SUPPRESSION OF SENSORY TO MOTOR SYNAPTIC … · rate of 3APsh~' to avoid low-frequency homosynaptic depression (Goldberg & Lukowiak, 1984). In one set of experiments, low-frequency

J. exp. Btol. 128, 47-62 (1987) 47Printed in Great Britain © The Company of Biologists Limited 1987

SUPPRESSION OF SENSORY TO MOTOR SYNAPTICTRANSMISSION AND NARROWING OF THE SENSORY

NEURONE ACTION POTENTIAL BY ARGININEVASOTOCIN IN APLYSIA CALIFORNICA

BY J. GOLDBERG, W. COLMERS, J. EDSTROM AND K. LUKOWIAK

Department of Medical Physiology, Faculty of Medicine, University of Calgary,Calgary, Alberta, Canada T2N 4N1

Accepted 8 October 1986

SUMMARY

The vertebrate neurohypophysial peptide arginine vasotocin (AVT), which maybe endogenous to the Aplysia central nervous system, was tested for its effect onsensory to motor neurone synaptic transmission. In the semi-intact preparation,superfusion of AVT (10~6moll~') over the abdominal ganglion decreased theamplitude of both the gill withdrawal reflex and the short-latency excitatorypostsynaptic potentials (EPSPs) evoked in gill and siphon motor neurones by singleaction potentials elicited in sensory neurones. AVT slowed the rate of rise of theEPSP, enhanced the rate of homosynaptic depression, and reversibly decreased theduration of the action potential of mechanosensory neurones in isolated, perfusedabdominal and pleural ganglia. Frequency-dependent prolongation of action poten-tials of pleural sensory cells was also decreased by application of AVT. Because thispeptide has been shown to modulate the gill withdrawal reflex and its subsequenthabituation, the hypothesis that AVT plays a physiological role in the expression ofthe suppressed behavioural state is proposed. In addition, it is proposed thatmodulation of the reflex by AVT occurs in part by shortening the duration of thesensory neurone action potential.

INTRODUCTION

The siphon, mantle, gill and abdominal ganglion preparation of Aplysiacalifornica has been used extensively as a model system to investigate of the neuronalmechanisms which underlie non-associative and associative learning behaviour (seeMpitsos & Lukowiak, 1985). More recently this preparation has also proved to beuseful in the study of the neuronal mechanisms which underlie behavioural state, orlevel of arousal. In Aplysia, three behavioural states (normal, facilitated and sup-pressed) have been defined on the basis of the amplitude of the defensive gillwithdrawal reflex (GWR) (Ruben et al. 1981). The suppressed state, characterizedby small-amplitude, rapidly habituating GWRs, reliably occurs in preparationstaken from animals which are sexually active (Lukowiak & Freedman, 1983) or

Key words: synaptic transmission, action potential, arginine vasotocin, Aplysia californica.

48 J. GOLDBERG AND OTHERS

food-satiated (Lukowiak, 1980). In preparations exhibiting the normal behaviouralstate, superfusion of the vertebrate neurohypophysial peptide arginine vasotocin(AVT) over the abdominal ganglion induces a behavioural state indistinguishablefrom the naturally occurring suppressed state (Thornhill et al. 1981). In this study,we explore the mechanisms by which exogenous AVT mediates this change inbehavioural state. Furthermore, we present the hypothesis that an endogenous AVT-like peptide (Moore et al. 1981) plays a similar role in mediating the naturallyoccurring suppressed behavioural state.

We have previously reported that the AVT-induced reduction in the amplitude ofthe GWR evoked by tactile stimulation is accompanied by an attenuation of theexcitatory responses evoked in identified central gill motor neurones by the tactilestimuli to the siphon (Thornhill et al. 1981). This suggests that the reduction inreflex amplitude was mediated at least in part by a change in neuronal activity withinthe central reflex arcs. Since the efficacy of transmission through the central sensoryto motor neurone synapses has been shown to be influenced by numerous en-dogenous (Goldberg & Lukowiak, 1984) and exogenous (Lukowiak, Goldberg,Colmers & Edstrom, 1986) modulators, we have examined both the possibility thatsynaptic plasticity at these sites is a component of the AVT-induced change inbehavioural state and the mechanisms underlying such plasticity. In addition, wehave extended the analysis to include a group of large, accessible, tail sensoryneurones in the pleural ganglion.

The results reported here are consistent with the hypothesis that an AVT-likepeptide is involved in the suppressed behavioural state olAplysia.

MATERIALS AND METHODS

Aplysia californica (150-250 g) were maintained in aerated 1350-1 aquaria inartificial sea water (ASW, Instant Ocean) at 15-16°C and pH7-9 and were fedseaweed (rehydrated) weekly. Animals were used for experiments at least 2 days afterfeeding, and thus were not in a food-satiated state. On the basis of their GWRamplitude, they were classified as normal (Ruben et al. 1981).

