Characterization of muscarinic receptor subtypes inhibiting Ca2+ ...

Proc. Nati. Acad. Sci. USA

Vol. 89, pp. 9544-9548, October 1992 Neurobiology

Characterization of muscarinic receptor subtypes inhibiting Ca2+ current and M current in rat sympathetic neurons (whole-cell voltage damp/oxotremorine/pertussis toin/guanine nudeotide binding protein)

LAURENT BERNHEIM*, ALISTAIR MATHIEt, AND BERTIL HILLE* Department of Physiology and Biophysics SJ-40, University of Washington School of Medicine, Seattle, WA 98195

Contributed by Bertil Hille, June 18, 1992

Barlow et al. (14) using 4-diphenylacetoxy-N-methylpiperidine, was established by Hammer et al. (15) in binding studies using the muscarinic receptor antagonist pirenzepine. Five muscarinic receptor subtypes have now been cloned (16) and affinity profiles for numerous semispecific antagonists have been determined (17, 18), providing effective tools for pharmacological subclassification of muscarinic effects. A primary role for M1 receptors in inhibiting IM was suggested by Marrion et al. (19) using pirenzepine and AF-DX 116 in work done before the discovery of all five muscarinic receptor subtypes. We confirm their conclusion and show here that Ml and M4 receptors are involved in the modulation of 'Ca.

ABSTRACT Muscarinic receptors mediating suppression of Ca+ current and of M-type K+ current in rat superior icervical ganglion neurons were sbclassified ph cally by using the muscarinic receptor antagonists pirenzepine and himbacine. Our voltage clamp experiments previously distinguished fast and slow intraceliular signaling pathways coupling muscarinic receptors to calcium channels. We now establish that the fast, pertussis toxin-sensitive suppression of Ca+ current is mediated primarily by muscarinic receptors of the M4 subtype, whereas the slow, bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetate (BAPTA)-sensitive suppression of Ca2+ current is mediated primarily by muscarinic receptors of the Ml subtype. Both actions on Ca2+ current are blocked by guanosine 5'-[13-thio]diphosphate. Muscarnic suppression of M current is slow, BAPTA-sensitive, and mediated by receptors of the Ml subtype. Hence the two musarinic pathways use different receptors and different guanine nucleotide binding proteins to produce different actions on channels.

MATERIALS AND METHODS Cell Preparation and Current Recording. Neurons were dissociated from the SCG of 5- to 6-week-old male SpragueDawley rats (3, 4). They were either kept at 40C and studied within 10 hr or plated on collagen-coated plastic dishes and incubated up to 30 hr in culture medium at 370C. Currents were recorded in the whole-cell configuration of the patchclamp technique (20) at 200C-230C (pipette resistances, 1-2 MU), low-pass filtered at 1 kHz, and sampled at 2 kHz. Potentials have been corrected for junction potentials of -2 mV (0.1 mM BAPTA) or -4 mV (20 mM BAPTA). To study ICa, we used Cs-based pipette solutions. The membrane potential was held at -80 mV and stepped every 4 s to -40 mV for 2.5 ms to reduce capacity transients (4) and then to 0 mV for 10 ms. Peak ICa was measured as the peak current during the step to 0 mV minus the current recorded after addition of 100 uM Cd2+ to the bath to block Ca channels. To study IM, we used a KCl-based pipette solution, held the membrane potential at -25 mV, where IM is activated, and stepped every 5 s to -55 mV for 500 ms to close M channels. Means are given with SEM in the text and figures. PTX and BAPTA Treatments. For PTX experiments, we left neurons 22-30 hr in culture medium with PTX (50 ng/ml) at 370C (5). To load neurons with the Ca chelator BAPTA, we used Cs-based 20 mM BAPTA. Mean series resistance and cell capacitance of acutely dissociated neurons loaded with BAPTA were 4.3 + 0.1 MU and 31 ± 2 pF (n = 69). Whole-cell dialysis for at least 5 min was allowed before drug application. We estimate that after 5 min of dialysis neurons may have contained -11 mM BAPTA (3). In six neurons treated with PTX and then loaded with high BAPTA, mean series resistance and cell capacitance were 4.0 ± 0.6 MU and

