A kainate binding protein in pigeon cerebellum: purification and localization by monoclonal antibody

Pharmacological characterization, anatomical distribution and sex differences of the non-NMDA excitatory amino acid receptors in the quail brain as identified by CNQX binding

The distribution of non-N-methyl-D-aspartate binding sites was studied in coronal and sagittal sections through the brain of adult Japanese quail by quantitative autoradiography, using tritiated 6-cyano-7-nitroquinoxaline-2,3-dione as a radioligand. Saturation binding experiments were, in addition, carried out in areas showing high levels of binding (cerebellar molecular layer, nucleus anterior medialis and nucleus infundibularis) and demonstrated that the binding of tritiated ligand was specific and saturable. Competition studies with alpha-amino-3-hydroxy-methyl-4-isoxazole propionic acid and kainic acid indicated that kainic acid strongly inhibited ligand binding in all brain areas. alpha-Amino-3-hydroxy-methyl-4-isoxazole propionic acid was only a weak inhibitor in the hypothalamic nuclei whereas in the cerebellar molecular layer both high and low affinity inhibitions were detected. The highest binding levels of tritiated ligand were observed in the molecular layer of the cerebellum. Very high levels of binding were detected in various preoptic/hypothalamic sites including the nucleus suprachiasmaticus pars medialis, nucleus anterior medialis hypothalami, nucleus infundibularis, nucleus mammillaris medialis, nucleus posteromediale hypothalami and nucleus hypothalami ventromedialis. High levels of binding were also detected in the bulbus olfactorius, bed nucleus commissuralis anterior, bed nucleus commissuralis pallii, nucleus accumbens, bed nucleus striae terminalis and nucleus interpeduncularis. In the preoptic area/hypothalamus, high levels of binding were clearly present in all areas that contain gonadotropin releasing hormone cells or fibers. In the pons and mesencephalon, moderate levels of binding were associated with catecholaminergic areas such as the area ventralis tegmentalis (area ventralis of Tsai) and the locus coeruleus. Saturation analysis demonstrated the presence of a higher number of binding sites in females than in males in the cerebellar molecular layer, nucleus infundibularis and nucleus anterior medialis. This latter difference was confirmed in the one point assays that also identified higher levels of specific binding in the nucleus suprachiasmaticus pars medialis of males as compared with females. These anatomical data suggest a possible implication of non-N-methyl-D-aspartate receptors in the synthesis and/or release of both gonadotropin releasing hormone and catecholaminergic neurotransmitters that should now be tested by pharmacological experiments.

Molecular biology of glutamate receptors in the central nervous system and their role in excitotoxicity, oxidative stress and aging

Forty years of research into the function of L-glutamic acid as a neurotransmitter in the vertebrate central nervous system (CNS) have uncovered a tremendous complexity in the actions of this excitatory neurotransmitter and an equally great complexity in the molecular structures of the receptors activated by L-glutamate. L-Glutamate is the most widespread excitatory transmitter system in the vertebrate CNS and in addition to its actions as a synaptic transmitter it produces long-lasting changes in neuronal excitability, synaptic structure and function, neuronal migration during development, and neuronal viability. These effects are produced through the activation of two general classes of receptors, those that form ion channels or "ionotropic" and those that are linked to G-proteins or "metabotropic". The pharmacological and physiological characterization of these various forms over the past two decades has led to the definition of three forms of ionotropic receptors, the kainate (KA), AMPA, and NMDA receptors, and three groups of metabotropic receptors. Twenty-seven genes are now identified for specific subunits of these receptors and another five proteins are likely to function as receptor subunits or receptor associated proteins. The regulation of expression of these protein subunits, their localization in neuronal and glial membranes, and their role in determining the physiological properties of glutamate receptors is a fertile field of current investigations into the cell and molecular biology of these receptors. Both ionotropic and metabotropic receptors are linked to multiple intracellular messengers, such as Ca2+, cyclic AMP, reactive oxygen species, and initiate multiple signaling cascades that determine neuronal growth, differentiation and survival. These cascades of complex molecular events are presented in this review.

