Long-lasting inhibition of neuron R 15 OF Aplysia: Role of the interneuron II network

Camp. Biochem.Physiol. Vol. 78A, No. I, pp. 83-89. 1984 Printed in Great Britain

LONG-LASTING APLYSIA:

ROLE J. D.

0300-9629184 $3.00 + 0.00 C 1984 Pergamon Press Ltd

INHIBITION

OF NEURON

OF THE INTERNEURON ROTH,’

K.

LUKOWIAK’

and R. W.

R15 OF

II NETWORK

BERRY*’

‘Department of Psychiatry, University of Chicago, Chicago, Illinois, USA *Department of Medical Physiology, Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada ‘Department of Cell Biology and Anatomy, Northwestern University, School of Medicine, 303 E. Chicago Ave., Chicago, IL 60611, USA

(Received 25 July 1983) of the branchial or either connective nerve of the abdominal ganglion of Aplysia responses in cells R15, L8, L9, and Ll 1 which are indistinguishable from those arising from spontaneous interneuron II (INT II) activity. 2. Threshold for the INT II-like response in all cells is identical, suggesting that the response is mediated by INT 11 activity. 3. The magnitude of the response in each cell increases with stimulus intensity and is subject to both temporal and spatial summation, implying the existence of multiple fibres in each nerve which converge on INT II. 4. Repetitive stimulation evokes long-lasting inhibition in R15. The onset of this phenomenon is always Abstract-l.

Stimulation

californica evokes simultaneous

accompanied by an INT II burst in the other follower cells. 5. Long-lasting inhibition in R15 is not accompanied by prolonged INT II activity, suggesting an endogenous mechanism of inhibition. 6. The phosphodiesterase inhibitor, IBMX, potentiates the response of R15 to nerve stimulation without affecting threshold for the resoonse. This is consistent with inhibition by a mechanism endogenous to R15.

-

INTRODUCTION

(INT II, nomenclature of Frazier et al., 1967) and postulated that the effects of nerve stimulation were mediated by activation of INT II. INT II is a central burst generator which makes extensive contacts with neurons of the abdominal ganglion and which functions as a command center for respiratory pumping (Byrne and Koester, 1978). INT II is probably a multineuronal network consisting of coupled “command” neurons which generate periodic bursts and which are presynaptic to “relay” cells, each of which is presynaptic to some, but not all, of the follower cells such as R15 (Byrne and Koester, 1978). Thus, there are actually four hypotheses which could account for the prolonged nature of the LLI. (1) It is due to a mechanism intrinsic to R15; (2) it is due to prolonged activity in the INT II burst generators; (3) it is due to prolonged activity in the INT II relay cell presynaptic to R15; (4) it is due to prolonged activity in some interneuron which is not part of the INT II network. The most direct test of these alternatives would require recording from the relay neuron onto R15 during an LLI. Unfortunately, this neuron has yet to be identified (Byrne and Koester, 1978). The only evidence yet available concerning the possibility of repetitive interneuronal activity comes from the report of Parnas and Strumwasser (1974) that tetrodotoxin-induced abolition of impulse activity in the ganglion within 15 min of LLI onset did not shorten LLI duration. This indirect evidence does not preclude the possibility of a tetrodotoxin-insensitive interneuron continuing to supply transmitter to R15, nor does it exclude the possible necessity for repetitive interneuronal activity early in the LLI.

Long-duration electrophysiological consequences of synaptic transmission are of interest both in terms of their cellular mechanisms and in terms of their implications for mechanisms of behavioral plasticity. One such effect which has been intensively studied is the so-called long-lasting inhibition, or LLI, which occurs in Aplysiu neuron Rl5 (Parnas et al., 1974) and its homologue in Helix, neuron F-l (Judge et al., 1978). Both cells are endogenous pacemaker neurons of the bursting type, and each can be hyperpolarized and inhibited for up to an hour by ten or fewer shocks applied to an appropriate nerve at a frequency of 1 Hz. Single stimuli evoke a depolarizing-hyperpolarizing biphasic postsynaptic potential (BPSP), and BPSP summation leads to the LLI. The initial hyperpolarization of the BPSP is due to an increase in Cl- conductance (Parnas and Strumwasser, 1974), and in the LLI this is followed first by an increase in K+ conductance, then by a decrease in conductance to Na+ or Ca*+ (Adams et al., 1980). A primary unresolved issue is whether the prolonged duration of the LLI in R15 arises from factors intrinsic to the postsynaptic cell, or from the continuous release of presynaptic transmitter. Resolution of this issue requires knowledge of the neural circuitry responsible for the LLI. Parnas et al. (1974) noted that the BPSP evoked in R15 by nerve stimulation was similar or identical to the BPSP evoked in this cell by spontaneous activity in interneuron II