Ionic channels are affected by many neurotransmitters in peripheral and central neurons. In sympathetic neurons from the superior cervical ganglion (SCG) of the rat, activation of muscarinic acetylcholine receptors inhibits Ca2+ currents (ICa) and a K+ current (1, 2). Our previous work in SCG neurons defined two intracellular pathways coupling muscarinic receptors to Ca channels (3-5). One, a slow process, is disrupted by high intracellular concentrations of the Ca2+ chelator bis(2-aminophenoxy)ethane-N, N, N', N '-tetraacetate (BAPTA) and involves a diffusible second messenger. The other, a fast process, is not sensitive to BAPTA, does not use a diffusible second messenger, and is greatly reduced by pertussis toxin (PTX). Whole-cell ICa is carried by w-conotoxin-GVIA-sensitive N-type Ca channels and dihydropyridine-sensitive L-type Ca channels in rat SCG neurons (6). About 85% of the current is carried by N-type channels (7). Using whole-cell and single-channel recording, we have shown that the slow muscarinic modulation affects both Nand L-type Ca channels, whereas the fast modulation inhibits only N-type Ca channels (ref. 8; but see also ref. 9). A timeand voltage-dependent K+ current called M current (IM) is suppressed by muscarinic agonists in frog sympathetic cells and in many other neurons (10, 11). In rat SCG neurons, the muscarinic suppression of IM is BAPTA sensitive (3), PTX insensitive (12), and slow (ref. 4; see also ref. 13). Activation of muscarinic receptors therefore has at least three effects on channels of rat SCG neurons: a fast PTXsensitive suppression of 'Ca, a slow BAPTA-sensitive suppression of Ich, and a slow BAPTA-sensitive suppression of IM. This paper investigates whether all these actions are mediated by the same muscarinic receptors. The existence of different subtypes of muscarinic receptors, first suggested by

Abbreviations: SCG, superior cervical ganglion; BAPTA, bis(2aminophenoxy)ethane-N,N,N',N'-tetraacetate; PTX, pertussis toxin; oxo-M, oxotremorine methiodide; GDPLBS], guanosine 5'-(#thioldiphosphate; G protein, guanine nucleotide binding protein. *Present address: Departement de Physiologie, Centre M6dical Universitaire, 9 avenue de Champel, 1211 Geneva 4, Switzerland. tPresent address: Department of Pharmacology, Royal Free Hospital School of Medicine, Rowland Hill Street, London NW3 2PF, United Kingdom. tTo whom reprint requests should be addressed.

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.

9544

Proc. Natl. Acad. Sci. USA 89 (1992)

Neurobiology: Bemheim et al. 69 ± 7 pF. In these neurons, 12 min of dialysis was allowed because cells are larger after 1 day of culture (3). Schild Plots. Antagonist affinities were analyzed by the method of Arunlakshana and Schild (21). Antagonists were perfused in the bath for at least 5 min before testing with the muscarinic agonist, oxotremorine methiodide (oxo-M). Dose ratios are ratios of oxo-M concentrations that elicit an equal suppression of ICa or IM in the presence or absence of antagonist. To calculate dose ratios, concentrations of oxo-M giving a certain percentage inhibition of current (e.g., 25% in Fig. 3B) were obtained from Hill equations fitted to the control mean data points. In the presence of antagonist, the concentration of oxo-M required to produce the same percentage inhibition in each cell was obtained by interpolation from the closest two data points. They were calculated for each individual cell, plotted as a Schild plot, and unconstrained or constrained-unity slope lines were fitted using the simplex algorithm (22). Intercepts of these lines yield empirical measures of antagonist affinities, pA2 and pKB, as described and defined by Jenkinson (23), which can be compared to published affinities pKB estimated from binding and functional studies in the literature. Confidence limits (95%) were calculated as described by Goldstein (24). Materials and Solutions. Chemical sources and solutions were as described (5) except for pirenzepine and guanosine 5'-[(3-thioldiphosphate (GDPLBS]) (Sigma) and himbacine (gift from W. C. Taylor, Sydney, Australia). Solutions were continuously perfused at 2-2.5 ml/min through a 100-lId chamber and were fully exchanged in -20 s. External solution: 150 mM NaCl/2.5 mM KCl/5 mM CaCI2/1 mM MgCl2/10 mM Hepes/8 mM glucose/1 AM propranolol/250500 nM tetrodotoxin, pH 7.4. Pipette solutions: control, 155 mM CsCl or KCl/5 mM MgCl2/5 mM Hepes/0.1 mM BAPTA; high BAPTA, 115 mM CsCl/5 mM MgCl2/5 mM Hepes/20 mM BAPTA; 3 mM Na2ATP, 0.1 mM Na3GTP, and 80 AuM leupeptin were added fresh each day (pH 7.4).

