Stimulation of cAMP and phosphomonoester production by ...

Proc. Natl. Acad. Sci. USA Vol. 88, pp. 1506-1510, February 1991 Biochemistry

Stimulation of cAMP and phosphomonoester production by melanotropin in melanoma cells: 31P NMR studies (signal transduction/M2R mouse melanoma cells/intraceilular ATP/phosphoethanolamine)

HADASSA DEGANI*, JOHN 0. DEJORDY*t,

AND

YORAM SALOMONtt

Departments of *Isotope and tHormone Research, The Weizmann Institute of Science, 76 100 Rehovot, Israel

Communicated by Mildred Cohn, November 8, 1990

A major part of the present understanding of ABSTRACT the molecular basis of signal transduction has been gained from in vitro studies using classical biochemical methods. In this study, we used 31P NMR spectroscopy to investigate the response of live M2R mouse melanoma cells to stimulation by melanocyte-stimulating hormone (MSH; melanotropin). In the presence of 3-isobutyl-1-methylxanthine and a synergistic dose of forskolin (1.67 ,uM), MSH induced a transient (=60-min) rise in the cellular concentration of 3',5'-cyclic adenosine monophosphate (cAMP), which coincided in time with an equivalent decrease (-40%) in ATP. However, no detectable change in phosphocreatine concentration was observed. Concomitantly, MSH induced a striking and unexpected increase in the concentration of three phosphomonoester (PME) metabolites ('-2-fold increase in total PME signal area); one signal has been assigned to phosphoethanolamine. The levels of the PMEs remained high for 24 hr and declined slowly (410 hr) to basal level, following perfusion with fresh culture medium. The increase in PME was also observed after stimulation with MSH alone. In contrast, stimulation with a high dose of forskolin (50 ,uM) and isobutylmethylxanthine (0.2 mM), although effective in stimulating the production ofcAMP, did not induce the PME response. Evaluation of the cells' energetics indicated that the enhanced production of phosphoethanolamine is probably not due to ethanolamine phosphorylation. Therefore, it is likely to result from hydrolysis of phosphatidylethanolamine by a specific phospholipase C. The response of the PMEs appears to be regulated by a cAMP-independent process, suggesting the existence of an alternative transduction pathway controlled by MSH.

presumably without significantly affecting cellular ATP levels (10, 11). However, the inclusion of IBMX and forskolin together with MSH in these studies results in an extreme elevation of cAMP. On the basis of this observation, the stimulation of cAMP in M2R cells by MSH was believed to be amenable to analysis by 31P NMR spectroscopy. In this report, we present results of experiments in which the response of M2R cells to MSH was studied by 31P NMR spectroscopy. Methods that allow monitoring of intact cells within the NMR spectrometer, under conditions essentially identical to those existing in a cell culture incubator, were applied (12-14). After hormonal stimulation, the time course of the changes in the intracellular phosphate metabolites was monitored. The stimulation with MSH resulted in an increase in cAMP and a decrease in ATP, observable in the 31P spectrum. In addition, MSH stimulated the production of three phosphomonoester (PME) metabolites; one signal has been assigned to phosphoethanolamine (PEtn). The findings suggest the existence of an alternative transduction pathway regulated by MSH in melanoma cells. This pathway may be analogous to the hormonally regulated phosphatidylinositol phospholipase C system (15) or to the phosphatidylcholinespecific phospholipase C, which in some cases has recently been shown to be under hormonal control (16).