Kainate-binding proteins: phylogeny, structures and possible functions

Recent advances have demonstrated that the family of [3H]kainate-binding proteins and kainate receptors comprise a number of related polypeptides. In all the cases so far investigated, the kainate-binding proteins from non-mammalian vertebrates have M(r) values in the range of 40-50 kDa whereas mammalian kainate receptors and kainate-binding proteins have M(r) values in the order of 100 kDa. There have not, as yet, been any reports of 40-50 kDa kainate-binding proteins in mammalian CNS and, despite the cloning of increasing numbers of cDNAs encoding new kainate-binding proteins, the relationships between these two general groups of polypeptides remain unclear. Nonetheless, there is now a wealth of phylogenetic, structural and molecular biological data available about these proteins. In this review, Jeremy Henley outlines the properties and structures of kainate-binding proteins and offers some possibilities as to the roles of these often hugely abundant proteins.

Quantitative analysis of the distributions of glutamatergic ligand binding sites in goldfish brain

Goldfish brain is a widely used model system for the study of the mechanisms involved in neuronal regeneration and synaptic plasticity. Because of the proposed role of glutamate receptors in these processes we have investigated the anatomical localisations of [3H]AMPA (alpha-amino-3-hydroxy-5-methylisoxazole-4-propionate), [3H]kainate, [3H]CNQX (6-cyano-7-nitroquinoxaline-2,3-dione) and [3H]L-glutamate binding sites in horizontal and sagittal sections. Binding sites for [3H]L-glutamate were the most widespread and both NMDA (N-methyl-D-aspartate) and non-NMDA sensitive components were detected. The density of [3H]kainate binding was very high in the cerebellum compared to other regions and in comparison with the other radioligands used. Conversely, relatively low amounts of [3H]AMPA binding were present with the telencephalon being the most densely labelled structure. [3H]CNQX binding was most densely localised in the tectum with the cerebellum also possessing high binding. In addition, there was a small population of [3H]CNQX binding sites located in the telencephalon and lobus vagi that appeared insensitive to AMPA and kainate.

Excitatory amino acid receptors: a gallery of new targets for pharmacological intervention

The excitatory amino acids (EAAs) L-glutamate and L-aspartate are the most abundant amino acids in brain and play a number of roles in maintaining neuronal function. Among these are their use as protein constituents, as key intermediates in ammonia metabolism, and as precursors for other neurotransmitters. Given the widespread distribution of EAA-containing neurons, these transmitters are likely to be involved in virtually all central nervous system functions, with abnormalities in neurotransmission contributing to the symptoms of a host of neurological and psychiatric disorders. Because of the importance of EAAs in maintaining the functional integrity of the central nervous system, efforts are underway to design agents capable of regulating the activity of these transmitters for therapeutic gain. Inasmuch as potential side effects preclude a generalized modification of this system, strategies must be found to alter EAA neurotransmission in selected brain regions. In this regard, pharmacological data suggest several functionally distinct EAA receptors, a finding confirmed by cloning studies which hint at an even larger family of sites. Moreover, it appears that some excitatory amino acid receptor complexes are composed of interacting sites which orchestrate receptor function, and there is evidence that EAA receptors may influence the activity of one another. Thus, there appear to be numerous sites that can be targeted to selectively modify excitatory amino acid neurotransmission in brain. Besides the agonist recognition site for each receptor subtype, other targets include regulatory subunits, ion channels and components of receptor-coupled second messenger systems.

Molecular cloning of the kainate-binding protein and calmodulin genes which are induced by an imprinting stimulus in ducklings

For the formation of imprinting in birds, protein synthesis is known to be essential in the medial hyperstriatum ventrale (MHV) of the forebrain after presentation of an imprinting stimulus. We have searched for the genes whose expressions are increased in duckling's MHV during formation of imprinting, and identified kainate-binding protein and calmodulin genes. This may reflect the formation of glutamatergic pathways in MHV.