RESULTS Muscarinic Suppression of Ica and IM. Application of oxo-M at increasing concentrations produces an increasing inhibition of ICa and IM. Fig. 1A shows peak ICa at 0 mV and representative current traces recorded as incremental concentrations of oxo-M (0.1-10 AM) were applied by bath

perfusion. With this cumulative dose-response protocol, 10 ,uM oxo-M suppressed ICa by 83% ± 3% (n = 12), virtually the same (84% ± 2%; n = 25; ref. 4) as when 10 ,LM oxo-M was tested in single applications. The current traces in Fig. 1B show IM recorded during voltage steps to -55 mV from a holding potential of -25 mV while oxo-M was bath perfused. Fig. lB, plots the standing outward current at -25 mV and Fig. 1B2 plots the mean current 10-20 ms after the beginning of the hyperpolarizing voltage step to -55 mV minus the current at the end ofthe voltage step. This relaxing component of current at -55 mV is our quantitative measure of IM. Oxo-M (10 AM) suppressed IM by 88% ± 2% (n = 13). Modulatory Pathways. Intracellular BAPTA and treatments with PTX can be used to distinguish two parallel signaling pathways coupling muscarinic receptors and Ca channels (4, 5). Thus, loading with BAPTA or treating with PTX attenuates the muscarinic response only partially (Fig. 2A); the mean suppression of ICa by 10 jLM oxo-M in control, BAPTA-loaded, and PTX-treated cells is 84% + 2%, 47% ± 31%, and 58% + 4%, respectively. But when both treatments are used simultaneously (Fig. 2A), muscarinic modulation is almost abolished (suppression of only 5% ± 2%), suggesting that the coupling between muscarinic receptors and Ca channels is primarily BAPTA sensitive or PIX sensitive. Only one pathway is PTX sensitive, but both seem to involve GTP-binding regulatory proteins (G proteins), as muscarinic modulation is virtually abolished (suppression of 7% ± 3%) in cells dialyzed with 1 mM GDP[I8S]. Fig. 2B shows dose-response relations for suppression of ICa by oxo-M in control, BAPTA-loaded, and PTX-treated cells. Continuous lines are fitted Hill equations (data points for BAPTA-loaded and PTX-treated cells are given in Figs. 3 and 5). Fig. 2B (curve mul.) shows a predicted doseresponse relation for the control condition, assuming that the individual BAPTA-sensitive and PTX-sensitive actions combine multiplicatively-i.e., that each acts independently to diminish the effective conductance of the same pool of channels (5). Surprisingly, this prediction underestimates the actual extent of suppression at low and middle oxo-M concentrations, suggesting a positive cooperativity between the two pathways. Fast, BAPTA-Resistant Muscarinic Suppression of Ica Is Mediated by Receptors of the M4 Subtype. Effects of the muscarinic receptor antagonists pirenzepine and himbacine

