Abstract: The central neurotoxicity of the excitatory amino acid neurotransmitter glutamate has been postulated to participate in the pathogenesis of the neuronal cell loss associated with several neurological disease states, but the complexity of the intact nervous system has impeded detailed analysis of the phenomenon. In the present study, glutamate neurotoxicity was studied with novel precision in dissociated cell cultures prepared from the fetal mouse neocortex. Brief exposure to glutamate was found to produce morphological changes in mature cortical neurons beginning as quickly as 90 sec after exposure, followed by widespread neuronal degeneration over the next hours. Quantitative dose-toxicity study suggested an ED50 of 50-100 microM for a 5 min exposure to glutamate. Immature cortical neurons and glia were not injured by such exposures to glutamate. Uptake processes probably do not limit GNT in culture, as the uptake inhibitor dihydrokainate did not potentiate GNT. Possibly reflecting the lack of uptake limitation, glutamate was found to be actually more potent than kainate as a neurotoxin in these cultures, a dramatic reversal of the in vivo potency rank order. Some neurons regularly survived brief glutamate exposure; these possibly glutamate-resistant neurons had electrophysiologic properties, including chemosensitivity to glutamate, that were grossly similar to those of the original population.

Abstract: 1. Spinal cord and hippocampal neurones in cell culture were voltage clamped using the tight-seal, whole-cell recording technique. The concentration of sodium and a series of divalent cations in the extracellular media was varied to study permeation through excitatory amino acid receptor channels activated by the selective agonists N-methyl-D-aspartic acid (NMDA), kainic acid and quisqualic acid. 2. On raising the extracellular calcium concentration, with [Na+]o held constant at 105 mM, the reversal potential of responses to NMDA shifted in the depolarizing direction. This shift was adequately described by the extended constant-field equation over the range 0.3-50 mM-calcium. Using ionic activity coefficients we calculate a value of PCa/PNa = 10.6. Under the same experimental conditions the reversal potential of responses to kainic and quisqualic acids was much less affected by raising the calcium concentration, such that PCa/PNa = 0.15. A depolarizing shift of the NMDA reversal potential was also recorded during application of 20 mM-barium, strontium or manganese, suggesting permeation of these ions. The permeability sequence was Ca2+ greater than Ba2+ greater than Sr2+ much greater than Mn2+. No depolarizing shift of the NMDA reversal potential occurred during application of 20 mM-cobalt, magnesium or nickel. 3. In experiments in which the extracellular Na+ concentration was varied the extended constant-field equation was adequate in predicting shifts of the NMDA reversal potential recorded on varying [Na+]o over the range 50-150 mM, but failed to accurately predict the reversal potential of responses to NMDA with 10 mM-[Ca2+]o and only 10 or 20 mM-[Na+]o. These results imply an apparent increase in PCa/PNa on lowering [Na+]o and may result from interaction of permeant ions within the channel. 4. Barium and to a lesser extent calcium, but not strontium (all 20 mM), reduced the slope conductance of responses to NMDA recorded within +/- 15 mV of the reversal potential; over this limited range of membrane potential the current-voltage relationship remained linear in the presence of each of these ions. In contrast manganese produced a strong, voltage-dependent block of responses to NMDA, similar to that produced by magnesium, such that even close to the reversal potential the NMDA current-voltage relationship was highly non-linear. Thus manganese both permeates and blocks the NMDA receptor channel. 5. Raising the extracellular calcium concentration, from 0.1 to 5 mM, had two effects on the conductance mechanism activated by NMDA.(ABSTRACT TRUNCATED AT 400 WORDS)

Abstract: Intracortical infusion of the "N-methyl-D-aspartate" (NMDA) receptor blocker D,L-2-amino-5-phosphonovaleric acid (APV) renders kitten striate cortex resistant to the effects of monocular deprivation. In addition, 1 week of continuous APV treatment (50 nanomoles per hour) produces a striking loss of orientation selectivity in area 17. These data support the hypothesis that crucial variables for the expression of activity-dependent synaptic modifications are a critical level of postsynaptic activation and calcium entry through ion channels linked to NMDA receptors.

