Neuropharmacological studies on the receptors mediating responses to carbachol, amino acids and octopamine on Limulus and Hirudo central neurons

Intracellular recordings were made from central neurons of Limulus polyphemus and from Retizu cells of the leech, Hirudo medicinalis. The effects of carbachol, amino acids and octopamine were examined on these neurons. Octopamine was found to have a mainly inhibitory effect on a few Limulus neurons. The effects of octopamine were mimicked by clonidine and napthazoline but not by xylazine. Both compounds were slightly more potent than octopamine. Yohimbine, metoclopramide, chlorpromazine and chlordimeform failed to antagonize this octopamine response. The excitatory effect of carbachol was blocked by alpha-bungarotoxin, 10(-7)M. Neither this concentration nor higher concentrations of alpha-bungarotoxin had any effect on L-glutamate excitation. m-Carboxyphenyl derivatives of alanine and glycine acted differentially on Limulus neurons responding to L-glutamate. m- Carboxyphenylglycine only inhibited neurones which showed a biphasic response to L-glutamate while m- carboxyphenylalanine only excited these neurons. Both compounds excited leech Retzius cells, with m- carboxyphenylalanine being about 20 times more potent than m- carboxyphenylglycine . The actions of alpha- ketokainate and allo-alpha- ketokainate were compared to kainate, dihydrokainate and L-glutamate on leech Retzius cells. The equipotent molar ratios for kainate, dihydrokainate , alpha- ketokainate and allo-alpha- ketokainate were 0.0029 +/- 0.0004, 0.021 +/- 0.047, 0.029 +/- 0.005 and 0.14 +/- 0.0093 respectively with L-glutamate as one. All the analogues were more potent than L-glutamate. Quinolinic acid had no glutamate-like activity on either Limulus or Hirudo neurons. Methyltetrahydrofolate was inactive on Limulus neurons but excited leech Retzius cells, being slightly less potent than L-glutamate. Dibutyl cAMP terminated the excitatory actions of kainate on both Limulus and Hirudo neurons. Anisatin , a putative GABA antagonist, was a potent antagonist of GABA inhibition on Limulus neurons.     

The pharmacology of Limulus central neurons

1. Intracellular recordings have been made from neurons in the central nervous system of the horse-shoe crab, Limulus polyphemus. Neurons possess resting potentials between -40 and -60 mV, with action potentials ranging from 2-3 mV up to 60 mV in amplitude. Neurons also have excitatory and inhibitory postsynaptic potentials. 2. All the neurons studied are inhibited by GABA and excited by cholinomimetics. The GABA response is chloride mediated and reversibly antagonised by picrotoxinin but not by bicuculline or bicuculline methochloride or methoiodide. The cholinergic response is nicotinic and blocked by pentolinium, hexamethonium, chlorisondamine and dihydro-beta-erythroidine. 3. L-Glutamate can excite some cells, inhibit others and have a biphasic action, inhibition followed by excitation, on other cells. The inhibitory effect is chloride mediated and blocked by picrotoxinin. Ibotenate mimics the action of glutamate both in terms of inhibition and excitation but kainate and quisqualate only mimic the excitatory action of L-glutamate. 4. Dopamine, octopamine, 5-hydroxytryptamine and histamine excite some neurons while inhibiting others or have a biphasic action. Dopamine and octopamine normally have different effects on the same cell, suggesting they act via different receptors. Octopamine shows stereospecificity for the (-) isomer which is more than 100 times more active than the (+) isomer and octopamine is reversibly antagonised by phentolamine and cyproheptadine. 5. Proctolin has an excitatory action on these neurons and this effect is long lasting and can be potentiated by dibutyl cyclic AMP. 6. The pharmacology of Limulus central neurons is compared to the pharmacology of insect and crustacean central neurons. It is concluded that GABA and acetylcholine are central transmitters throughout the arthropods. It is also probable that L-glutamate and octopamine have a physiological role in the arthropod central nervous system. Proctolin appears to modify neuronal and muscle activity in the arthropods and has a modulatory or transmitter function.     