MATERIALS AND METHODS Materials. [Nle4,DPhe]caMSH, a potent analog of aMSH (17), was kindly provided by A. Lerner of Yale University. Forskolin was obtained from Calbiochem. IBMX was obtained from Sigma. All other chemicals were of analytical grade. Cell Culture. M2R mouse melanoma cells were cultured routinely as monolayers in a 1:1 Ham's F12 and Dulbecco's modified Eagle's medium (GIBCO), supplemented with 10% heat-inactivated horse serum (6). For NMR experiments, the cells were subcultured for 72 hr to a density of 1.5-2 x 107 cells per ml on agarose/polyacrolein microspheres (Galisar; Ramat Gan, Israel) treated with polylysine according to a protocol described previously (12). The cells on microspheres were then transferred into a 10-mm tube which was placed in the NMR spectrometer. In the spectrometer the cells were perfused at a rate of 1.5 ml/min, using 60 ml of culture medium circulating in a closed sterile loop at 37°C, under 95% 02/5% CO2 at atmospheric pressure (13, 14). [Nle4,DPhe7]aMSH, forskolin, and IBMX were dissolved in 2 ml of culture medium and added into the perfusate bottle to give a final concentration specified in each experiment. Extracts. Extraction of cells was initiated by adding 10 ml of 5% ice-cold perchloric acid to cell monolayers (2 x 107

cAMP serves as a second messenger for numerous hormonally regulated processes, and its levels are tightly controlled by the relative rates of the adenylate cyclase and cAMP phosphodiesterase reactions (1-3). Melanocyte-stimulating hormone (MSH; melanotropin) controls melanin synthesis in pigment cells by cAMP-mediated (4) regulation of the levels and activity of tyrosinase, the rate-limiting enzyme in melanin synthesis (5). The cAMP response to stimulation by MSH was studied in M2R mouse melanoma cells (6), using [3H]adenine labeling (7) and conventional techniques (8). cAMP accumulation was stimulated to a surprisingly high extent by MSH (10-1000 nM) and a low dose (2 ,uM) of forskolin under conditions where cAMP phosphodiesterase activity was inhibited by 3-isobutyl-1-methylxanthine (IBMX). This dose of forskolin has a negligible effect when used alone but induces a strong synergistic effect in combination with hormones (9), including MSH (6). Twenty minutes after stimulation, approximately 45% of the cell's adenine nucleotide pool was converted into cAMP (6). Intracellular cAMP concentrations are generally believed to fluctuate in the micromolar range,

Abbreviations: MSH, melanocyte-stimulating hormone (melanotropin); IBMX, 3-isobutyl-1-methylxanthine; PME, phosphomonoester; PEtn, phosphoethanolamine; PCho, phosphocholine; PCr,

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

phhosphocreatine . STo whom reprint requests should be addressed at t.

1506

Biochemistry: Degani et al. cells per 15-cm dish), at 40C. After 30 min the extracts were collected and centrifuged (500 x g) at 40C. K2CO3 was then added to the supernatant to precipitate the perchlorate and neutralize the solution. The perchlorate-free samples were treated with Chelex (Bio-Rad) to remove divalent ions and were lyophilized. The dried samples were resuspended in approximately 1 ml of solution containing 10% 2H20, 10 mM ethylenediaminetetraacetate, and 10 mM Hepes buffer. Final pH was approximately 8.5. NMR Measurements. NMR experiments were performed with a Bruker AM-500 NMR spectrometer. 31p spectra were recorded at 202.5 MHz. The temperature in the living cell samples was maintained at (37 1)'C. For each spectrum 180 transients were accumulated by applying 900 pulses, 10-sec repetition delay, and composite pulse proton decoupling (1 W) (13, 14). In extract studies, 9000 transients were accumulated by applying 450 pulses, 2-sec repetition delay, and composite pulse proton decoupling. By comparing with spectra recorded with 90° pulses and 20-sec repetition delay we have verified that under these conditions only the Pi signal is saturated (decreasing the integrated signal intensity of Pi by 20%). The extracts were studied at room temperature. The assignment of the signals was based on their chemical shift (12-14). Phosphocreatine (PCr), at -2.5 ppm, served as an internal reference for calibrating the chemical shift relative to 85% H3PO4. The areas of the signals were determined either by the instrument integration mode or by multiplying the peak height by the width at half-height. The experimental error in the area was generally less than 20%. However, since the intensity of the cAMP signal was low, the error of its area was 30-40%. The areas of the signals were normalized to the area of the Pi signal, which is predominantly due to the medium Pi (1 mM).