Localization of AMPA receptors in the hippocampus and cerebellum of the rat using an anti-receptor monoclonal antibody

The primary amino acid sequences of the kainate binding proteins from the amphibian and avian central nervous systems are homologous with the functional alpha-amino-3-hydroxyl-5-methyl-isoxazole-4-propionate receptors that have been cloned from rat brain. In this study, we have analysed the anatomical and subcellular distribution of the alpha-amino-3-hydroxyl-5-methyl-isoxazole-4-propionate receptors in the rat hippocampus and cerebellum, using a monoclonal antibody that was raised against a kainate binding protein purified from frog brain. Immunoblots of rat hippocampus and cerebellum, and membranes from COS cells transfected with rat brain alpha-amino-3-hydroxyl-5-methyl-isoxazole-4-propionate receptor cDNAs (GluR1, GluR2, or GluR3) showed a major immunoreactive band migrating at a relative molecular weight of 107,000. In the cerebellum, an additional immunoreactive protein of approximately 128,000 mol. wt was also seen on immunoblots probed with the antibody. The distribution of this protein is apparently restricted to the cerebellum since the 128,000 mol. wt band was not present in other brain areas examined. The identity of the 128,000 mol. wt cerebellar protein is not known. Immunocytochemical analyses of the hippocampus demonstrated that alpha-amino-3-hydroxy-5-methyl-isoxazole-4-propionate receptor subunits are present in the cell bodies and dendrites of pyramidal cells. The granule cells were also immunostained. All of the pyramidal cell subfields were heavily labeled. In the pyramidal cell bodies, a high level of immunoreactivity was observed throughout the cytoplasm. In the cerebellum, the Purkinje cell bodies and dendrites also displayed very high levels of immunoreactivity. In addition to the Purkinje neurons, the Bergmann glia and some Golgi neurons were clearly immunostained. Subcellular fractionation and lesioning experiments using the excitotoxin domoic acid indicated that the alpha-amino-3-hydroxyl-5-methyl-isoxazole-4-propionate receptor subunits were associated with postsynaptic membranes. Direct visualization of the immunoreactivity using electron microscopy confirmed the postsynaptic localization of the staining in the dendritic areas in both the hippocampus and the cerebellum. Thus, unlike the kainate binding proteins, which are found primarily extrasynaptically in the frog and on glial cells in the chicken cerebellum, the GluR1, GluR2, and GluR3 receptor subunits are localized to the postsynaptic membrane in the dendrites of neurons in the rat central nervous system.

Evidence for hybrid NMDA/kainate receptors from protein reconstitution studies and expression of vertebrate CNS RNAs in xenopus oocytes

1. A review is presented of recent advances in glutamate receptor research with particular emphasis on studies which show that some glutamate receptors in the central nervous systems (CNS) of Xenopus and rat contain a mixture of N-methyl-D-aspartate-sensitive and kainate-sensitive subunits. 2. Protein isolated from Xenopus CNS using a domoic acid affinity column exhibits complex pharmacological properties. It binds both [3H]kainate and [3H]glycine: the binding of the latter is strychnine-insensitive. 3. When reconstituted into lipid bilayers, channels gated by kainate and NMDA can be elicited and the properties of these channels are similar to those gated by kainate receptors and NMDA receptors, respectively, in studies of vertebrate central neurones in situ. 4. The protein can be fractionated into two components; one of which is sensitive only to kainate and AMPA, the other exhibiting sensitivity to both kainate and NMDA. 5. When RNA isolated from Xenopus and rat CNS is injected into Xenopus oocytes, responses to kainate and NMDA can be seen within 2-3 days. The responses to co-application of these agonists support the contention that some of the glutamate receptors expressed in oocytes contain both kainate-sensitive and NMDA-sensitive subunits.