Neuromuscular synaptic transmission in Limulus polyphemus--I. Actions of aspartate, glutamate and the natural transmitter

Aspartate and glutamate were examined as excitatory transmitter candidates for the tibia flexor muscle of the chelicerate arthropod, Limulus polyphemus. Bath application of aspartate or glutamate caused dose-dependent depolarizations of Limulus muscle fibers and contractions of the whole muscle. Glutamate was about 10 times more potent than aspartate. Aspartate and glutamate depolarizations were associated with a conductance increase in muscle fibers, although aspartate depolarizations were dependent on external sodium, while glutamate depolarizations persisted in the absence of sodium. Although the Limulus excitatory postsynaptic potential (epsp) was associated with a conductance increase the ionic basis of the epsp could not be determined. If, however, the Limulus epsp, like other arthropod epsps, is sodium-dependent then the sodium-dependence of the aspartate depolarization is consistent with the action of the natural excitatory transmitter. The sodium-independence of glutamate action, however, is not consistent with generally accepted models of arthropod neuromuscular transmitter action. The rank order of potency for amino acid agonists indicates that the Limulus neuromuscular junction is pharmacologically very similar to other arthropod junctions which are well-accepted to be glutamatergic. Pentobarbital reversibly attenuated the amplitudes of the epsp and aspartate and glutamate depolarizations, and it was found to be the only useful antagonist in Limulus.     

Gliotoxic action of glutamate on cultured astrocytes

Because of the well-documented importance of glutamate clearance by astrocytes in protecting neurons against excitotoxicity, it was interesting to examine whether L-glutamate exerts a toxic action on cultured astrocytes. Cell damage was evaluated by measuring activity of lactate dehydrogenase (LDH) released into the culture medium. Exposure of astrocyte cultures of the neonatal rat cerebral cortex to L-glutamate resulted in a concentration- and time-dependent increase in the release of LDH. L-Glutamate-induced gliotoxicity appeared to be mediated predominantly by the increase of oxidative stress because the reduced glutathione content and its effects were almost completely blocked by vitamin E and pyrrolidinedithiocarbamate. To support this notion further, the supplementation or depletion of intracellular reduced glutathione content attenuated or worsened L-glutamate toxicity, respectively. Activation of the glutamate transporter mimicked the action of L-glutamate on astrocytes. In addition, degrees of cell damage were not directly correlated to the levels of glutamate uptake. Moreover, the mechanism of this toxicity required energy and macromolecular synthesis. Taken together, brief exposure to L-glutamate resulted in glutamate uptake and cell swelling, whereas sustained exposure injured astrocytes via oxidative stress instead of the excitatory mechanism.     

The effect of local application of homocysteine on neuronal activity in the central nervous system of the rat

Homocysteine, a monocarboxylic, sulfur-containing amino acid, produces convulsions in rats and mice when administered systemically. Convulsions and high serum concentrations of homocysteine are among the symptoms that characterize patients with homocystinuria, a hereditary disorder of amino acid metabolism. In order to evaluate the effects of homocysteine on the central nervous system directly, extracellular recordings were made from neurons in rat cerebral cortex, cerebellum and midbrain during local application of homocysteine by pressure ejection or iontophoresis. Both methods of drug delivery produced dose-dependent increases in the activity of neurons in every area tested. Activity was increased by D, L-homocysteine and L-glutamate in 67 percent of cells tested with both drugs. The doses required to produce equivalent excitations in this group of cells were similar, suggesting that homocysteine is at least as potent as glutamate. The excitatory effects of both homocysteine and glutamate were antagonized by local application of betaine, a biological methyl donor which blocks convulsions produced by systematic administration of pentylenetetrazol and electroshock as well as homocysteine. The effects of local application of homocysteine were also blocked by local application of the glutamate antagonist glutamate diethylester (GDEE). In 6 of 7 cells tested, GDEE appeared to preferentially affect homocysteine-induced excitations. These data indicate that homocysteine has an excitatory action on neurons, a finding which may account for some of the symptoms associated with certain disorders of amino acid metabolism.     

The mechanism of block of glutamate synapses by dipicolinic acid

The effect of dipicolinic acid (2,6-pyridine dicarboxylic acid) on the mealworm neuromuscular junction was studied using conventional microelectrode recording techniques. Dipicolinic acid (10^?5-10^?3 M) added to the bathing solution reversibly blocked neuromuscular transmission. The depolarization in response to iontophoretically applied L-glutamate (glutamate potential) was not affected by dipicolinic acid even when the neurally evoked excitatory postsynaptic potential (EPSP) was totally abolished. Focal extracellular recordings from single synaptic sites revealed that in the presence of 1 x 10^?4 M dipicolinic acid the presynaptic spike was unchanged, but the quantal content for evoked transmitter release was reduced. The calcium-dependent action potential elicited by direct stimulation of the muscle fiber was not impaired by dipicolinic acid. These results suggest that dipicolinic acid interferes with the transmitter-releasing mechanism from the presynaptic terminal.     