Three preparations were used in the study: (1) a semi-intact preparation; (2) theisolated abdominal ganglion preparation; and (3) the isolated pleural ganglion. Thesemi-intact preparation consisted of the siphon, mantle, gill and abdominal ganglion(or CNS). The siphon, ctenidial and branchial nerves (by which the abdominalganglion innervates the siphon, mantle and gill) were left intact; all other nerves andconnectives were severed. The preparation was pinned on Sylgard (Dow Corning) ina 500-ml chamber filled with ASW maintained at 15-16°C. Gill and siphon con-tractions were measured using force transducers connected by fine surgical thread toa single gill pinnule and the siphon. Tactile stimuli to the siphon were delivered by acalibrated mechanical tapper (Peretz & Lukowiak, 1975). In all semi-intactpreparations, the amplitude of the GWR evoked by a 1-g stimulus was at least 35 %of the measured amplitude of the spontaneous gill respiratory movements (SGMs).Thus, these preparations met the minimal response criteria of Carew, Castelluci,

Suppression of synoptic transmission 49

Byrne & Kandel (1979) and displayed the normal behavioural state (Ruben et al.1981). The abdominal ganglion was pinned to a raised Sylgard platform. A smallplastic ring, placed on a bed of petroleum jelly surrounding the ganglion, isolated thedesheathed ganglion from contraction- and tapper-induced turbulence, and allowedthe superfusion of peptides specifically over the abdominal ganglion. The left ventralsurface of the ganglion was desheathed to facilitate the penetration of LE clustermechanoreceptors and motor neurones (Frazier et al. 1967). For intracellularrecording, single-barrelled micropipettes were filled with 3moll~1 KC1; tipresistance ranged between 6 and 40 MQ. A bridge circuit in the electrometer(Dagan8700) allowed simultaneous recording and current injection. A calomelreference electrode was immersed in an ASW-filled container coupled to the ASWbathing the ganglion by way of a saturated NaCl—agar bridge.

The isolated abdominal ganglion preparation was pinned out and desheathed in asmall (^-lml) recording chamber where it was continuously perfused at a rate of2—3 ml min"1 and maintained at 15—16°C. Recordings were made using glassmicroelectrodes with a resistance between 15 and 50 MQ, rilled with a mixturecontaining O^moll"1 potassium acetate and 0-1 moll"1 KC1. These electrodes wereshielded (see Edstrom & Lukowiak, 1985). A Dagan8100 single electrode voltageclamp was employed as an electrometer for intracellular recording and currentpassage. In the isolated abdominal ganglion, action potentials (APs) were evoked bypassing depolarizing current into the mechanosensory cells in brief (3—5 ms) pulses;the amplitude of the resulting excitatory postsynaptic potential (EPSP) in thefollower motor neurone was measured. The follower cell was maintained at aconstant resting potential throughout a given experiment to permit comparison ofEPSP amplitudes within the experiment (before, during and after peptide super-fusion). In the experiments in which sensory neurone AP duration was examined,APs were evoked at a frequency of 1 per 5 min. There was no change in AP durationat this stimulation rate.

In the isolated pleural ganglion, VC cluster sensory cells were identified byposition according to the description given by Walters, Byrne, Carew & Kandel(1983). The ganglion was constantly perfused with ASW in a 0-75-ml chamber at arate between 3 and 6 ml min"1. Trains of 40 APs were evoked at an interstimulusinterval (ISI) of 140ms to produce the action potential prolongation referred to asfrequency-dependent spike broadening (FDSB) (Edstrom & Lukowiak, 1985). Anintertrial interval of 5 min was found to give reproducible results. APs were digitizedand subsequently superimposed by aligning the peak of dV/dt of the rising phase ofeach AP (see Edstrom & Lukowiak, 1985, for full details). Several control trains wererecorded to verify repeatability before any peptides were perfused.

In most experiments, sensory neurones were stimulated to produce single APs at arate of 3APsh~' to avoid low-frequency homosynaptic depression (Goldberg &Lukowiak, 1984). In one set of experiments, low-frequency homosynaptic de-pression was produced by stimulating the sensory neurone to fire 2 APs min"1.

Data were collected and stored either on FM tape (Racal) or on floppy diskettescompatible with a digital storage oscilloscope. Resting potentials were recorded


continuously on a chart recorder. For FDSB experiments, data were digitized,stored and analysed on a DEC MINC 11/23 computer.

Arginine vasotocin (AVT, Bachem) and tetraethylammonium (TEA, Sigma)solutions (in ASW) were prepared just prior to application. Solutions in the semi-intact preparations were exchanged using syringes; isolated ganglia were con-tinuously perfused. Continuous perfusion was used in all experiments involvingTEA.

