Neuroscience Vol. 36, No. 2, pp. 299 304, 1990 Printed in Great Britain

0306-4522/90$3.00+ 0.00 Pergamon Press plc :ti: 1990IBRO

HYPERPOLARIZATION OF MYENTERIC NEURONS BY OPIOIDS DOES NOT INVOLVE CYCLIC ADENOSINE-3',5'-MONOPHOSPHATE S. M. JOHNSON* and N. P. PILLAI NH & MRC Addiction Research Unit, Department of Physiology, Centre for Neuroscience, Flinders University of South Australia, Bedford Park, 5042 Australia Abstract--To investigate the role of cyclic adenosine-3',5'-monophosphate on the inhibitory actions of

opioids in guinea-pig ileum, we made intracellular recordings from the two electrophysiologically defined classes of neurons (S and AH) in the myenteric plexus. The selective opioid mu agonisI (o-Ala2,N-Me Phe4,GlyS-ol)-enkephalin caused a membrane hyperpolarization in 34 out of 67 S neurons but did not affect the membrane potential of AH neurons. The mean amplitude (_+ S.E.M.) of the hyperpolarization was 8.2__0.8mV. Forskolin, which activates adenylate cyclase and increases intracellular cyclic adenosine-Y,5'-monophosphate levels, caused a membrane depolarization in AH neurons (9.4 __ 1.9 mV) but did not alter the resting membrane potential of S neurons. Similarly, neither the phosphodiesterase inhibitor, isobutylmethylxanthine, nor the membrane permeable analogue of cyclic adenosine-Y,5'monophosphate, dibutyryl cyclic adenosine-Y,5'-monophosphate, altered the resting membrane properties of S neurons. Furthermore, none of these agents affected significantly the amplitude of the hyperpolarization of S neurons by (o-Ala2,N-Me-Phe4,GlyS-ol)-enkephalin. The experiments indicate that changes in intracellular cyclic adenosine-3',5'-monophosphate are not important in the processes that link occupation of mu receptors to the opening of potassium channels on myenteric neurons.

Cyclic adenosine-Y,5'-monophosphate (cyclic AMP) has been proposed as an intracellular messenger for receptor-mediated neuronal actions of several drugs and putative neurotransmitters. 3~7'25"29This hypothesis has depended largely on extensive biochemical observations that a variety of agents such as noradrenaline, dopamine, adenosine, G A B A and opioids alter adenylate cyclase activity or cyclic A M P levels in neuroblastoma-glioma hybrid cells and in a variety of neurons. Experiments which have examined the effects of these agents on neurotransmitter release have not, however, consistently supported the hypothesis, which is therefore still regarded as speculative for some classes of drugs, including the opioids) For example, opioid-induced decrease in noradrenaline release from rat brain cortical slices was independent of alterations in cyclic AMP. 24 More recent experiments showed that opioid-induced depression of excitatory cholinergic transmission to the longitudinal muscle of the guinea-pig ileum was not influenced by several agents which alter intracellular levels of cyclic AMP. l° On the other hand, a role for cyclic A M P in opioid actions has been proposed from electrophysiological *To whom correspondence should be addressed. Abbreviations: Cyclic AMP, cyclic adenosine-3',5'monophosphate; DAGO, (D-Ala2,N-Me-Phe4,GlyS-ol)enkephalin; db~zyclic AMP, dibutyryl cyclic adenosine-Y,5'-monophosphate; EPSP, excitatory postsynaptic potential; G-protein, guanine nucleotide binding protein; IBMX, isobutylmethylxanthine; MP LM, myenteric plexus-longitudinal muscle.

studies on neurons in rat mesencephalic reticular formation, 7 mouse spinal cord cultures, 4 and in the rat locus coeruleus. 1'2 For example, opioids hyperpolarize rat locus coeruleus neurons by interaction with mu receptors and subsequent activation of a potassium conductance. 2° The hyperpolarization was prevented by pretreatment of rats with pertussis toxin ~ and reversed by membrane permeable analogues of cyclic AMP, 2 suggesting that it resulted from a reduction in intracellular cyclic A M P following coupling of the opioid mu receptor to an inhibitory guanine nucleotide binding protein (G-protein) linked to adenylate cyclase. Opioids also hyperpolarize a proportion of a subpopulation of neurons in the myenteric plexus of the guinea-pig ileumfl This hyperpolarization, like that observed in the locus coeruleus, also involves a mu receptor linked to a potassium conductance/6'28 To further our investigation of the possible role of cyclic A M P in opioid actions in the myenteric plexus, ~° we examined whether the hyperpolarization of myenteric neurons by the selective mu agonist, (o-Ala2, N - Me-Phe4,GlyS- ol)-enkephalin (DAGO) was influenced by agents that have been reported to elevate or mimic cyclic A M P levels. A preliminary account of the study was reported at the International Narcotics Research Conference. ~°