Proc. Natl. Acad. Sci. USA 88 (1991) A

1507

Pi

PME PE

5.0

PC

PME

Pwr

y-ATP

a-ATP

4.0

±

RESULTS NMR of perfused M2R cells cultured on microspheres revealed the presence of intracellular nucleoside triphosphates (mainly ATP), PCr, PME metabolites [including PEtn, phosphocholine (PCho), and unassigned species], a UDP-sugar derivative, and Pi, primarily extracellular (Fig. 1, spectrum A). To maximize cAMP production, we stimulated the cells with [Nle4,DPhe7]aMSH (0.67 ,uM), forskolin (1.67 ,uM), and IBMX (0.1 mM). In the NMR spectrum recorded for 30 min after this stimulation, a peak identified as cAMP and a concurrent reduction in ATP (44%) were observed (Fig. 1, spectra B and C). The ratio between the increase in cAMP and the decrease in ATP (AcAMP/AATP) was 0.92. In addition, a marked increase in the total PME signal area (154%) could be seen (Fig. 1, spectra B and C). In an 31p

expanded scale of the PME region, at least four PMEs can be resolved; three were markedly enhanced by the stimulation, PCho was not. Due to the extended overlap of signals in this region (Fig. 1 Insets), the extent of stimulation was determined from measurements of the total integrated intensities of all PME signals. Results of experiments in which MSH stimulation was tested under different conditions are summarized in Table 1. Increased levels of PME were observed in all experiments with saturating levels of MSH (0.067-1.0 ,uM) in the presence or absence of forskolin and IBMX (Table 1, exps. 1-8). The mean increase in PME was 109% ± 29% (±SD; n = 8). The appearance of cAMP and the reduction in ATP were seen only when both forskolin and IBMX were present in addition to MSH (Table 1, exps. 1-3, 5). The average decline in ATP was 36% ± 7% (±SD; n = 4), while the ratio AcAMP/AATP was found to be approximately 1 (0.98 0.3; ±SD; n = 4). In control experiments, when cells were perfused with either IBMX (0.2 mM) alone or IBMX (0.1, 0.2 mM) and a low dose of forskolin (0.5, 1.67 gM), no response in cAMP, ATP, or

50

4.0

l...

c

-

..

5.0 4.0 ppm

5

0

-5

-10

-15

-20

-25

ppm

FIG. 1. 31P NMR spectra of resting and MSH-stimulated perfused M2R mouse melanoma cells. Spectrum A, resting M2R cells. Spectrum B, M2R cells after 30-min stimulation with 0.67 AM [Nle4,DPhe7]aMSH, 1.67 AM forskolin, and 0.1 mM IBMX. Spectrum C, difference spectrum B - A. Insets present the PME region with an expanded scale. M2R cells were cultured on beads and perfused in the NMR spectrometer; 180 transients (30 min) were accumulated for each spectrum. An exponential multiplication with 25-Hz line-broadening was applied. The peaks were assigned as described in the text. PE = PEtn, PC = PCho, and UDPS = UDP-sugar derivative. The identities of PEtn and cAMP were confirmed in extracts, as described in Results. Pi represents primarily the extracellular concentration (1.0 mM) in the culture medium. The a, and y phosphates of ATP are indicated a-ATP, etc. /3,

PME level was detected. A high concentration of forskolin, in the presence of IBMX, induced a significant elevation in cAMP along with a decline in ATP (Table 1, exps. 9 and 10). Table 1. Effects of MSH stimulation of phosphate metabolites Changes within 30 min Stimulation mixture of stimulation MSH, Forskolin, IBMX, PME, ATP, AcAMP/ Exp. AtM AM mM % increase % decrease AATP 1 0.67 1.67 0.1 118 36 1.05 2* 0.67 1.67 0.1 85 0.64 28 3 0.67 1.67 0.2 154 44 0.92 t 4 0.67 1.67 t 0.1 112 5 0.067 1.67 0.1 98 1.5 31 6 0.67 0 0.1 62 7 1.0 0 0 107 8* 1.0 0 0 140 9 0 50 0.2 -21t 0.75 46 10 0 50 0.2 -28f 42 0.83 No change observed. *Restimulation after 8-10 hr of perfusion with fresh medium. tIn this experiment, 15-min spectra were accumulated and the signal-to-noise ratio of cAMP and ATP was too low to permit accurate integration.

tPME decreased.

