Characterization of long-lasting histaminergic inhibition in a beating pacemaker neuron of Onchidium

Brain Research, 332 (1985) 1-14

1

Elsevier BRE 10657

Research Reports

Characterization of Long-Lasting Histaminergic Inhibition in a Beating Pacemaker Neuron of Onchidium TSUKASA GOTOW

Department of Physiology, School of Medicine, Kagoshima University, Kagoshima890 (Japan) (Accepted July 17th, 1984)

Key words: molluscan neuron - - prolonged inhibiiion - - metabolic inhibitors - - electrogenic pump A single BPSP (excitatory-inhibitory postsynaptic potential) was monosynaptically produced in an identified Onchidium neuron, Be-l, with a beating rhythm upon stimulation of the cardiac nerve. The BPSPs summated to produce an inhibition of long duration (ILD) upon blockage of the beating rhythm after repeated stimulation, so that the BPSPs seemed to be functionally inhibitory. Ten stimuli (1-2 Hz) applied to the cardiac nerve usually evoked an ILD (0.5-1 min) of about 10 mV. The early and middle phases of this ILD reversed near -80 to -85 mV, but the late phase did not reverse at more negative potentials. None of the phases was significantly affected by low CI or Na solutions or by high Ca solutions. However, by changing the external K, the shift of the reversal potentials for the early and middle phases reached about 65% of that predicted for the K electrode, although the late phase was insensitive to the external K. Intraeeilular tetraethylammonium (TEA) attenuated the amplitude of the ILD but did not shorten the duration. These suggest that the ILD has another conductance-independent mechanism simultaneously with the increase in K conductance. Several lines of evidence suggested that a ouabain-sensitive Na pump does not contribute to the ILD. Inhibitors of energy supply, 2,4-dinitrophenoi sodium salt (DNP) and cyanide, selectively and reversibly reduced the ILD. Simultaneous applications of intracellular TEA and DNP completely abolished the ILD. As for the ionic basis, the histamine-induced inhibitory response in Be-1 was closely related to the ILD. Cimetidine specifically blocked the ILD and histamine-induced inhibitory response, which were mimicked by 2-methylhistamine, but not by dimaprit. It is concluded that the ILD, mediated by some histamine receptor other than the H 1or H 2 type, results from an increase in K conductance and a hyperpolarizing ion pump insensitive to ouabain. INTRODUCTION Long-lasting cholinergic and d o p a m i n e r g i c inhibitory responses are o b s e r v e d in identified Aplysia burst firing neurons, and the ionic mechanisms of these responses, which cannot be inverted and are insensitive to potassium changes, have b e e n investigated by several workerst1,12,18, 22. The following 3 mechanisms have b e e n p r o p o s e d to explain these long-lasting inhibitory responses: (1) an activation of the electrogenic N a pumpt8, (2) an increase in K conductance occurring in a r e m o t e region away from the cell somali, 12 and (3) a blockage of regenerative slow inward current, p e r h a p s carried by N a and Ca ions22. On the o t h e r hand, ai~ identified Onchidium neuron, B e - l , located a p p r o x i m a t e l y in the middle of the

dorsal surface of the right pleuro-parietal ganglion shows a beating rhythm, consisting of a sustained train of e n d o g e n o u s spikes 6. In this B e - l , a putative neurotransmitter, histamine produces a long-lasting hyperpolarization with inhibition of the beating spikes, and it has been suggested that this inhibitory histamine response is p r o d u c e d mainly by an electrogenic ion p u m p insensitive to ouabain, although an increase in K conductance was not excluded8, 9. In addition, cyclic nucleotides are involved in the histamine response of Be-17. Thus, the mechanism of the inhibitory histamine response seems to differ from the above 3 alternative mechanisms p r o p o s e d for the long-lasting inhibitory responses in the Aplysia bursting neurons. F u r t h e r m o r e , the histamine response of Be-1 is somewhat different from histaminergic inhibi-

Correspondence: T. Gotow. Present address: Department of Physiology, School of Medicine, Kagoshima University, Kagoshima 890, Japan. 0006-8993/85/$03.30 © 1985 Elsevier Science Publishers B.V. (Biomedical Division)

tory responses produced by an increase in K or C1 conductance in the same Aplysia neurons ~0. The present study was undertaken to clarify whether such a characteristic, inhibitory response to exogenous histamine reflects normal physiological events (postsynaptic potentials) in Be-1. The results provide evidence that the ionic mechanism of a synaptic inhibition of long duration evoked by stimulation of the cardiac nerve may be equivalent to that of the inhibitory histamine response, and that this synaptic inhibition is mediated by a histamine receptor different from the typical H 1- or H2-receptor. MATERIALS AND METHODS Experiments were made on the isolated circumesophageal ganglia of the marine gastropod mollusc, Onchidium verruculatum, collected from Kinko Bay in Kagoshima, Japan. The ganglia were pinned in a 1 ml lucite bath and perfused with saline at rate of 3 ml/min. The temperature of the bath ranged from 20 to 24 °C. T h e Be-1 neuron, the cardiac nerve and the abdominal nerve 1 in the present ganglia have been identified previouslyS,6. For intracellular recording, one or two single microelectrodes filled with 2.5 M KCI, having a resistance of less than 10 M ~ , were inserted into Be-1. Current was applied to the cell through the second electrode, and a bridge circuit was used when a single electrode served for both current and recording. A 2.5 M K C L - a g a r , A g - A g C I bridge was placed in the bath as an indifferent electrode. The methods used to record membrane potential and measure current intensity have been described s. The cardiac nerve or the abdominal nerve 1 was sucked into a tightly fitting glass tube containing a silver wire for stimulation with isolated 0.5-1 ms voltage pulses. Intracellular injection of tetraethylammonium (TEA) was made by passing current pulses between the barrels of double microelectrodes filled with 2 M TEA-Br. The normal saline had the following composition (mM): NaC1 450, KCI 10, CaCI2 10, MgC12 50 (ref. 6). The pH was adjusted to 7.8 with 10 mM trishydroxymethylaminomethane (THAM) and HC1. Variations in external K were made by adding or omitting KC1. Low Na solutions were made by replacing Na with THAM. Isethionate salts or sucrose