299

EXPERIMENTAL PROCEDURES

Preparation of myenterie ganglia Guinea-pigs (MVS strain; Dept Agriculture, Gilles Plains, South Australia) of either sex, weighing between 250 and

300

S.M. JOHNSONand N. P. PILLAI

400 g, were killed by a blow to the head and bled from the carotid arteries. A l-cm segment of ileum, 10cm from the ileo~aecal junction, was removed and placed in a modified Krebs solution of the following composition (in raM): NaCI 118, KCI 4.75, NaH2PO 4 1.0, NaHCO 3 25, MgSO 4 1.2, glucose 1l and CaC12 1.5. The solution was bubbled with a gas mixture of 95% 02-5% CO 2. The segment of ileum was slit open adjacent to the mesenteric connections and pinned out as a flap, with the mucosa facing upwards. The mucosa and submucosa were then removed to expose the circular muscle layer. Using fine forceps, a band of the circular muscle, approximately 1 2 mm in width, was stripped away. This exposed two to five rows of ganglia of the myenteric plexus, aligned parallel to the direction of the circular muscle and remaining attached to the longitudinal muscle. Pins made from 25-/~m diameter steel wire were used to anchor the strip of myenteric plexus-longitudinal muscle (MP-LM) to the base of a tissue chamber, after which the remainder of the flap was dissected away and discarded. Each MP-LM preparation contained between 20 and 50 ganglia.

Intracellular recording The tissue chamber was mounted on the stage of an inverted microscope (Olympus IMT2) and the ganglia visualized at 200-fold magnification using differential interference contrast (Nomarski) optics. Krebs solution, maintained at 37°C and continuously bubbled with 95% 02-5% CO:, was perfused by gravity into the tissue chamber at a rate of 4-5 ml/min. Fibre-filled glass microelectrodes (CG 100F-15; Clark Electromedical Instruments), containing 2M KCI, were used to impale and record from neurons within ganglia of the myenteric plexus. Electrode resistances ranged from 60 to 120M~. Signals were amplified (Axoclamp 2, Axon Instruments Inc.) and displayed on an oscilloscope (Tektronic Model 5113). Membrane potential was monitored continuously on a chart recorder (Linear Instruments Inc.). To measure input resistance, hyperpolarizing current pulses (0.02-0.2 nA) were injected through the recording electrode and the peak amplitudes of the resulting electrotonic potentials measured. Input resistances were determined from the slopes of the current-voltage relationships which were linear in each neuron for hyperpolarizations up to 20 mV. The two electrophysiological classes of neurons in the myenteric plexus were defined according to the classification of Hirst et al.6 Cells were classified as S neurons if electrical stimulation, using a focal electrode positioned on an internodal strand or on the ganglion surface, evoked a fast cholinergic nicotinic excitatory postsynaptic potential (EPSP). Neurons were classified as AH if a long-lasting (5-20s) hyperpolarization followed an action potential evoked in the soma by passing a depolarizing current pulse through the recording electrode. In AH neurons, focal electrical stimulation of fibre tracts did not evoke a fast EPSP. Drugs were applied by adding known concentrations to a reservoir containing the Krebs perfusate. Due to the fluid volume of the perfusion system, approximately l min was required for the drug to reach the tissue chamber from the Krebs reservoir. Contractions of the longitudinal muscle were minimized by addition of hyoscine (1 ~M) and nicardipine (3/~M) to the Krebs perfusate.

Materials DAGO (Sigma, St Louis, MO, U.S.A.) was prepared as a stock solution in 0.1 N HC1 and added to the Krebs perfusate to give a final concentration of 500 riM. Forskolin (Sigma), isobutylmethylxanthine (IBMX; Sigma), and nicardipine (Sigma) were prepared as stock solutions in ethanol and added to the perfusate so that the concentration of ethanol never exceeded 1 : 2000. Other drugs used were

hyoscine (David Bull Laboratories, Victoria, Australia) and dibutyryl cyclic adenosine-3',5'-monophosphate (dl~cyclic AMP; Sigma). RESULTS

