Nicotinic Receptors of Frog Ganglia Resemble Pharmacologically Those of Skeletal Muscle

The Journal of Neuroscience, September 1988, 8(9): 3258-3265

Diane Lipscombe and Humphrey P. Rang” M.R.C. Receptor Mechanisms Research Group, Department of Pharmacology, University College London, London WClE 6BT, United Kingdom

The actions of ACh antagonists were studied on synaptic the amino acid sequencesof their a-subunits (Boulter et al., currents of autonomic ganglia of the frog. Fast excitatory 1986) and, more generally, their different sensitivitiesto various synaptic currents (ESCs) were recorded from cardiac and ACh antagonists.For example, ol-bungarotoxin (cu-BuTx) is an paravertebral neurons with the use of the 2-microelectrode irreversible inhibitor of synaptic transmission at the neuro- voltage-clamp method. The actions of 4 ACh antagonists, muscularjunction (Miledi and Potter, 197l), whereasit is with- tubocurarine, hexamethonium, trimetaphan, and decame- out effect in several ganglion preparations, including the rat thonium were studied. Tubocurarine was effective at reduc- superior cervical ganglion (Brown and Fumagalli, 1977), guinea ing the peak amplitude of ESCs (50% inhibition at 3 PM). In pig myenteric plexus (Bursztajn and Gershon, 1977), and rat contrast, tubocurarine (l-30 MM) reduced the time constant submandibular ganglion (Ascher et al., 1979). Neuronal and of ESC decay by only 9% compared with controls. Both of muscleACh receptor-channel complexesalso differ in their sen- these effects of tubocurarine were independent of mem- sitivity to ion channel block. Hexamethonium blocks synaptic brane potential. Hexamethonium was a weak inhibitor of transmissionin the superior cervical ganglion of the cat (Paton ESCs; at 600 WM peak amplitude was reduced only to about and Zaimis, 1951; Lee and Nishi, 1972) and submandibular 60% of controls and decay time constants were unaffected ganglion of the rat (Rang, 1982) but has little effect on trans- at concentrations between 10 and 600 PM. These effects of missionat the neuromuscularjunction (Milne and Byrne, 1981). tubocurarine and hexamethonium are consistent with these Ion channel block is the main mechanismby which hexame- drugs being receptor antagonists with no evidence of ion thonium, and certain other drugs, blocks transmission in the channel block. Trimetaphan (3-100 PM) and decamethonium rat submandibular ganglion (Ascher et al., 1979; Rang, 1982; (100 PM) reduced the peak amplitude of ESCs. In the pres- Gurney and Rang, 1984) and rabbit sympathetic ganglia(Skok ence of 100 PM trimetaphan or 10 PM decamethonium, ESC et al., 1983; Skok, 1986), whereas at the end-plate receptor decays were biexponential. The 2 exponential components antagonismis, in general, a more prominent mechanismof in- induced by the presence of these drugs were faster and hibition. slower, respectively, than the single-exponential component There is also good evidence that more than one type of neu- of control ESC decays. The effects of these 2 drugs were ronal nicotinic receptor exists. a-BuTx has been found to block more pronounced at hyperpolarized potentials and are con- cholinergic transmission at certain vertebrate synapsessuch as sistent with a channel-blocking action. The actions of the 4 the retinotectal synapseof the toad (Freeman, 1977)and goldfish ACh antagonists on frog autonomic ganglia are similar to (Freeman et al., 1980), in the chick ciliary ganglion (Chiappinelli their effects at the neuromuscular junction but dissimilar to and Zigmond, 1978; Ravdin and Berg, 1979) and in sympa- their effects on the rat submandibular ganglion. We conclude thetic ganglia of the bullfrog (Marshall, 1981). Diversity in the that the ACh receptor-ion channel complex present on gan- neuronal nicotinic receptor is also implied by the isolation of glia on the frog is pharmacologically similar to that at the distinct cDNA clones encoding different a-subunits in various end-plate and represents a distinct class of neuronal nico- regions of the mammalian brain (seeGoldman et al., 1987). tinic receptor responsible for mediating fast excitatory trans- We were interestedto determine whetherthe ability of a-BuTx mission. to block synaptic responsesin certain ganglia might be due to the presenceof a classof nicotinic receptor different from that mediating fast excitatory transmission in mammalian ganglia. The nicotinic ACh receptor found on mammalian neurons is Others have suggestedthat the ability of a-BuTx to block syn- clearly distinct, both pharmacologically and kinetically, from aptic transmission in some ganglia may depend on the unde- the nicotinic receptor of skeletalmuscle (seeColquhoun, 1981). tected presenceof different amountsof 2 other a-toxins, referred The nicotinic receptor of neuronsdiffers from that of musclein to as fractions 3.1 and 3.3, which appear to be more effective inhibitors of synaptic transmissionin chick ciliary gangliathan oc-BuTx (Ravdin and Berg, 1979). a-Toxins from fraction 3 can Received July 20, 1987; revised Dec. 10, 1987; accepted Feb. 8, 1988. also inhibit synaptic transmissionat the neuromuscularjunction We thank the M.R.C. for providing the financial support for this study. We are (Lee et al., 1972). most grateful to David Colquhoun for critical reading of the manuscript. Correspondence should be addressed to Diane Lipscombe, Ph.D., Yale Uni- We chose to characterize the action of certain ACh antago- versity, Department of Cellular and Molecular Physiology, School of Medicine, nists, already extensively studied in mammalian ganglia (rat 333 Cedar Street, New Haven, CT 06510. = Present address: Sandoz Institute London, Gower Place, London WClE 6BT, submandibular: Ascher et al., 1979; Rang, 1982; Gurney and U.K. Rang, 1984; rabbit sympathetic: Skok et al., 1983; Skok, 1986), Copyright 0 1988 Society for Neuroscience 0270-6474/88/093258-08$02.00/O in frog ganglia, which from studieswith ol-BuTx were likely to The Journal of Neuroscience, September 1988, B(9) 3259 have a different type of neuronal nicotinic receptor. We show its binding to receptor, an ACh receptor antagonist (Jenkinson, 1960), that the actions of ACh antagonists on synaptic currents re- or by blocking the open channel, an ion channel blocker (Adams, 1976), and these 2 mechanisms have predictable and distinct effects on synaptic corded from frog ganglia are indeed different from their actions currents (see, for example, Colquhoun et al., 1979; Rang, 1982). Open previously reported on mammalian ganglia, and more similar channel block can be represented according to the following simplified to their actions at the neuromuscular junction. kinetic scheme (Adams, 1976; Colquhoun and Hawkes, 1983): A preliminary account of this work has been published in abstract form (Lipscombe and Rang, 1984). VW+, Shuts ol, Own CkTB B1ocked, Materials and Methods Preparation and recording. Cardiac and paravertebral ganglia were re- where p’ is the effective forward rate constant which increases with the moved from frogs (Rana pipiens) and prepared for intracellular record- concentration of ACh (xA); l/a’ is the apparent mean open lifetime of ing as described by others (sympathetic: Nishi and Koketsu, 1960; car- the channel in the absence of a channel blocker, so (Y’ is the apparent diac: McMahan and Kuffler, 1971). Ganglia were pinned to a Sylgard rate constant for channel closing from the open state; [B] is the con- resin-coated recording chamber. Preganglionic stimulation was achieved centration of the blocker; and k,, and k-, are, respectively, the asso- via a suction electrode into which the vagosympathetic nerve branches ciation and dissociation rate constants for blocking and unblocking of of the cardiac ganglion-or in the case of the sympathetic