were used as CI substitutes. High Ca solutions were made by adding Ca without osmotic compensation. High Mg saline6 containing 4 times normal Mg (200 mM) was used to examine synaptic transmission to Be-1. Possible transmitter substances tested were: acetylcholine-Cl (ACh-CI), dopamine-HCl, L-noradrenaline-HC1 (Sigma); histamine-2HCl, serotonine creatine sulfate (5-HT) (Merck); L-glutamic acid sodium salt, v-amino-n-butyric acid (GABA), and glycine (Nakarai Chemicals). These drugs were applied to Be-1 by iontophoresis or by superfusion as described previously 8. Dimaprit and 2-methylhistamine used as histamine agonists were gifts from Smith, Kline and French (Welwyn, U.K.). Antagonists tested were: cimetidine, mepyramine maleate (Smith, Kline and French); curare, atropine sulfate, hexamethonium-Br, picrotoxin, TEA-C1 (Nakarai Chemicals); phentolamine (Ciba-Geigy). Other agents used were ouabain (Merck), 2,4-dinitrophenol sodium salt (DNP) (Pfaltz and Baer), and picric acid (trinitrophenol; TNP), sodium cyanide (Nakarai Chemicals). RESULTS Synaptic responses in Be-1 to stimulation of the cardiac nerve The neuron Be-1 showed biphasic responses, consisting of a short depolarizing component followed by a long hyperpolarizing one after stimulation of the cardiac nerve. Fig. 1A shows an example of a biphasic postsynaptic potential (BPSP) in Be-l, produced by a single stimulus to the cardiac nerve, not containing the axon branches of Be-16. This BPSP was first sharply activated at relatively low stimulus intensities, and its threshold was remarkably constant between preparations (for 0.5 ms duration, 2 - 3 V applied to the suction electrode). Furthermore, changing stimulus intensities or intervals between stimuli failed to activate selectively one of the components of the BPSP, so that the biphasic form always appeared. This suggests that the biphasic form is specific to the BPSP but is not due to a composite of two individual synaptic potentials, and that the BPSp is mediated by a single presynaptic axon within the cardiac nerve. At resting potentials in the range of - 4 0 to -50 mV, the

10 mV

A 400 msec I Hz •

•

•

°

•

•

.

~ •

•

•

2

LI,~, e

o

°

=

*

o

o

2 Hz J ° ~

G •

•

•

.

.

•

•

•

•

__.__I20my oo°~*~o.

•

I ,Hz '

8

4

4 Hz

8

/••.7 oel.°...ol

.

.

.

.

.

.

.

.

.

.

SffC

]

Hz

20mY

2.5sec

Fig. 1. Biphasic postsynaptic potentials (BPSPs) of Be-1 to postsynaptic stimulation of the cardiac nerve. A: an oscilloscope trace showing the BPSP produced by a single presynaptic stimulus. In this trace, Be-1 was hyperpolarized by about 5 mV (to -50 mV) to weaken the spontaneous spike activity. B: penwriter records showing a summation of the BPSPs in Be-1 having the initial membrane potential of -45 mV, produced by 10 presynaptic stimuli with 0.5 ms duration pulses. The stimulating voltage was 2 V in 1 and 4-8, 20 V in 2 and 60 V in 3. The stimulating frequency delivered at dotted marks are shown in each record. Sharp and upward deflections superimposed on the BPSPs show the stimulus artifacts. Note that BPSPs were less influenced by higher stimulus intensities (2, 3) above the threshold (1). Note also the change in time speeds between 1-6 and 7 and 8. Top portions of action potentials are cut in this and subsequent penwriter records. depolarizing c o m p o n e n t of the B P S P was less than 5 mV in amplitude and 100-400 msec duration, and sometimes was barely detectable. In response to r e p e a t e d stimulation, when the frequency of stimuli was increased, each depolarizing c o m p o n e n t of the BPSPs did not s u m m a t e but was m a s k e d by each h y p e r p o l a r i z a t i o n ( c o m p a r e Fig. 1B1 with B4_S). On the o t h e r hand, the hyperpolarizing component of the BPSP by a single presynaptic shock was less than 2 - 3 m V in a m p l i t u d e and about 3 s in duration (Fig. 1A). W h e n the cardiac nerve was stimulated r e p e a t e d l y (Fig. 1B), the hyperpolarizing components s u m m a t e d to p r o d u c e an intense and l o n g -

lasting hyperpolarization with inhibition of the endogenous firing, so that the BPSPs were functionally d o m i n a t e d by t h e hyperpolarizing c o m p o n e n t s , although the depolarizing ones following the first one or two stimuli were sometimes sufficient to elicit a spike. H e r e , such a s u m m a t i o n of the BPSPs dominated b y the long-lasting inhibitory action was designated an inhibition of long duration ( I L D ) , similar to that in Aplysia neurons, first described by Tauc 21. In response to stimulus trains of 10 shocks, the I L D increased with increasing frequency of stimuli up to a m a x i m u m p e a k amplitude of 20 m V of 1 - 2 min (Fig. 1B6_8). Ten stimuli delivered at 1 - 2 Hz usually gave an 1LD of 10 m V lasting for 0 . 5 - 1 min

Control

S l i m e s normal Cai *

/, l i m e s flor n-4tl Mg2**lO rain

Wash

S se¢

Fig. 2. Effects of salines containing 5 times normal Ca (b) and 4 times normal Mg (c and d) on the BPSPs and their summation followingthe cardiac nerve stimulation, a: control record of the BPSPs to 10 stimuli at 1 Hz delivered as indicated by the dotted marks, b: 15 rain after exposure to high Ca saline. Control normal saline was changed to high Mg saline, c after 10 rain and d after 15 rain. e: recovery of the BPSPs after washing with normal saline. Membrane potential was -45 mV in a and e, -50 mV in b, and -40 mV in c and d. All records are from the same Be-1.