Effects o f (o-Ala2,N-Me-Phe 4, Glytol)-enkephalin on S and A H neurons S neurons. Intracellular recordings were made from 67 S neurons, with membrane potentials ranging from - 4 5 to - 6 0 mV. When administered by bath perfusion, D A G O (500 nM) produced a hyperpolarization in 34 out of the 67 neurons (Fig. 1). The remaining 33 neurons were unaffected. The mean amplitude of the hyperpolarization ( ± S . E . M . ) was 8.2 + 0.8 mV. The hyperpolarization reversed when D A G O was removed by washing and was prevented by naloxone (1/~M). Repeated applications of D A G O at intervals of 15min did not result in any consistent change in the amplitude of the hyperpolarization. To measure input resistance during perfusion with D A G O , the hyperpolarization was offset by passing constant depolarizing current through the recording electrode. Satisfactory measurements of input resistance were possible in 22 out of the 34 cells that were hyperpolarized. In 10 cells, the hyperpolarization was associated with a decrease in input resistance, from 158 + 29 to 115 _ 22 Mf~, while in the remaining 12 cells, no clear change was detected. A H neurons. Only one out of 24 A H neurons responded to bath application of D A G O (500 nM) with a small hyperpolarization (4 mV), an observation consistent with previously reported paucity of effects of opioids on the resting membrane potentials of A H neurons, "'t9 (but see Ref. 23). An example of the selective effect of D A G O on S compared with A H neurons is shown in Fig. 1. Mean input resistance ( + S.E.M.), measured satisfactorily in 18 out of the 2 4 A H neurons, was 1 6 8 + 2 1 M f l prior to application of D A G O and 170 + 20 Mf~ after application of D A G O . D A G O also did not affect the amplitude or duration of the afterhyperpolarization following three to 10 action potentials evoked in the soma by passing depolarizing current pulses through the recording electrode. Effects o f forskolin, isobutylmethylxanthine and dibutyryl cyclic adenosine-3',5'-monophosphate A H neurons. In 24 out of 30 A H neurons, forskolin (1 # M ) caused a depolarization ranging in amplitude from 3 to 24mV, with a mean (_+S.E.M.) of 9.4 + 1.9 mV. In 14 of these neurons, the depolarization was accompanied by an increase in input resistance (Fig. 1), from 140 + 19 to 167 +__19 Mfl. In six out of 30 neurons, forskolin did not affect the resting membrane potential or input resistance. When a higher concentration of forskolin ( 1 0 # M ) was administered, the magnitude of the depolarization, observed in six out of 10 neurons, was not greater but rather somewhat less (6.8 + 1.2mV) than that

Opioids and cyclic AMP

'J

a

301 DG

W2

DAGO (500 nM)

FORSKOLIN(I~M) J

l0 mV

1 min d

e

DAGO (500 nM)

FORSKOLIN (1~uM)

Fig. 1. Intracellular recordings from AH neurons (top traces; a, b) and from S neurons (bottom traces; c, d) showing the effects of DAGO and forskolin. Each agent was added to the perfusate during the periods indicated by the bars. DAGO hyperpolarized the S neuron (c) but not the AH neuron (a). The broken line in c indicates the level of the resting membrane potential. The delay in onset of the effect of DAGO is due to the fluid space of the perfusion system. Forskolin depolarized the AH neuron (b) but not the S neuron (d). The downward deflectionsin b are electrotonic potentials evoked by passing hyperpolarizing current pulses (0.I nA) through the recording electrode. The increase in amplitude of the responses, recorded when the depolarization by forskolin was offset by direct current (DC), indicates that the depolarization was associated with an increase in input resistance. Also shown in b are recordings 10 rain (%) and 20 min (w2) after washout of forskolin. observed with 1 ktM forskolin. In three neurons, the depolarization evoked by forskolin was accompanied by the discharge of action potentials. In all neurons, irrespective of whether they were depolarized, forskolin greatly reduced or abolished the afterhyperpolarization that followed an action potential evoked in the soma by passing a depolarizing current pulse through the recording electrode. S neurons. Unlike its effects on AH neurons, forskolin (1 #M) never depolarized or altered the input resistance of S neurons (Fig. 1). Two out of 15 cells were hyperpolarized, by 2 and 6 mV, while the remaining 13 cells were unaffected by forskolin. The effects of the phosphodiesterase inhibitor, IBMX (50/tM) and the membrane permeable analogue of cyclic AMP, db-cyclic AMP (50/~M) on S neurons were also examined. At these concentrations, both IBMX and db-cyclic AMP produce effects on AH neurons consistent with elevation of intracellular cyclic AMP. 22 However, neither IBMX nor db-cyclic AMP altered the resting membrane potential or input resistance in any of the 23 S neurons examined.