Biochemistry: Degani et al.

1508

Proc. NatL. Acad. Sci. USA 88 (1991)

However, this elevation in cAMP was not accompanied by an increase in PME. In contrast, 1 ILM MSH, added alone, elicited a 107-140% increase in PME levels, compared with unstimulated control, but in the absence of IBMX and forskolin the changes in cAMP and ATP levels were presumably too small to be detected by 31P NMR (Table 1, exps. 7 and 8). The drop in ATP seen in these experiments was not accompanied by an observable change in ADP (obtained by subtracting the area of the ATP ( phosphate signal from that of the ATP y phosphate) or in PCr and Pi levels. However, the Pi signal is mainly due to the medium Pi (1 mM) and changes in intracellular Pi, less than 20% of the total Pi, could not be detected. The changes in phosphate metabolites were monitored over a period of several hours. The results of a time course experiment are illustrated'in Fig. 2. 31p spectra were initially recorded under resting conditions for 6 hr. At 6 hr cells were stimulated as described above (exp. 1, Table 1). cAMP levels reached peak values within 30 min and declined to an undetectable level 60 min later, apparently due to desensitization. The levels of the PMEs also increased rapidly in parallel to the increase in cAMP and reached their maximum within 30 min, where they remained for w150 min in spite of the complete decline in cAMP. At 8 hr from the start of the experiment, the perfusion medium was replaced by fresh medium devoid of stimulants (Fig. 2, open arrow). Thereafter, PME levels declined slowly over a period of 10 hr, reaching nearly basal levels by 18 hr. At this time (Fig. 2, right arrow), restimulation of the cells (exp. 2, Table 1) resulted in similar changes, showing that hormone responsiveness was fully regained. The decline in ATP in response to hormonal stimulation was observed to coincide in time and magnitude with the temporary rise in cAMP at 6 and 18 hr (AcAMP/ AATP 1). The steady slow rise in PCr over the entire period of the experiment appears to represent an increase in cell PCr_

mass due to cell proliferation (doubling time on microspheres determined by cell counts was approximately 24 hr). Improved resolution of the spectrum can be obtained in cell extracts. Unstimulated cells and cells stimulated for 30 min with [Nle4,DPhe IaMSH (1 pAM), forskolin (2 QgM), and IBMX (0.1 mM) were extracted as described in Materials and Methods. The resulting extracts were studied by 31P NMR (Fig. 3). As expected, the signals were much better resolved (cf. Figs. 3 and 1). The changes in the spectra ofthese extracts were similar to those recorded from intact cells (except for two unidentified phosphodiester signals). The appearance of cAMP, the reduction in ATP, and the increase in PMEs were all clearly resolved. The small decrease in PCho and PCr signals observed in extracts but not in the intact stimulated cells could result from inherent limitations of the extraction procedure. Consequently, no quantitative analysis of the concentrations of the metabolites in the extracts was performed. We confirmed the assignments of cAMP and PEtn by adding the authentic compounds to the cell extracts, resulting in the augmentation of the respective signals (not shown). The following substances were also tested in this way, but were ruled out as candidates for the MSH-stimulated PMEs: 1-a-glycerophosphate, 5'-AMP, and 5'-GMP. Furthermore, we excluded phosphate intermediates of the glycolytic pathway as likely candidates of the stimulated PME signals. This was achieved by recording alternate 31P and 13C spectra and following the PME stimulation and the

A

PDE

y -ATP

I

-~~~~~~~~~~~~~~~

0

0.2 _

B

PMEtT U.