(Fig. 1Bt_5), and this stimulus condition was used for most of the following experiments. Higher stimulus intensities above the threshold of the BPSP may also activate other inputs in addition to the BPSP in Be-1. However, the BPSPs produced by the lower stimulus intensities to the cardiac nerve were unaffected by stimulation with current pulses of amplitudes 10 (Fig. 1B2) and 30 (Fig. 1B3) times larger than those required to activate the BPSPs. Furthermore, the BPSP could not be elicited by stimulation of the other main peripheral nerves except for the cardiac nerve leaving the ganglia. These suggest that the BPSPs are mediated only by a single presynaptic axon in the cardiac nerve. The depolarizing component of the BPSP had short, constant latency (about 23 ms, Fig. 1A), and could follow high-frequency stimulation (5-10 Hz) of the cardiac nerve without loss of transmission. These observations suggest that the depolarizing component is mediated monosynaptically, according to the criteria 3 of latency and high-frequency response for monosynaptic connections. However, these criteria could not be applied to the hyperpola-

rizing component which developed slowly and was long lasting. The best evidence for monosynapticity of such a long-lasting component of the BPSP came from experiments with high Ca and high Mg salines (Fig. 2). When the external Ca concentration was increased from 10 (normal) to 50 mM, both components of the BPSPs or the ILD persisted or were slightly augmented (Fig. 2b). A change in the perfusion fluid from normal saline to high Mg saline (4 times higher than normal) suppressed the BPSPs gradually (Fig. 2c, d). Finally, the BPSPs were completely eliminated after 20 min exposure to this high Mg saline. After return to normal saline, the BPSPs recovered (Fig. 2e). As predicted from the criteria 3 used as evidence of monosynaptic connections, if the BPSPs were produced by a polysynaptic pathway, the BPSPs should have been reduced after exposure to high Ca saline and abruptly blocked after high Mg saline. Thus, the BPSPs or the ILD in Be-1 are apparently activated monosynaptically through the cardiac nerve.

Ionic mechanisms of the BPSPs (1) Current-voltage characteristics of the ILD The above results showed that the BPSPs were functionally inhibitory and the summation of the BPSPs was characterized as the ILD. When the effect of membrane potential on a given synaptic response is investigated, a common procedure is to elicit the response repeatedly, each time at a different level of injected membrane current. However, this procedure for the ILD was precluded for several reasons, among them, the prolonged duration of the ILD and the long recovery time (5-10 min) for obtaining reproducible responses of the ILD. To avoid this difficulty, during the resting state and a single ILD evoked by the cardiac nerve stimulation, the membrane potential response to 3 levels of injected hyperpolarizing currents was determined at regular intervals. The responses were measured at 12 s intervals to a 3-step current pulse lasting 3.3 s (Fig. 3). Each step pulse of 1.1 s was long enough for reaching the steady-state membrane potential at the end of the step. An alternative graphic representation of Fig. 3B 1 under normal (10 mM K) saline is shown in Fig. 3B2. The slope of the current-voltage ( I - V )

line obtained from the m e m b r a n e response to the

was more negative ( - 8 6 mV). The late period (phase

3-step current at the early period (phase 1) of the ILD was smaller than at the resting period (control) before the stimulus train, indicating the occurrence of an increase in m e m b r a n e conductance. The resting and early conductance lines intersected at -

3) of the ILD, 12 s later than phase 2, showed a negative shift in voltage but did not intersect with the control in the measured range. The further negative shift of.the reversal potential with time suggests that the ILD is activated by more than two mechanisms.

81.5 mV, the reversal potential for phase 1. The I - V line for the middle period (phase 2), 12 s later than

(2) Effects of K and CI ions

phase 1, also indicated an increase in m e m b r a n e conductance, but the intersection with the control line BI

AI S mid I(*

In normal saline the I L D reversed at - 8 1 . 5 m V for

Kt(normal)

I 0 mM

C!

15 mM K"

t

15 s e c

-2

-I

0 -60

-2

-1

!

!

nA

-2

-40 ,

-I ,

I

0 ;-40

Control,~

Control

c2

B2

A2

-60

/

-60

P h a s e 2

- 80

-80

Phas~

-80 Phase

Phase2

1

-100 '/'qk" Phase3

~ P h a s e3

~

-100 Y

"Pha~3

-100 my

Fig. 3. Effects of membrane potential and changing external K concentration on the ILD (summation of the BPSPs) in the same Be-1 produced by 10 stimuli at 1 Hz to the cardiac nerve. A 1, B], C1: the membrane potential responses (upper traces) to 3-step hyperpolarizing current pulses (lower traces) before (control) and at 3 different periods after stimulation, in 5 mM (A1), 10 mM (Bl) and 15 mM (C1) K solutions. The 3-step current pulses were applied at 12 s intervals and were set so that the second 3-step current came to approximately the peak (early period) of the ILD following the end of the stimulus train. Initial membrane potential was -47 mV in A], 45 mV in B~ and -47 mV in Cr Lower current traces also contain stimulation marks of the presynaptic nerve. A2, B2, C2: current-voltage (I-V) characteristics at the 3 different periods of the ILD. Each I-V line in A2, B2 and C2 is constructed from the corresponding 3-step pulse shown in A1, B1 and Cl, respectively. Phase 1 (filled circles), phase 2 (open squares) and phase 3 (filled squares) in each graph correspond to each period of the second, third and last 3-step pulses injected during the ILD, respectively. The I-V lines for control (open circles) and phase 1 of the ILD intersect at -92.5 mV in 5 mM K (A2), -81.5 mV in 10 mM K (B2) and -76.5 mV in 15 mM K (C2). During phase 2, the intersection point (reversal potential) shifted to a negative value, -97 mV in A2, -86 mV in B2and -80 mV in Cz. No intersection was found during phase 3 in any K solution.

-70

A

,,"

E

Fig. 3 A l , A2), the reversal potentials for phases 1 and 2 were shifted to - 9 2 . 5 and - 9 7 mV, respectively. W h e n potassium was increased to 15 m M (Fig. 3C1, C2) the reversal potential became - 7 6 . 5 mV for phase 1 and - 8 0 mV for phase 2. However, phase 3 did not reverse in either 5 or 15 m M K saline or in the normal 10 m M K saline (Fig. 3 A 2,

/////,///

".~

C •

-80

e~ i11

C2). o

.J

-1004

i

|

i

1

5

10

15

20

External K+ concentration(mM) -70

-70 B

E v

200mM

Na +

C

5 0 mM Ca a+

2 E 0a

-80

•

-80

O Q.

• >

-90

|

-90

In Fig. 4A, the results obtained from 5 experiments are presented in graphic form, and the slope of shifts in reversal potential for a 3-fold change in [K]o shows 17 mV for phase 1 and 19 mV for phase 2. These represent about 65% of the theoretical slope value (28 mV/3-fold change in [K]o) predicted by the Nernst equation, suggesting that an increase in K conductance contributes to phases 1 and 2 of the ILD. Moreover, when the [Cl]o was reduced to 20% (200 mM) by replacing with isethionate, no significant changes in the I L D or the resting m e m b r a n e potential were observed. Thus, a change in Cl conductance does not seem to be involved in the I L D .

2

I-

a ...I

• -100

. 5

(3) Effects of Na and Ca ions

6 .