Influence of forskolin, isobutylmethylxanthine and db-c),clic adenosine-3',5'-rnonophosphate on the actions o1"(D-Ala 0.5, paired t-test). Similarly, neither IBMX nor db-cyclic AMP affected significantly the amplitude of the hyperpolarization by DAGO. An example of the hyperpolarizations induced by DAGO before and after addition of db-cyclic AMP is shown in Fig. 2. Table 1 summarizes the effects of DAGO on the resting membrane

S. M. JOHNSONand N. P. PILLAI

302 DC

a

DC

DAGO(500 nM)

10 mV

DAGO (500 nM) 1 min

DC

b

DC

DAGO (500 nM)

DAGO (500 nM)

FORSKOLIN (1/JM)

DIBUTYRYL cAMP (50.uM)

Fig. 2. Intracellular recordings from S neurons showing hyperpolarizations by DAGO before, and 15 min after, exposure to forskolin (a, b) or db-cyclic AMP (c, d). In each panel, the broken line represents the level of the resting membrane potential. During the periods indicated by DC, the hyperpolarizations were offset by direct current. Downward deflections in c and d are electrotonic potentials evoked by hyperpolarizing current pulses (0.1 nA). DAGO-induced hyperpolarization was associated with a decrease in input resistance, as indicated by the decrease in amplitude of the electrotonic potentials. Neither forskolin nor db~zyclic AMP affected the magnitude of the hyperpolarizations induced by DAGO (see also Table 1). potentials of S neurons in control preparations and in those treated with forskolin, IBMX and db-cyclic AMP. Thirty-three out of the 67 S neurons were not hyperpolarized by DAGO. In 20 of these, the application of DAGO was repeated 15 min after addition of either forskolin, IBMX or db-cyclic AMP. DAGO was still without effect on the resting membrane potential or input resistance.

DISCUSSION The principal findings in these experiments are that neither the resting membrane properties of myenteric S neurons, nor the hyperpolarization of these cells by the selective mu agonist, DAGO, are altered significantly by forskolin, IBMX or db-cyclic AMP. Forskolin directly activates adenylate cyclase, elevating intracellular levels of cyclic AMP. 27 More recently, biochemical studies have demonstrated cyclic AMP in myenteric neurons, and its elevation by forskolin at concentrations similar to those used in the present study. 9 IBMX increases intraceltular

Table 1. Hyperpolarization of myenteric S neurons by (D-AIa2,N-Me-Phe4,Gly-ol)-enkephalin (500nM) mV ±S.E.M. Control + forskolin (1 #M)

9 9

___2 +2

n =6, P >0.8"

Control + IBMX (50#M)

9 8

+ 1 +2

n =6, P >0.5"

Control + d~cyclic AMP (50 p M)

8 7

+ 1 ___1

n = 4, P > 0.5*

*Comparisons were made using paired t-tests.

levels of cyclic AMP by inhibiting phosphodiesterase. 5'26db-cyclic AMP is a membrane permeable analogue of cyclic AMP and mimics the intracellular actions of the latter. The lack of effect of forskolin, IBMX and db~yclic AMP on S neurons suggests either that the adenylate cyclase/cyclic AMP system is maximally active in these neurons under resting conditions, or that the resting membrane potential and conductance of the somata of S neurons are not controlled by intracellular cyclic AMP. In AH neurons, however, the membrane depolarization and increase in input resistance produced by forskolin is consistent with an elevation in cyclic AMP. This observation is essentially in agreement with those of Nemeth et al., ~8 although in the latter study, forskolin depolarized all AH neurons examined and the magnitude of the depolarization was somewhat greater than that observed here. In our experiments, forskolin was more likely to evoke a depolarization when the resting conductance of the AH neuron was high. Although resting conductances of AH neurons were not reported by Nemeth et al., Is it is likely that variation in the resting conductance is at least one factor contributing to the differences in magnitudes of the effects of forskolin observed. As suggested previously, TM variations in the basal intracellular status of impaled neurons may strongly influence their responses to drugs. The depolarization by forskolin is mimicked by membrane permeable analogues of cyclic AMP and potentiated by IBMX. 18,22Based on these observations, it has been proposed that cyclic AMP may be important in influencing the excitation of AH neurons. Thus whereas DAGO affected the resting membrane properties of S neurons, cyclic AMP appears to