RESULTS

Semi-intact preparationsSuperfusion of AVT (10~6 mol 1~') over the abdominal ganglion for 5 min brought

about a significant reduction in the amplitude of the GWR evoked by a 1-g tactilestimulus to the siphon (Fig. 1). With ASW bathing the ganglion, the tactile stimulusevoked a GWR that met the minimal response criteria for preparations displaying thenormal behavioural state (Ruben et al. 1981; Carew et al. 1979). One hour after thecontrol stimulus, AVT was superfused over the ganglion for 5 min and the GWRevoked by the siphon stimulus was reduced. Following AVT wash-out and a further1 h rest, the GWR returned to control levels. The suppressive effects of AVT on theGWR could be immediately, but transiently, reversed by the interposition of a novelstimulus (gill pinch) (Thornhill et al. 1981). The gill pinch results in sensitization ofthe GWR; thus while AVT brought about a suppressive effect on the GWR, thiseffect could be masked upon activation of the appropriate neuronal pathways. These

400 T _ _

30-0-

20-0-

10-0- T

ASW AVT ASW

Fig. 1. A summary of the effect on the gill withdrawal reflex (GWR) amplitude ofarginine vasotocin (AVT) when superfused over the abdominal ganglion. The GWR isevoked by a 1-g tactile stimulus to the siphon. The GWR was first evoked with artificialsea water (ASW) bathing the ganglion. Following a 1-h rest, the peptide-containing ASWwas perfused over the ganglion (AVT, 10~6moll~') for 5 min and the GWR evoked.Finally the peptide was washed out and replaced by ASW and the preparation rested anadditional hour before a further stimulus was applied (Ar= 15).


results are quantitatively similar to those reported earlier for the effects of AVT inpicomolar concentrations on gill reflex behaviour (Thornhill et al. 1981).

We next determined the effect of AVT superfusion upon the efficacy of synaptictransmission between sensory and motor neurones of the CNS. Single actionpotentials were evoked in an LE sensory neurone by passing depolarizing currentpulses and the resulting motor neurone EPSPs were measured. To control for thepossibility of low-frequency homosynaptic depression, single APs in the LE sensory'neurone were evoked at a frequency of 3 APs h~'. At this frequency, the amplitude ofthe evoked EPSP was stable (Goldberg & Lukowiak, 1984).

AVT superfusion of the abdominal ganglion reversibly suppressed the EPSPevoked in L7 by an LE sensory neurone AP (Fig. 2A). A similar action was alsoobserved in LDG1, L9 and siphon motor neurone LBS and the results werecombined (Fig. 2B,C). AVT also slowed the rate of rise of the EPSP waveform in L7(Fig. 2D) and other cells tested. Input resistance of follower neurones was un-affected by AVT, in agreement with earlier findings (Thornhill et al. 1981).

AVT is known to increase the rate of gill reflex habituation (Thornhill et al. 1981)so we determined whether AVT also affected the rate of low-frequency homo-synaptic depression at the sensory to motor neurone synapses. The LE sensoryneurone was depolarized to produce an AP at a frequency of ZAPsmin"1 (seeGoldberg, 1983) and homosynaptic depression was observed as a decrease in theamplitude of the evoked EPSP in successive trials (Fig. 3). In ASW, the EPSPamplitude decreased by approximately 40% over the course of six trials, and a least-squares regression yielded a slope of —0-44 (see Peretz & Lukowiak, 1975).Following a rest of 50min, superfusion with 10~6moll~1 AVT reduced theamplitude of the initial EPSP compared to the control EPSP and virtually doubledthe rate of low-frequency homosynaptic depression, which showed a slope of —0-84.

Isolated ganglia

Further investigations of the effects of AVT were carried out in isolated abdominaland pleural ganglia. In the isolated abdominal ganglion, we recorded mainly fromLE sensory neurones, gill motor neurone L7 and siphon motor neurone LBS. Theganglion was perfused continuously and, because of the greater stability inherentwhen using the isolated ganglion preparation, recovery from the effects of AVTcould be more closely examined (e.g. Fig. 4). In these experiments, AVT reducedthe EPSP amplitude to 13-3 % of the control value; the EPSP recovered to 110% ofthe control value after 20 min washing with ASW. Thus, a suppressive effect of AVTwas observed in semi-intact and isolated ganglion preparations both with staticbathing solutions and under constant perfusion.

Previous work (Thornhill et al. 1981) and our present measurements showed thatsuperfusion of AVT did not significantly affect the passive membrane properties ofthe gill and siphon motor neurones. We therefore hypothesized that AVT broughtabout its suppressive effects by directly or indirectly acting on the presynaptic LEsensory neurone. To test this presynaptic hypothesis, we studied the effects pro-duced by AVT superfusion on the duration of the action potential in both the LE and


12 T B

150-0 -i

1 1 2 - 5 -

'o. 75-0 -CO

on0-

T

T

T>E

D.

E

ASW ASW AVT ASW ASW AVT

D

D.