. . . 10 15 20

100

,

' 5

10

' • 15 20

External K* c o n c e n t r a t i o n ( r a M ) Fig. 4. A: relationship between the log of external K concentration and the reversal potential for phase 1 (open circles) or for phase 2 (filled circles) of the ILD. In this and subsequent B and C, the ILD was evoked by 10 stimuli at 1 Hz to the cardiac nerve, and the reversal potentials for phases 1 and 2 were estimated by using the same procedure as in Fig. 3. Each point for phases 1 and 2 was averaged from 5 Be-1 neurons, with one S.E. of the mean shown by bars. The solid lines drawn through each point of phases 1 and 2 show 17 mV and 19 mV slopes for a 3-fold change in K, respectively. The dotted line shows a slope of 28 mV/3-fold change in K predicted b3;the Nernst equation. B and C: effects of low Na (B) or high Ca (C) solution on the relationship between the external K and the reversal potential of the ILD in another Be-l. Open and filled circles are the same as shown in A. Open and filled triangles show reversal potentials for phases 1 and 2 in 200 mM Na (B) or in 50 mM Ca (C) solution, respectively. phase 1 and - 8 6 m V for phase 2, suggesting that at least either K or C1, or both, contributes to these phases. The effects of changing the external K concentration, [K]o, on the I L D are shown in Fig. 3A, C. W h e n the amount of external potassium was halved (5 mM,

As shown above, the reversal potential for phase 2 of the I L D shifted toward m o r e negative values from that for phase 1, until finally no reversal potentials were observed for phase 3. Such a behavior of the I L D could be explained by a mechanism in which the increase in K conductance, dominant in phase 1 was accompanied by a simultaneous decrease in membrane ionic conductance either to N a or Ca, or both, during phases 2 and 3, as predicted from theoretical considerations 4. To test this hypothesis, the effect of changes in [Na]o or [Ca]o on the I L D reversal potential was analyzed at different concentrations of external K saline. W h e n the [Na]o was reduced to 200 m M by replacing with T H A M , the reversal potential for phase 1, estimated by using the p r o c e d u r e as in Fig. 3, was not significantly affected in [K]o of 5, 10 and 15 m M (Fig. 4B), but in all of the three [K]o, the reversal potential for phase 2 showed a small negative shift ( 2 - 3 mV). On the other hand, exposure to this 200 m M Na saline shifted the p e a k potential for phase 3 of the I L D to depolarizing levels only slightly in all of 5, 10 and 15 m M [K]o (not shown). If the negative shift c,f *.he reveral potentials p r o d u c e d during

phases 1 and 2 were due to a decrease in Na conductance following an increase in K conductance, the reduction of [Na]o would make the reversal potentials for their phases more negative because of decrease in the Na equilibrium potential. In addition, if phase 3 were involved in a decrease in Na conductance, the 200 mM Na saline would shift its peak potential to hyperpolarizing levels. When the [Ca]o was increased from 10 (normal) to 50 mM, no significant changes in the reversal potentials during either phase 1 or 2 were observed in [K]o of 5, 10 and 15 mM (Fig. 4C), contrary to expectation. Phase 3 of the I L D was only slightly altered by the increase of [Ca]o at these three [K]o. These results suggest that the negative shift and elimination (at phase 3) of the reversal potential during the time course of I L D are not due to the ionic conductance mechanism but may be related to an electrogenic ion pump.

lnA l Hista mine~t0"SM)

t

TEA injection

I' Glutamite( 10-4 M)

t

(4) Effects of intracellular TEA The above findings suggested that the I L D had a conductance-dependent component mediated by an increase in K conductance and a second hyperpolarizing component. Several workers2,11A3,16 have shown that T E A injected intracellularly blocks slow K-dependent synaptic hyperpolarizations in molluscan neurons. Thus, T E A injection should selectively eliminate the conductance-dependent component of the I L D but not the second component. When T E A was injected into the Be-1 by current pulses (20-40 nA, 0.5 s) with i Hz for 5 - 1 0 min, the duration of the action potentials was prolonged to about 100 ms (5 times normal) and the resting membrane resistance was slightly increased, but the membrane potential remained nearly normal. As shown in Fig. 5A, the I L D evoked by cardiac nerve stimulation was attenuated after injection of T E A , but the stimulation still caused a detectable hyperpolarization of long duration and a decrease in the spontaneous spike frequency, as expected. On the other hand, the reduction of input resistance observed at phases 1 and 2 of the I L D was completely eliminated by the T E A injection (compare Fig. 5A 1 with A2). Fig. 5 also shows effects of T E A on the histamine (Fig. 5B) and glutamate (Fig. 5C) responses obtained from bath application to the same Be-/. Part of the histamine-induced slow hyperpolarization was

C

2

TEA in ection ~

~1

20mV

15 sec

Fig. 5. Effects of intracellularly injected TEA on the ILD (A), histamine (B) and glutamate (C) responses in the same Be-1. AI: control ILD produced by 10 stimuli at 1.4 Hz to the cardiac nerve. In this and subsequent A2, constant current pulses (1.2 s duration) shown in A 3 w e r e applied to monitor input resistance. Trace A 3 also contains stimulation marks. B1: control response by bath application of histamine (10-5 M) shown at arrow in the trace. C1: control response by bath application of glutamate (10-4 M). A 2, B2, C2: responses after intracellular TEA injection. Note that the ILD and histamine responses were only attenuated, but glutamate response was completely blocked by TEA. The order of each test was carried forward from A1, B1, C 1 to A 2, B2, C2. Initial control membrane potential was -45 mV in A 1, -47 mV in B t and C 1, and -46 mV in A 2, B2 and C2.

clearly observed even after the injection of T E A , although it was largely reduced (Fig. 5B2). However, the glutamate-induced hyperpolarization, which was adjusted to nearly the same level as the histamine response in amplitude, was completely blocked by T E A (Fig. 5C2). Usually these actions of T E A appeared immediately after the 5 - 1 0 min injection, remained constant for over 2 h and did not reverse. It has been reported that the histamine response is produced by an active ion-transport mechanism, superimposed partly on an increase in K conductance8,9, whereas the glutamate response is explained only by a K conductance mechanismlS.