Opioids and cyclic AMP influence only AH neurons, which were not affected by DAGO. It seems likely, therefore, that the link between mu receptors and the opening of potassium channels on myenteric S neurons does not involve cyclic A M P as an intraceilular effector. In neurons of the submucous plexus, a variety of agents such as opioids, alpha 2 adrenoceptor agonists and somatostatin also act by increasing potassium conductance, leading to a membrane hyperpolarization.~4 Biochemical studies have shown that all these agents decrease cyclic A M P production. Furthermore, forskolin depolarizes submucous neurons by decreasing a potassium conductance, indicating a link between cyclic A M P and potassium channels. '4 However, forskolin did not alter the agonist-induced hyperpolarizations of the same neurons, further emphasizing that the opening of potassium channels on enteric neurons is quite independent of cyclic AMP. In addition to the hyperpolarization of myenteric neurons recorded with intracellular microelectrodes, other inhibitory actions of opioids have been investigated in the guinea-pig ileum. Both mu and kappa agonists inhibit the firing of myenteric neurons recorded with extracellular electrodes. 12 These effects were unaltered by IBMX or db-cyclic AMP. Furthermore, neither cholera toxin, forskolin, nor db-cyclic A M P altered the inhibitory effects of mu and kappa agonists on excitatory cholinergic transmission to the longitudinal muscle. ~°13 Thus although the relationships among the membrane hyperpolarization recorded intracellularly, the decrease in firing rate recorded extracellularly, and the inhibition of nerve-

303

mediated contractions are uncertain, none of these actions involves changes in intracellular cyclic AMP. In rat locus coeruleus neurons, hyperpolarizations by mu agonists were reversed by membrane permeable analogues of cyclic AMP. 2 It was suggested that the hyperpolarization resulted from inhibition of adenylate cyclase and consequent decrease in intracellular cyclic A M P levels. Subsequent observations, however, have failed to substantiate this proposal. Neither N6,O2-db-cyclic A M P nor forskolin altered the hyperpolarization of locus coeruleus neurons produced by mu agonists. 2~ More recently, mu agonists have been shown to open potassium channels in cell attached patches of locus coeruleus neurons. 15 Responses were recorded only when the agonist was added to the patch pipette and not when administered on the outside, implying that coupling between the opioid receptor and potassium channels did not involve a freely diffusable intracellular effector system. Although numerous biochemical investigations have demonstrated that opioids decrease adenylate cyclase activity or cyclic A M P production in a variety of neurons, the functional importance of this action has not been established. The work presented here suggests that intracellular cyclic A M P is not involved in coupling mu receptors to potassium channels in myenteric S neurons. Acknowledgements--This work was funded by the National

Health and Medical Research Council of Australia. We would like to thank Mrs Kirsty Dick and Joy Davis for the careful preparation of the manuscript.

REFERENCES

1. Aghajanian G. K. and Wang Y. Y. (1986) Pertussis toxin blocks the outward currents evoked by opiate and alpha2-agonists in locus coeruleus neurons. Brain Res. 371, 390-394. 2. Andrade R. and Aghajanian G. K. (1985) Opiate- and alpha2-adrenoceptor-induced hyperpolarizations of locus coeruleus neurons in brain slices: reversal by cyclic adenosine-3':5'-monophosphate analogues. J. Neurosci. 5, 2359-2364. 3. Collier H. O. J. (1980) Cellular site of opiate dependence. Nature 283, 625 629. 4. Crain S. M., Crain B. and Peterson E. R. (1986) Cyclic AMP or forskolin rapidly attenuates the depressant effects of opioids on sensory-evoked dorsal-horn responses in mouse spinal cord-ganglion explants. Brain Res. 3"/0, 61--72. 5. Helfman D. M. and Kuo J. F. (1982) Differential effects of various phosphodiesterase inhibitors, pyrimidine and purine compounds, and inorganic phosphates on cyclic GMP, cyclic AMP and cyclic GMP phosphodiesterase. Biochem. Pharmac. 31, 43 47. 6. Hirst G. D. S., Holman M. E. and Spence I. (1974) Two types of neurons in the myenteric plexus of duodenum in the guinea-pig. J. Physiol. 236, 303-326. 7. Hosford D. A. and Haigler H. J. (1981) Cyclic AMP, morphine, met-enkephalin and neuronal firing. J. Pharmac. exp. Ther. 219, 496 504. 8. llles P. (1986) Mechanisms of receptor-mediated modulation of transmitter release in noradrenergic, cholinergic and sensory neurons. Neuroscience 17, 909 928. 9. Jeitner T. and Costa M. (1989) Isolation of myenteric ganglia from guinea-pig small intestine. Neurosci. Lett., Suppl. 34, I01. 10. Johnson S. M., Pillai N. P. and Costa M. (1989) Lack of effect of forskolin on the inhibitory actions of (D-Ala2,N-Me-Phe4,Glytol)-enkephalin on myenteric neurons of guinea-pig ileum. Adv. Biosci. 75, 169-172. 11. Johnson S. M., Williams J. T., Costa M. and Furness J. B. (1987) Naloxone-induced depolarization and synaptic activation of myenteric neurons in morphine-dependent guinea pig ileum. Neuroscience 21, 595-602. 12. Karras P. J. and North R. A. (1979) Inhibition of neuronal firing by opiates: evidence against the involvement of cyclic nucleotides. Br. J. Pharrnac. 65, 647-652. 13. Lux B. and Schulz R. (1986) Effect of cholera toxin and pertussis toxin on opioid tolerance and dependence in the guinea-pig myenteric plexus. J. Pharrnac. exp. Ther. 237, 995 I000. 14. Mihara S., North R. A. and Suprenant A. (1987) Somatostatin increases an inwardly rectifying potassium conductance in guinea-pig submucous plexus neurons. J. Physiol. 390, 335 355.

