D-Arabinofuranosyladenine 800 25 9-fl-D-Xylofuranosyl adenine Class C 8 240 2-Chloroadenosine 30 250 3'-Deoxyadenosine 30 6 5'-Deoxyadenosine Class D NA NI 7-Deazaadenosine NA NI 8-Azaadenosine NA NI 8-Bromoadenosine NA NI 9-f3-L-Ribofuranosyl adenine NA NI Adenine, inosine, guanosine

Ka(app), concentration required for half-maximal activation; NA indicates that no activation was observed. Ki(app), concentration required to inhibit activity by 50%; NI indicates that no inhibition was observed at the highest concentration tested, which was at least 100 MM. Adenylate cyclase activity in Leydig cell membranes varied from 0.3 to 1.2 nmol of cAMP per mg of protein per 10 min in the absence of test compounds; all active analogs increased activity by approximately 3- to 5-fold with the exception of 5'-deoxyadenosine which achieved 60-70% of the maximal adenosine-mediated stimulation.

Proc. Natl. Acad. Sci. USA 74 (1977)


mm Mg2+

150 V 140 F 130


X 120




u C)

> 110






1..I -1



.uM FIG. 1. Effects of varying magnesium concentrations on the response to adenosine by platelet adenylate cyclase. Activity is expressed as the % of activity in the absence of adenosine at each of the three magnesium concentrations tested. These activities were 13, 39, and 62 pmol of cAMP per mg of protein per 10 min for 0.5, 2, and 5 mM Mg2 respectively. With either 1 mM Mn2+ and no Mg2+ or with 10 M prostaglandin E1 and 1 mM Mg2+, the only effect of adenosine was inhibition; under both conditions 30,MM adenosine produced 50% Adenosine,



These differences have permitted a rough division of the analogs into four classes. One group of compounds (class A) were nearly equipotent with adenosine as activators of the Leydig enzyme yet were inactive on the liver system. The class A compounds differ from the native nucleoside by virtue of alterations in the purine ring. Another group of analogs (class B) differ from adenosine in the ribose ring. These compounds were generally as potent as adenosine in inhibiting the hepatic enzyme yet were either inactive or very weakly active on the Leydig cell adenylate cyclase. Note, in particular, that 2',5'-dideoxyadenosine was far more effective on the liver enzyme than adenosine but was inactive on the Leydig cell membranes. Two other groups of analogs did not distinguish between the two enzyme systems. One group, class C, consists of those analogs that were active on both systems. Finally, those compounds that-had no effect on either enzyme are designated class D; the analogs altered in the "imidazole" portion of the purine ring fall into this category. Several adenylate cyclase systems exhibit biphasic responses to adenosine, with activation occurring at low concentrations and inhibition at higher concentrations of the nucleoside (12, 14, 15, 27-29). It seemed reasonable to ask whether the biphasic effects of these enzymes are mediated through different sites and whether the specificity of the sites mediating the opposite effects would conform to the specificities of the apparently different sites in the liver and Leydig cell adenylate cyclase

systems. To test the above, we selected-the platelet system (12) which is both stimulated and inhibited by adenosine and 2chloroadenosine. The data in Fig. 1 show that the effects of adenosine on the platelet enzyme are highly dependent on the concentration of Mg2+ in the assay medium. At 0.5 mM Mg2+, activation was readily evident with 0.4 AM adenosine whereas inhibition was rather weak at 10 1uM. When the concentration of Mg2+ was increased, activation by adenosine was nearly abolished whereas inhibition became more apparent. In order to permit detection of both activation and inhibition by adenosine, further experiments with the platelet system were performed in the presence of 1 mM Mg2+. The biphasic effect of adenosine is-obvious in Table 2, with stimulation occurring at submicromolar concentrations of adenosine and inhibition (relative to the peak activity) evident at 10-40 AM adenosine. Class A compounds stimulated the platelet enzyme; none was inhibitory even when tested at very high concentrations. The stimulatory effects. of the class A compounds were abolished with theophylline (data not shown). Class B compounds (inhibitory to liver, inactive or -very weakly .active on Leydig cell cyclase) were only inhibitory on the platelet enzyme. Class C compounds were for the most part inhibitory except for 2-chloroadenosine which yielded the biphasic response pattern. These data thus provide strong evidence that the biphasic effect of adenosine on the platelet cyclase system is mediated by different sites: (i) a site for activation


Biochemistry: Londos and Wolff

Proc. Natl. Acad. Sci. USA 74 (1977)