PE

B

_zj

(n,

cAMP

TIME OF PERFUSION (Hrs) FIG. 2. Time course of the changes in phosphate metabolites before, during, and after stimulation of M2R melanoma cells with MSH. The sample preparation and the NMR experimental conditions were as for Fig. 1. At 6 hr (left arrow), 2 ml of the mixture containing the hormone was added into the medium circuit to give final reagent concentrations as in exp. 1, Table 1. At 8 hr (open arrow), the medium was replaced with fresh medium without added reagents. At 18 hr (right arrow), stimulation with the same hormone mixture was repeated. Each point in the plots represents the area of the corresponding signal relative to the area of the Pi signal, in a 30-min NMR spectrum, as shown in Fig. 1.

8

4

0

-4

-8

PPm FIG. 3. 31p NMR spectra of perchloric acid extracts of resting and MSH-stimulated M2R melanoma cells. Spectrum A, extract of resting M2R cells. Spectrum B, extract of M2R cells after 30-min stimulation with [Nle4,DPhe7]aMSH (1 IAM), forskolin (2 AtM), and IBMX (0.1 mM) at 370C. In recording the spectra, 9000 transients were accumulated, using exponential multiplication with a 5-Hz line broadening. Abbreviations as in Fig. 1; phosphodiesters are indicated as PDE.

Biochemistry: Degani et al. production of [3-13C]lactate from 1-'3C- and 6-13C-labeled D-glucose, respectively, using methods described previously (13, 14). Although a marked increase in PMEs was observed after stimulation (exp. 4, Table 1), no augmentation of the level of any P-containing 13C-labeled glycolytic intermediate could be detected. This experiment also revealed that glycolysis is the major source of energy (-90%o of the added glucose is converted into lactate under aerobic conditions) in M2R cells and its rate is not affected by stimulation with MSH.

DISCUSSION

Using 31P NMR spectroscopy, we have shown that, in the presence of IBMX and forskolin, it is possible to observe an inverse relation between the levels of ATP and cAMP, the substrate and product of the adenylate cyclase reaction. Under these experimental conditions (Figs. 1 and 2), cAMP synthesis resulted in an -40%6 reduction in cellular ATP concentration. This observation confirms our earlier findings obtained by [3Hladenine prelabeling of M2R cells, in which 45% of the total [3H]adenine pool of the cells was found to be converted into cAMP under the same incubation conditions (6). Furthermore, the 31P NMR results provided evidence that the 3H-labeled ATP compartment indeed represents the major component of the 3H-labeled adenine pool. The synthesis of cAMP by adenylate cyclase is accompanied by the formation of pyrophosphate, which is usually hydrolyzed rapidly to Pi. Within the signal-to-noise ratio and the time resolution of this experiment, neither a pyrophosphate signal nor a change in Pi was observed. Considering the 1 mM concentration of the medium Pi and the extracellularto-cellular volume ratio (>9:1), it is reasonable to assume that the observed Pi signal was predominantly extracellular. Therefore, expected changes in the intracellular Pi appear to be within the experimental error (20%) of the total signal and thus were not detectable. In addition, the possibility of Pi exchange between the intra- and extracellular compartments could damp expected changes in intracellular Pi concentration. PCr, by means of the creatine kinase reaction, is generally considered to buffer cellular ATP concentrations. In muscle, for instance, enhanced conversion of ATP to ADP is initially reflected by a reduction in PCr as regulated by the creatine kinase reaction (18). It was, therefore, unexpected to find that cellular ATP decreased without an observable change in PCr. This is probably due to the fact that a sizable portion (-40%) of the consumed ATP was trapped as cAMP. Because IBMX inhibits the reconversion of cAMP to ADP via 5'-AMP, no ADP becomes available for the creatine kinase reaction. If we assume that in M2R cells the creatine kinase reaction is maintained close to equilibrium during stimulation, then the reduction in ATP should be accompanied by changes in the concentration of other substrates of this enzyme. No significant changes were observed in PCr or in the pH. A reduction in ADP would have indicated retainment ofthe equilibrium; however, ADP was too low to be observed in the 31p spectra. Thus the possibility that the stimulation moved the creatine kinase reaction far from equilibrium cannot be excluded. It appears that the large changes in cAMP and in ATP resulted exclusively from the presence of IBMX, since this phenomenon was not observed in cells stimulated by MSH in the absence of the phosphodiesterase inhibitor (Table 1, exps. 7 and 8). These observations may provide insight into possibilities of cell-specific manipulation of ATP levels. The stimulated state with low ATP and high cAMP was maintained for approximately 60 min. Then ATP and cAMP returned close to their initial values within 60 min. There are several mechanisms for deactivating the adenylate cyclase