ganglia, the the open channel by B. paravertebral chain above ganglion 7 -were secured. Only B-type neu- Strictly, this form is valid provided that the agonist binding reaction rons are excited by stimulating the chain anterior to the seventh ganglion (not shown explicitly above) is much faster than the subsequent (Skok, 1965: Libet et al.. 1968). Individual neurons were identified with shut-open conformation change. This is probably not true, however, a Nomarski’optical system and a x 40 water-immersion objective. Re- for the nicotinic receptor, where a burst of several channel openings, cording methods were essentially the same as those described by Ascher separated by brief shuttings, occurs during a single activation of the et al. (1979) and Rang (198 1). Nerve-evoked synaptic currents were channel (Colquhoun and Sakmann, 198 1). However, the fast binding elicited by suprathreshold stimuli between 0.2 and 0.5 msec in duration assumption will hold as a good approximation even when there are delivered from a Grass stimulator (S44) and isolation unit. Synaptic multiple channel openings as long as l/cu’ is interpreted as the mean currents were recorded from individual neurons with the use of the length of a single channel activation (the mean burst length; see Ogden 2-microelectrode voltage-clamp method. Electrodes (40-60 Mn) were and Colquhoun, 1985, Appendix 2 for details). The form of the depen- filled with 4 M potassium acetate solution for voltage recording and 0.5 dence of p’ on agonist concentration depends on the details of the binding M potassium sulfate solution for current injection. After impalement reaction, but it is not critical for the results presented here. with 2 electrodes, neurons were held under voltage clamp at a potential In the absence of a blocker, synaptic current decays are monoexpo- close to the resting potential, generally between -40 and -80 mV. nential and LY’approximates to the rate constant (kJ or reciprocal time When necessary, the voltage was stepped to the required potential (usu- constant (l/r,) of synaptic current decays, provided that ACh is not ally between -30 and - 100 mV) for about 500 msec, prior to pregan- present during the decay (i.e., p’ = 0) and that channel opening occurs glionic stimulation. The frequency response of the voltage clamp was synchronously and much faster than channel closing (Magleby and Ste- optimized for each cell by adjusting the clamp gain and the negative vens, 1972a, b). In the presence of a channel blocker, B, the synaptic capacitance feedback to the input stage. A driven shield was also used current decay, is predicted to show 2 exponential components, the rate to screen the voltage electrode in some recordings. Currents settled constants k, and k, being, respectively, faster and slower than the control within about 2-2.5 msec in response to a voltage step. For recording rate constant, k,. The faster rate constant, k,, will be given approximately synaptic currents, the clamp gain was increased as much as possible by (a’ + [B]k+, + k-,) as long as there is a small fraction of channels without precipitating oscillations and adjusted when necessary during open (p’ < 01) and the 2 observed rate constants are well separated (k, the experiment. However, even with the clamp gain optimally adjusted > k,). If the.rate of dissociation of B is slow compared with k, (i.e., there was often a small voltage escape during the synaptic response. k_, (< k,). then k, = (01’ + IBlk,,). Thus. k, increases as IBl is increased. With the membrane potential clamped at -80 mV, this error did not When B’dissociates from the channel, the’channel ret&s to the open exceed 2-3 mV and no correction was made for it. state before finally shutting, and this produces a slow tail, the rate Analysis. Voltage and current signals were continually viewed on an constant for which is denoted k,. oscilloscope and stored for later analysis on magnetic tape (Racal Store On the basis of this simplified kinetic scheme, an open channel blocker 4) the current signal having first been low-pass-filtered between 300 will have no effect on ESC amplitudes (in so far as channels are opened and 600 Hz (-3 dB). Analysis was carried out on computer (PDP 1 l/ synchronously), since the channels must open before the blocker can 73) and currents were digitized, usually at 1024 Hz. Sampling by the act, but the blocker will speed the initial rate of synaptic current decay. computer was initiated by a stimulus trigger pulse recorded on tape The effect of many charged channel blockers is also known to be voltage during the experiment. Three to four synaptic currents were averaged dependent (procaine: Adams, 1976; tubocurarine: Manalis, 1977; hex- at each holding potential, the peaks being lined up exactly before av- amethonium: Ascher et al., 1978, 1979), as expected for a binding site eraging to correct for small variations in the synaptic delay. Synaptic within the membrane potential field. current peak amplitudes and decay time constants were measured from An ACh receptor antagonist, by definition, prevents the binding of the averaged signals, a nonlinear least-squares fitting procedure being ACh to its receptor and has no direct effect on subsequent processes, used to fit a single exponential or the sum of 2 exponentials to the decay such as channel opening or closing. On the basis of the simplified kinetic phase. The effects of drugs on synaptic currents recorded from both scheme, which assumes that ACh disappears from the synaptic cleft ganglion preparations were similar, and results have been pooled for very rapidly, a receptor antagonist is predicted to reduce the peak ESC presentation. amplitude without affecting the rate of decay. This is undoubtedly an Solutions and drugs. The recording chamber, 0.5 ml capacity, was oversimplification, since binding of ACh to receptors is known to be continually perfused at 3 ml/min with a frog Ringer solution containing one of the factors controlling the rate at which it is hydrolyzed or diffuses (in mM): NaCl, 116.5; KCl, 2.5; CaCl,, 1.8 or 4.0; NaHEPES, 5; pH out of the synaptic cleft (Katz and Miledi, 1973; Pennefather and Quas- adjusted to 7.2 with NaOH and maintained at 20°C. Increasing the tel, 198 1). In the absence of anticholinesterase drugs, however, this effect calcium concentration to 4 mM tended to prolong the duration of cell is likely to be small and can, as a first approximation, be neglected. survival after impalement and was used for about half the recordings. The criteria outlined above are generally useful for distinguishing Antagonists were added to the perfusing solution at least 5 min prior between receptor antagonism and ion channel block as a mechanism to recording and until their effects on synaptic currents reached a steady by which a drug can antagonize the effect of ACh (see Rang, 1982). It state. should be pointed out, however, that several drugs are known to be The following drugs were used: decamethonium iodide (Koch Light), both receptor antagonists and channel blockers at the same receptor ion hexamethonium bromide (Sigma), trimetaphan camsylate (Roche), and channel complexes (e.g., tubocurarine: Manalis, 1977; Katz and Miledi, (+)-tubocurarine chloride (Burroughs Wellcome). 1978; Colquhoun et al., 1979). In addition, the term ACh receptor Analysis of the site of drug action. In this paper an ACh antagonist is antagonist is taken to include different forms of receptor block such as used in the broadest sense to describe a drug that inhibits the action of competitive and noncompetitive antagonism since this study cannot ACh. ACh antagonists may inhibit the response to ACh by inhibiting distinguish these different mechanisms. Lipscombe and Rang - Nicotinic Receptors of Frog Ganglia