A

0 mV mV

-400 %..O

-

-800 -1 200

500 pA

__

0.1 0.3

Bi

j _

1

5 ms

10

600 F 400 F 200

-55 mV ---w-

B2

Z I

9545

-

300 F 200

100

FIG. 1. Muscarinic inhibition of Ica and IM in SCG neurons. (A) (Left) Peak inward currents (primarily ICa, but not corrected for current remaining in Cd2+) elicited in steps to 0 mV and plotted against time. Oxo-M (0.1-10 ,uM) and Cd2+ (100 &M) were bath applied as indicated. (Right) Representative leak-subtracted (by subtracting Cd2+ record) ICa traces in control conditions and during application of 0.1, 0.3, 1, 3, and 10 ,uM oxo-M. Cell mb401. (B,) (Left) Standing outward current (primarily IM) recorded at a -25 mV holding potential of -25 mV plotted against time. (B2) (Left) Amplitude of time-dependent current relaxation (IM) during steps to -55 mV. Oxo-M (0.1-10 ,uM) was applied as indicated. (B) (Right) Representative IM traces in control condition and during application of 0.1, 0.3, 1, 3, and 10 uM oxo-M. 200 pA Cell mb442. Dashed lines indicate zerocurrent levels and short horizontal lines 200 ms indicate current levels used in calculating dose-response relations.

j

0

400

200 time (S)

6S00

9546

Proc. Natl. Acad Sci. USA 89 (1992)

Neurobiology: Bernheim et al.

B loo

A loo p 80 s 60 N 40 t 20 0

~-j 5

jIi
11 for each data point). Fitted Hill equation has a maximum of 52%, EC50 of 1.2 tM, and Hill coefficient of 0.97. (A) (Left) Competitive shift of oxo-M doseresponse curve by 10 nM (n = 7), 100 nM (n = 7), and 1 ,uM (n = 6) pirenzepine. (Right) Schild plot for antagonism of the oxo-M inhibition by pirenzepine (pz). Dose ratios (dr) were calculated for 20% inhibition of Ica- Slope of the unconstrained fitted line is 0.9 (solid line), and curves define the 95% confidence interval. Dashed line shows the fit with slope constrained to 1. (B) (Left) Competitive shift of oxo-M dose-response curves by 10 nM (n = 6), 100 nM (n = 7), and 1 ,uM (n = 7) himbacine. (Right) Schild plot for antagonism of oxo-M by himbacine (him). Dose ratios were calculated for 25% inhibition of Ica. Slope of unconstrained fitted line is 0.9.

9

MI M2 M3 M4 MS CaB CaP KM FIG. 4. Antagonist potencies of pirenzepine and himbacine compared. Values for pirenzepine are solid lines and solid circles. Values for himbacine are dashed lines and open circles. M1-M5, average values and range of pKB from binding and functional studies in the literature (17, 18, 25-28) for the five muscarinic subtypes. Measurements were on mammalian cells at room temperature. On average n = 4 for pirenzepine and n = 2 for himbacine. CaB, CaP, and KM, average values and 95% confidence intervals (23) of pA2 values determined in our experiments from Schild plots. Antagonism of modulation of Ica with BAPTA in the pipette (CaB), of ICa after PTX (CaP), and of IM (KM).

Neurobiology: Bemheim et al. A 60

Proc. Natl. Acad. Sci. USA 89 (1992)

A1400_ a' 80-

3

40

i

*0

10-7 104 1 0- 10-4 oxo-M (M)

Ba 60 *0

O--

0

I -1

3 Ck

0

20

a

-1

oxo-M (w 00 g0 60 40 :20

-1

3

Im w4

-

2

I§ 1

-

0

0

10-7 10-6 10-5 10-4 oxo-M (M)

0

104 10-7 10-6 10-S 10-4

*3

8 40 m

3

6040:20-_

ne 20

.,4

9547

-8 -7 -6 -5 -4 log him (M)