Table 2. Effects of adenosine and adenosine analogs on platelet

adenylate cyclase activity Conc., Test compound None (n






Activity (mean ± SD)


71.2 + 1.0 82.1 + 5.2 80.6 ± 5.5 77.4 ± 0.2 59.0 : 0.3 45.8 ± 0.8

10 50 1000

80.8 + 3.2 88.8 I 5.3 89.8 + 2.5

12 120

81.4 ± 0.4 89.9+0.9


2.5 50 250 1000

65.5 + 2.8 83.8 ± 2.3 95.9 + 2.1 104 ± 4.1


15 150 1500

86.1 93.9 89.4

0.1 i


10 40

Class A


9-B-D-Ribofuranosyl purine

2.5 3.4 3.5

1 10 100

65.0 2.7 71.3 3.4 85.6 A: 2.8

5 50 250

59.8 ± 4.6 51.1±0.3 34.0 + 1.3


2.5 25 250

38.0 ± 1.9 17.5 1.7 7.2 +0.6


10 100

44.5 25.3


2.5 25 250


Class B 2'-Deoxyadenosine

2',3'-Isopropylidene adenosine

Class C 2-Chloroadenosine


Test compound 5'-Deoxyadenosine

60.1 + 3.0



Table 2. Continued

10 100 1000

16.6 0.2 8.0 0.3 59.7 2.3 57.0 ± 2.6 54.3 ± 1.3










10 100

2.9 2.6

+ 2.7 + 2.1 + 7.0 + 2.4

47.2 ± 2.4 25.1 + 0.2 12.6±0.1

Class D 9-fl-L-Ribofuranosyladenine


Activity (mean ± SD)

10 100 1000

47.5 ± 2.9 33.6 ± 0.8 16.9 ± 0.4

6 60

61.3 ± 0.8 57.9 0.8

Platelet adenylate cyclase activity (prnol/mg of protein per 10 min) was assayed in the presence of 1 mM Mg2+. Activities in the presence of test compounds are the means of duplicate determinations. Compounds are designated class A, B, C, and D according to their effects on liver and Leydig cell adenylate cyclase (see Table 1). In separate experiments, 2-chloroadenosine at concentrations higher than 50 IAM was inhibitory.

that strongly resembles the site through which adenosine activated the Leydig cell enzyme; and (ii) a site that closely resembles the site through which adenosine inhibited the liver adenylate cyclase system. The foregoing permit a rough classification into two adenosine sites: (i) those that have strict structural requirements in the purine moiety (designated "P" sites), and (ii) those, that have strict structural requirements in the ribose moiety (designated "R" sites). Because the liver enzyme does not tolerate alterations in the purine moiety of adenosine, this system contains the "P" type of adenosine site. By contrast, the Leydig cell cyclase system contains an adenosine site that is sensitive to changes in the ribose moiety and can thus be designated as representing an "R" type of adenosine site; modifications in the 2 and N6 positions are well tolerated. A salient feature of the "R" site is the requirement for the 2'-hydroxyl group on the ribose ring. Omission of this hydroxyl (as in 2'-deoxyadenosine), a change in configuration (9-3-D-arabinofuranosyladenine, 9-f3-D-xylofuranosyladenine), and chemical modification (2',3'-isopropylidene adenosine) all result in a marked loss of activity at the "R" site. On the other hand, for the "P" site the 2'-hydroxyl is clearly not critical. A particular feature of the "P" site is the increased potency, with respect to adenosine, of compounds lacking the 5'-hydroxyl group. The great potency of 2',5'-dideoxyadenosine in inhibiting adenylate cyclase systems has been observed by others (19, 22, 23). It has recently been suggested (30) that analogs of adenosine that stimulate adenylate cyclase exist in the glycosidic high anti conformation, whereas the 9-(3-D-arabinofuranosyladenine and 2'-deoxyadenine are not stable in this conformation. Although not enough compounds have been analyzed for their conformation, the data thus far are consistent with a "high anti" requitement for the "R" site and an anti or syn conformation for the "P"' site. This possibility is being investigated currently. The classification of adenosine-responsive "PI' and "R" sites can be extended to other tissues. Huang et al. (6) have shown in brain slices that there is an "R"-type site at which purinesubstituted adenosine derivatives stimulate cAMP accumulation, whereas analogs altered in the ribose ring have been described an "antagonists" of adenosine-, histamine-, and veratridine-stimulated cAMP accumulation. We examined the effects of class B compounds on the cyclase in a 20,000 X g fraction of rat brain homogenates and found that they inhibit activity with a spectrum of apparent affinities that closely parallels that seen with the hepatic cyclase. That is, 2',5'-dideoxyadenosine, 5'-deoxyadenosine, and the arabinose and xylose analogs were as, or more, potent than adenosine, whereas 2'-deoxy-