A CONTROL B TUBOCURARINE 1 PM C TUBOCURARINE 3 PM D TUBOCURARINE (l-30 PM) I;, HP = -‘TV 0 1.2 1 .

0.8 7 = 11 .o ms

0.08 1 \i

o.sJ , 1 -100 -80 -60 -40 -20 HOLDING POTENTIAL (mV)

Figure 1. Action of tubocurarineon ESCsrecorded from cardiacneurons. ESCs recorded from a cardiacneuron at the indicatedholding potentials are shownsuperimposed in A and in the presenceof 1 PM (B) and 3 PM (C’) tubocurarine.Tubocurarine reversibly reduced the peakcurrent amplitudesimilarly at all potentials.Decays of synapticcurrents recorded at - 100mV are plotted on a semilogarithmicscale together with values of the fitted time constantsto showthat the decaykinetics are unaffectedby tubocurarine.In D the effectof tubocurarine(l-30 FM) on decaytime constantsis shownat differentholding potentials. Each point representsa measurementfrom an individual neuronfrom a total of 5 neurons. Valuesshown are exnressedrelative to control, indicatedby the dashedline (1.O). The averagereduction in rdecayinduced by tubocurarinewas 9 i 7% (5 cells).

ganglia, respectively. The decay time constants of ESCs(TV,..,) Results increased slightly with hyperpolarization, the voltage change In the absenceof ACh antagonistsESCs of cardiac and sym- corresponding to an e-fold changein rdecay,being -417 f 60 pathetic ganglia decayed with a single exponential component mV (n = 16) and -304 f 18 mV (n = 14) in cardiac and as describedin sympathetic and cardiac ganglia of the bullfrog sympathetic ganglia, respectively. (Kuba and Nishi, 1979; MacDermott et al., 1980; Connor and Figure 1 showsa seriesof ESCsof a cardiac neuron recorded Parsons, 1983). The averagetime constant fitted to ESC decays in the absenceand presenceof 1 and 3 PM tubocurarine. In the of cardiac ganglia at -80 mV was 9.5 + 0.7 msec (n = 16), a presenceof tubocurarine (l-30 PM), synaptic current amplitudes value rather longer than that measuredfrom bullfrog ganglia were reduced and rdecayonly very slightly shortened. At 3 PM (6-7 msec at -80 mV, value extrapolated from authors’ value the peak amplitude of currents was reduced by 54 f 3% (4 of 7 at -50 mV and their estimate of H; Connor and Parsons, cardiac neurons) at all potentials. 7decay,in contrast, was reduced 1983). The mean ESC amplitude measuredfrom frog cardiac by only 9 + 7% (5 cardiac neurons) in the presenceof between gangliaat -50 mV was -4.9 * 0.5 nA, which is very similar 1 and 30 KM tubocurarine. Figure 1D shows that the small to that measuredin bullfrog cardiac ganglia (-4.3 nA at -50 reduction of 7decayinduced by tubocurarine was independent of mV; Connor and Parsons, 1983). potential. Recordings from frog sympathetic ganglia were mainly ob- Figure 2 showsthe effect of 10, 30, and 600 PM hexamethoni- tained in solution containing 4 mM calcium (seeMaterials and urn on ESCs recorded from a sympathetic neuron. Hexame- Methods). Raisingthe external calcium concentration increased thonium in concentrations of up to 100 I.IM had no effect on the time constant of ESC decays and also the peak amplitude. either ESC amplitude or ~~~~~~(6 cells: 3 cardiac, 3 sympathetic): The averageT of ESC decays in 1.8 mM calcium was 5.3 * 0.2 In the presenceof 10 PM hexamethonium the average peak (n = 2) at -50 mV and a peak amplitude of -8 f 3 nA (n = amplitude was 101 f 0.3% and ~~~~~~was 99 -t 0.7% of control 2) in good agreementwith those values measuredfrom bullfrog values (3 cells). At concentrations greater than 100 PM hexa- ganglia under the same conditions (T = 5-6 msec; Kuba and methonium reduced ESC peak amplitudes, but even at 600 PM Nishi, 1979; MacDermott et al., 1980; Connor and Parsons, there was only a 40% reduction with no marked effect on decay 1983). The average7 of ESC decays in 4 mM external calcium kinetics (Fig. 2). The average 7decaymeasured from ESCsin the was increasedto 8.2 f 0.6 msec (n = 7) at -50 mV and the presenceof between 10 and 600 PM hexamethonium was 94 f peak amplitude was - 15 rf: 3 nA (n = 7) in good agreement 2% of control values (6 cells). with those values measuredin bullfrog ganglia in similar con- Trimetaphan reduced ESC amplitude, and, at a higher conditions (7 msec, - 14 nA, Connor et al., 1985). centration, it also affected the decay kinetics (5 cells: 4 cardiac, Increasing external calcium also prolongs the decay of both 1 sympathetic). In the presenceof 3-100 PM trimetaphan, ESC evoked and miniature synaptic currents in rat submandibular amplitudes were reduced, there being approximately 50% in- ganglia (Rang, 1981) as well as miniature end-plate currents hibition at 10 PM. In the presenceof 100 FM trimetaphan, current (Cohen and Van der Kloot, 1978).The increasedsize of synaptic decays were split into 2 componentsand the time constants of currents observed in high calcium is likely to be the result of these were voltage dependent. At -80 mV ~~~~~~of the fast increasedtransmitter release.The prolongation of the synaptic component was reduced by 40% relative to control, whereasat current decay induced by raising calcium may be explained by - 50 mV there was only a 12% reduction. The inhibition of ESC a direct effect on channel open time since calcium has been amplitude produced by 100 PM trimetaphan was also voltage found to increase the mean burst length at the frog end-plate dependent, with reductions of 8 1 and 58% at - 80 and - 50 (Colquhoun & Sakmann, 1985). mV, respectively. ESC peak amplitudes increased,as expected, with membrane Figure 3 highlights the different effects of tubocurarine, hex- potential hyperpolarization, the extrapolated reversal potential amethonium, and trimetaphan on ESCsof frog gangliaand on for ESCs being -4 and -8 mV in cardiac and sympathetic ESCsof the rat submandibularganglion (Rang, 1982). Control The Journal of Neuroscience, September 1988, 8(9) 3261

A HP = -80 mV B HP = -80 mV

2Or 10r

(nA) 2 = 12.2 ms 2 = 10.8 ms 2 \I’

0.2 1 CONTROL HEXAMETHONIUM 600 pM

If=-==-

Figure 2. Action of hexamethonium -80 mV-v 50 ms on ESCsrecorded from a sympathetic neuron.Control ESCs(A) and thosein the presenceof 10 /IM (c), 300PM (O), and 600 PM (B) hexamethoniumat the C HEXAMETHONIUM 10 pM D HEXAMETHONIUM 300 pM indicatedpotentials. At 10 PM, hexa- methoniumdid not effect the synaptic response,whereas at 300and 600I.IM it reducedthe peakamplitude. The semi- logarithmicplot of the decay of synaptic currents at -80 mV and values of the fitted time constants show that the decay kinetics are unaffected by hexamethonium.