*To whom

correspondence

should

be addressed 83

J. D. ROTH et al.

84

An alternative strategy to provide indirect evidence concerning the necessity of repetitive INT II activity for LLI generation is feasible and forms the basis for the studies reported here. By recording simultaneously from R15 and other INT II follower cells during LLI and in situations where INT II activity is altered by experimental manipulations, correlations between INT II activity and LLI can be developed which can throw light on the role of INT II in LLI production. We have taken this approach, and the results presented here support the contention that activation of the INT II network is necessary for LLI generation, but that prolonged INT II activity is not a requirement for prolonged inhibition of R15. MATERIALS

AND METHODS

Aplysia californica, weighing lSG300 g, were obtained

from Pacific Bio-Marine Laboratories

(Venice, CA) and

held at 15°C in Instant Ocean (Aquarium Systems, Mentor, OH). The animals were not exposed to a controlled 1ight:dark cycle, and experiments were performed between 9 a.m. and 9 p.m. Following dissection, the abdominal ganglion was pinned out in a recording chamber holding

335 ml of artificial seawater (Berry, 1975) at 15-C unless otherwise noted. Standard intracellular techniques were used to obtain simultaneous recordings from R15 and either L8, L9, or Lll. Impalements were accomplished through the intact connective tissue sheath using microelectrodes filled with 3 M KC1 and having tip resistances of 5-20 MQ. Stimuli were delivered to the branchial, left pleuroabdominal connective, and/or right connective nerves via suction electrodes. Pulses of 1.5-2.0 msec duration, ranging from 5-6OV, were normally used. Some experiments were done in a 1 mM Car+ 72mM Mg’+ artificial seawater described previously (Berry, 1975). 3-Isobutyl-l-methylxanthine (IBMX) was obtained from Sigma Chemical Co. (St. Louis, MO) and added as a concentrated stock solution to the incubation chamber to achieve final concentrations ranging from O.l&l.OmM. Cells Lll and RI5 could be identified unambiguously by virtue of their size, color, location and electrophysiological characteristics (Frazier et al., 1967). Cells L8, L9,, and L9,, are not readily distinguishable on the basis of these properties, and since all receive a burst of ipsp’s from INT II, we did not attempt such a distinction. They will be referred to as L8/9 in this report. All cells are labeled according to the nomenclature of Frazier et al. (1967).

A

.

R15

B

R15

Fig. 1. Simultaneous recording of the response of R15 and other INT II followers to branchial nerve stimuli. (A) Recording from Rl5 and Lll. Lll has been hyperpolarized by passing 1OnA through the recording electrode to abolish action potential generation. Following a spontaneous INT II burst (arrow), the branchial nerve was stimulated with 2 msec shocks of increasing voltage. (B) Recording from R15 and L8/9. Calibration: horizontal-12 set in (A), 24 set in B; vertical-10 MV (Ll 1). 20 mV (R15 in A), 25 mV for both traces in (B).

85

Long-lasting neural inhibition

Fig. 2. The interval between INT II bursts is plotted against sequential burst number. Single suprathreshold shocks to the branchial nerve (arrows) cause burst frequency to increase briefly. Repetitive stimulation (3 set’ for 30 set -bold arrow) induces an LLI in R15 (horizontal bar) and reduces burst frequency.

RESULTS

Spontaneous

INT II activity

In the isolated ganglion INT II fires periodic bursts at rates ranging from 21 to 7.5 per hour (mean = 30/hr). These bursts produce characteristic post-synaptic effects in follower neurons R1.5, L8/9 and Ll 1 (Fig. 1). In accord with previous descriptions (Frazier et al., 1967; Kandel et al., 1967) we observed nearly simultaneous bursts of epsp’s in Ll 1, ipsp’s in L8/9, and biphasic depolarizing-hyperpolarizing effects in R15. In L8/9 and Ll 1, the response is clearly composed of multiple summating psp’s but the response in R15 is too slow for discrete unitary potentials to be observed. While INT II input occurs in each of these cells on a one-for-one basis, the inputs to a pair of cells are neither exactly simultaneous in onset nor equal in duration. The onset of the response in L8/9 or Ll l typically precedes the response in Rl5 by 2-4 sec. The duration of the response in RI 5 averages 15 set, whereas the duration in L8/9 and Ll 1 hardly ever exceeds 10 sec. Responses