Horseshoe crab (Limulus polyphemus) lectin but not garden snail (Helix pomatia) lectin elicits insulin-like activities in vitro

Lectins from the horseshoe crab (Limulus polyphemus) and the garden snail (Helix pomatia) were tested for insulinomimetic activities in isolated rat epididymal adipocytes. The sialic acid binding horseshoe crab lectin suppressed epinephrine-induced lipolysis and augmented lipogenesis from D-[3-3H]-glucose while the N-acetylgalactosamine binding snail lectin was inactive. The results suggest that the insulin receptor on rat adipocytes contains sialic acid in its carbohydrate moiety but does not possess non-reducing alpha-D-galactopyranosyl or 2-acetamido-2-deoxy-alpha-D-galactopyranosyl end groups.     

Conformational aspects of the actions of some piperidine dicarboxylic acids at excitatory amino acid receptors in the mammalian and amphibian spinal cord

A series of piperidine dicarboxylates (PDA) have been tested for excitatory amino acid agonist and antagonist activity and for synaptic depressant properties in the spinal cords of frogs and immature rats in vitro and of cats in vivo. The substances tested comprised (±)-cis-2,3-PDA, (±)-cis-2,4-PDA, (±)-cis-2,5-PDA, (±)-cis-2,6-PDA, (±)-trans-2,3-PDA, (±)-trans-2,3-PDA and both (+) and (–) forms ofcis-2,3-PDA. Peak excitatory amino acid agonist activity was observed with (±)-trans-2,3- and (±)-trans-2,4-PDA. Excitatory amino acid antagonism and synaptic depressant activity was observed only withcis-dicarboxylates, this activity being greatest in the 2,3-analogue. The agonist actions of piperidine dicarboxylates were effectively depressed by the specific NMDA receptor antagonist, (–)-2-amino-5-phosphonovalerate and, where tested, also byd--aminoadipate and low concentrations of Mg²⁺. It was concluded that the major part of these agonist actions were mediated by NMDA receptors. The main structural feature of the NMDA agonist actions of these substances was considered to be their close relationship to N-alkyl-aspartic and glutamic acid molecules, with thetrans arrangement of the respective 2,3- and 2,4-situated carboxyl groups promoting most effective interaction with the active sites of the NMDA receptor. (±)-Cis-2,3-PDA depressed excitatory responses induced by NMDA, kainate, quisqualate, (±)-trans-2,3-PDA and (±)-trans-2,4-PDA, or evoked by dorsal root stimulation. Both monosynaptic and polysynaptic excitation were susceptible to the depressant action of this substance. The (–) isomer ofcis-2,3-PDA carried both excitatory amino acid agonist and antagonist activity and also the synaptic depressant properties observed with the racemic form of this substance. The (+) isomer showed little pharmacological activity. It is proposed that the structure-activity features of these heterocyclic amino acids indicate some of the conformational requirements for interaction with physiological excitatory amino acid receptors.This paper is dedicated to Dr. Derek Richter on his seventy-fifth birthday.     

Glutamate receptors on the somata of dorsal unpaired median neurons in cockroach,Periplaneta americana,thoracic ganglia

Effects of application of glutamate and glutamatergic ligands were studied to characterize the receptors for glutamate present on the soma membrane of the dorsal unpaired median (DUM) neurons in the thoracic ganglia of the cockroach, Periplaneta americana, using the intracellular recording technique. Application of L-glutamate did not block the GABA-response, and application of beta-guanidino-propionic acid, a competitive antagonist for GABA, failed to block the response to L-glutamate. These results indicate that most of L-glutamate action may not be mediated by a GABA-activated channel. To examine glutamate receptor types on the DUM neurons, glutamate receptor agonists were applied. The ionotropic glutamate receptor (iGluR) agonists evoked depolarizations with the following relative rank of order of potency: kainate > AMPA > quisqualate. Metabotropic glutamate receptor (mGluR) agonists also elicited membrane depolarizations or hyperpolarizations associated with an increase in membrane conductance. The mGluR agonists evoked depolarizations or hyperpolarizations with the following relative rank of order: L-CCG-1 > 1S, 3R-ACPD > L-AP4. Depolarization of the same DUM neuron was detected following exposure of kainate and L-CCG-I, suggesting the coexistence of distinct iGluR and mGluR types. A membrane permeable cAMP analog, CPT-cAMP, could not mimic the effect of mGluR agonists. The mGluR selective antagonists, MCCG and MCPG, failed to antagonize the response to mGluR agonists. The involvement of cAMP in the mGluR response was not confirmed in DUM neurons. Although the functional roles of these receptors are unknown, it might be possible then that these extrasynaptic receptors have a modulatory effect on the excitability of the DUM neurons.     