ESCs of rat submandibular ganglia are described by 2 exponen- decamethonium is increased, and the effects are more pro- tials corresponding to “fast” and “slow” ACh-activated synaptic nounced at -80 mV than at -40 mV. These results are con- channels (Rang, 198 1). In Figure 3 the effects of the drugs on the sistent with decamethonium being an ion channel blocker, as “fast” component of the rat ESCs are compared with the frog hasbeen shown in the rat submandibularganglion (Gurney and because of the similar time constants. However, the effects of Rang, 1984), and at the neuromuscular junction (Adams and the drugs on the “fast” component shown in Figure 3 are very Sakmann, 1978; Milne and Byrne, 1981). The slow component similar to those on the “slow” component (Rang, 1982). Where- of ESC decays did not vary with decamethonium concentration as in the frog ganglia tubocurarine (l-30 PM) reduced ESC am- as expectedfor open channelblock (seeColquhoun and Hawkes, plitudes in a dose-dependent manner, effects on rat ganglionic 1983). A similar result was obtained by Adams and Sakmann current amplitudes were only observed at concentrations ex- (1978) at the frog end-plate, the slow component only decreasing ceeding 20 FM. In the rat ganglion, tubocurarine acts predom- inantly as an ion channel blocker of the “slow” channels at these concentrations, speeding the decay of the “slow” component with little effect on the peak amplitude. Hexamethonium had Table 1. Effect of decamethonium on ESCs of cardiac and no effect on the decay kinetics of ESCs recorded from frog ganglia sympathetic neurons (up to 600 PM), whereas it is a potent channel blocker in rat ganglia, speeding the initial rate of decay of synaptic currents affecting both the “fast” and “slow” components of the decay. rdecayof ESCs of rat ganglia was reduced by x 50% in the presence -80 mV of 10 KM hexamethonium (Rang, 1982). Trimetaphan reduced 10 3 7.9 k 0.7 6.8 k 0.7 21.2 k 0.4 0.99 k 0.06 the peak current amplitudes in both the frog and rat ganglion 30 5 8.6 2 0.6 5.9 k 0.7 18.7 3~ 1.7 0.88 f 0.06 preparations but was effective at slightly lower concentrations 100 2 8.7 f 1.2 1.7 x!r 0.1 23.4 f 6.7 0.31 t 0.06 in the rat. -40 mV The effects of decamethonium were also studied. Figure 4 10 4 8.2 k 0.5 9.9 + 1.0 - 0.92 +- 0.08 shows the action of 100 PM decamethonium on synaptic currents 30 7 8.1kO.6 11.5 + 1.0 - 0.79 k 0.07 recorded from a cardiac neuron. Decamethonium (1 O-l 00 FM) 100 2 7.9 + 1.0 4.0 + 1.7 18.7 k 1.9 0.27 k 0.09 consistently split the decay of ESCs, recorded at potentials more Average data on the effect of 10, 30, and 100 MM decamethonium on ESC decay negative than about -50 mV, into 2 components with decay time constants and peak current amplitudes at -80 and -40 mV. Control ESC rates, respectively, faster and slower than control. Table 1 shows dews ~~~~~~~~~~are compared with those in the presence of decamethonium (rtaJ, and r,,,J. The peak current amplitude in the presence ofdecamethonium is expressed that the reduction in 7decayis concentration and voltage depen- as a fraction of the control value (&J&J. Values are means k SE calculated dent; the initial rate of decay increases as the concentration of from n number of cells. 3262 Lipscombe and Rang - Nicotinic Receptors of Frog Ganglia

FROG l effectsat the neuromuscularjunction (seeTable 2 for summary A TUBOCURARINE RAT 0 and references). 0 A In frog ganglia, tubocurarine and, in higher concentrations, hexamethonium reduced the amplitude of synaptic currents to a degreethat was independent of membrane potential and had no appreciableeffect on rdecaY,results consistent with thesedrugs acting as receptor antagonists.Similar effectsof thesedrugs are observed at the neuromuscular junction, where tubocurarine reducesend-plate current amplitude at concentrations similar o.oL 1 I I to those effective at the frog ganglia (Colquhoun et al., 1979; 1 10 100 Shaker et al., 1982), and near millimolar concentrations of hexamethonium are required to produce an equivalent inhibition R HEXAMETHONIUM (Milne and Byrne, 1981). Inhibition of end-plate currents in- -5 1.0 -----r-*---‘3<-r duced by hexamethonium and tubocurarine is independent of 4 I Q 0 l * membrane potential unless the fibers are strongly hyperpolar- >O-J 0 0 ized, when the additional channel-blocking effect becomesno- i= z 0.5 0 0 ticeable. In contrast, in the rat ganglion these drugs predomi- 0 nantly act as ion channel blockers, their effects being highly 0 voltage-dependent and accompanied by a marked speedingof 2 o.oL 1 I I I 1 10 100 1000 current decays at hyperpolarized potentials (Rang, 1982). The small but consistentreduction of rdecayinduced by tubocurarine in frog ganglia was unlikely to be the result of channel C TRIMETAPHAN ,.p------block since the effect was independentof voltage. Tubocurarine alsohas a similar effect on ~~~~~of end-plate currents (Colquhoun 0 et al., 1977; Feltz et al., 1977), and here it was suggestedthat if the rate of ACh removal during the decay of end-plate currents was rate limiting under these conditions tubocurarine (or any ACh receptor antagonist) by reducing the number of available binding sitesfor ACh would speedthe removal of ACh from the synaptic cleft and thus speedthe rate of synaptic current decay (seealso Katz and Miledi, 1973). A similar explanation CONCENTRATION (PM) could account for the action of tubocurarine in the frog ganglia, Figure 3. ACh antagonistactions on synapticcurrents of frog sym- if it is assumedthat the decay rate of synaptic currents is not patheticand cardiacganglia compared with thoseof the rat subman- solely determined by the rate of channel closing. dibularganglion. The actionsof tubocurarine(A), hexamethonium(B), and trimetaphan(C) are compared.The plotted points representre- Decamethonium was found to be an ion channel blocker in sponsesfrom separatecells (- 80 mV; 20°C).The actionsof eachof the frog ganglia, similar to its effects at the end-plate (Adams and 3 drugswere tested on at least5 differentfrog neurons.Data on the rat Sakmann, 1978) and the rat submandibularganglion (Ascher et gangliaare reproducedfrom Rang(1982). Effects on rat ESCdecays are al., 1979; Gurney and Rang, 1984). At the end-plate, decame- shownfor the “fast” component,though similar results are obtained thonium is an agonist (Adams and Sakmann, 1978), but it is for theseactions on the “slow” component;control ESCdecays of sub- mandibularganglionic currents are biexponential(Rang, 1981). The not an agonist in mammalian ganglia (Paton and Perry, 1953; reductionin eitherpeak amplitude or ‘T,,=~~induced by the antagonists Ascher et al., 1979; Gurney and Rang, 1984). If the frog neuronal is shownrelative to the control values(indicated by the broken he). receptor is like that of skeletal muscle,decamethonium should also be an agonist at this site. Although in the present study no agonist action of decamethonium was apparent, patch-clamp significantly with decamethoniumconcentrations at much more recordings from freshly dissociated frog sympathetic neurons negative potentials (- 130 mV). Table 1 also showsthat 10 and (Lipscombe, 1986a, b) have shown that decamethonium is an 30 I.LM decamethonium had little effect on the peak synaptic agonist, inducing channel openingssimilar to those seenat the current amplitude; however, at 100 FM peak amplitudes were end-plate. The actions of decamethonium are therefore consis- reduced to 27 f 9 and 31 f 6% at -40 and -80 mV, respec- tent with the generalization that the neuronal nicotinic receptor tively. The reduction in the peak amplitude of ESCsindicates of frog gangliaresembles that of skeletal muscle. that at this concentration (100 PM) decamethonium may be an Although the presentresults suggest that the neuronal receptor ACh receptor antagonist as well as a channel blocker. Alter- of the frog is qualitatively similar to the end-plate receptor, the natively, a channel blocker would reducethe peak amplitude of following observations establishthat they are not identical. ESCsif channel opening was not synchronous (seeappendix to 1. The effect of a-BuTx. Marshall (198 1) found that c~-BuTx Rang, 1982). did not totally abolishthe synaptic response,its effect was slowly reversible, and rather higher concentrations than thoseeffective Discussion at the end-plate were required for inhibition. The results demonstrate the presenceof a neuronal receptor in 2. The channel open times are different. The time courseof frog distinct from that described in mammalian ganglia. The synaptic currents in frog ganglia (Kuba and Nishi, 1979; actions of ACh antagonistsat frog ganglia-in particular, tubo- MacDermott et al., 1980; Lipscombe, 1986a, b) are about 5 curarine and hexamethonium - were very unlike their reported times slower than those at the end-plate (Magleby and Steven, effects in the rat submandibular ganglion and similar to their 1972b). This suggeststhat the channelopen time of the neuronal The Journal of Neuroscience, September 1988, 8(9) 3263