Purified unitary kainate/alpha-amino-3-hydroxy-5-methylisooxazole-propionate (AMPA) and kainate/AMPA/N-methyl-D-aspartate receptors with interchangeable subunits

We have purified and characterized two vertebrate excitatory amino acid ionotropic receptors from the Xenopus central nervous system. Each is a unitary receptor (i.e., having more than one class of excitatory amino acid agonist specificity within one protein oligomer). The first is a unitary non-N-methyl-D-aspartate (non-NMDA) receptor and the second is a unitary NMDA/non-NMDA receptor. The specific agonist-activated channel activity and pharmacology of each type were recognized by patch-clamping lipid bilayers in which the isolated protein was reconstituted. In the second case, the NMDA and the non-NMDA sites could not be physically separated and exhibited functional interaction. Parallel evidence for this was obtained when poly(A) RNA from Xenopus brain was translated in oocytes: a noncompetitive inhibition of the response to L-kainate is produced by NMDA to a maximum depression of 30% at 1 mM NMDA. Each isolated oligomer contains 42-kDa subunits of the non-NMDA ligand binding type, but the second type has an additional NMDA-receptor-specific 100-kDa subunit. Thus, a subunit-exchange hypothesis can account for the known multiplicity of excitatory amino acid receptor types.

Solubilization and purification of an alpha-amino-3-hydroxy-5-methylisoxazole-4-propionic acid binding protein from bovine brain

alpha-Amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA) is a selective ligand for an excitatory amino acid receptor subtype in mammalian brain. We have solubilized an AMPA binding protein from bovine brain membranes with 1% Triton X-100 in 0.5 M phosphate buffer and 20% glycerol at 37 degrees C and purified the stable binding sites using a series of chromatographic steps. Scatchard analysis of the purified preparation showed a curvilinear plot with dissociation constants of 10.6 and 323 nM and Bmax values of 670 and 1,073 pmol/mg of protein for the high- and low-affinity sites, respectively. Inhibition constants for several excitatory amino acid analogues were similar to those obtained for other membrane and solubilized preparations. Gel filtration of the soluble AMPA binding protein showed a single peak of [3H]AMPA binding activity at Mr approximately 500,000. With sodium dodecyl sulfate-polyacrylamide gel electrophoresis, the purified AMPA binding protein showed a single major band at Mr = 110,000. Previously, we have shown that a monoclonal antibody (KAR-B1) against a frog brain kainate binding protein selectively recognizes an unknown protein in mammalian brain migrating at Mr approximately 100,000. We now show that this antibody recognizes the major component of the purified AMPA binding protein, supporting a structural similarity between the frog brain kainate binding protein and the mammalian AMPA binding protein.

Glutamic acid-insensitive [3H]kainic acid binding in goldfish brain

Kainic acid is supposed to be a specific agonist for a subclass of excitatory glutamate receptors in the vertebrate CNS. An investigation of (2 nM) [3H]kainic acid binding sites in goldfish brain, using quantitative autoradiography, has revealed evidence for two types of kainic acid receptors which differ in sensitivity to glutamic acid. L-Glutamic acid (0.1-1 mM) displaced over 95% of specific [3H]kainic acid binding elsewhere in the brain but only 10-50% in the cerebellum and cerebellar crest. These structures apparently contain [3H]kainic acid binding sites that are extremely insensitive to glutamic acid. The glutamic acid-insensitive [3H]kainic acid binding was not displaced by quisqualic acid, kynurenic acid, alpha-amino-3-hydroxy-5-methylisoxazolepropionic acid (AMPA), or N-methyl-D-aspartatic acid, but was completely displaced by the kainic acid analogue domoic acid. The data indicate that two types of high affinity binding sites for [3H]kainic acid exist in the goldfish brain: glutamic acid-sensitive and glutamic acid-insensitive. High affinity [3H]kainic acid binding may therefore not always represent binding to subsets of glutamic acid receptors.