CO

LU

100

80

60

40

207mV

3-7ms

ASW AVT

Fig. 2. The effect of arginine vasotocin (AVT) on the EPSP recorded in gill motorneurone L7 elicited by an action potential evoked in an LE sensory neurone.(A) Summary, showing the effect of AVT on EPSP amplitude evoked by an actionpotential in an LE sensory neurone. An action potential was elicited in the LE sensoryneurone and this evoked an EPSP in the follower motor neurone. The intracellularstimulus was applied at a frequency of 3h~' to avoid low-frequency homosynapticdepression. Two control stimuli were delivered (ASW) before the peptide was superfusedover the ganglion (AVT, 10~ 6 moir ' ) . Following wash-out, the EPSP was again evoked(ASW,iV = 7). Mean values ± S.E.M. are shown for A,B and C. (B) Data as in A. AverageEPSP amplitudes in control (ASW) and during AVT application (AVT) expressed inmV. (C) The data in B, replotted to express the EPSP amplitude in AVT as a percentageof the control EPSP amplitude for each experiment. Data for B and C were obtained fromexperiments on gill motor neurones LDG1 and L9 and siphon motor neurone LBScombined. As can be seen, there is qualitatively no difference between these data andthose obtained in L7. N = 9 in each case. (D) Data obtained from a single experiment inwhich the action potential (not shown) was evoked in the sensory neurone and the EPSPwas recorded in L7. (1) With artificial sea water bathing the ganglion, (2) in 10~6moll~'AVT and (3) following wash-out of the peptide. EPSPs were evoked once every 20min.As can be seen, the peptide not only reduces the amplitude but also slows the rate ofrise of the EPSP recorded in L7. These observations were common throughout allexperiments.


pleural ganglion sensory neurones. Initially we found that AVT had no discernibleeffect on LE AP duration in ASW. However, in the presence of a low concentrationof TEA (lOmmolF1), we found that AVT decreased the duration of the AP in asimilar manner in both LE and pleural ganglion sensory neurones. The constantperfusion of TEA-containing ASW caused an increase in the duration of the AP(Fig. 5A); in the experiment shown in Fig. 5A the AP duration increased from 1-61to 2-24ms (at half-amplitude). When AVT (10"6moir1) was then added to theperfusate, the AP duration decreased from 2-24 to 1-88ms (Fig. 5B). This was a16-3 % decrease. Following wash-out of the peptide, the AP duration increased backto 2-24 ms. Finally, following wash-out of TEA, the duration of the AP returned tocontrol values (not shown). We found similar results when recording from thepleural ganglion sensory neurones (Fig. 5C). Thus, the pleural ganglion sensoryneurones behave in a similar way to the abdominal ganglion LE sensory neurones intheir response to TEA and AVT. In nine such experiments the AP duration inlOmmol P 1 TEA was taken as 100% and the duration of the action potential in AVTwas normalized accordingly. AVT elicited a 10-7 ± 4-2% reduction in AP duration(P < 0-005). Following wash-out of the AVT, the AP duration was 108-6 ± 14-9 % ofthe control value. This, however, was not significantly different from control values

1000

10-32 3

Trial number

Fig. 3. Effect of arginine vasotocin (AVT) on low-frequency homosynaptic depression.Single LE sensory neurone action potentials were evoked at a frequency of 2min ' toproduce low-frequency homosynaptic depression at the sensory to motor neuronesynapse. Control (in artificial sea water, ASW, D) and test (X, AVT 10~6moll~1) runswere separated by 60min. The rate of homosynaptic depression was measured as theslope of the least-squares fit on the log-log plot: EPSP amplitude vs trials. In eachexperiment, the data were normalized relative to trial 1 of the control run. In ASW, theslope obtained was —0-44. When AVT was superfused over the abdominal ganglion, therate of decrement was increased to —0-84. In addition to increasing the rate of synapticdepression almost two-fold, AVT also brought about suppression of the EPSP amplitudeas shown in Fig. 2.


LE-L7 EPSP

Fig. 4. The effects of 10 6moll ' arginine vasotocin (AVT) on synaptic transmissionbetween an LE sensory neurone and gill motor neurone (L7), in the isolated, con-tinuously perfused abdominal ganglion. The interstimulus interval was 20min. Actionpotentials were evoked by passing brief depolarizing current pulses into the LE neurone.The artificial sea water (ASW) was perfused at a rate of 3 ml min"', the control EPSPamplitude was 13-6 mV. Application of 10~6moll~' AVT at the same rate of perfusioncaused a reduction of the EPSP amplitude to 1-8 mV. Wash-out of the peptide with ASWfor 20min resulted in recovery of the EPSP amplitude to 15 mV. L7 was held hyper-polarized to — 80 mV to prevent spontaneous action potentials from occurring. Thesensory neurone action potential (not illustrated) was not measurably affected by peptideapplication.

Fig. 5. Effects of lOmmoll"1 tetraethylammonium ion (TEA) and 10~6moll~1 argininevasotocin (AVT) on action potential duration of sensory neurones in the abdominal andpleural ganglia. (A) Action potentials evoked in LE sensory neurones in artificial seawater (ASW) (control) and after 20 min constant perfusion (2-0 ml min~') of TEA. TEAcaused a slight increase in action potential (AP) amplitude, an increased duration of theAP peak, and caused an increase in AP duration, at the AP half-amplitude, from 1-61 to2-24ras. (B) Same neurone and same traces as in A on a different time scale. 10~6moll~'AVT was added to the TEA-containing ASW and perfused at the same rate for 20 min.The action potential evoked was reduced in its duration at half-amplitude from 2-24 tol-88ms. Wash-out of the peptide for IS min with TEA-containing ASW caused anincrease in the duration to 2-22 ms. Subsequent wash-out of TEA with ASW caused arecovery of AP duration to near control values. (C) Experiments similar to that in Bperformed with a sensory neurone of the VC cluster of the pleural ganglion. Half-amplitude duration for control action potential was l-32ms; 20 min constant perfusion ofTEA at 2-5 ml min"1 increased duration to 2-10 ms. 20 min perfusion of 10~6moll~1

AVT and TEA-ASW reduced duration to 1-71 ms; wash-out of peptide with TEA-ASWincreased duration to 2-05 ms. The duration returned to control values after wash-outwith ASW (not shown).