FIG. 5. Muscarinic inhibition of Ica in neurons treated 22-30 hr in culture medium supplemented with PTX (50 ng/ml). A, Mean suppression of ICa in the absence of antagonist (n = 11). Fitted Hill equation has a maximum of62%, EC5o of 1.1 !4M, and Hill coefficient of 1.18. (A) Shift of the oxo-M dose-response curve by 10 nM (n = 13), 100 nM (n = 7), and 1 AM (n = 5) pirenzepine (pz) and corresponding Schild plot. Dose ratios were calculated for 25% inhibition of ICa and the slope of the unconstrained fitted line is 0.8. (B) Shift of the oxo-M dose-response curve by 100 nM (n = 6), 1 ,uM (n = 7), and 10 ,uM (n = 7) himbacine (him) and corresponding Schild plot. Dose ratios were calculated for 25% inhibition of ICa and slope of the unconstrained fitted line is 0.9.

confidence limits, 8.0-9.1; n = 25) and an estimated pKB value of 8.2. A Schild plot of the himbacine shift gave a pA2 value of 7.2 (95% confidence limits, 6.6-7.7; n = 20) and an estimated pKB value of 7.1. The high affinity for pirenzepine suggests an M1 receptor subtype, which is confirmed by the low affinity for himbacine. These observations indicate that the receptors coupled to the slow, PTX-resistant BAPTAsensitive pathway are of the M1 subtype. Muscarinic Suppression of Im Is Also Mediated by Receptors of the Ml Subtype. Fig. 6 shows studies of Im in neurons neither loaded with BAPTA nor treated with PTX. The dose-response curve for suppression of IM by oxo-M had an EC50 of 0.3 AuM and a maximum inhibition of 93%. The Schild plots gave a pA2 value of 8.0 (95% confidence limits, 7.8-8.2; n = 19), an estimated pKB value of 7.8 for pirenzepine and a pA2 value of 7.2 (95% confidence limits, 6.9-7.6; n = 20), and an estimated pKB value of 6.9 for himbacine. These results, very similar to those obtained for the slow PTX-resistant BAPTA-sensitive suppression of 'Ca, again point toward an M1 receptor subtype.

DISCUSSION We have asked whether more than one muscarinic receptor underlies channel modulation in SCG cells because transduction of muscarinic stimulation seems to use multiple G proteins. The transduction mechanism coupling muscarinic receptors and M channels of sympathetic neurons remains to be understood (refs. 3, 31, and 32; but see ref. 33), but it is slow (4, 13), sensitive to intracellular BAPTA (3), and uses a PTX-insensitive G protein (12, 34). Similarly, one pathway coupling muscarinic receptors to suppression of both N- and L-type Ca channels is slow, sensitive to intracellular BAPTA, and insensitive to PTX (3-5, 8). Another pathway, acting only on N-type Ca channels, is faster, insensitive to intracellular BAPTA, and abolished by PTX (1, 3-5, 8). The cDNAs encoding five different muscarinic receptor subtypes have been identified (29, 35-37). In general, m1, M3,

10-8 10-7 10-6 1 0-5 10-4

oxo-M (M)

-1

-8 -7 -6 -5 -4

log him(M)

FiG. 6. Muscarinic inhibition of IM. Neurons kept for 22-30 hr in culture medium at 370C. A, Mean suppression of IM in the absence of antagonist as in Fig. 1B2 (n = 13). Fitted Hill equation has a maximum of 93%, EC5o of 0.3 AtM, and Hill coefficient of 1.03. (A) Shift of oxo-M dose-response curve by 10 nM (n = 6), 100 nM (n = 6), and 1 ,uM (n = 7) pirenzepine (pz) and corresponding Schild plot. Dose ratios were calculated for 50% inhibition of IM and slope of the unconstrained fitted line is 0.8. (B) Shift of oxo-M dose-response curve by 100 nM (n = 7), 1 A&M (n = 7), and 10 A&M (n = 6) himbacine (him) and corresponding Schild plot. Dose ratios were calculated for 50%o inhibition of IM and slope of the unconstrained fitted line is 0.8.