Londos and Wolff

Table 3. Effect of adenosine analogs on Y-1 adrenal cell membrane adenylate.cyclase Stimulation Adenosine 2-Chloroadenosine N6-Methyladenosine 2-Methyladenosine 2-Aminoadenosine

Inhibition Adenosine 2-Chloroadenosine

2'-Deoxyadenosine 2',5'-Dideoxyadenosine 9-0--Xylofuranosyl adenine Membranes were assayed with 0.1 mM ATP. The basal activity was -0.05 nmol/mg per 10 min and with 0.1 gM corticotropin (1-24) it was 0.21 nmol/mg per 10 min. Adenosine and stimulatory analogs yielded maximal rates 2-3.5 times basal activity. Sensitivity to inhibitory effects was increased in the presence of corticotropin (1-29) and


adenosine-, 2-chloroadenosine, and 2',3'-isopropylidene adenosine were less potent than adenosine. Stimulation by class A compounds (13, 29, 31) and inhibition by class B analogs has also been observed in neuroblastoma and striatal adenylate cyclase preparations. Thus, it seems probable that brain cyclase contains both "P" and "R" sites, and that at least some of the "antagonistic" effects of certain analogs on brain slices are mediated by the "P" site. Previous studies (17) have shown that adenosine and 2chloroadenosine stimulate adenylate cyclase activity in membranes prepared from Y-1 mouse adrenal tumor cells maintained in culture. We have now found that class A compounds stimulate and class B compounds inhibit adenylate cyclase in membrane preparations from these cells (Table 3). As with the platelet enzyme, adenosine and 2-chloroadenosine produced a biphasic effect. In published studies on lung (21), intestinal epithelial (22), and adipocyte (19) cyclase systems in membrane preparations, in which adenosine inhibits activity, class A compounds were generally ineffective whereas analogs containing modifications of the ribose ring were inhibitory. Finally we examined the effect of adenosine on the thyrotropin-sensitive adenylate cyclase from thyroid; the nucleoside was inhibitory and the inhibitory potency was enhanced by the addition of 0.2 mM Mn2+ to the assay medium (see below). Class A compounds had no effects on the thyroid enzyme, but class B compounds were active and the order of potency for the class B compounds resembled that seen with liver and brain. It seems clear, therefore, that the division of adenosine reactive sites into "P' and "R" types is widely applicable. The "P" and "R" sites identified in this report may be compared to adenosine "receptors" in other tissues by characteristics other than their reactivity with the various analogs. A feature shared by all adenylate cyclases that are activated by adenosine ("R" sites) is that theophylline antagonizes the effects of the nucleoside and its active analogs (4, 7, 12-15, 17, 28, 29). In general, the potency of adenosine in activation is not affected by the Mg2+ concentration in the assay medium. On the other hand, the inhibitory potency of the nucleoside on the lung and hepatic (2, 23) enzymes is quite sensitive to the concentration of divalent cation, particularly Mn2 , in the assay medium; inhibitory potency increases with increasing cation concentration. A Mg2+-dependent shift in adenosine potency at the "P" site in the platelet enzyme would explain our finding (Fig. 1) that low Mg2+ concentrations were required in order to detect adenosine activation of that enzyme. Thus, both theophylline and cation effects provide parallels between the "P"

Proc. Natl. Acad. Sci. USA 74 (1977)


