JOURNALOF NEUROPHYSIOLOGY Vol. 70, No. 3, September 1993. Printed
in U.S.A.
Excitatory Amino Acid Neurotransmission at Sensory-Motor and Interneuronal Synapses of Aplysia caZij?mica LOUIS-ERIC
TRUDEAU
AND
VINCENT
Laboratoire de Neurobiologie et Comportement, Sciences Neurologiques, Universite’ de Montrkal, SUMMARY
AND
F. CASTELLUCCI
Institut de Recherches Cliniques de MontGal, Montreal, Quebec H2 W 1R 7, Canada
CONCLUSIONS
1. Although
the gill and siphon withdrawal reflex of Aplysia has been used as a model system to study learning-associated changes in synaptic transmission, the identity of the neurotransmitter released by the sensory neurons and excitatory interneurons of the network mediating this behavior is still unknown. The identification of the putative neurotransmitter of these neurons should facilitate further studies of synaptic plasticity in Aplysia. 2. We report that sensory-motor transmission within this circuit is mediated through the activation of an excitatory amino acid receptor that is blocked by the non-N-methyl-D-aspartate excitatory amino acid receptor antagonists 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX ) and 1- (4-chlorobenzoyl) -piperazine-2,3dicarboxylic acid (CBPD). Compound postsynaptic potentials evoked in motor neurons by electrical stimulation of the siphon nerve were blocked by 92% with CNQX (75 PM) and 89% with CBPD (75 PM). 3. Simultaneous intracellular recordings were obtained from sensory neurons, excitatory interneurons, and motor neurons. Monosynaptic excitatory postsynaptic potentials (EPSPs) evoked in motor neurons by an action potential in a sensory neuron were blocked by 86% with CNQX (75 PM) and 7 1% with CBPD (75 PM). The two antagonists also blocked monosynaptic interneuronal EPSPs onto motor neurons by 65% and 67%, respectively. 4. Potential agonists of the synaptic receptors were puff-applied in the intact abdominal ganglion. Homocysteic acid (HCA) was found to mimic the action of the synaptically released transmitter because it strongly excites motor neurons. This effect was blocked by CNQX. Kainate and domoic acid were also effective agonists. 5. The actions of L- and D-glutamate as well as quisqualate were found to be mainly hyperpolarizing, whereas aspartate and ( +) -amino-3-hydroxy-5-methylisoxazole-4-proprionic acid had no effect. 6. Several reasons may be proposed to explain the inability of puff-applied glutamate to excite effectively the postsynaptic neurons in the intact ganglion. It is possible nonetheless that other endogenous amino acids such as HCA act as neurotransmitters at these synapses. INTRODUCTION
Studies of synaptic transmission within the small neuronal networks mediating withdrawal reflexes in Aplysia have identified some cellular and molecular mechanisms that may take part in modifications of the CNS in relation to learning and memory (Braha et al. 1990; Byrne et al. 1993; Castellucci and Kandel 1976; Dash et al. 1990; Goldsmith and Abrams 199 1; Klein 1993; Klein et al. 1982; Mayford et al. 1992; Montarolo et al. 1986). The central component of the gill and siphon withdrawal ( GSW) reflex involves I )
Centre de Recherches en
excitatory synapses from sensory neurons to excitatory interneurons and to motor neurons, 2) excitatory synapses from excitatory interneurons to inhibitory interneurons and to motor neurons, and 3) inhibitory synapses that produce feedback inhibition onto excitatory interneurons and in some cases onto sensory neurons (Castellucci et al. 1970; Frost et al. 1988; Hawkins et al. 198 1; Trudeau and Castellucci 1992). Although it appears that acetylcholine ( ACh) activates transmitter-gated Cl- channels at many inhibitory synapses of the circuit ( Trudeau and Castellucci 1993 ), the identity of the transmitter( s) released at the synaptic terminals of sensory neurons and excitatory interneurons is still unknown. In Aplysia CNS, the only fast excitatory synapses where the neurotransmitter is known are 1) the monosynaptic excitatory postsynaptic potential (EPSP) that is recorded in neuron R15 of the abdominal ganglion on threshold stimulation of an axon in the right pleuroabdominal connective (Frazier et al. 1967; Woodson et al. 1975) and 2) some of the synapses made by neuron LlO (Giller and Schwartz 197 1; Kandel et al. 1967; Segal and Koester 1980, 1982; Waziri and Kandel 1969). These EPSPs are produced by the activation of cationic cholinergic receptors with nicotinic pharmacology. Previous work has shown, however, that sensory-motor EPSPs are not blocked by cholinergic antagonists such as d-tubocurarine and hexamethonium ( Segal and Koester 1982; Trudeau and Castellucci 1993 ). Together with the observation that most motor neurons of the GSW reflex are inhibited by ACh (Kandel et al. 1967), this makes it unlikely that ACh is the transmitter of sensory neurons. Many central excitatory synapses of vertebrates appear to utilize amino acids such as glutamate and possibly aspartate as their neurotransmitter (Brodin et al. 1988; Dryer 1988; Gasic and Hollmann 1992; Nicholls and Attwell 1990; Tsumoto 1990). This has been suggested to be the case for excitatory transmission from primary afferents in several vertebrate species (Brodin et al. 1987; Christenson and Grillner 199 1; Jahr and Yoshioka 1986; Takeuchi et al. 1983 ) . There is also some evidence that glutamate could be released at some excitatory synapses in the CNS of invertebrates (De Santis and Messenger 1988; Quinlan and Murphy 199 1). Other substances such as ATP (Edwards et al. 1992) and serotonin (Sugita et al. 1992) appear to be released at some excitatory synapses in the CNS of the rat. Homocysteic acid (HCA) has also been proposed as an endogenous neuroexcitant in the vertebrate CNS. It has been shown that in various brain regions HCA excites neurons