and m5 code for receptors regulating inositol phospholipid hydrolysis through a PTX-insensitive G protein, whereas m2 and m4 code for those linked to adenylyl cyclase inhibition through a PTX-sensitive G protein (16, 38, 39). According to Hulme et al. (16), most autonomic ganglion cells coexpress M1, M2, M3, and m4 genes, but, using subtype-specific antisera, Dorje et al. (40) found only ml, M2, and, in smaller amounts, m4 receptor protein subtypes in rabbit sympathetic ganglia. The combined use of pirenzepine and himbacine in our work demonstrated that the slow, PTX-insensitive suppression of IM and ICa is primarily due to activation of receptors of the M1 subtype, whereas the fast, PTX-sensitive modulation ICa is primarily due to activation of receptors of the M4 subtype. Hence, there is no need here to postulate that one receptor subtype activates two quite different G proteins. Pirenzepine and himbacine are both established competitive antagonists at muscarinic receptors and will cause a parallel shift to the right of agonist dose-response curves with no reduction in the maximum response attainable. This holds true for all the results in this study with the possible exception of the effect of pirenzepine on BAPTA-resistant inhibition of ICa (Fig. 3A). Here, at the highest concentration (1 ,uM), there is a suggestion that pirenzepine's action may deviate from that expected of a normal competitive antagonist. Omitting these data from the Schild analysis makes no significant difference to the pA2 and pKB values obtained. The pharmacological profile of muscarinic receptors responsible for the fast inhibition of ICa in rat SCG is very similar to the pharmacological profile of those responsible for inhibition of ICa in neuroblastoma-glioma hybrid (NG108-15) cells, where the inhibition has been attributed to M4 muscarinic receptors (28). In NG108-15 cells transfected with muscarinic receptors, both m2 and m4 subtypes are capable of inhibiting ICa but neither ml nor m3 couples well (41). It is surprising that we have implicated M4 but not M2 receptors in the fast inhibition of ICa in rat SCG, as m2 receptors can inhibit ICa as well as m4 receptors (41) and m2 receptor subtype proteins are present in rabbit sympathetic ganglia (40).

9548

Neurobiology: Bernheim et al.

Possibly this discrepancy reflects species differences, as we do not know which receptors are expressed at the protein level in the rat SCG. Alternatively, functional m2 receptor subtypes may not be present in all SCG neurons and the dissociation procedure may select a population of neurons without them. A third possibility would be that enzyme treatment during the dissociation procedure has damaged m2 receptors. There is another suggestion of functional M4 receptors in sympathetic neurons. Eltze (42) studied the antagonist affinities of muscarinic receptors mediating a presynaptic inhibition of sympathetic nerve action on rabbit vas deferens, finding a moderate to high affinity for pirenzepine and a high affinity for himbacine. Although Eltze originally concluded that the receptors have the M1 subtype, newer work would suggest that this profile is that of an M4 receptor (28). Our assignment of M1 receptors to inhibition of IM agrees with other evidence. In transfected NG108-15 cells, activation of either ml or m3 receptors inhibits IM-the inhibition with m3 receptors was full versus 50% with ml (43). The inhibition was mediated by a PTX-insensitive G protein, as in sympathetic neurons. No inhibition was seen in m2- or m4-transfected cells or in control NG108-15 cells, which have endogenous M4 receptors. At least in the rabbit, SCG neurons do not seem to express m3 receptors (40). In rat, we find that stimulation of M1 receptors suppresses IM by -900%. What physiological roles can be attributed to these Ml and M4 receptors? N-type Ca channels seem to play a major role in the release of norepinephrine by sympathetic neurons (44), and trans-synaptic inhibition of norepinephrine secretion by acetylcholine released from parasympathetic fibers has been suggested (45). It is possible that this inhibitory mechanism at nerve terminals is activated by M4 receptors, which couple solely to N-type Ca channels (this study and ref. 8); however, the relative abundance of the various muscarinic receptor subtypes at sympathetic nerve terminals remains unknown. On the other hand, M1 and M4 receptors, both said to be present on the cell body, should be able to affect cell