L-glutamate as an excitatory transmitter at the neuromuscular junction of a beetle larva

The effects of L-glutamate and acetylcholine on the ventral muscle fibres of the larval mealworm Tenebrio molitor were studied by means of microelectrodes. Bath application of L-glutamate at concentrations higher than 1 × 10 ⁴M suppressed excitatory postsynaptic potentials (EPSPs) and evoked both a depolarisation and a reduction in the input resistance of the muscle fibre. In contrast, acetylcholine chloride (up to 1 mM) had no effect at all. Circumscribed spots could be detected on the fibre surface where iontophoretic applications of L-glutamate caused transient depolarizations (glutamate potentials). Focal extracellular recordings revealed that the glutamate sensitive spots were identical with synaptic sites. The reversal potentials of the EPSP and the L-glutamate potential were identical. These results are compatible with the hypothesis that L-glutamate is an excitatory transmitter at the neuromuscular junction.     

Molecular organization of glutamate-sensitive chemo-stimulated nerve cell membranes. Interaction of monoclonal antibodies with glutamate-binding membrane proteins in the rat brain, human neuroblastoma and molluscan neurons

Hybrid cells obtained by fusion of myeloma PX63-Ag8-653 with immune splenocytes of BALB/c mice were found to produce monoclonal antibodies with a high degree of specificity to rat and human brain. The kinetics of specific IgG binding to purified fractions of glutamate-binding membrane proteins from rat and human brain were analyzed in Scatchard plots. The presence of a single type of binding sites with Kd = 100 nM was demonstrated. The monoclonal antibodies were shown to inhibit the specific binding of tritium-labeled L-glutamate to different brain synaptic membranes. Addition of monoclonal antibodies to the incubation medium induced a modulating effect of physiological responses to L-glutamate in Planorbarius corneus neurons. The possible use of specific antibodies to glutamate-binding proteins as immunochemical markers for the study of glutamate receptor topography on membrane surface was demonstrated with the aid of neuroblastoma cells N18 Tg2a and rat brain tissue slices. An analysis of glutamate receptor binding sites with the use of monoclonal antibodies revealed that these antibodies specifically recognize the active center in the receptor molecules which have identical antigen determinant sites in different biological systems.     

Synthesis and neuroprotective effects of serofendic acid analogues

Analogues of serofendic acid were prepared and their protective effects against L-glutamate (Glu)-induced neurotoxicity were examined using primary cultures of rat cortical neurons. Some analogues exhibited similar neuroprotective activity to that of serofendic acid.     

L-glutamate suppresses amyloid beta-protein-induced stellation of cultured rat cortical astrocytes

Alzheimer's amyloid beta-protein (Abeta) has been reported to potentiate glutamate toxicity in neurons, but very little is known about interaction between Abeta and glutamate in astrocytes. Therefore, in the present study, we investigated the effects of Abeta and glutamate on morphology of astrocytes. Cultured rat cortical astrocytes exhibited polygonal morphology in the absence of stimulation and differentiated into process-bearing stellate cells following exposure to Abeta (20 microM). L-Glutamate (30-1,000 microM) had no effect on astrocyte morphology in the absence of stimulation but strongly suppressed Abeta-induced stellation. The suppressive effect of L-glutamate on Abeta-induced stellation was not mimicked by glutamate receptor agonists and not blocked by glutamate receptor antagonists. In contrast, the suppressive effect of L-glutamate was mimicked by D- and L-aspartate and transportable glutamate uptake inhibitors. These results suggest that Abeta-induced astrocyte stellation is suppressed by a mechanism related to glutamate transporters.     