to stimulution

of nerve trunks

As others have noted (Parnas et al., 1974) stimulation of the branchial nerve or either pleuroabdominal connective nerve evokes a biphasic psp (BPSP) in R1.5 which is similar in magnitude, duration and waveform to the input from INT II (Fig. 1). In agreement with previous studies, we find that stimulation of the right connective also evokes excitatory input to R15, but that branchial nerve stimulation does not. Similarly, left connective nerve stimulation results in a BPSP only, although this is usually smaller than those evoked by stimulation of the other nerves. The latency to BPSP onset is about 50 msec with branchial nerve stimulation. Stimulation of either nerve at intensities below threshold for the BPSP in R15 excites both L8/9 and L 11. With branchial nerve stimulation, the latency to the first epsp in both cells is less than lOmsec, suggesting a direct connection. At higher stimulus intensities, single shocks to a nerve produce a barrage of epsps in LI 1 which is similar in magnitude and

duration to spontaneous INT II input (Fig. 1A). Higher stimulus intensities also produce a brief barrage of epsps in L8/9, but this is followed by an ipsp burst (Fig. 1B). In both cells, the presence or absence of INT II-like input is correlated with the presence or absence of a BPSP in R15 (Fig. 1). This suggests that the observed effects of nerve stimulation at higher intensities are due to triggering an INT II burst. Stimulation at intensities which are subthreshold for BPSP generation does not affect the regular rhythm of INT II activity. However, repeated suprathreshold stimulation causes an increase in INT II frequency for several minutes, so long as this stimulation is subthreshold for a LLI (Fig. 2). This constitutes evidence for excitatory input to INT II in each of the nerves examined. As is the case for spontaneous INT II input, the postsynaptic potentials evoked in the follower cells by nerve stimulation are not unitary, but are graded functions of stimulus intensity (Fig. 1). This is in accord with the results of Adams el af. (1980) for R15. In addition, both temporal summation to two shocks to the same nerve and spatial summation to closely-spaced shocks in two nerves can be demonstrated (Fig. 3). It would appear from these data that there exist populations of fibres in each nerve which converge on the INT II network. The magnitude of the response to nerve stimulation recorded in R15 parallels that of the other cells (Fig. 1). This also supports the idea that nerve stimulation acts via the INT II network. Long-fasting

in~i~~ti~~

As others have noted (Parnas et al., 1974), a brief bout of repetitive stimulation to the branchial or connective nerves at intensities sufficient to activate INT II causes a dramatic and prolonged hyperpolarization of R15 to approximately -70 mV, accompanied by a cessation of spike activity (Fig. 4). In our experiments, this long-lasting inhibition, or LLI, had a mean duration of 20 min and could last as long as 40 min. We also confirmed previous reports of a membrane conductance increase at the onset of LLI which progressively declines with time (Fig. 4A). LLI’s were evoked in R15 while simultaneously recording from L8/9 or Lll to monitor INT II activity, and the results indicate that INT II activation always accompanies the development of an LLI: in no instance was it possible to evoke an LLI without activating INT II. Although INT II activation appears to be a prerequisite for LLI, INT II is not continuously active throughout the LLI, nor even throughout the period of increased conductance in RI 5 (Fig. 4). Indeed, the regular rhythm of INT II bursting is interrupted by the repetitive stimulation (Fig. 2). Although several INT II bursts may occur during the course of an LLI, as in Fig. 4, this is not always the case. To assess the possibility that the INT II network fires asynchronousiy during the LLI, we also evoked LLI’s while hyperpolarizing Ll 1 to make any enhancement of epsp activity more visible and could fmd no evidence of increased excitatory input (data not shown).

86

J. D. ROTH et al.

B

2b Br Fig. 3. Temporal and spatial summation of nerve inputs to INT II. (A) Temporal summation recorded in R15. In trace 1, a single 2 msec 30 V shock to the branchial nerve elicited two spikes and a small hyperpolarization. In trace 2, two pulses of the same intensity as above and separated by 10 msec evoke one spike and a greater hyperpolarization. Calibration: horizontal-l sec.; vertical-10 mV. (B) Spatial summation recorded simultaneously from Ll 1 (hyperpolarized as in Fig. 1) and R15. Responses to 2 msec pulses, at the voltage indicated, to the left connective nerve (LCN) or the branchial nerve (Br) are smaller than the response when both are stimulated simultaneously. Calibration: horizontal-12 set; verticalIOmV (LII), 20mV (R15).