Suf$wession of synaptic rransmission 5 5

'(P > 0- 1) (Fig. 6). Thus, when TEA was u s d to increase the duration of the sensory neurone's AP, A V T brought about a significant reduction in AP duration.

Interpretation of the above results is hampered by the unclear specificity of TEA. which might be interacting with the peptide. Accordingly, an alternative method of

LE neurone

2.5 rns

LE neurone

10 mmoll- ' TEA (pre- and post-AVT) 25 mV I

Fig. 5


enhancing AP duration, frequency-dependent spike broadening (FDSB), was em-ployed to corroborate the above findings. Previously, it had been demonstrated thatFDSB is a very sensitive indicator of the ability of a neural active agent to affectbroadening (Hicks, Edstrom & Lukowiak, 1984; Edstrom & Lukowiak, 1985).When trains of APs are evoked in the sensory neurones (see Materials and Methods;Edstrom & Lukowiak, 1985), the duration of the AP increases so that by the fortiethAP in the train the duration of the AP is significantly increased (Fig. 7). Severalcontrol runs were performed to demonstrate reproducibility of FDSB in a sensoryneurone. Following the control runs, FDSB was elicited in the presence of AVT, thepeptide was then washed off and FDSB was again evoked. We found that AVTsuperfusion caused a reduction in FDSB (Fig. 7). This can be clearly seen when onetrain of APs is digitally subtracted from the one following (Fig. 8). A negativedeflection in Fig. 8A,B indicates that an AP of the second train repolarized earlierthan did its counterpart in the first train. The differences between trains are plottedagainst AP number in Fig. 8C. AVT had significant effects on the duration of the APfrom action potential number 10 to action potential number 40 (P< 0-001), whilethere are no significant differences between the control runs (P>0-05). Thus, AVT

130 T

2 120- •

110 - -

100 • •

o 9 0 - -

Control AVT Recovery

Fig. 6. The effects of arginine vasotocin (AVT) on the duration of the action potentialof sensory neurones from abdominal and pleural abdominal ganglia treated, in simi-lar experiments, with lOmmoll"1 tetraethylammonium (TEA). The duration, athalf-amplitude of each neurone's action potential after 20min constant perfusion inTEA-ASW (artificial sea water) was normalized to 100% (Ar = 9). The duration ofsubsequent action potentials was expressed as a percentage of this. Constant perfusion of10"6moir' AVT for 20min in the TEA-ASW caused a decrease of 10-7±4-7%(mean± S.E.M.) in the duration of the action potential of these neurones (P< 0-005).Subsequent wash-out of the peptide by TEA-ASW caused an increase to 108 ± 14-9% ofcontrol (P>01) .

Suppression of synaptic transmission 57

reduced AP duration when the AP was prolonged without the use of TEA. Theseresults were obtained primarily from the large pleural ganglion sensory neurones andsimilar results were obtained from LE sensory neurones.

55-0-T-

2 5 - 0 - -

Control

28-2

-65-0

0-0 11-3 16-9

Time (ms)

22-5 28-2

Fig. 7. The effects of arginine vasotocin (AVT) on frequency-dependent spike broaden-ing (FDSB) in pleural ganglion VC cluster sensory neurones. An action potential (AP)was evoked by passing short (3 ms) high-intensity (4-8 nA) stimulus pulses into thesensory neurones. Up to 40 APs were evoked, separated by 140-ms intervals. A completedescription of this technique is given in Edstrom & Lukowiak (1985). (A) Normal FDSBevoked in artificial sea water (ASW). Every fifth AP is shown. (B) 20min later,10~6mol I"1 AVT was continuously perfused over the ganglion and FDSB evoked. AVTbrought about a reduction in FDSB. Note the reduction in duration, best seen bycomparing the bases of the last action potentials in A with those in B.