(X)22-3077/93 $2.00 Copyright 0 1993 The American Physiological Society
1221
1222
L.-E. TRUDEAU
AND V. F. CASTELLUCCI
by activating excitatory amino acid receptors of the N- glass pipettes with a tip diameter of 5-6 pm. Pipettes were backmethyl-D-aspartate (NMDA) and non-NMDA types filled with light paraffin oil and tip-filled with the agonist. Brief pulses were applied with a pressure injection system (Medical Sys(Cuenod et al. 1990; Do et al. 1986a; Ito et al. 199 1; Neal and Cunningham 1989 ) . These latter observations raise the tems). Two to five puffs of 20 ms in duration with a pressure of 35 possibility that excitatory amino acid receptors may have as kPa were usually used. Statistical analyses consisted of either analyses of variance (ANOVAs) (with post-hoc Tukey tests) or Stunatural ligands amino acids other than glutamate or aspar- dent’s t test, where appropriate. Data are presented as means ~fiSE. tate. A recent report has provided some indirect support for such a notion by demonstrating in the goldfish cerebellum Drugs the existence of high-affinity kainate binding sites that are L-HCA, L-aspartic acid, D-glutamic acid, N-acetyl-aspartyl-gluglutamate insensitive (Davis et al. 1992). In the cockroach, acid (AP5 ) were obPeriplaneta americana, Wafford et al. ( 1992) have also tamate, and 2-amino-5-phosphonovaleric demonstrated that an identified motoneuron could be tained from Sigma (St. Louis, MO). Ketamine hydrochloride, NMDA, D-glutamylamino methanesulfonic acid strongly excited by kainate and domoate but inhibited by AMPA, (GAMS), kainic acid, and ( + )-MK-80 1 were from Research BioL-glutamate. chemicals (Natick, MA). L-Glutamate was from Schwartz BioreHere we report that the non-NMDA excitatory amino search (New York, NY). Tetrodotoxin (TTX) and L-cysteic acid acid receptor antagonists 6-cyano-7-nitroquinoxaline-2,3were from Calbiochem (La Jolla, CA). L-Cysteic sulfinic acid was dione (CNQX) (Honore et al. 1988) and I-( 4-chlorobenfrom Aldrich (Milwaukee, WI). CNQX, CBPD, philanthotoxinzoyl) -piperazine-2,3-dicarboxylic acid ( CBPD ) ( Davies et 343 (PhTX), quisqualate, domoic acid, and HA-966 were from To& Neuramin (Bristol, UK). al. 1984; Ganong et al. 1986) are effective blockers of monosynaptic sensory-motor and interneuronal EPSPs of the GSW network. Additionally, HCA as well as kainic and RESULTS domoic acid have a strong excitatory effect on neurons of Efect of CNQX and CBPD on compound EPSPs in motor the GSW reflex network. This effect is also blocked by neurons CNQX. On the other hand, puff-applied glutamate, quisqualate, N-acetyl-aspartyl-glutamate, cysteic acid, and cysWe first verified the effect of various excitatory amino teic sulfinic acid have a predominantly hyperpolarizing ef- acid receptor blockers on EPSPs evoked in motor neurons fect, whereas aspartate, ( +) -amino-3-hydroxy+methyliof the GSW reflex by electrical stimulation of the siphon soxazole-4-proprionic acid (AMPA) and NMDA have no nerve. The interstimulus interval was 2 min. At this freeffect. These results suggest that the neurotransmitter re- quency of stimulation, the EPSPs are stable during the conleased at the synaptic terminals of sensory neurons of this trol period. A compound EPSP was recorded in a motor circuit is an excitatory amino acid. HCA may be considered neuron of the LFS group or in motor neuron L7 polarized a candidate. to -80 mV to prevent firing. This allowed us to screen simultaneously for effects of the blockers on sensory-motor METHODS and interneuronal transmission because these responses are m 75% polysynaptic ( Trudeau and Castellucci 1992). Preparation In preliminary experiments, excitatory amino acid recepAplysia calzfornica were obtained from Marinus (Venice, CA). tor antagonists of the NMDA subtype (AP5, MK-80 1, keExperiments were performed on the isolated and desheathed abtamine, and HA-966) were found to have no effect on syndominal ganglion, which was dissected from the animal along aptic transmission (results not shown). On the other hand, with long sections of the siphon nerve, branchid nerve, and pleurabath application of the non-NMDA antagonist CNQX proabdominal connectives. The nerves were aspirated into suction duced a powerful blockade of compound EPSPs. At a conelectrodes for extracellular stimulation. The ganglion was percentration of 75 PM, CNQX reduced the magnitude of fused with artificial seawater ( ASW) that was composed of (in mM) 460 NaCl, 30 MgCI,, 25 MgSO,, 11 CaCl,, 10 KCl, and 10 compound EPSPs (as quantified by their surface area) by N-2-hydroxyethylpiperazine-l\r-2-ethanesulfonic acid (HEPES) , 91.9&2.8%(mean-tSE)(n=9)(t=5.3,iP