Parnas et al. (1974) were unable to block the BPSP evoked in R15 by connective nerve stimulation in solutions of high Mg2 + /low Ca2 + concentration and raised the possibility that the nerve trunks contain INT II axons. We were unable to completely block the BPSP with high Mg2+/low Ca*+ solutions, but this treatment did attentuate the BPSP to such an extent that single nerve shocks produced no response in R15 (Fig. 5). However, during repetitive stimulation, BPSP’s were evoked and grew progressively larger in size, suggesting facilitation at a synapse which is relatively insensitive to the Mg* + /Ca’ + ratio

of the medium (Fig. 5). We could not evoke a full-blown LLI in high Mg’+/low Ca’+ media, but with repetitive stimulation, we could induce minuteslong inhibition of R 15. Psp’s recorded from L8/9 and Lll behaved similarly to the BPSP in R15. In this respect then, all cells studied are similar in terms of their response to nerve stimulation in high Mg2 + /low Ca2 + , and the BPSP and the LLI in R15 are attenuated in parallel. Efects

of IBMX

The phosphodiesterase

inhibitor,

isobutylmethyl-

Fig. 4. Effects of repetitive nerve stimulation. (A) Simultaneous recordings from RI5 and Ll 1 during an LLI. Hyperpolarizing current pulses are delivered to RI5 via the recording electrode to monitor the conductance change during the LLI. Solid circles denote spontaneous TNT II bursts. At the arrow, a train of 1.5 msec, 50 V pulses was delivered to the branchial nerve at 3 set-’ for 40 sec. Traces are continuous records. Calibration: horizontal-l min; vertical-25 mV. (B) Simultaneous recordings from RI5 and L8/9. Stimulation as in (A). In this instance, the LLI developed gradually. The upper and lower pair of traces are separated by a period of 8 min in which RI5 remained silent and one spontaneous INT II burst occurred. Calibration: horizontal-l min; vertical-30 mV.

Long-lasting

neural

A R15

11

Rl5

Fig. 4.

inhibition

87

J. D. ROTH et al.

88

I I 1

.

1,

I.,

. . . . . . .

*

-I

Fig. 5. Effect of low Ca2+/high Mg2+ on the responses of R15 and L8/9 to nerve stimulation. After one hour in low Ca* + /high Mg* + medium, a 1.5msec 60V pulse to the branchial nerve (arrow) produces no response in either cell. However, repetitive pulies at 2 set-’ (&rows) facilitate to give typical INT II responses. Calibration: horizontal1 set; vertical-l0 mV (RIS), 20 mV (L8/9).

xanthine (IBMX), has been reported to enhance both the depolarizing and hyperpolarizing phases of the bursting rhythm in R15 (Levitan et al., 1979). Thus it was of interest to determine its effects on the BPSP and LLI. IBMX potentiates the hyperpolarizing phase of the BPSP, an effect which can be seen most easily when a series of shocks of increasing intensity is given (Fig. 6). However, the threshold for evoking a BPSP is unaltered. The most obvious explanation for these observations is that IBMX is potentiating either the release of transmitter by neurons presynaptic to R15 or the response of R15 to that transmitter. LLI’s could be evoked in the presence of IBMX, and their duration was not altered by the drug. IBMX depolarized and excited L8/9 and Ll 1, and we could detect no change in spontaneous INT II activity. DISCUSSION

The foregoing

A

data provide

further support

‘60

for the

:

contention that the LLI evoked in R15 by nerve stimulation is initiated by transient INT II activity. Certainly, axons in the connective and branchial nerves provide excitatory input to INT II: in all cells studied, the response to nerve stimulation is identical to spontaneous INT II input, and threshold for this response is identical in pairs of cells recorded from simultaneously. Intermittent stimulation at intensities which are below threshold for LLI generation increase the rate of INT II bursting, suggesting a persistence of excitatory drive in the network, while stimulation massive enough to evoke an LLI reduces burst frequency, possibly reflecting fatigue in the network. Finally, the latency to onset of the response to nerve stimulation, and its non-unitary nature in L8/9 and Ll 1 also argue for excitation of ganglionic interneurons. Since the BPSP persists in low Ca* +/high Mg’+ media, Parnas er al. (1974) suggested that the effects of nerve stimulation result from the antidromic activation of INT II. However, our results (Fig. 5) imply the existence of at least one chemical synapse in the circuit, albeit one which is resistant to changes in the divalent cation concentration of the medium. This does not rule out the possibility of antidromic activation in part of the circuit, but the existence of spatial and temporal summation suggests that any INT II axons in the nerve trunks must be numerous. Byrne and Koester (1978) have proposed that INT II is actually a multineuronal circuit consisting of a population of burst-generating neurons which control the follower cells, such as those we have recorded from, through a population of intermediary relay cells. Since our results indicate differences in the time of onset and duration of spontaneous INT II inputs to the various followers, it seems likely that these inputs are not mediated by the same relay cells. Thus, it is important to ask if the responses to nerve stimulation are due to the activation of the common burst generators, or to direct activation of the relay cells. Certainly, all of our observations are consistent