A

B

"1 .,'M 6.p;1o,;,i.,

3

3

I

j

3 mM TNP.10

UU

5 $ec

,

C

I m M D N P , 1 0 rain

'll U

,

L

i

,IJ U'

'

min

•

1

JlnA

Fig. 6. Effects of DNP (A and B) and TNP (C) on the ILD or EPSPs in Be-1. Dotted marks below each voltage trace show the presynaptic stimulation. A: ILD produced by 10 stimuli at 1.4 Hz to the cardiac nerve. A 1and A2: control responses. A 3 and An: responses at 5 min (A3) and 10 min (A4) after exposure to 1 mM DNP. As: recovery after 30 min washing. In A 2, A 4, A 5 to measure changes in input resistance, constant current pulses (1.2 s duration) were applied as shown in trace A 6. B: EPSPs in the different Be-1 from A produced by stimulation of the abdominal nerve 1. Each EPSP to stimulation was amplified by conditioning hyperpolarizing current injection shown in trace B 3. BI: control. B2: EPSPs at 10 min after exposure to 1 mM DNP. C 6 control ILD in another Be-1. C2: ILD at 10 min after exposure to 3 mM TNP. C3: current trace injected in C 1, C 2. Initial membrane potential was -45 mV in A~, A2, B~ and C 1, -42 mV in m 3, -44 mV in m4, -47 mV in As, and -43 mV in B 2 and C 2.

T h e s e suggest that the r e m a i n d e r of the I L D resis-

W h e n d i n i t r o p h e n o l ( D N P ; 1 m M ) was a p p l i e d in

tant to T E A reflects the s e c o n d h y p e r p o l a r i z i n g c o m p o n e n t which m a y be r e l a t e d to a p u m p - d e p e n d e n t

n o r m a l saline, for the first 5 min the m e m b r a n e gradually d e p o l a r i z e d up to 3 - 5 m V and an increase in

process.

the s p o n t a n e o u s spike f r e q u e n c y e n s u e d . H o w e v e r , 5 - 1 0 min later the m e m b r a n e slowly r e p o l a r i z e d , fol-

(5) Effects of metabolic inhibitors T o test m o r e directly w h e t h e r this s e c o n d c o m p o nent could be a p u m p - d e p e n d e n t process, the following e x p e r i m e n t s w e r e c o n d u c t e d .

l o w e d by a d e c r e a s e in the spike f r e q u e n c y . D u r i n g b o t h the d e p o l a r i z i n g and r e p o l a r i z i n g phases, the m e m b r a n e input resistance was slightly r e d u c e d . A s s h o w n in Fig. 6A3, A4, the a m p l i t u d e and d u r a t i o n of

ILD decreased and shortened both after 5 and 10 min perfusions with 1 mM DNP. However, when the change in membrane resistance was measured by passing constant-current pulses before and during the ILD, the degree of the resistance change was not significantly affected by the DNP application (compare Fig. 6A2 with 6A4). This contrasts with the finding in Fig. 5A2 that the reduction of input resistance during the ILD was completely eliminated by TEA injection. By subsequent washing with DNP-free solutions the ILD returned to the control value within 30 min (Fig. 6A5). On the other hand, the DNP concentration did not affect the short depolarizing components, which occurred prior to the development of ILD following repeated stimulation of the cardiac nerve as described in Fig. 1. Likewise, DNP (1 mM) had little effect on excitatory synaptic potentials (EPSPs) in the same Be-l, which are elicited by stimulation of different inputs (abdominal nerve 1) from the ILD input (Fig. 6B). Furthermore, the K-dependent glutamate response in Be-1 was not affected by 1 mM DNP perfusions greater than 10 min, though the histamine-induced hyperpolarizing response in the same Be-1 was greatly reduced (not shown). It has been shown that trinitrophenol (TNP; a structural analogue of DNP) is as effective as DNP in cellular phenomena (such as endocytosis of lymphocyte) which involve shape changes of erythrocyte or lymphocyte membrane, but TNP is incapable of uncoupling oxidative phosphorylation in intact mitochondria19. In the present study, TNP (1-3 mM) had no effect on the ILD even following application of over 30 min (Fig. 6C). However, the depressant actions of DNP on the ILD were mimicked after 15-20 min perfusions with another metabolic inhibitor, cyanide (3 mM), and the drug actions were easily reversed after washing (not shown). On the other hand, the membrane resistance before and during the ILD was not significantly affected by this cyanide, as in the case of DNP. These results suggest that inhibitors (DNP, cyanide) of energy supply have a specific depressant effect on the ILD. This also supports the idea that a pump-dependent process, which requires energy supply, contributes to the ILD in combination with the conductance-dependent component. This idea was further supported by the effects of

both TEA and DNP applied simultaneously on the ILD. In Fig. 7b, the effects of intracellular TEA injection on the ILD were similar to those shown in Fig. 5. Extracellular addition of 1 mM DNP to Be-1 treated with intracellular TEA substantially abolished the ILD (Fig. 7c). Washing with DNP-free saline returned the ILD level to that of treatment with TEA alone (Fig. 7d). A synaptic activation mechanism of an electrogenic Na-pump has been proposed for the slow inhibitory synaptic potentials in sympathetic ganglia 14 and

Control

intracellular TEA

lntracellular TEA& DNP(ImM) c

~. . . .

i

!

"

Wash(intracellular TEA ) d ~

~

,

]20j mV

15

sec

Fig. 7. Simultaneouseffects of intracellular TEA and extracellular DNP on the ILD in the same Be-]. Constant current pulses shown in trace c were applied to monitor resistance. Trace e also contains marks of the presynapticstimulation, a: normal ILD produced by ]0 stimuli at ].4 Hz to the cardiac n e r v e , b: responseafter intraceHularTEA injection for ]0 rain with 1 Hz of 20 nA and 0.5 s duration, c: responseat 5 rain after bath application of 1 mM DNP following the intraccllular TEA injection. Note that the [LD was substantially blocked by the simultaneous application of TEA and DNP. d: response after 30 rain washing with DNP-frec saline. Note that the effect of the first TEA injection made on trace b remained. Initial control membrane potential was -45 mV in a and b, -43 mV in c, and -46 mV in d.

10

Aplysia neurons~8. Thus, the contribution to the pump process from an electrogenic Na pump was tested. After ouabain (10 -4 M) was applied in normal (450 mM Na) saline, the I L D was suppressed within 10 min (Fig. 8A). Moreover, ouabain added to the normal salines also reduced both EPSPs elicited by the abdominal nerve 1 stimulation and K-dependent glutamate response in the same Be-1. However, the I L D in 150 mM Na saline ( T H A M substitution of Na) was almost not influenced even after more than 20 min perfusion with ouabain (Fig. 8B), although ouabain depolarized resting membrane potentials by 3 - 5 mV and slightly reduced input resistances both in the normal and 150 mM Na salines. On the other hand, D N P added to the 150 mM Na saline reversibly attenuated the I L D (Fig. 8C), as was seen with the normal saline (Fig. 6A). Thus, an electrogenic Na pump does not seem to be a candidate for the above pump-dependent process.