Presence and distribution of immunoreactive and bioactive FMRFamide-like peptides in the nervous system of the horseshoe crab, Limulus polyphemus

FMRFamide immunoreactivity was detected in all regions of the Limulus nervous system, including the brain (6.5 +/- 0.6 pg FMRFamide/mg), cardiac ganglion (2.06 +/- 0.67 pg FMRFamide/mg), and ventral nerve cord (5.8 +/- 0.7 pg FMRFamide/mg). The distribution of immunoreactive FMRFamide (irFMRFamide) was mapped by immunofluorescence and the distribution corresponded to regional RIA data. A good proportion of the CNS and cardiac ganglion neuropile contained irFMRFamide, and fluorescent cell bodies were observed in several areas. High performance liquid chromatography (HPLC) was employed to separate and characterize the FMRFamide-like peptides from extracts of Limulus brains. HPLC fractions were analyzed using coincidental radioimmunoassay and bioassay (the radula protractor muscle of Busycon contrarium). There appear to be at least three FMRFamide-like peptides in the Limulus brain, including one similar to clam FMRFamide. FMRFamide acts on Limulus heart in a biphasic manner at relatively high concentrations (10(-5)M), but has no effect on the activity of the isolated ventral nerve cord. These data suggest that in Limulus FMRFamide-like peptides are acting as neurotransmitters, or neuromodulators.     

The ionic mechanisms associated with the excitatory response of kainate, L-glutamate, quisqualate, ibotenate, AMPA and methyltetrahydrofolate on leech Retzius cells

Intracellular recordings were made from Retzius cells from segmental ganglia of the leech, Hirudo medicinalis. The ionic mechanisms of the following compounds were examined: L-glutamate, ibotenate, quisqualate, AMPA, kainate, methyltetrahydrofolate and carbachol. All these compounds depolarise and excite Retzius cells. In sodium-free Ringer, the responses to L-glutamate, kainate, ibotenate and AMPA were greatly reduced, the response to quisqualate was reduced, the response to methyltetrahydrofolate was normal while the response to carbachol was abolished. In sodium-free high calcium Ringer the responses to L-glutamate, ibotenate and carbachol were absent, the responses to quisqualate and AMPA greatly reduced, the responses to methyltetrahydrofolate and kainate were normal. The methyltetrahydrofolate and kainate responses in sodium-free high calcium Ringer were greatly reduced on addition of cobalt. All the responses are associated with an increase in conductance, the increase being the largest in the case of kainate. It is concluded that the response to L-glutamate, ibotenate and carbachol are dependent on sodium, the responses to quisqualate and AMPA are mainly sodium dependent, possibly with a small calcium component. The kainate response in normal Ringer is largely sodium dependent but in sodium-free Ringer calcium can completely substitute for sodium. The methyltetrahydrofolate response appears to be sodium independent but at least partly calcium dependent. These studies provide further evidence that L-glutamate and ibotenate act on a common receptor on leech Retzius cells while kainate acts on a separate receptor which can activate a calcium ionophore. It is probable that methyltetrahydrofolate acts on a different ionophore system to kainate. N-Methyl-D-aspartate has no agonist activity on any of these receptors.     

The T-axial membrane system in striated muscles of the horseshoe crab

Cardiac muscle and skeletal muscles powering walking leg, tailspine and gill movements in Limulus polyphemus were studied by transmission electron microscopy. All muscles examined had extensive invaginations of the sarcolemma at the Z disc and at the lateral margins of the A band. Invaginations at the Z disc were often branched in the transverse and longitudinal planes, but branching was not observed at other locations within a sarcomere. Dyads, and occasional triads, were observed at the A band in all muscles examined.     

Interactions of phencyclidine with ion channels of nerve and muscle: behavioral implications