excitability, calcium-dependent enzyme activity, or gene expression (46). Activation of M1 receptors, which inhibit three different channel types through a slow process involving a diffusible second messenger, may affect neuronal excitability more profoundly than activation of the M4 receptors directed only toward N-type Ca channels. We thank Prof. D. A. Brown and Drs. M. Caulfield, J. B. Herrington, K. Mackie, Y. B. Park, M. S. Shapiro, A. Tse, F. W. Tse, and L. P. Wollmuth for reading the manuscript and D. Anderson and L. Miller for technical help. We also thank Prof. W. C. Taylorfor the generous gift of himbacine. This work was supported by National Institutes of Health Grant NS08174, a McKnight Neuroscience Research Award, fellowships from La Fondation Suisse de Bourses en M6ddcine et Biologie and the Muscular Dystrophy Association, and a Fogarty International Research Fellowship (F05-TW04457). 1.

Wanke, E., Ferroni, A., Malgaroli, A., Ambrosini, A., Pozzan,

Nati. Acad. Sci. USA 84, 4313-4317. Constanti, A. & Brown, D. A. (1981) Neurosci. Lett. 24, 289-294. Beech, D. J., Bernheim, L., Mathie, A. & Hille, B. (1991) Proc. Natd. Acad. Sci. USA 88, 652-656. Bernheim, L., Beech, D. J. & Hille, B. (1991) Neuron 6, 859-867. Beech, D. J., Bernheim, L. & Hille, B. (1992) Neuron 8, 97-106. Plummer, M. R., Logothetis, D. E. & Hess, P. (1989) Neuron 2, 1453-1463. Regan, L. J., Sah, D. W. Y. & Bean, B. P. (1991) Neuron 6, 269-280. Mathie, A., Bernheim, L. & Hille, B. (1992) Neuron 8, 907-914. Plummer, M. R., Rittenhouse, A., Kanevsky, M. & Hess, P. (1991) J. Neurosci. 11, 2339-2348. T. & Meldolesi, J. (1987) Proc.

2. 3. 4.

5. 6. 7.

8. 9.