304

S.M. JOHNSON and N. P. PILLAI

15. Miyake M., Christie M. J. and North R. A. (1989) Single potassium channels opened by opioids in rat locus coeruleus neurons. Proc. natn. Acad. Sci. U.S.A. 86, 3419 3422. 16. Morita K. and North R. A. (1982) Opiate activation of potassium conductance in myenteric neurons: inhibition by calcium ion. Brain Res. 242, 145-150. 17. Neher E. (1988) The use of the patch clamp technique to study second messenger-mediated cellular events. Neuroscience 26, 727 734. 18. Nemeth P. R., Palmer J. M., Wood J. D. and Zafirov D. H. (1986) Effects of forskolin on electrical behaviour of myenteric neurons in guinea-pig small intestine. J. Physiol. 376, 439 450. 19. North R. A. and Tonini M. (1977) The mechanism of action of narcotic analgesics in the guinea-pig ileum. Br. J. Pharmac. 61, 541-549. 20. North R. A. and Williams J. T. (1985) On the potassium conductance increased by opioids in rat locus coeruleus neurones. J. Physiol. 364, 265~80. 21. North R. A., Williams J. T., Suprenant A. and Christie M. J. (1987) Mu and delta receptors belong to a family of receptors that are coupled to potassium channels. Proc. natn. Acad. Sci. U.S.A. 84, 5487 5491. 22. Palmer J. M., Wood J. D. and Zafirov D. H. (1986) Elevation of adenosine 3',5'-monophosphate mimics slow synaptic excitation in myenteric neurones of the guinea-pig. J. Physiol. 376, 451 460. 23. Palmer J. M., Wood J. D. and Zafirov D. H. (1987) Purinergic inhibition in the small intestinal myenteric plexus of the guinea-pig. J. Physiol., Lond. 387, 357-369. 24. Schoffelmeer A. N., Wierenga E. A. and Mulder A. H. (1986) Role of adenylate cyclase in presynaptic alpha 2-adrenoceptor- and mu-opioid receptor-mediated inhibition of [3H] : noradrenaline release from rat brain cortex slices. J. Neurochem. 46, 1711-1717. 25. Schramm M. and Selinger Z. (1984) Message transmission: receptor controlled adenylate cyclase system. Science 225, 1350-1356. 26. Schultz J. and Daly J. W. (1973) Accumulation of cyclic adenosine 3',5'-monophosphate in cerebral cortical slices from rat and mouse: stimulatory effect of ct- and fl-adrenergic agents and adenosine. J. Neurochem. 21, 1319-1326. 27. Seamon K. B. and Daly J. W. (1981) Forskolin: a unique diterpene activator of cyclic AMP-generating system. J. Cyclic Nucleotide Res. 1, 201-224. 28. Suprenant A. and North R. A. (1985) #-Opioid receptors and ~t2-adrenoceptors coexist on myenteric but not on submucous neurones. Neuroscience 16, 425-430. 29. Worley P. F., Baraban J. M. and Snyder S. H. (1987) Beyond receptors: multiple second-messenger systems in brain. Ann. Neurol. 21, 217~29. (Accepted 1 February 1990)