Phencyclidine (1-(1-phenylcyclohexyl)piperidine [PCP]), a behaviorally active analogue (1-(1-m-aminophenylcyclohexyl)piperidine [m-amino-PCP]), and two behaviorally inactive analogues (1-(1-m-nitrophenylcyclohexyl)piperidine and 1-piperidinocyclohexanecarbonitrile) block neuromuscular transmission, depress the amplitude and rate of rise of directly elicited action potentials in frog sartorius muscle, and cause voltage- and concentration-dependent decreases of the peak end-plate current amplitude. This implies that all four compounds block the ion channel of the acetylcholine (ACh) receptors. Only PCP and m-amino-PCP prolong the action potential, block delayed rectification, potentiate muscle twitch, increase quantal content of end-plate potentials, and block K+-induced 86Rb+ efflux from rat brain synaptosomes. PCP also possesses central and peripheral antimuscarinic activity but is much less potent than 3-quinuclidinyl benzilate (QNB). Atropine, scopolamine, and QNB require much higher concentrations to induce behavioral alterations than to block muscarinic receptors. Thus PCP and some of its behaviorally active and inactive derivatives share two common effects, blockade of the nicotinic ACh receptor-ion channel complex and blockade of central and peripheral muscarinic receptors. The feature that apparently separates behaviorally active from inactive derivatives of PCP is their ability to block K+ conductance (gK) and thereby potentiate muscle twitch and increase the release of transmitters from central and peripheral synapses. The similarity between PCP-induced behavioral alterations and primary schizophrenia in humans raises the possibility of involvement of an altered gK in the human disease.     

A comparison of the effect of kainate and some related amino acids on locomotor activity in cockroaches and electrical activity recorded from locust ventral nerve cord

The ED50 for loss of righting behaviour of cockroaches induced by kainate (43 mumol/kg body weight) indicated the toxicity of kainate to be much greater than would have been predicted from the excitatory action of this amino acid at insect skeletal muscle fibres. N-Methyl-D-aspartate had little effect on righting behaviour (ED50 greater than 3500 mumol/g body weight). Electrical recordings from the locust ventral nerve cord showed kainate (0.1-2 mM) to have a depolarizing action on neurons within the metathoracic ganglion. The depolarizing action of kainate was partially resistant to tetrodotoxin. The kainate-induced abolition of rostrally evoked potentials recorded in the abdominal connectives from the metathoracic ganglion suggests that the giant fibres are sensitive to kainate. Domoic acid was 46 times more potent than kainate. The lack of effect of N-methyl-D-aspartate (2 mM), dihydrokainate (2 mM), quisqualate (2 mM) and L-glutamate (20 mM) on nerve cords in the present experiments suggests that the kainate receptors in this preparation show a chemical selectivity comparable to that observed at vertebrate central neurones.     

Direct Evidence That Excitotoxicity in Cultured Neurons Is Mediated via N-Methyl-D-Aspartate (NMDA) as well as Non-NMDA Receptors

Cultured GABAergic cerebral cortex neurons were exposed to the excitatory amino acid (EAA) L-glutamate, kainate (KA), N-methyl-D-aspartate (NMDA), or RS-alpha-amino-3-hydroxy-5-methyl-4-isoxazolopropionate (AMPA). To ensure a constant glutamate concentration in the culture media during the exposure periods, the glutamate uptake inhibitor L-aspartic acid beta-hydroxamate was added at 500 microM to the cultures that were exposed to glutamate. Each of these EAAs was able to induce neurotoxicity. It was not possible to reduce or prevent glutamate-induced cytotoxicity by blocking only one of the glutamate receptor subtypes with either the NMDA receptor antagonist D-(-)-2-amino-5-phosphonopentanoate (APV) or with one of the specific non-NMDA antagonists 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and 6,7-dinitroquinoxaline-2,3-dione (DNQX). However, if the cultures were exposed simultaneously to glutamate and the antagonists in combination, i.e., APV plus CNQX or APV plus DNQX, the toxicity was completely prevented. Furthermore, CNQX and DNQX were shown to be selective blockers of cytotoxic phenomena induced by non-NMDA glutamate agonists with no effect on NMDA-induced cell death. Likewise, APV prevented NMDA-induced cell death without affecting the KA- or AMPA-induced neurotoxicity. It is concluded that EAA-dependent neurotoxicity is induced by NMDA as well as non-NMDA receptors.     

Abolition of the NMDA-mediated responses by a specific glycine antagonist, 6,7-dichloroquinoxaline-2,3-dione (DCQX)

Among various quinoxaline derivatives examined, only 6,7-dichloroquinoxaline-2,3-dione (DCQX) competitively displaced the strychnine-insensitive binding of [3H]glycine, without affecting the other binding sites on the N-methyl-D-aspartate (NMDA) receptor complex. This novel specific antagonist abolished the ability of L-glutamate to potentiate [3H]MK-801 binding activity in brain synaptic membranes treated with Triton X-100. Inclusion of glycine reversed this preventive action of DCQX on the potentiation induced by glutamate.