E

DISCUSSION

Superfusion of the abdominal ganglion with AVT significantly suppressed theamplitude of the GWR evoked by tactile stimulation of the siphon, as reportedpreviously (Thornhill et al. 1981). We have, moreover, demonstrated that AVTsuppresses synaptic transmission at the central sensory to motor neurone synapse and

50 -r A Control-Control 50-r B Control2-AVT

o.E

-15-0

-100-•

0-0 5-6 113 16-9 22-5 28-2 0-0

Time (ms)

11-3 16-9 22-5 28-2

0 - 0 - r C

-50 - -

-10-0 - -

-15-0

10 20 30

Action potential number

40

Fig. 8. The effects of arginine vasotocin (AVT) application on frequency-dependentspike broadening (FDSB) in pleura! ganglion sensory neurones. (A) A preceding controlseries of action potentials (Control), not shown in Fig. 7A) was subtracted point forpoint, action potential for action potential, from the control series illustrated in Fig. 7A(Control). Every fifth trace is displayed. Some differences can be seen between thesecontrol series. However, they were not significant (P>0-05). (B) The series of actionpotentials in Fig. 7B (with 10~6moll~' AVT superfusion) was subtracted point forpoint, action potential for action potential, from Control- These data illustrate theeffects of 10~6moll~1 AVT on the action potential in this neurone. Effects weresignificant (P<00001). (C) Plot of amplitude differences at 8-8ms (the time of peakeffect of AVT at the fortieth action potential) against action potential number for theseries in A and B. Differences were significant after the tenth action potential in eachseries (P<001-00001).


Increases the rate of low-frequency homosynaptic depression observed at thesesynapses. Furthermore, AVT was shown to affect the duration of the AP in thesensory neurones, both when TEA was present and when FDSB was evoked. Theseresults suggest that an endogenous AVT-like peptide (Moore et al. 1981) may play arole in the mediation of the suppressed behavioural state \nAplysia.

In these experiments, we have used a higher concentration of AVT (10-6moll~1)than in our previous experiments ( lO^-lO^^molF1; Thornhill et al. 1981).Although the lower doses were sufficient to suppress gill reflex amplitude, they wereunable to elicit consistent changes in the duration of the sensory neurone actionpotentials. It is worth noting that, even at the higher AVT concentrations used inthis study, no changes in passive membrane properties of the follower cells wereobserved.

In the semi-intact preparation and the isolated ganglion, AVT brought about areversible suppression of the EPSP evoked in follower motor neurones by a sensor)'neurone action potential. Not only did AVT bring about a reduction in the size of theevoked EPSP but it also slowed the rate of its rise. These data are consistent withearlier reports that showed that AVT suppressed the compound EPSP evoked by aweak tactile stimulus applied to the siphon (Thornhill et al. 1981). Moreover, theysuggest that a modulation of the sensory to motor neurone synaptic transmission,rather than the attenuation of the sensory neurone generator potential, wasresponsible for the reduction in activity within these central reflex arcs. Again, as inthe previous study, we found AVT to be selective in its ability to reduce the EPSP.Spontaneous EPSPs and IPSPs associated with the activity of interneurone I (L10)and interneurone II networks, which control the spontaneous gill respiratory move-ments (Byrne & Koester, 1980), were unaffected by this peptide. Therefore, theAVT-induced suppression of synaptic transmission was observed specifically at thesensory to motor neurone synapses.

Presynaptic hypothesis

Several lines of evidence indicate that AVT acts presynaptically in mediating itseffects on sensory to motor neurone synaptic transmission. First, AVT did not affectthe membrane potential or passive membrane properties of the postsynaptic motorneurones. Second, AVT not only suppressed EPSP amplitude but also increased therate of low-frequency homosynaptic depression seen at the LE-L7 synapse, which isknown to be a presynaptic mechanism (Klein & Kandel, 1980). Previously, it hasbeen demonstrated that AVT increases the rate of gill reflex habituation evoked byrepeated tactile stimulation of the siphon (Thornhill et al. 1981) but, as far as weknow, this is the first time that the rate of low-frequency homosynaptic depressionhas been shown to be increased by a neuroactive agent. Although both serotonin(Klein & Kandel, 1980) and SCPB (Lukowiak, Edstrom & Colmers, 1984) affectEPSP amplitude at these synapses (both increase it), it has not yet been determinedwhether these neuroactive agents affect the rate of EPSP decrement or, for thatmatter, whether they might even prevent low-frequency homosynaptic depressionfrom occurring. It would appear that these two agents, as well as met-enkephalin


(Lukowiak, Thornhill & Edstrom, 1982) and FMRFamide (Lukowiak et al. 1986)'should be tested in a similar manner.

The third line of evidence indicating a presynaptic site of action comes from ourstudies of sensory neurone action potentials. We first examined the effect of AVT onsingle APs evoked in a sensory neurone and found that we could not detect, whilerecording in the cell body, any measurable difference between the AP waveform inASW and in AVT-containing ASW. Two methods of action potential prolongationwere thus employed to examine the actions of AVT with enhanced resolution. Whensensory neurone action potential waveforms were prolonged either by TEA perfusionor through the technique of FDSB, AVT did have a significant and reversible effecton AP duration. AVT caused, on average, a 10 % decrease in the duration of the APmeasured at half-amplitude. The decrease in AP duration seen here is similar, bothas regards conditions and results, to previous studies which demonstrated alterationsof an inward Ca2+ current (Dunlap & Fischbach, 1978; Mudge, Leeman &Fischbach, 1979; Shapiro, Castellucci & Kandel, 1980). A decrease in the inwardCa2+ current would result ultimately in less neurotransmitter being released from thepresynaptic terminal. Our working hypothesis is that AVT similarly affects aninward Ca + current when the AP is not initially prolonged and that the decreasedinflux of Ca + results in less transmitter being released. Somewhat similar resultshave also been obtained with met-enkephalin (Lukowiak et al. 1986). More recentevidence (Abrams et al. 1985) has shown that another neuroactive peptide, SCPB,has the opposite effect; that is, it prolongs the sensory neurone action potential whichresults in an increase in EPSP amplitude. Thus, other evidence shows that a peptidecan affect AP duration. However, although SCPB increased EPSP amplitude, wehave not been able to demonstrate that SCPB causes a reversible alteration in APduration, either with TEA or with FDSB (Lukowiak et al. 1984). Further exper-iments are now under way to clear up this situation and to test more directly the Ca2+