1,o 1,5

2,o 3,o

B

Fig. 6. Effect of IBMX on the response of R15 to branchial nerve stimulation. Responses to 1.5 msec pulses at various voltages are indicated: (A) in artificial seawater; (B) after 1 hr in 0.5mM IBMX. Threshold for the BPSP is 15V in both cases. Calibration: horizontal-24sec; vertical-30mV (A). 25 mV (B).

Long-lasting

neural

with a direct input to the burst generators: The threshold for the INT II-like response is the same in each pair of cells tested, and the magnitudes of the responses seen in pairs of cells increase in parallel as stimulus intensity is raised. Thus, while we cannot eliminate the possibility of direct activation of relay cells, particularly with repetitive stimulation, we have no evidence to support this possibility. Our observations further indicate that INT II activation is necessary to produce the LLI, as an LLI was only seen in R15 in those instances when stimulation evoked INT II input to the other cells. Moreover, continuous INT II activity is not required for the entire duration of the LLI. Repetitive stimulation did evoke a more prolonged INT II-like response in L8/9 and Ll 1 than did single stimuli, but in no case did this persist more than a few minutes. Furthermore, in some cases, INT II bursts continued at reduced frequency during the LLI, suggesting that the inhibition in R15 runs its course more or less independently of INT II activity. The absence of observable epsp’s in Lll after LLI onset would appear to preclude the possibility of continuing desynchronized activity in the INT II generators. It remains possible that the LLI is due to continuing activity in a portion which is presynaptic to Rl5,

of the INT II network but not the other cells.

However, the preponderance of the evidence suggests that the mechanism responsible for the prolonged inhibition is endogenous to R15. An extensive series of investigations by Levitan and colleagues (Levitan et al., 1979) suggests that this mechanism may involve the cyclic nucleotide system. Since an increase in the cAMP/cGMP ratio in this cell leads to inhibition, it might be hypothesized that INT II input increases CAMP levels. That IBMX enhances the hyperpolarizing phase of the BPSP is consistent with this notion.

inhibition

89

Acknowledgements-This work was supported by grants from MRC Canada and the Alberta Heritage Foundation for Medical Research (KL), and a Visiting Scientist Award from the Alberta Heritage Foundation for Medical Research (RWB).

REFERENCES Adams W. B., Pamas I. and Levitan I. B. (1980) Mechanism of long-lasting inhibition in Aplysia neuron R15. J. Neurophysiol. 44, 1148-l 160. Berry R. W. (1975) Functional correlates of low molecular weight peptide synthesis in Aplysia neurons. Brain Res. 86, 323-333. Byrne J. H. and Koester J. (1978) Respiratory pumping: neuronal control of a centrally commanded behavior in Aplysia. Brain Res. 143, 87-105. Frazier W. T., Kandel E. R., Kupfermann I., Waziri R. and Coggeshall R. E. (1967) Moruholoeical and functional properties of identified neurons -in the abdominal ganglion of Aplysia calijornica. J. Neurophysiol. 30, 1288-1351. Judge S. E., Kerkut G. A. and Walker R. J. (1978) Long-lasting hyperpolarization in a pacemaker neurone. Comp. Biochem. Phvsiol. 61A. 475481. Kandei E. R., Frazier W. T., Waziri R. and Coggeshall R. E. (1967) Direct and common connections among identified neurons in Aplysia. J. Neurophysiol. 30, 1352-1376. Levitan I. B., Harmar A. J. and Adams W. B. (1979) Synaptic and hormonal modulation of a neuronal oscillator: A search for molecular mechanisms. J. exp. Eiol. 81, 131-151. Pamas I. and Strumwasser F. (1974) Mechanisms of longlasting inhibition of a bursting pacemaker neuron, J. Neurophysiol. 37, 609-620. Pamas I., Armstrong D. and Strumwasser F. (1974) Prolonged excitatory and inhibitory synaptic modulation of a bursting pacemaker neuron. J. Neurophysiol. 37, 594608.