Abstract: NMDA (N-methyl-D-aspartate) receptors serve as modulators of synaptic transmission in the mammalian central nervous system (CNS) with both short-term and long-lasting effects. Divalent cations are pivotal in determining this behaviour in that Mg2+ blocks the ion channel in a voltage-dependent manner, and Ca2+ permeates NMDA channels. Zn2+ could also modulate neuronal excitability because it is present at high concentrations in brain, especially the synaptic vesicles of mossy fibers in the hippocampus and is released with neuronal activity. Both proconvulsant and depressant actions of Zn2+ have been reported. We have found that zinc is a potent non-competitive antagonist of NMDA responses on cultured hippocampal neurons. Unlike Mg2+, the effect of Zn2+ is not voltage-sensitive between -40 and +60 mV, suggesting that Zn2+ and Mg2+ act at distinct sites. In addition, we have found that Zn2+ antagonizes responses to the inhibitory transmitter GABA (gamma-aminobutyric acid). Our results provide evidence for an additional metal-binding site on the NMDA receptor channel, and suggest that Zn2+ may regulate both excitatory and inhibitory synaptic transmission in the hippocampus.

Abstract: There is considerable evidence that glutamate is the principal neurotransmitter that mediates fast excitatory synaptic transmission in the vertebrate central nervous system1–3. This single transmitter seems to activate two or three distinct types of receptors, defined by their affinities for three selective structural analogues of glutamate, NMDA (N-methyl-D-aspartate), quisqualate and kainate1–6. All these agonists increase membrane permeability to monovalent cations7–9, but NMDA also activates a conductance that permits significant calcium influx10,11 and is blocked in a voltage-dependent manner by extracellular magnesium12,13. Fast synaptic excitation seems to be mediated mainly by kainate/quisqualate receptors14–18, although NMDA receptors are sometimes activated 19–21. We have investigated the properties of these conductances using single-channel recording22 in primary cultures of hippocampal neurons, because the hippocampus contains all subtypes of glutamate receptors23,24 and because long-term potentiation of synaptic transmission occurs in this structure25,26. We find that four or more distinct single-channel currents are evoked by applying glutamate to each outside-out membrane patch. These conductances vary in their ionic permeability and in the agonist most effective in causing them to open. Clear transitions between all the conductance levels are observed. Our observations are compatible with the model that all the single channel conductances activated by glutamate reflect the operation of one or two complex molecular entities.

Abstract: The optic tecta of surgically produced three-eyed tadpoles were chronically exposed to the N-methyl-D-aspartate (NMDA) receptor antagonist aminophosphonovaleric acid (APV), or to NMDA itself, to assess the influence of NMDA receptor/channels on the eye-specific segregation of retinal ganglion cell (RGC) terminals that occurs whenever two retinas innervate one tectal lobe. Exposure of the tectum to the active isomer of APV produces desegregation of the RGC terminals without blocking electrical activity in the afferents or altering their terminal arbor morphology. Exposure to the inactive isomer of APV causes no perturbation of the normal stripe pattern. APV-induced desegregation is completely reversible within 2 weeks of removal of the APV. In addition, exposure of the optic tectum to NMDA results in stripes with sharper borders and fewer forks and fusions than untreated animals. These results suggest that the NMDA receptor/channel plays a role in eye-specific segregation in the three-eyed tadpole.

Abstract: The influence of endogenous and exogenous acidic amino acids on the binding of [3H]-MK-801, a selective, non-competitive antagonist of N-methyl-D-aspartate (NMDA) receptors, has been investigated in rat cerebral cortex crude synaptic membranes (CSM). Removal of endogenous glutamate and aspartate from CSM by repeated washing reduced the affinity of [3H]-MK-801 for its binding site (with no change in the total number of binding sites) and increased NMDA-sensitive L-[3H]-glutamate binding. In washed CSM, competitive NMDA antagonists of the DL-alpha-amino-omega-phosphonocarboxylate series reduced [3H]-MK-801 binding and NMDA-sensitive L-[3H]-glutamate binding, the most active compounds being 2-amino-5-phosphonovalerate (AP5) and 2-amino-7-phosphono-heptanoate (AP7). Exogenous excitatory amino acid agonists enhanced the binding of [3H]-MK-801 to washed CSM by up to 700%. A selective involvement of NMDA receptors in these effects was indicated by the excellent correlation between EC50s for stimulation of [3H]-MK-801 binding and IC50s for inhibition of NMDA-sensitive L-[3H]-glutamate binding in the same membranes. The selective, competitive NMDA receptor antagonist D-AP5 blocked the L-glutamate-induced increase in [3H]-MK-801 binding in a competitive manner with a pA2 value of 6.0. These results seem to reflect a molecular interaction between two distinct components of the NMDA receptor complex: the transmitter recognition site and the site through which MK-801 exerts its antagonist effects, possibly the ion channel.