Oua bain (10"4 M)

DNP (IO-3M)

Possible transmitters for the ILD An attempt was made to examine responses in the Be-1 receiving the BPSP input to 8 substances (dopamine, glycine, noradrenaline, 5;HT, ACh, G A B A , glutamate and histamine), which are regarded as possible transmitters in at least molluscan neurons, and to compare these responses with the ILD. When the drugs were individually applied to Be-1 by iontophoresis or by superfusion, Be-1 was depolarized by dopamine, glycine, noradrenaline and 5-HT, and hyperpolarized (inhibited) by A C h , G A B A , glutamate and histamine (not shown except responses to A C h , glutamate and histamine). It has been reported that ACh- and glutamate-induced inhibitory responses in 4 0 n c h i d i u m G - H neurons, one of which is Be-i, are mediated by an increase in Ci and K conductances, respectively15. On the other hand, the findings obtained by changing [K]o and/or [Cl]o suggested that the G A B A response results from increases in K and CI conductances, similar to those found in the identified Onchidium cerebral neuron 20. Thus, ionic mechanisms of these responses differ from those of the I L D , indicating that ACh, glutamate and G A B A can be eliminated as inhibitory transmitters proposed for the ILD. The mechanism of the histamine-induced hyperpolarization reported previouslyS,9 is in general

Fig. 8. Comparison between effects of ouabain (10-4 M) (A, B) and DNP (10-3 M) (C) on the ILD responses to 10 presynaptic stimuli at 1.4 Hz delivered at dotted marks. AI: control ILD in normal saline. A2:10 rain after addition of ouabain to the normal saline. BI: control ILD in 150 mM Na (THAM substitute) saline. B2:20min after addition of ouabain to the 150 mM Na saline. Note that a blocking action of ouabain for the ILD was eliminated by reduction of [Na]o to 150 raM. C1: control in 150 mM Na (THAM substitute) saline. C2: ILD at 10 min after addition of DNP to the 150 mM Na saline. C3: recovery after returning to the 150 mM Na saline without DNP. A-C are from the different Be-l, respectively. In each trace, downward deflections before stimulation are electrotonic potentials to injection of constant current pulses. The membrane potential was -45 mV in A1, -42 mV in A 2 and C3, -40 mV in B l, C 1 and C2, and -37 mV in B2. agreement with the present finding that the I L D was composed of an increase in K conductance and a pump-dependent process. In addition, histamine applied to the axonal region of Be-1 by iontophoresis produced a slowly developing inhibitory response, similar to that evoked by nerve stimulation (compare Fig. 9A 1 with B1). Thus, to determine whether histamine is the more likely inhibitory transmitter for the ILD, histamine-blocking drugs were tested. H2-antihistamine, cimetidine, specifically and reversibly blocks histamine-induced slow inhibitory re-

11 1' HIstamlne(lO'SM)

sponses or slow inhibitory synaptic potentials by activation of histaminergic n e u r o n in

Aplysia

neu-

rons 10:3. In the present experiments, the addition of 1' 2-tdet hylhis tamine(I 0-t' M I

2 × 10-4 M cimetidine to the bath substantially blocked both the histamine-induced response (Fig. 9A2) and the ILD (Fig. 9B2). After washing,

b

~

,.~

~

these inhibitory responses readily r e t u r n e d to control values (Fig. 9A 3, B3). However, in addition to H r a n t i h i s t a m i n e , m e p y r a m i n e (Fig. 9B4), some oth-

t, 2-Methylhi

stamine

Cimetidine

er antagonists, curare, atropine, h e x a m e t h o n i u m , picrotoxin and p h e n t o l a m i n e (ca. 5 x 10 -4 M), had no significant effects on either the ILD or the depolarizing components of the BPSPs.

t' 2-Me t h y l h i s t a m i n e

Wash

On the other hand, it has been pointed out that a class of histamine receptors distinct from H 1- and H2-receptors may be present in Aplysia neurons

2-Methylhistamine

1' ACh(10"4M)

Mepyramine

1' ~ u t a m a t e ( l O - 4 M )

A 20 mV 30sec

Cimetidine(2 x 10"~' M )

B Cime tidine (2x10 "t' M )

4

~

pyramtne(§xlO'.M) ~ . . . ~ [ 5

20 my

se¢

Fig. 9. Effects of H2-antihistamine, cimetidine (2 x 10-4 M) on histamine-induced (A) and ILD (B) responses. A: iontophoretic current pulses for histamine application shown by arrows were 1000 nA and 3 s. AI: control histamine response in Be-1. A2:10 min after addition of cimetidine. A3: recovery after 15 min washing. B: 10 stimuli at 1.4 Hz to the cardiac nerve delivered at dotted marks produced ILD response in another Be-1. BI: control ILD. B2: ILD at 10 min after addition of cimetidine. B3: recovery after 15 min washing. Moreover, Hi-antihistamine, mepyramine (5 x 10-4 M) had no effects on the ILD after 15 min exposure (B4). Initial control membrane potential was between -45 mV and -47 mV through all records in A and B.

Fig. 10. a-f: responses to histamine (10-5 M), H r and H2-histamine agonists, 2-methylhistamine (10-4 M) and dimaprit (10-3 M) applied by superfusion and effects of cimetidine (2 x 10-4 M) or mepyramine (5 x 10-4 M) on 2-methylhistamine responses. 2-methylhistamine (b) as well as histamine (a) produced a slow hyperpolarization, but dimaprit produced no detectable inhibition even at the high concentration of 10-3 M (c). The 2-methylhistamine response was significantly reduced after a 10 min exposure to cimetidine (d) and reversed after 15 min washing (e), but unaffected after exposure to mepyramine (f). g and h: responses to ACh (10-4 M) and glutamate (10-4 M) applied by superfusion, and effects of cimetidine (2 x 10-4 M) on their responses. ACh and glutamate inhibitory responses (g) were unaffected after 10 min or 15 min exposure to cimetidine (h). Arrows above each record indicate application of drugs shown. All records are from the same Be-l, which had membrane potentials between -40 and -42 mV. To suppress presynaptic actions of drugs, high Mg saline containing 4 times normal Mg was used for all records. since the histamine-induced slow response blocked by cimetidine is mimicked by a Ht-agonist but not by a H2-agonisttO. Therefore, characterization of cimetidine-sensitive histamine receptors in Be-1 was examined. As shown in Fig. 10, the histamine-induced hyperpolarization sensitive to cimetidine was mimicked by the H r a g o n i s t , 2-methylhistamine (Fig. 10b), but not by the H2-agonist, dimaprit (Fig. 10c), which is highly specific for the H2-receptor 17. The response