Proc. NatL. Acad. Sci. USA 88 (1991)

1509

system and thereby turning off signal transmission (19-21). Such processes, collectively referred to as desensitization, are most likely responsible for the observed reduction in cAMP and recovery of ATP. cAMP, definitely a major mediator of MSH action in M2R melanoma cells (5, 6), may not necessarily be involved in the pathway that stimulates PME production. This is supported by the finding that forskolin, at a high dose (50 .uM) and in the presence of IBMX, stimulated cAMP formation and reduced ATP but failed to induce PME production. However, it is not possible to rule out a mechanism by which MSH elicits the formation of an additional second messenger that enhances PME production in concert with cAMP. The discovery that PME production is stimulated by MSH deserves additional attention. The possibility that the enhanced production of PME resulted from ATP-dependent phosphorylation reactions (e.g., phosphorylation of ethanolamine by ethanolamine kinase) appears unlikely, as the decline in ATP can be fully accounted for by the increase in cAMP (AcAMP/AATP 1). In addition, there was no apparent increase in ATP production, since the rate of glucose consumption, the main source for ATP, was unaffected by the hormonal stimulation (data not shown). Considering the magnitude of the PEtn response, it appears reasonable that this metabolite may be derived from hydrolysis of ethanolamine-containing lipids (phosphatidylethanolamine and sphingosine phosphate) (22), which are known sources of cellular PEtn (ethanolamine is not provided in the culture medium but may be present in low concentrations in the serum). Other possibilities that cannot be excluded presently are MSH-mediated inhibition of the CTP:phosphocytidyltransferase reaction, the rate-limiting step in the synthesis of phosphatidylethanolamine from PEtn, and inhibition of the degradation of PEtn by phosphatase. It is most tempting to speculate that the enhanced PEtn production results from an MSH-dependent activation of a specific phospholipase-C analogous to the activation ofphosphatidylinositol 4,5-bisphosphate phospholipase C (15). Alternatively, the response could result from activation of a phosphatidylethanolamine phospholipase C by protein kinase C or Ca2+-mediated mechanism similar to that proposed recently for phosphatidylcholine phospholipase C activation (16). It should, however, be noted that no observable changes in PCho levels were induced by MSH in M2R cells (Fig. 1). The possible existence of two MSH receptor subtypes in M2R cells has recently been proposed (23, 24). The results presented here are consistent with the possibility that one subtype may be involved in the regulation of adenylate cyclase and the other in the pathway that enhances PME

production.

In summary, this study demonstrates that, by using 31P NMR spectroscopy, it is possible to monitor the time course of the MSH activation of adenylate cyclase in living M2R melanoma cells. Unexpectedly, MSH also stimulates the production of PEtn and two other PME metabolites. This stimulation was initiated in parallel to the activation of adenylate cyclase, but PMEs remained elevated for a longer period than cAMP. The prolonged formation of the PMEs may be important in control of cellular mechanisms that require long-term activation. The high concentrations of the PMEs and their prolonged presence in the stimulated cells suggest that they may have physiological importance. Further experiments will be required to determine the universality of these observations, to identify the putative physiological role of PME metabolites, and to elucidate the nature of the enzymatic pathway(s) responsible for the observed effects of MSH. We thank Rachel Benjamin for excellent secretarial assistance and Yosepha Schmidt-Sole and Anat Azrad for devoted technical help.

1510

Biochemistry: Degani et al.