Abstract: Acidic amino acids, such as glutamate and aspartate, are thought to be excitatory transmitters in the cerebral neocortex and hippocampus1–8. Receptors for these amino acids can be classified into at least three types on the basis of their agonists. Quisqualate-preferring receptors and kainate-preferring receptors are implicated in the mediation of synaptic transmission in many regions including the hippocampus9,10 and visual cortex11, whereas N-methyl-D-aspartate (NMDA)-preferring receptors are thought to be involved in modulating synaptic efficacy, for example in long-term potentiation, a form of synaptic plasticity in the hippocampus12–14. In the visual cortex of the cat and monkey, it is well established that synaptic plasticity, estimated by susceptibility of binocular responsiveness of cortical neurons to monocular visual deprivation, disappears after the 'critical' period of postnatal development15–17. Here we report that during the critical period in young kittens, a selective NMDA-receptor antagonist blocks visual responses of cortical neurons much more effectively than it does in the adult cat. This suggests that NMDA receptors may be involved in establishing synaptic plasticity in the kitten visual cortex.

Abstract: Intrathecal administration of the excitatory amino acid (EAA) agonists, N-methyl-D-aspartate (NMDA), quisqualate (Quis) or kainic acid (KA), in the spinal subarachnoid space of mice produced a dose-related biting and scratching behavior. Higher doses appeared aversive, suggesting a nociceptive action for EAAs in the spinal cord. Intrathecally administered NMDA, but not Quis or KA, produced a hyperalgesic effect in the tail-flick and hot-plate tests. To test the hypothesis that EAA agonists are involved in transmission of nociceptive information in the spinal cord, we tested the effect of various opioid, sigma and phencyclidine compounds on the action of NMDA in the tail-flick, hot-plate and biting and scratching nociceptive tests. Our results indicated that the involvement of mu, sigma and phencyclidine receptors was predominant in blockade of the behavioral and hyperalgesic effects of intrathecally administered NMDA. Delta receptors appeared less involved, and involvement of kappa receptors was not detectable in blockade of the behavioral and hyperalgesic effects of intrathecally administered NMDA. Quis and KA effects were not altered by any of these agonists. Agonist doses required to inhibit NMDA-induced hyperalgesia in the tail-flick and hot-plate tests were significantly less than those needed to inhibit biting and scratching behavior. The adrenergic agonist norepinephrine inhibited NMDA- but not Quis- or KA-induced biting and scratching behavior. This action appeared to be alpha-1 mediated because it was reversed by phentolamine but not by yohimbine. These results suggest that the actions of EAAs in the spinal cord are differentially affected by various opioid phencyclidine, sigma and adrenergic receptor agonists and support the hypothesis that EAAs are involved in the transmission of nociceptive information in the spinal cord.

Abstract: We have studied the binding of the excitatory amino acid antagonist 3H-labeled MK-801 [(+)-5-methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohept-5,10-imine maleate] to extensively washed rat brain membranes. Binding of 3H-labeled MK-801 was inhibited by phencyclidine, Mg2+, and 3-(2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid in a manner dependent on the presence of L-glutamate or N-methyl-D-aspartate, suggesting that it labeled a site linked to the N-methyl-D-aspartate subtype of the glutamate receptor. Glycine also regulated 3H-labeled MK-801 binding, enhancing it in the concentration range of 0.01-10 microM. The actions of glutamate and glycine involved increases in binding affinity, without altering the number of 3H-labeled MK-801 binding sites. The effects of glycine in this system were mimicked by L- and D-alanine and L- and D-serine. However, beta-alanine and taurine were much less effective, and strychnine did not block the actions of glycine indicating that they were not mediated by the classical glycine receptors. Glycine also enhanced the ability of N-methyl-D-aspartate to increase Ca2+ influx into primary cultures of mouse striatal neurons measured using the Ca2+-sensitive fluorescent dye fura-2. These results support the suggestion that glycine may be an important regulator of the physiological actions of glutamate in vivo.