12 elicited by this Hl-agonist was reversib~ suppressed by cimetidine (Fig. 10d, e) but not by mepyramine (Fig. 10f). Furthermore, cimetidine had no effect on the hyperpolarizing response by ACh and glutamate in the same Be-1 (Fig. 10g, h). These suggest that histamine receptor in the Onchidium Be-1 is different from H1- and H2-receptors, similar to that found in

Aplysia lo. DISCUSSION In response to stimulation of the cardiac nerve, a single BPSP consisting of a short depolarizing component followed by a long hyperpolarizing one is observed in Be-1 of Onchidium, which has a beatingtype endogenous discharge. This BPSP appears only in an all-or-none manner, and it is the only postsynaptic potential evoked in Be-1 by cardiac nerve stimulation. This suggests that the BPSP or the following ILD is activated by only a single presynaptic axon (input) within the cardiac nerve, although the soma of this axon has not been identified. On repeated stimulation, each BPSP summates to produce the ILD formed by long-lasting hyperpolarization with inhibition of the endogenous firing, so that these hyperpolarizing components of the BPSPs seem to play a dominant physiological role. The present study suggests that the ILD is mediated by two different mechanisms, an increase in K conductance and a pump-dependent process, and that histamine is a more likely transmitter for the ILD.

Mechanisms of the ILD With normal saline, the reversal potentials for the early and middle periods (phases 1 and 2) of the ILD show approximately -80 and -85 mV, respectively, which are much more negative than the resting potential (-45 mV). The shift of the reversal potentials for phases 1 and 2 caused by a 3-fold change in the [K]o was about 65% of the theoretical value (28 mV) predicted by the Nernst equation. However, the effects of change in [Cl]o, [Na]o or [Ca]o on the ILD are different from those to be expected from the changes in these ionic conductances. These findings imply that an increase in K conductance contributes significantly to phases 1 and 2 of the ILD, but their phases may be not accompanied by changes in conductance to CI, Na or Ca.

On the other hand, the K-dependent glutamate hyperpolarizing response in the identified G-H neurons containing Be-1 has been reported to reverse at about -60 mV which is less negative than the reversal potentials for the ILD 15. Furthermore, the reversal potential of the ILD continues to shift toward more negative potentials with time, until finally no further reversal potentials are observed (phase 3). These suggest that the ILD appears to have a second hyperpolarizing component simultaneously with the increase in K conductance contributing to at least phases 1 and 2 of the ILD. In fact, after injection of TEA blocks synaptically induced increases in K conductance, the amplitude of the ILD is attenuated, but its duration is not significantly affected, and the remaining ILD, which still shows a clear inhibition of membrane firing, is associated with no detectable changes in input resistance. Inhibitors of energy supply, DNP and cyanide reversibly reduce the amplitude and duration of the ILD but have no effect on the change in input resistance during the ILD. This should be expected if these drugs selectively abolish the second hyperpolarizing component without detectable changes in input resistance, but not the K-dependent component of the ILD. As expected further, simultaneous applications of intracellular TEA and DNP almost abolish the ILD. However, in the same Be-1 neither the EPSP (evoked by different inputs from the ILD input) nor the K-dependent glutamate response were affected by the same DNP which reduced the ILD. Furthermore, TNP has no effect on the ILD. This TNP has been reported to exert the same effect as DNP on the surface membranes, such as erythrocyte or lymphocyte, but does not have the uncoupling effect of DNP on the oxidative phosphorylation 19. Thus, these results strongly support the proposal that in addition to the K-dependent component, a second hyperpolarizing one, which is independent of ionic conductances and requires energy supply, contributes to the ILD. An electrogenic Na pump may be a suitable candidate for this component, but this possibility seems unlikely for the following reasons. DNP added to one-third normal Na saline reversibly reduces the ILD, as well as when added to normal saline. However, ouabain applied to the one-third normal Na saline had no effect on the ILD, although when added to

13 the normal saline it reduced the ILD. The reduction of the ILD in the normal saline by ouabain may be a result of an indirect effect due to the large intracellular accumulation of Na, as suggested previously for the effects of ouabain on the histamine-induced hyperpolarization 9. Furthermore, in normal saline ouabain also reduces both the EPSP and K-dependent glutamate response. These suggest that ouabain has no specific effect on the ILD. On the other hand, the ILD, particularly at phase 3 (or the second hyperpolarizing component) without detectable changes in resistance may result from an increase in K conductance and/or a decrease in Na or Ca conductance in remote synaptic regions away from the Be-1 soma. In this case, electrotonic synaptic potentials transmitted to the soma from the synaptic region may be greatly attenuated by a reduction in membrane resistance. Thus, the attenuation of the ILD by DNP and cyanide may be an indirect effect due to reduction of membrane resistance. However, this also seems unlikely for two reasons. Although exposure to 15 mM K saline and addition of ouabain to one-third normal Na saline reduced the membrane resistance, respectively, as well as the DNP treatment, they did not affect the ILD or its phase 3. Inversely, although the reduction of the [Na]o, increase of the [Ca]o or intracellular injection of T E A increased the membrane resistance, none of them prolonged the ILD. Furthermore, the Na pump may be also blocked by DNP or cyanide. This blockage of the Na pump may evoke K-accumulation in the remote synaptic regions, suggesting a depolarizing shift in the K equilibrium potential. Thus, if the second hyperpolarizing component of the ILD is due to the increased K conductance occurring at the remote synaptic region, this component may be indirectly blocked by DNP or cyanide. However, this is unlikely since ouabain (a specific inhibitor of the Na pump) applied in the onethird normal saline had no effects on the ILD. In conclusion, it is suggested that the ILD may be composed of an increase in K conductance and some hyperpolarizing ion pump process, different from a Na pump. It seems that the combination of both components make the reversal potentials for the ILD more negative or eliminate them. Although the depolarizing component of the BPSP has not been investigated thoroughly, this compo-

nent was reduced by exposure to low Na saline but was insensitive to DNP and ouabain, suggesting a contribution of an increase in Na conductance to the depolarizing component.