B DECAMETHONIUM 100pM

8 yast = 2.0 ms (0.4)

Figure 4. Action of decamethonium on ESCsof a sympatheticneuron. Syn- aptic currents recordedat -60 and - 100mV in the absence(A) and presenceof 100PM decamethonium(B). The semilogarithmicplots of the decayof currents recordedat - 100 mV are showntogether with the fitted timecon- stants.In the presenceof decamethonium the decayis biexponential.The peakamplitudes of currentsare also re- ducedat this concentration.The effect of decamethoniumon both the time courseand peakamplitude of the cur- l rentsis greater at the morehyperpolar- 50 ms ized potentialof - 100mV. receptor is longer. This has been confirmed from patch-clamp channelconductance of the frog neuronal channel is smaller, 18 recordings from frog sympathetic neurons (Lipscombe, 1986a, pS (Lipscombe, 1986a, b; also see Marshall, 1985), than the b). end-plate, 30 pS (Colquhoun and Sakmann, 1981); however, 3. Voltage dependenceof synaptic current decays. The voltage this does not account for the large difference in the size of syn- dependenceof the time constant of end-plate current decays is aptic currents. The sensitivity of the neuronal receptor for ACh more pronounced than that measuredin frog ganglia. At 20- hasnot been measuredand therefore remainsas another possible 25°C about 100 mV hyperpolarization is required to produce difference between neuronal and skeletal muscle receptors. an e-fold increase in T at the end-plate (Magleby and Stevens, In conclusion, we have shown that the neuronal nicotinic 1972a)compared with estimatesof - 150 to - 500 mV required receptor responsiblefor mediating fast excitatory transmission to produce a similar increase in r at frog and bullfrog ganglia in the cardiac and sympathetic gangliaof the frog is similar in (frog: present study; bullfrog: Kuba and Nishi, 1979; Mac- its pharmacological characteristics to that found at the end- Dermott et al., 1980; Connor and Parsons, 1983). plate. This neuronal nicotinic receptor is not the sameas that 4. The size of synaptic currents. The amplitude of the synaptic so far studied in mammalian ganglia; both the ACh receptor responseof ganglia is about 20 times smaller than end-plate antagonist binding site and ion channel blocker binding site are currents. This could be the result of lessACh releaseduring a different for the 2 receptors. Although similar, the frog neuronal synaptic event or a lower density of postsynaptic nicotinic re- ACh receptor channel is not identical to that found at the end- ceptors in the ganglia compared with the end-plate. Alterna- plate, having quite distinct channel gating kinetics. An impor- tively, the smallersynaptic responsemay reflect a lower intrinsic tant considerationis whether this type of receptor isoffunctional sensitivity of the neuronal receptor for ACh or that the neuronal importance only in the nervous system of the frog. There is an channel has a smaller single-channelconductance. The single- abundanceof evidence for the presenceof high-affinity binding

Table 2. Relative potencies of ACh antagonists on the rat submandibular ganglion (“fast” and “slow” components), frog cardiac, and sympathetic ganglia, and frog and rat neuromuscular junction