Proc. Nad. Acad Sci. USA 89 (1992) 10. Brown, D. A. & Adams, P. R. (1980) Nature (London) 283, 673-676. 11. Brown, D. A. (1988) Trends Neurosci. 11, 294-299. 12. Brown, D. A., Marnion, N. V. & Smart, T. G. (1989) J. Physiol. (London) 413, 469-488. 13. Jones, S. W. (1991) Biophys. J. 60, 502-507. 14. Barlow, R. B., Berry, K. J., Glenton, P. A. M., Nikolaou, N. M. & Soh, K. S. (1976) Br. J. Pharmacol. 58, 613-620. 15. Hammer, R., Berrie, C. P., Birdsall, N. J. M., Burgen, A. S. V. & Hulme, E. C. (1980) Nature (London) 283, 90-92. 16. Hulme, E. C., Birdsall, N. J. M. & Buckley, N. J. (1990) Annu. Rev. Pharmacol. Toxicol. 30, 633-673. 17. Lazareno, S., Buckley, N. J. & Roberts, F. F. (1990) Mol. Pharmacol. 38, 805-815. 18. Dorje, F., Wess, J., Lambrecht, G., Tacke, R., Mutschler, E. & Brann, M. R. (1991) J. Pharmacol. Exp. Ther. 256, 727-733. 19. Marrion, N. V., Smart, T. G., Marsh, S. J. & Brown, D. A. (1989) Br. J. Pharmacol. 98, 557-573. 20. Hamill, 0. P., Marty, A., Neher, E., Sakmann, B. & Sigworth, F. J. (1981) Pflfigers Arch. 391, 85-100. 21. Arunlakshana, 0. & Schild, H. 0. (1959) Br. J. Pharmacol. 14, 48-58. 22. Caceci, M. S. & Cacheris, W. P. (1984) Byte 9, 340-362. 23. Jenkinson, D. H. (1991) Trends Pharmacol. Sci. 12, 53-54. 24. Goldstein, A. (1964) Biostatistics: An Introductory Text (Macmillan, New York), pp. 135-143. 25. Fukuda, K., Kubo, T., Akiba, I., Maeda, A., Mishina, M. & Numa, S. (1987) Nature (London) 327, 623-625. 26. Buckley, N. J., Bonner, T. I., Buckley, C. M. & Brann, M. R. (1989) Mol. Pharmacol. 35, 469-476. 27. Akiba, I., Kubo, T., Maeda, A., Bujo, H., Nakai, J., Mishina, M. & Numa, S. (1988) FEBS Lett. 235, 257-261. 28. Caulfield, M. P. & Brown, D. A. (1991) Br. J. Pharmacol. 104, 39-44. 29. Bonner, T. I., Buckley, N. J., Young, A. C. & Brann, M. R. (1987) Science 237, 527-532. 30. Anwar-ul, S., Gilani, H. & Cobbin, L. B. (1986) NaunymSchmiedeberg's Arch. Pharmacol. 332, 16-20. 31. Bley, K. R. & Tsien, R. W. (1990) Neuron 2, 379-391. 32. Bosma, M. M., Bernheim, L., Leibowitz, M. D., Pfaffinger, P. J. & Hille, B. (1990) in G-Proteins and Signal Transduction, Society ofGeneral Physiologists Series, eds. Nathanson, N. M. & Harden, T. K. (Rockefeller Univ. Press, New York), Vol. 45, pp. 43-59. 33. Kirkwood, A., Simmons, M. A., Mather, R. J. & Lisman, J. (1991) Neuron 6, 1009-1014. 34. Pfaffinger, P. (1988) J. Neurosci. 8, 3343-3353. 35. Kubo, T., Fukuda, K., Mikami, A., Maeda, A., Takahashi, H., Mishina, M., Haga, T., Haga, K., Ichiyama, A., Kangawa, K., Kojima, M., Matsuo, H., Hirose, T. & Numa, S. (1986) Nature (London) 323, 411-416. 36. Kubo, T., Maeda, A., Sugimoto, K., Akiba, I., Mikami, A., Takahashi, H., Haga, T., Haga, K., Ichiyama, A., Kangawa, K., Matsuo, H., Hirose, T. & Numa, S. (1986) FEBS Lett. 209, 367-372. 37. Bonner, T. I., Young, A. C., Brann, M. R. & Buckley, N. J. (1988) Neuron 1, 403-410. 38. Fukuda, K., Higashida, H., Kubo, T., Maeda, A., Akiba, I., Bujo, H., Mishina, M. & Numa, S. (1988) Nature (London) 335, 355-358. 39. Peralta, E. G., Ashkenazi, A., Winslow, J. W., Ramachandran, J. & Capon, D. J. (i988) Nature (London) 334, 434-437. 40. D6rje, F., Levey, A. I. & Brann, M. R. (1991) Mol. Pharmacol. 40, 459-462. 41. Higashida, H., Hashii, M., Fukuda, K., Caulfield, M. P., Numa, S. & Brown, D. A. (1990) Proc. R. Soc. London Ser. B 242, 68-74. 42. Eltze, M. (1988) Eur. J. Pharmacol. 151, 205-221. 43. Robbins, J., Caulfield, M. P., Higashida, H. & Brown, D. A. (1991) Eur. J. Neurosci. 3, 820-824. 44. Hirming, L. D., Fox, A. P., McCleskey, E. W., Olivera, B. M., Thayer, S. A., Miller, R. J. & Tsien, R. W. (1988) Science 239, 57-61. 45. Langer, S. Z. (1981) Pharmacol. Rev. 32, 337-362. 46. Murphy, T. H., Worley, P. F. & Baraban, J. M. (1991) Neuron 7, 625-635.