hypothesis.It is very tempting to speculate that the suppressed behavioural state observed in

Aplysia is mediated, at least in part, by peptides such as AVT. The suppressedbehavioural state in Aplysia has been associated both with food satiation (Lukowiak,1980) and sexual activity (Lukowiak & Freedman, 1983) and two peptides (AVT andmet-enkephalin) have been shown to bring about a similar suppression. We nowknow (Leonard, Colebrook & Lukowiak, 1984) that nalaxone, a relatively specificopioid antagonist, reverses the suppression of gill reflex behaviour in Aplysia actingas sperm recipients (females) but does not reverse the suppression associated withfood satiation. Thus, it is possible that more than one peptidergic pathway isinvolved in the mediation of the suppressed behavioural state.

Although this study has focused on the modulatory effects of AVT on the centralsensory to motor neurone synapses, previous reports on interactions between thecentral and peripheral nervous systems in Aplysia (Lukowiak & Peretz, 1977) suggestthat AVT must exert additional actions to yield the observed behavioural results. Insupport of this notion are recent data showing that the superfusion of AVT over theabdominal ganglion reduces the amplitude of the gill movement elicited by


activity in both L7 and LDG1 (Lukowiak & Murphy, 1985). Perhaps there aremultiple descending peptidergic pathways that impinge upon a network of inter-neurones in the abdominal ganglion, which in turn produces the set of effectsrequired to suppress gill reflex behaviour (see Lukowiak & Peretz, 1980; Lukowiak &Murphy, 1985). This possibility has received support from preliminary experimentsin which a saline containing a high concentration of divalent cations, which lowersneuronal excitability, blocked the suppression by AVT of synaptic transmission.Further experiments, however, are required to delineate between direct and indirecteffects of AVT.

In summary, we have shown that a peptide, AVT, which may be endogenous (orhave similar properties to an endogenous peptide) in Aplysia, brings about sup-pression of the excitatory pathway between the central sensory and central motorneurones. This suppression is probably the result of a presynaptic effect of AVT onthe sensory neurone action potential. Whether AVT is the peptide which mediatesthe suppression associated with food satiation remains to be determined.

REFERENCES

ABRAMS, T . W., CASTELLUCCI, V., CARMARDO, J. S., KANDEL, E. R. & LLOYD, P. E. (1985). Two

endogenous neuropeptides modulate the gill and siphon withdrawal reflex in Aplysia bypresynaptic facilitation involving cAMP-dependent closure of a serotonin-sensitive potassiumchannel. Proc. natn. Acad. Set. U.SA. 81, 7956-7960.

BYRNE, J. & KoESTER, J. (1978). Respiratory pumping: Neuronal control of a central commandedbehaviour in Aplysia. Brain Res. 143, 87-105.

CAREW, T. J., CASTELLUCCI, V., BYRNE, J. & KANDEL, E. (1979). Quantitative analysis of relativecontribution of central and peripheral neurones to gill withdrawal reflex in Aplysia californica.J. Neumphysiol. 42, 497-509.

DUNLAP, K. & FlSCHBACH, G. D. (1978). Neurotransmitters decrease the calcium component ofsensory neurone action potentials. Nature, Lond. 276, 836—839.

EDSTROM, J. P. & LUKOWIAK, K. (1985). Frequency-dependent action potential prolongation inAplysia pleural sensory neurones. Neurosci. 16, 451-465.

FRAZIER, W. T., KANDEL, E. R., KUPFERMANN, I., WAZIRI, R. & COGGESHALL, R. E. (1967).

Morphological and functional properties of identified neurons in the abdominal ganglion ofAplysia californica. J. Neurophysiol. 30, 1288-1351.

GOLDBERG, J. I. (1983). Transfer of habituation in Aplysia californica: Evidence for aheterosynaptic mechanism underlying habituation of the gill-withdrawal reflex. Ph.D. thesis,University of Calgary, Alberta.

GOLDBERG, J. I. & LUKOWIAK, K. (1984). Transfer of habituation in Aplysia: Contribution ofheterosynaptic pathways in habituation of the gill withdrawal reflex. J. Neurobiol. 15, 395-411.