Abstract: The ability of several glutamate receptor antagonists to reduce hypoxic cortical neuronal injury was quantitatively examined in cell cultures derived from fetal mice. Cultures exposed to hypoxia for 8 hr showed by the following day widespread neuronal injury, which was substantially attenuated by addition of the specific N-methyl-D-aspartate (NMDA) receptor antagonist 2-amino-5-phosphonovalerate (APV). The protective effect of APV was concentration dependent (ED50 about 2 microM) and stereospecific (D-APV approximately 100 times more potent that L-APV). Neuron-protective effects were also observed with several other NMDA antagonists: 2-amino-7-phosphonoheptanoate, phencyclidine and (+)-SKF 10,047 [(+)-N-allylnormetazocine]--as well as with the nonspecific glutamate antagonists D-glutamylglycine and kynurenate. In addition, a similar antagonist profile was observed with a chemical model of hypoxic neuronal injury, produced by brief exposure to high concentrations of cyanide. In contrast, 1 mM concentrations of glutamate diethylester and gamma-aminomethyl sulfonate, compounds reported in some studies to preferentially antagonize non-NMDA glutamate receptors, failed to protect neurons against either hypoxia or cyanide. These results are consistent with the hypothesis that NMDA receptors are preferentially involved in the pathogenesis of hypoxic cortical neuronal injury and suggest that cortical cell culture may be a useful system in which to quantitatively characterize the pharmacology of that injury.

Abstract: Glutamate or a related excitatory amino acid is thought to be the major excitatory neurotransmitter of hippocampal afferents, intrinsic neurons, and efferents. We have used an autoradiographic technique to investigate the status of excitatory amino acid receptors in the hippocampal formation of patients dying with dementia of the Alzheimer type (DAT). We examined L-[3H]glutamate binding to sections from the hippocampal formation of six patients dying of DAT and six patients without DAT and found marked reductions in total [3H]glutamate binding in all regions of hippocampus and adjacent parahippocampal cortex in DAT brains as compared to controls. When subtypes of excitatory amino acid receptors were assayed, it was found that binding to the N-methyl-D-aspartate (NMDA)-sensitive receptor was reduced by 75-87%, with the greatest loss found in stratum moleculare and stratum pyramidale of CA1. Binding to quisqualate (QA)-sensitive receptors was reduced by 45-69%. There were smaller reductions (21-46%) in GABAA receptors in DAT cases. Muscarinic cholinergic receptors assayed in adjacent sections of hippocampal formation were unchanged in DAT. Benzodiazepine receptors were reduced significantly only in parahippocampal cortex by 44%. These results suggest that glutamatergic neurotransmission within the hippocampal formation is likely to be severely impaired in Alzheimer's disease. Such impairment may account for some of the cognitive decline and memory deficits that characterize DAT.

Abstract: 3-(2-Carboxypiperazin-4-yl)propyl-1-phosphonic acid (CPP) was synthesized as a rigid analog of 2-amino-7-phosphonoheptanoate, a previously known antagonist at the N-methyl-D-aspartate (NMDA) preferring, or NMDA-type, of excitatory amino acid receptor. CPP was found to be a potent, selective and competitive antagonist of NMDA-type receptors. CPP antagonized with an IC50 of 8 muM [3H]ACh release which was evoked from rat striatal brain slices by NMDA (50 muM). In contrast, the release of [3H]ACh evoked by elevated KCI was not inhibited by CPP even at a concentration of 100 muM. The antagonism by CPP of NMDA-evoked [3H]ACh release was competitive, with a pA2 of 5.66 for CPP, compared with a pA2 value of 5.22 for 2-amino-7-phosphonoheptanoate. CPP affected neither the uptake of L-[3H]glutamate nor the inhibition by aconitine of L-[3H]glutamate uptake, suggesting a lack of membrane-stabilizing or local anesthetic effects, and also suggesting that CPP itself may not be taken up through the L-glutamate membrane transporter. Moreover, [3H] CPP was not accumulated by synaptosomes (P2 fraction) which avidly accumulate L-[3H]glutamate, supporting the concept that this NMDA-type receptor antagonist acts at an NMDA-type receptor on the external surface of the plasma membrane. CPP (10 muM) failed to interact with any of 21 other putative neurotransmitter receptors including alpha-[3H]amino-3-hydroxy-5-methylisoxazole-4-propionic acid binding (quisqualate-type receptor) and [3H]kainate binding (kainate-type receptor). Audiogenic convulsions in DBA/2 mice were blocked by CPP (ED50 = 1.5 mg/kg i.p.) as were NMDA-induced seizures in CF-1 mice (ED50 = 1.9 mg/kg i.p.). In both strains, CPP impaired the traction reflex at higher doses (ED50 = 6.8 mg/kg and 6.1 mg/kg and 6.1 mg/kg i.p. for DBA/2 and CF-1, respectively). The traction reflex impairment by CPP may be due to muscle relaxant effects of the compound, an explanation supported by the finding that CPP reduced muscle tone as assessed by electromyogram measurement in animals whose muscle tone had been increased by opiate administration. Finally, cerebellar cyclic GMP levels, known to be sensitive to neurotransmission via NMDA-type receptors, were decreased by CPP (ED50 = 4.7 mg/kg i.p.) in mice. In conclusion, based upon the competitive antagonism by CPP of NMDA-evoked [3H] ACh release in vitro and the antagonism of NMDA-induced convulsions in vivo, the data presented are consistent with competitive antagonism of NMDA-type receptors.