This work was supported in part by a research grant to Y.S. from the Crown Endowment Fund for Immunological Research at the Weizmann Institute of Science and by research grants to H.D. from the German-Israeli Foundation for Scientific Research and Development and the U.S. Public Health Service (CA42238). Y.S. is the Charles and Tillie Lubin Professor of Hormone Research. 1. Sutherland, E. W., 0ye, I. & Butcher, R. W. (1965) Rec. Prog. Hormone Res. 21, 623-646. 2. Greengard, P. & Robison, G. A., eds. (1984) Advances in Cyclic Nucleotide and Protein Phosphorylation Research (Raven, New York), Vol. 18. 3. Rosen, M. 0. & Krebs, E. G., eds. (1981) Protein Phosphorylation, Cold Spring Harbor Conferences on Cell Proliferation (Cold Spring Harbor Lab., Cold Spring Harbor, NY), Vol. 8, Books A and B. 4. Bitensky, M. W., Demopoulos, H. B. & Russell, V. (1973) in Pigmentation, Its Genesis and Biologic Control, ed. Riley, V. (Appleton-Century-Crofts, New York), pp. 247-255. 5. Lerner, A. B. (1959) Nature (London) 184, 674-677. 6. Gerst, J., Sole, J., Mather, J. P. & Salomon, Y. (1986) Mol. Cell. Endocrinol. 46, 137-147. 7. Humes, J. L., Rounbehler, M. & Kuel, F. L., Jr. (1969) Anal. Biochem. 32, 210-217. 8. Salomon, Y. (1979) Adv. Cyclic Nucleic Res. 10, 35-57. 9. Seaman, K. B. & Daly, J. W. (1986) Adv. Cyclic Nucleic Res. 20, 1-150. 10. Shimizu, H., Creveling, C. R. & Daly, J. W. (1970) inAdvances in Biochemical Psychopharmacology, eds. Greengard, P. & Costa, E. (Raven, New York), Vol. 3, pp. 135-154.

Proc. Natl. Acad. Sci. USA 88 (1991) 11. Krishna, G., Forn, J., Voigt, K., Paul, M. & Gessa, G. L. (1970) in Advances in Biochemical Psychopharmacology, eds. Greengard, P. & Costa, E. (Raven, New York), Vol. 3, pp. 155-172. 12. Neeman, M., Rushkin, E., Kadouri, A. & Degani, H. (1988) Mag. Reson. Med. 7, 236-242. 13. Neeman, M. & Degani, H. (1989) Cancer Res. 49, 589-594. 14. Neeman, M. & Degani, H. (1989) Proc. Natl. Acad. Sci. USA 86, 5585-5589. 15. Berridge, M. J. & Irvine, R. F. (1989) Nature (London) 341, 197-205. 16. Exton, J. H. (1990) J. Biol. Chem. 265, 1-4. 17. Sawyer, T. K., Sanfilippo, P. J., Hruby, V. J., Engel, M. H., Heward, C.-B., Burnett, J. B. & Hadley, M. E. (1980) Proc. Natl. Acad. Sci. USA 77, 5754-5758. 18. Shoubridge, E. A., Bland, J. & Radda, G. K. (1984) Biochim. Biophys. Acta 805, 72-78. 19. Sibley, D. R. & Lefkowitz, R. J. (1985) Nature (London) 317, 124-129. 20. Gilman, A. G. (1987) Annu. Rev. Biochem. 56, 615-649. 21. Birnbaumer, L., Abramowitz, J. & Brown, A. M. (1990) Biochim. Biophys. Acta 1031, 163-224. 22. Ansell, G. B. & Spanner, S. (1982) in Phospholipids, New Comprehensive Biochemistry, eds. Hawthrone, J. N. & Ansell, G. B. (Elsevier, New York), Vol. 4, pp. 1-49. 23. Gerst, J., Sole, J., Hazum, E. & Salomon, Y. (1988) Endocrinology 123, 1792-1797. 24. Solca, F. F. A., Salomon, Y. & Eberle, A. N. (1991) Receptor Res., in press.