HICKS, T. P., EDSTROM, J. P. & LUKOWIAK, K. (1984). Effects of octopamine and serotonin onneurones of Aplysia californica. In Neurobiology of the Trace Amines (ed. A. A. Boulton, G. B.Baker, W. G. Dewhurst & M. Sandier), pp. 225-246. Clifton, NJ: Humana Press.

KLEIN, M. & KANDEL, E. R. (1980). Mechanism of calcium current modulation underlyingpresynaptic facilitation and behavioural sensitization in Aplysia. Proc. natn. Acad. Set. U.SA.77, 6912-6916.

LEONARD, J. L., COLEBROOK, E. & LUKOWIAK, K. (1984). Unitary drives or general arousal?Models and physiological evidence in Aplysia. Soc. Neurosci. Abstr. 10, 509.

LUKOWIAK, K. (1980). CNS control over gill reflex behaviors in Aplysia: Satiation causes anincrease in the suppressive control in older but not younger animals..7. Neurobiol. 11, 591—611.

LUKOWIAK, K., EDSTROM, J. P. & COLMERS, W. F. (1984). Peptidergic (SCPB and FMRFamide)modulation of gill reflex behaviors and associated neuronal activity in Aplysia. Soc. Neurosci.Abstr. 10, 510.


LUKOWIAK, K. & FREEDMAN, L. (1983). The gill withdrawal reflex in sexually active Aplysia issuppressed. Can.J. Physiol. Pharmac. 61, 743-748.

LUKOWIAK, K., GOLDBERG, J. I., COLMERS, W. F. & EDSTROM, J. P. (1986). Peptide modulationof neuronal activity and behaviour in Aplysia. In Comparative Aspects of Opioid and RelatedNeumpeptide Mechanisms (ed. G. Stefano & M. Leung). Boca Raton: CRC Press (in press).

LUKOWIAK, K. & MURPHY, A. D. (1985). Neurally active peptides activate central modulators ofthe gill withdrawal response in Aplysia. Soc. Neurosci. Abstr. 15, 642.

LUKOWIAK, K. & PERETZ, B. (1977). The interactions between the central and peripheral nervoussystems in the mediation of gill withdrawal reflex behaviour in Aplysia. J. comp. Physiol. 117,219-244.

LUKOWIAK, K. & PERETZ, B. (1980). Control of gill reflex habitation and the rate of EPSPdecrement of L7 by a common source in the CNS of Aplysia. J. Neurobiol. 11, 425-433.

LUKOWIAK, K., THORNHILL, J. A. & EDSTROM, J. P. (1982). Methionine enkephalin increasesCNS suppressive control exerted over gill reflex behaviors and associated neural activity inAplysia californica. Reg. Peptides 3, 303-312.

MOORE, G. J., THORNHILL, J. A., GIKLL, V., LEDERIS, K. & LUKOWIAK, K. (1981). An arginine

vasotocin-like neuropeptide is present in the nervous system of the marine mollusc Aplysiacalifornica. Brain Res. 206, 213-218.

MPITSOS, G. J. & LUKOWIAK, K. (1985). Learning in gastropod molluscs. In TheMollusca, vol. 8,Neurobiology and Behavior (ed. A. 0 . D. Willows), pp. 95-267. New York: Academic Press.

MUDGE, A. W., LEEMAN, S. E. & FISCHBACH, G. D. (1979). Enkephalin inhibits release ofsubstance P from sensory neurones and decreases action potential duration. Proc. natn. Acad.Sri. U.SA. 76, 526-530.

PERETZ, B. & LUKOWIAK, K. (1975). Age-dependent CNS control of the habituating gillwithdrawal reflex and of correlated activity in identified neurones of Aplysia. J. comp. Physiol.103, 1-7.

RUBEN, P., GOLDBERG, J., EDSTROM, J., VOSHART, K. & LUKOWIAK, K. (1981). What the marinemollusc Aplysia can tell the neurologist about behavioural neurophysiology. Can.J. neurol. Sri.8, 275-280.

SHAPIRO, E., CASTELLUCCI, V. F. & KANDEL, E. R. (1980). Pre-synaptic inhibition in Aplysiainvolves a decrease in the Ca+ + current of the presynaptic neurone. Proc. natn. Acad. Sri.U.SA. 77, 1185-1189.

THORNHILL, J. A., LUKOWIAK, K., COOPER, K. E., VEALE, W. L. & EDSTROM, J. P. (1981).

Arginine vasotocin, an endogenous neuropeptide of Aplysia suppresses the gill withdrawal reflexand reduces the evoked synaptic input to central gill motor neurones. J. Neurobiol. 12, 533-544.

WALTERS, E. T., BYRNE, J., CAREW, T. & KANDEL, E. (1983). Mechanoafferent neuronesinnervating the tail of Aplysia. II . Modulation by sensitizing stimulation. J?. Neurophysiol. 50,1543-1559.

Documents

SUPPRESSION OF SENSORY TO MOTOR SYNAPTIC … · rate of 3APsh~' to avoid low-frequency homosynaptic depression (Goldberg & Lukowiak, 1984). In one set of experiments, low-frequency