Abstract: A newly developed continuous superfusion model was used for studies of 3H-GABA release from cultured mouse cerebral cortex neurons. It was found that a series of excitatory amino acids (EAAs) representing all receptor subtypes evoked Ca2+- dependent release of 3H-GABA from the neurons. Quisqualate was the most potent agonist tested, with an EC50 value of 75 nM. L-Glutamate, N-methyl-D-aspartate (NMDA), and kainate showed EC50 values of 12, 16 and 29 microM, respectively. The EAA-evoked 3H-GABA release could be blocked by a series of EAA antagonists. The highly selective NMDA antagonist D-2-amino-5-phosphonovaleric acid (D-APV) was found to block NMDA responses, whereas the nonselective antagonists cis-2,3-piperidine dicarboxylic acid (PDA) and gamma-D-glutamyl-aminomethyl sulphonic acid (GAMS) blocked responses to all agonists. NMDA responses were found to be sensitive to Mg+ blockade. EAA- as well as potassium-induced 3H-GABA release from the neurons could be detected as early as day 5 in culture. However, during the culture period up to 12 d, the responses to K+, quisqualate, and NMDA were increased. The ontogenetic development of binding sites for quisqualate, kainate, and NMDA in mouse cortex was studied using the radioligands 3H-alpha-amino-3-hydroxy-5-methyl-4-isoxasole propionate (3H-AMPA), 3H-kainate, and 3H-L-glutamate, respectively. The development of binding sites for the different EAA-receptor subtypes showed a good correlation with the development of neuronal 3H-GABA release evoked by the excitatory amino acids.(ABSTRACT TRUNCATED AT 250 WORDS)

Abstract: 3-(2-Carboxypiperazin-4-yl)propyl-1-phosphonic acid (CPP), a rigid analog of 2-amino-7-phosphonoheptanoic acid, has been reported as a selective N-methyl-D-aspartate (NMDA) antagonist. [3H]CPP bound with relatively high affinity (Kd = 201 nM) to Triton-treated rat brain crude synaptic membranes using a centrifugation assay. Binding was saturable, reversible, heat sensitive and dependent on protein concentration. Specific binding, which represented 75 to 85% of the total counts bound, was enriched in synaptosomal and microsomal fractions of rat brain, suggesting an involvement in events related to synaptic transmission. On a regional basis, binding was highest in hippocampus, followed by cortex greater than striatum greater than cerebellum = thalamus. No specific binding could be detected in pons medulla or in liver, kidney, heart, lung and adrenal tissue. [3H]CPP binding was stereoselective for the isomers of glutamate, 2-amino-5-phosphonopentanoic acid, homocysteic acid, alpha-aminoadipic acid and N-methyl-aspartate. The most potent compounds tested were L-glutamate and CPP, which were equiactive in displacing [3H]CPP. The order of activity of other excitatory amino acid receptor ligands was D-2-amino-5-phosphonopentanoic acid greater than L-homocysteic acid greater than or equal to DL-2-amino-7-phosphonoheptanoic acid = D-aspartate = L-aspartate greater than L-serine-O-sulfate = D-alpha-aminoadipic acid = ibotenate greater than NMDA greater than DL-2-amino-6-phosphonohexanoic acid greater than quisqualate greater than N-methyl-L-aspartate. The quisqualate- and kainate-type receptor agonists DL-alpha-amino-3-hydroxy-5-methylisoxazole-4-propionate and kainic acid, respectively, had negligible activity at 100 microM.(ABSTRACT TRUNCATED AT 250 WORDS)

Abstract: There is convincing evidence that acidic amino acids, in particular L-glutamate, or substances containing them serve as the major excitatory neurotransmitters in the brain. At least three distinct receptors mediate the excitatory effects of this class of neurotransmitters. Pharmacological studies with agonists and antagonists of these receptors suggest that they may mediate the neurodegenerative consequences of Huntington's disease, status epilepticus, and hypoxemia, and that glutamate receptor antagonists have clinical potential as anticonvulsants, analgesics, and neuroprotective agents.