Froglratd Rat ganglion Frog ganglia muscle Dm (PM) AMP r AMP 7 AMP r Tubocurarine >40 >206 =3 >30 =l >3e.f Hexamethonium >2@ =lOb 300-600 >600 > 1000 = 1ooog Decamethonium >20 l@ =lOO 30-100 >lOO 25-509.h Trimetaphan =5 -156 -109 100 100 25d Concentrations (in WM) of drugs producing a 50% reduction in peak amplitude of synaptic currents, Amp, and a 50% reduction in decay time constants, 7, are shown. Data for responses at - 80 mV and 20°C were used where possible. > , Indicates that higher concentrations were not used, and the value given represents a lower limit. y This value is much higher if single responses are used; inhibition of peak current amplitude by hexamethonium is highly use-dependent (Gurney and Rang, 1984). b Rang, 1982. i Gurney and Rang, 1984. d Gibb and Marshall, 1984. e Colquhoun et al., 1979. ‘Shaker et al., 1982. g Milne and Byrne, 198 1. h Adams and Sakmann, 1978. 3264 Lipscombe and Rang * Nicotinic Receptors of Frog Ganglia sites for ol-BuTx in the mammalian brain, but their function ol-bungarotoxin on retinotectal synaptic transmission in the goldfish remains unclear (see Clarke et al., 1985). There are also indi- and the toad. Neuroscience 5: 929-942. Gibb, A. J., and I. G. Marshall (1984) Pre- and post junctional effects cations from molecular biological studies that the mammalian of tubocurarine and other nicotinic antagonists during repetitive stim- brain possesses different types of nicotinic receptor since cDNAs ulation in the rat. J. Phvsiol. (Lond.) 351: 275-297. encoding for different a-subunits of the neuronal nicotinic re- Goldman, D., E. Deneris; W. Luyten; A. Kochhar, J. Patrick, and S. ceptor have now been isolated. Further studies are necessary, Heinemann (1987) Members of a nicotinic acetylcholine receptor however, to determine whether a class of neuronal nicotinic gene family are expressed in different regions of the mammalian cen- tral nervous system. Cell 48: 965-973. receptor like that found in frog ganglia is of physiological im- Gurney, A. M., and H. P. Rang (1984) The channel-blocking action portance in the mammalian nervous system. of methonium compounds on rat submandibular ganglion cells. Br. J. Pharmacol. 82: 623-642. Jenkinson, D. H. (1960) The antagonism between tubocurarine and References substances which depolarise the motor end-plate. J. Physiol. (Lond.) Adams, P. R. (1976) Drug blockade of open end-plate channels. J. 152: 309-324. Physiol. (Lond.) 260: 531-552. Katz, B., and R. Miledi (1973) The binding of acetylcholine to recep- Adams. P. R.. and B. Sakmann (1978) Decamethonium both ouens tors and its removal from the synaptic cleft. J. Physiol. (Lond.) 231: and blocks kndplate channels. Proc. Natl. Acad. Sci. USA 75: 2994- 549-574. 2998. Katz, B., and R. Miledi (1978) A re-examination of curare action at Ascher, P., A. Marty, and T. 0. Neild (1978) The mode of action of the motor end-plate. Proc. R. Sot. London [Biol.] 203: 119-133. antagonists of the excitatory response to acetylcholine in Aplysiu neu- Kuba, K., and S. Nishi (1979) Characteristics of fast excitatory post- rones. J. Physiol. (Lond.) 278: 207-235. synaptic current in bullfrog sympathetic ganglion cells. Pfluegers Arch. Ascher, P., W. A. Large, and H. P. Rang (1979) Studies on the mech- 378: 205-212. anism of action of acetylcholine antagonists on rat parasympathetic Lee, C. Y., S. L. Chang, S. T. Kau, and Shing-Hui Luh (1972) Chro- ganglion cells. J. Physiol. (Lond.) 295: 139-170. matographic separation of the venom of Bungarus multicinctus and Boulter, J., K. Evans, D. Goldman, G. Martin, D. Treco, S. Heinemann, characterisation of its components. J. Chromatogr. 72: 6 l-70. and J. Patrick (1986) Isolation of a cDNA clone for a possible neural Lee. G. M.. and S. Nishi (1972) Analysis of the mechanism of action nicotinic acetylcholine receptor a-subunit. Nature 319: 368-374. of some ganglion-blocking drugs in the rabbit superior cervical gan- Brown, D. A., and L. Fumagalli (1977) Dissociation of alpha-bun- glion. Br. J. Pharmacol. 