Does histamine mediate the ILD ? In view of the criteria of Berry and Pentreath presented as evidence for a chemical monosynaptic connection 3, the present data suggest that the link between Be-1 and the BPSP or ILD input (the presynaptic axon identified within the cardiac nerve) is monosynaptic. In addition to the previous report 8, histamine applied to Be-1 by iontophoresis produces a long-lasting inhibitory response similar to the ILD or the hyperpolarizing component of the BPSP. Cimetidine, an antihistamine, specifically and reversibly suppresses the ILD and histamine-induced inhibitory response but has no significant effect on the Be-1 resting potential or membrane resistance. In Aplysia ganglia, it has been reported that cimetidine selectively blocks slow inhibitory responses produced by histamine or by activation of the histaminergic neuron, which is also the only agent that can be classified as a histamine receptor blocking agent10,13. Furthermore, the present mechanism generating the ILD is compatible with that of the histamine-induced inhibition shown previouslyS,9. These findings thus support the possibility that the ILD in Be-1 is mediated by histamine released from the presynaptic axon terminals. In the present study, histamine applied iontophoretically on the axonal regions of Be-1 failed to simulate the depolarizing components of the BPSPs preceding the ILD. However, this does not exclude the possibility that histamine is the transmitter for the ILD or BPSP since some depolarizing components are not observed following cardiac nerve stimulation, and others are masked with increasing the stimulus frequency. Furthermore, it has been shown that the excitatory effects of dopamine on a dopaminergic biphasic response in Aplysia neurons depend critically on the method of dopamine application, and are often difficult to detect 1. In the present study, the ILD blocked selectively by cimetidine (H2-antagonist) was mimicked by 2-methylhistamine (Hl-agonist), but not by dimaprit (H2-agonist). Thus, the histamine receptor media-

14 ting the I L D s e e m s similar to that m e d i a t i n g the slow

ACKNOWLEDGEMENT

histaminergic inhibitory synaptic p o t e n t i a l s in A p l y -

sia t3. H o w e v e r , since t h e s e inhibitory synaptic p o t e n tials are due exclusively to an increase in K c o n d u c -

T h e a u t h o r wishes to t h a n k D. M r o z e k for r e v i e w ing the manuscript.

tance, the ionic m e c h a n i s m of the h i s t a m i n e r g i c I L D in O n c h i d i u m differs f r o m that of the h i s t a m i n e r g i c synaptic potentials of the Aplysia. REFERENCES 1 Ascher, P., Inhibitory and excitatory effects of dopamine on Aplysia neurones, J. Physiol. (Lond.), 225 (1972) 173-209. 2 Berry, M. S. and Cottrell, G. A., Ionic basis of different synaptic potential mediated by an identified dopamine-containing neuron in Planorbis, Proc. roy. Soc. B, 203 (1979) 427-444. 3 Berry, M. S. and Pentreath, V. W., Criteria for distinguishing between monosynaptic and polysynaptic transmission, Brain Research, 105 (1976) 1-20. 4 Brown, J. E., Muller, K. J. and Murray, G., Reversal potential for an electrophysiological event generated by conductance changes: mathematical analysis, Science, 174 (1971)318. 5 Gotow, T., Morphology and function of the photoexcitable neurones in the central ganglia of Onchidium verruculatum, J. comp. Physiol., 99 (1975) 139-152. 6 Gotow, T. and Hashimura, S., Spontaneous activity and action potential of an identified Onchidium pacemaker neuron, Comp. Biochem. Physiol., 69 (1981) 745-757. 7 Gotow, T. and Hashimura, S., Modulation of the histamine-induced inhibitory response in an identified Onchidium neuron by cyclic nucleotides, Brain Research, 239 (1982) 634-638. 8 Gotow, T., Kirkpatrick, C. T. and Tomita, T., Excitatory and inhibitory effects of histamine on molluscan neurons, Brain Research, 196 (1980) 151-167. 9 Gotow, T., Kirkpatrick, C. T. and Tomita, T., An analysis of histamine-induced inhibitory response in molluscan neurons, Brain Research, 196 (1980) 169-182. 10 Gruol, D. L. and Weinreich, D., Two pharmacologically distinct histamine receptors mediating membrane hyperpolarization on identified neurons of Aplysia californica, Brain Research, 162 (1979) 281-301. I 1 Kehoe, J. S., Ionic mechanisms of a two-component cholinergic inhibition in Aplysia neurones, J. Physiol. (Lond.),

225 (1972) 85-114. 12 Kehoe, J. S. and Ascher, P., Reevaluation of the synaptic activation of an electrogenic sodium pump, Nature (Lond.), 225 (1970) 820-823. 13 McCaman, R. E. and Weinreich, D., On the nature of histamine-mediated slow hyperpolarizing synaptic potentials in identified molluscan neurones, J. Physiol. (Lond.), 328 (1982) 485-506. 14 Nishi, S. and Koketsu, K., Origin of ganglionic inhibitory postsynaptic potential of bullfrog sympathetic ganglion, J. Neurophysiol., 31 (1968) 717-728. 15 Oomura, Y., Ooyama, H. and Sawada, M., Analysis of hyperpolarizations induced by glutamate and acetylcholine on Onchidium neurones, J. Physiol. (Lond.), 243 (1974) 321-341. 16 Parnas, I. and Strumwasser, F., Mechanisms of long-lasting inhibition of a bursting pacemaker neuron, J. Neurophysiol., 37 (1974) 609-620. 17 Parsons, M. E., Owen, D. A. A., Ganellin, C. R. and Durant, G. J., Dimaprit-S-[3-(N,N-dimethylamino)propyl] isothiourea - - a highly specific histamine H2-receptor agonist. Part 1. Pharmacology, Agents Actions, 7 (1977) 31-37. 18 Pinsker, H. and Kandel, E. R., Synaptic activation of an electrogenic sodium pump, Science, 163 (1969) 931-935. 19 Sheetz, M. P., Painter, R. G. and Singer, S. J., Biological membranes as bilayer couples. III. Compensatory shape changes induced in membranes, J. Cell Biol., 70 (1976) 193-203. 20 Shimizu, N., Akaike, N., Oomura, Y., Maruhashi, J. and Klee, M. R., GABA and lioresal actions on the identified Onchidium neuron, Jpn. J. Physiol., 33 (1983) 459-467. 21 Tauc, L., Processus postsynaptiques d'excitation et d'inhibition dans le soma neuronique de l'Aplysie et de l'escargot, Arch. ital. Biol., 96 (1958) 78-110. 22 Wilson, W. A. and Wachtel, H., Prolonged inhibition in burst firing neurons: synaptic inactivation of the slow regenerative inward current, Science, 202 (1978) 772-775.