Abstract: The purpose of this study was to identify central nervous system pathways and synaptic receptors that participate in baroreflex control of arterial pressure. Microinjections of excitatory amino acids into the caudal ventrolateral medulla (CVM) of anesthetized rats evoked depressor responses analogous to baroreceptor reflexes. Functional inactivation of CVM neurons produced by microinjection of the gamma-aminobutyric acid receptor agonist muscimol completely abolished baroreflex-mediated decreases in arterial pressure elicited by electrical stimulation of the aortic nerve and markedly reduced depressor responses produced by the excitatory amino acid L-glutamate. In contrast, selective blockade of N-methyl-D-aspartic acid (NMDA) receptors in the CVM abolished synaptically mediated depressor responses evoked by aortic nerve stimulation but not those elicited by L-glutamate, kainic acid, or quisqualic acid injected at the same site. These results indicate that the CVM contains an obligatory synapse in the central aortic baroreflex pathway; neural transmission of aortic baroreceptor information in the CVM is mediated by activation of NMDA receptors; and the neurotransmitter released at CVM synapses may be an excitatory amino acid.

Abstract: Neocortical neurons in cell cultures prepared from fetal mice were impaled for intracellular recording. Dextrorphan (DX), a clinically tested dextrorotatory morphinan lacking opioid action, did not alter neuronal membrane potential or conductance, but produced a selective attenuation of N-methyl-D-aspartate (NMDA) responses; kainate and quisqualate responses were not affected. DX also antagonized morphological and chemical (lactate dehydrogenase efflux) evidence of cortical neuronal cell injury produced by toxic bath exposure to NMDA, quinolinate or glutamate, but did not affect toxic exposure to quisqualate or kainate. This selective antagonism of neurotoxicity was apparent at micromolar concentrations of DX with an ED50 of 13 to 17 microM. A similar, but less potent neuron-protective effect, was seen with the opioid levorotatory enantiomer of DX, levorphanol (ED50, 40 microM). The O-methyl derivative of DX, dextromethorphan, also antagonized NMDA and glutamate neurotoxicity, but with possibly lower efficacy than DX. The higher potency of DX over levorphanol suggests that this novel neuron-protective action is not mediated by classic opiate receptors; it may be mediated at the "sigma opiate"/phencyclidine site. If further studies establish that DX and related compounds retain neuron-protective efficacy in appropriate animal models, the established clinical safety record of DX and dextromethorphan may allow prompt investigation of the NMDA receptor-blockade strategy in certain neurological disease states.

Abstract: Primary cultures of cerebellar granule cells have been used in pharmacologically and functionally characterizing excitatory amino acid recognition sites coupled with guanylate cyclase. When granule cells were incubated in physiological culture conditions (Locke's solution, pH 7.4), only kainate and, to a lesser extent, L-glutamate increased cyclic GMP (cGMP) levels. Under these conditions, L-aspartate, N-methyl- D-aspartate (NMDA), and quisqualate were inactive. When granule cells were incubated in the absence of extracellular Mg2+ or in the presence of the depolarizing agent veratrine, L-glutamate, L-aspartate, and NMDA became as effective as kainate in enhancing cGMP formation. The action of kainate was preferentially antagonized by 2,3-cis- piperidindicarboxylate, whereas the action of L-glutamate was preferentially antagonized by (+/-)2-amino-5-phosphonovalerate. These data suggest that 2 different excitatory amino acid recognition sites (activated by kainate or by L-glutamate, L-aspartate, and NMDA, respectively) are coupled with guanylate cyclase in primary cultures of cerebellar granule cells: While the coupling of the recognition site for kainate with guanylate cyclase operates under resting conditions and in the presence of Mg2+, the coupling of the recognition site for L- glutamate, L-aspartate, and NMDA with guanylate cyclase requires depolarizing conditions or the absence of extracellular Mg2+.