46: 78-88. garotoxin binding and receptor block in the rat superior cervical gan- Libet, B., S. Chichibu, and T. Tosaka (1968) Slow synaptic responses glion of the rat. Brain Res. 129: 165-168. and excitability in sympathetic ganglia of the bullfrog. J. Neurophys- Bursztajn, S., and M. D. Gershon (1977) Discrimination between iol. 31: 383-395. nicotinic receptors in vertebrate ganglia and skeletal muscle by alpha- Lipscombe, D. (1986a) Pharmacology of nicotinic receptors in frog bungarotoxin and cobra venoms. J. Physiol. (Lond.) 269: 17-3 1. neurones: Electrophysiological studies on intact and dissociated gan- Chiappinelli, V. A., and R. E. Zigmond (1978) oc-Bungarotoxin blocks glia. Ph.D. thesis, University of London. nicotinic transmission in the avian ciliary ganglion. Proc. Natl. Acad. Lioscombe. D. (1986b) Pronerties of ACh recentor channels in adult Sci. USA 75: 2999-3003. frog sympathetic neurons: Patch clamp studies:Soc. Neurosci. Abstr. Clarke, P. B. S., R. D. Schwartz, S. M. Paul, C. B. Pert, and A. Pert 12: 14a. (1985) Nicotinic binding in rat brain: Autoradiographic comparison Lipscombe, D., and H. P. Rang (1984) Acetylcholine antagonists in of 13Hlacetylcholine, 13Hlnicotine, and 112511-oc-bungarotoxin._ _ J. Neu- frog cardiac ganglion cells. IUPHAR 9th Int. Cong. Pharmacol., 6 17a. rosci. 3: 1307-1315. - - MacDermott, A. B., E. A. Connor, V. E. Dionne, and R. L. Parsons Cohen. I.. and W. Van der Kloot (1978) Effects of lCa*+l and IMe*+ (1980) Voltage clamp study of fast excitatory synaptic currents in on the decay of miniature endplate currents. Nature 27i: 77-?9.- A bullfrog sympathetic ganglion cells. J. Gen. Physiol. 75: 39-60. Colquhoun, D. (198 1) The kinetics of conductance changes at nicotinic Magleby, K. L., and C. F. Stevens (1972a) The effect of voltage on receptors of the muscle end-plate and of ganglia. In Drug Receptors the time course of end-plate currents. J. Physiol. (Lond.) 233: 15 l- and Their Effecters. N. Birdsall, ed., pp. 107-127, Macmillan Press, 171. London. Magleby, K. L., and C. F. Stevens (1972b) A quantitative description Colquhoun, D., and A. G. Hawkes (1983) The principles of the sto- of end-plate currents. J. Physiol. (Lond.) 233: 173-l 97. chastic interpretation of ion-channel mechanisms. In Single-Channel Manalis, R. S. (1977) Voltage-dependent effect of curare at the frog Recording, B. Sakmann and E. Neher, eds., p. 135, Plenum, New neuromuscular junction. Nature 267: 366-368. York. Marshall, L. M. (198 1) Synaptic localization of Lu-bungarotoxin bind- Colquhoun, D., and B. Sakmann (1981) Fluctuations in the micro- ing which blocks nicotinic transmission at frog sympathetic neurons. second time range of the current through single acetylcholine receptor Proc. Natl. Acad. Sci. USA 78: 1948-1952. ion channels. Nature 294: 464-466. Marshall, L. M. (1985) Presynaptic control of synaptic channel kinetics Colquhoun, D., and B. Sakmann (1985) Fast events in single-channel in sympathetic neurones. Nature 317: 621-623. currents activated by acetylcholine and its analogues at the frog muscle McMahan, I-J. J., and S. W. Kuffler (1971) Visual identification of end-plate. J. Physidl. (L&d.) 369: 501-557. synaptic boutons on living ganglion cells and of varicosities in post- Colquhoun, D., W. A. Large, and H. P. Rang (1977) An analysis of ganglionic axons in the heart of the frog. Proc. R. Sot. London [Biol.] the action of a false transmitter at the neuromuscular junction. J. 177: 485-508. Physiol. (Lond). 266: 361-395. Miledi, R., and L. T. Potter (197 1) Acetylcholine receptors in muscle Colauhoun. D.. F. Drever. and R. E. Sheridan (1979) The actions of fibres. Nature 233: 599-603. tubocurahne’at the frog’neuromuscular junction. J.‘Physiol. (Lond.) Milne, R. J., and J. H. Byrne (1981) Effects of hexamethonium and 293: 247-284. decamethonium on end-plate current parameters. Mol. Pharmacol. Connor, E. A., and R. L. Parsons (1983) Analysis of fast excitatory 19: 276-281. postsynaptic currents in bullfrog parasympathetic ganglion cells. J. Nishi, S., and K. Koketsu (1960) Electrical properties and activities Neurosci. 3: 2 164-2 17 1. of single sympathetic neurones on frogs. J. Cell. Comp. Physiol. 55: Connor, E. A., S. M. Neel, and R. L. Parsons (1985) Influence of the 15-30. extracellular ionic environment on ganglionic fast excitatory post- Ogden, D. C., and D. Colquhoun (1985) Ion channel block by ace- synaptic currents. Brain Res. 339: 227-235. tylcholine, carbachol and suberyldicholine at the frog neuromuscular Feltz, A., W. A. Large, and A. Trautmann (1977) Analysis of atropine junction. Proc. R. Sot. London [Biol.] 225: 329-355. action at the frog neuromuscular junction. J. Physiol. (Lond.) 269: Paton. W. D. M.. and W. L. M. Perrv (1953) The relationshin between 109-130. depblarisation and block in the cat’s superior cervical ganglion. J. Freeman, J. A. (1977) Possible regulatory function of acetylcholine in Physiol. (Lond.) 119: 43-57. maintenance of retinotectal synapses. Nature 269: 2 18-222. Paton, W. D. M., and E. J. Zaimis (1951) Paralysis of autonomic Freeman, J. A., J. T. Schmidt, and R. E. Oswald (1980) Effect of ganglia by methonium salts. Br. J. Pharmacol. 6: 155-168. The Journal of Neuroscience, September 1988, 8(9) 3265