and "R" sites we describe and the adenosine-reactive sites reported for several other tissues. The location of the "P" and "R" sites has not been established, but indirect evidence has suggested that the "R" site has the properties of an extracellular membrane receptor (8, 17, 32, 33). Although occupancy of this receptor usually leads to stimulation of adenylate cyclase, there is evidence for "R" site-mediated inhibition of cAMP accumulation in intact fat cells (19, 34, 35). Moreover, several class A compounds have been found to inhibit adenylate cyclase in membrane preparations from bone cells (S. Rodan, G. Rodan, and J. J. Egan, personal communication). On the other hand, occupancy of the "P" site appears to result in lowered cyclase activity in all cells studied. Preliminary studies indicate that the "P" site may be intracellular because detergent-solubilized adenylate cyclase from liver and thyroid (unpublished data) membranes are inhibited by class B compounds in the same order of potency as when these enzyme reside in the membrane. The information in this report should facilitate the effort to determine the locations of the adenosine-reactive sites and their possible role in mediating the biologic effects of adenosine. 1. Burnstock, G. (1972) Pharmacol. Rev. 24,509-581. 2. Dobson, J. G., Rubio, R. & Berne, R. M. (1971) Circ. Res. 29, 375-384. 3. Schwabe, V., Ebert, R. & Ebler, H. C. (1975) Adv. Cyclic Nucleotide Res. 5, 569-584. 4. Sattin, A. & Rail, T. W. (1970) Mol. Pharmacol. 6, 13-24. 5. Mills, D. C. B. & Smith, J. B. (1971) Biochem. J. 121, 185196. 6. Huang, M., Shimizu, H. & Daly, J. W. (1972) J. Med. Chem. 15, 462-466. 7. Blume, A. J., Dalton, C. & Sheppard, H. (1973) Proc. Nati. Acad. Sci. USA 70,3099-3102. 8. Clark, R. B., Gross, R., Su, Y. F. & Perkins, J. R. (1974) J. Biol. Chem. 249,5296-5303. 9. Wolberg, G., Zimmerman, T. P., Hiemstra, K., Winston, M. & Chu, L.-C. (1975) Science 187,957-959. 10. Sturgill, T. W., Schrier, B. K. & Gilman, A. G. (1975) J. Cyclic Nucleotide Res. 1, 21-30. 11. Peck, W. A., Carpenter, S. & Messinger, K. (1974) Endocrinology 94, 148-154. 12. Haslam, R. J. & Lynham, J. A. (1972) Life Sci. 11, 1143-1154. 13. Blume, A. J. & Foster, C. T. (1975) J. Biol. Cherpi. 250, 50035008. 14. Clark, R. B. & Seney, M. N. (1976) J. Biol. Chem. 251, 42394246. 15. Peck, W. A., Carpenter, J. G. & Schuster, R. J. (1976) Endocrinology 99,901-909. 16. Blume, A. J. & Foster, C. J. (1976) J. Biol. Chem. 251, 33993404. 17. Wolff, J. & Cook, G. H. (1977) J. Biol. Chem. 252, 687-693. 18. Moriwaki, K. & Foa, P. P. (1970) Experientia 26,22. 19. Fain, J. N., Pointer, R. H. & Ward, W. F. -(1972) J. Biol. Chem. 247,6866-6872. 20. McKenzie, S. G. & Bar, H. P. (1973) Can. J. Physiol. Pharmacol. 51, 190-196. 21. Weinryb, I. & Michel, I. M. (1974) Biochim. Biophys. Acta 334, 218-225. 22. Zenser, T. V. (1976) Proc. Soc. Exp. Biol. Med. 152, 126-129. 23. Londos, C. & Preston, M. S. (1977) J. Biol. Chem. 252, 59515956. 24. Salomon, Y., Londos, C. & Rodbell, M. (1974) Anal. Biochem.

58,541-548. 25. Pohl, S. C., Birnbaumer, L. & Rodbell, M. (1971) J. Biol. Chem. 246, 1849-1856. 26. Wolff, J. & Cook, G. H. (1975) J. Biol. Chem. 250, 68976903.


Biochemistry: Londos and Wolff

27. Birnbaumer, L., Nakahara, T. & Yang, P.-C. (1975) J. Biol. Chem. 249,7857-7866. 28. Penit, J., Hout, J. & Jard, S. (1976) J. Neurochem. 26, 265273. 29. Premont, J., Perez, M. & Bockaert, J. (1977) FEBS Lett. 75, 209-212. 30. Miles, D. L., Miles, D. W. & Eyring, H. (1977) Proc. Nati. Acad. Sci. USA 74,2194-2198.

Proc. Nati. Acad. Sci. USA 74 (1977) 31. Green, R. D. & Stanberry, L. R. (1977) Biochem. Pharmacol. 26, 37-43. 32. Schultz, J. & Daly, J. W. (1973) J. Biol. Chem. 248,853-859. 33. Huang, M. & Daly, J. W. (1974) Life Sci. 14,489-503. 34. Trost, T. & Stock, K. (1977) Third International Congress of Cyclic Nucleotides (Abstr.), p. 79. 35. Fain, J. N. & Wieser, P. B. (1975) J. Biol. Chem. 250, 10271034.

Two distinct adenosine-sensitive sites on adenylate cyclase.

Proc. Natl. Acad. Sci. USA Vol. 74, No. 12, pp. 5482-5486, December 1977 Biochemistry Two distinct adenosine-sensitive sites on adenylate cyclase (p...
841KB Sizes 0 Downloads 0 Views