Pennefather, P. and D. M. J. Quastel (1981) Relation between sub- Shaker, N., A. T. Eldefrawi, L. G. Aguayo, J. E. Wamick. E. X. Al- synaptic receptor blockade and response to quanta1 transmitter at the buquerque, and M. E. Eldefrawi (1982) Interactions of d-tubocu- mouse neuromuscular iunction. J. Gen. Phvsiol. 78: 3 13-344. rarine with the nicotinic acetylcholine receptor/channel molecule. J. Rang, H. P. (198 1) The characteristics of synaptic currents and re- Pharmacol. Exp. Ther. 220: 172-l 77. sponses to acetylcholine of rat submandibular ganglion cells. J. Phys- Skok. V. I. (1965) Conduction in tenth ganglion ofthe frog sympathetic iol. (Lond.) 311: 23-55. trunk. Fed. Prdc. Trans. Suppl. 24: T36<-367. - - - Rang, H. P. (1982) The action of ganglionic blocking drugs on the Skok, V. I. (1986) Channel blocking mechanism ensures specific block- synaptic responses of rat submandibular ganglion cells. Br. J. Phar- ade of synaptic transmission. Neuroscience 17: l-9. macol. 75: 151-168. Skok, V. I., A. A. Selyanko, and V. A. Derkach (1983) Channel block- Ravdin, P. M., and D. K. Berg (1979) Inhibition of neuronal acetyl- ing activity is a possible mechanism for selective ganglion blockade. choline sensitivity by a-toxins from Bungarusmulticinctus venom. Pfluegers Arch. 398: 169-17 1. Proc. Natl. Acad. Sci. USA 76: 2072-2076.