Excitatory and inhibitory amino acid neurotransmitter binding sites in the cerebellar cortex of the pigeon (Columba livia)

We used receptor autoradiography to determine the distribution of excitatory and inhibitory amino acid neurotransmitter binding sites in the cerebellar cortex of the pigeon (Columba livia). alpha-Amino-3-hydroxy-5-methylisoxazole-4-propionic acid, kainate and metabotropic binding sites had highest levels in the molecular layer. N-methyl-D-aspartate binding sites, assayed with both [3H]glutamate under selective conditions and with [3H]glycine binding to the associated strychnine-insensitive glycine site, had highest levels in the granule cell layer. There was little specific binding of the non-competitive N-methyl-D-aspartate antagonist, [3H]MK-801. The level of gamma-aminobutyric acid (GABA)-A binding sites was higher than GABA-B binding sites in both molecular and granule cell layers with the highest level of GABA-A sites in the granule cell layer. The highest level of GABA-B binding sites was in the molecular layer. [3H]Flunitrazepam binding levels were approximately the same in both molecular and granule cell layers. With the exception of kainate binding sites, the distribution of binding sites was identical to that seen in the cerebellar cortex of mammals. Our results support the concept that the chemoarchitecture of the cerebellar cortex has been conserved in the course of vertebrate evolution.

Comparison of solubilised kainate and alpha-amino-3-hydroxy-5- methylisoxazolepropionate binding sites in chick cerebellum

[3H]Kainate bound to chick cerebellar membranes with a KD of 0.6 microM and with an exceptionally high Bmax of 165 pmol/mg of protein. In octylglucoside-solubilised extracts, the affinity of [3H]kainate was reduced (KD = 2.7 microM), but the Bmax was relatively unchanged (130 pmol/mg of protein). The rank potency of competitive ligands was domoate greater than kainate greater than 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) greater than glutamate. Binding sites for alpha-[3H]amino-3- hydroxy-5-methylisoxazolepropionate ([3H]AMPA) were much less abundant, with KD and Bmax values in membranes of 86 nM and 1 pmol/mg of protein, respectively. The affinity of [3H]AMPA binding was also reduced on solubilisation (KD = 465 nM), but there was an increase in the Bmax (1.7 pmol/mg of protein). Quisqualate and CNQX were the most effective displacers of [3H]AMPA binding, but kainate was also a relatively potent inhibitor. However, in contrast to the displacement profile for [3H]kainate, domoate was markedly less potent than kainate at displacing [3H]AMPA. These results suggest that [3H]AMPA binds to a small subset of the kainate sites that, unlike the majority of the [3H]kainate binding protein, which has been reported to be located in the Bergmann glia, may represent neuronal unitary non-N-methyl-D-aspartate receptors.

Properties of kainate receptor/channels on cultured Bergmann glia

Following the localization, at the electron microscope level, of the immunoreactivity towards a putative kainate receptor on Bergmann glial cells in the chick cerebellar cortex, cultures of Bergmann glia were used to establish the presence of functional kainate receptor/channels and study their properties. Bergmann glia were identified by their fusiform morphology and their ability to bind an anti-kainate binding protein monoclonal antibody, a kainate receptor high affinity ligand--kainyl-bovine serum albumin--and a glial marker--anti-vimentin monoclonal antibody. Membranes prepared from the culture cells displayed, using 25 nM [3H]kainate, the binding of 4.1 pmol of [3H]kainate/mg protein and showed the presence in Western blots of the two polypeptides of 49 and 93 kDa attributed to the kainate binding protein. Kainate, at concentrations above 0.1 mM, was found to increase the influx into cultured Bergmann glia of 22Na+, 86Rb+, 45Ca2+ and 36Cl- ions. The traffic of 22Na+, induced by kainate and glutamate, observed only in the presence of 1 mM ouabain, was blocked by kainate receptor antagonists and by 0.01 mM quisqualate. Analysis of the kinetics of incorporation of 22Na+ and 45Ca2+ ions showed an initial accumulation of 22Na+ and 45Ca2+ ions followed by their total dissipation. The results indicate that the kainate-induced influx of Na+ ions through the kainate receptor/channel causes the reverse transport of Na+ ions, by activation of the Na+/Ca2+ and Na+/H+ exchangers which remove intracellular Na+ ions. Pre-exposure of the cells to 0.5 mM dibutyryl cAMP was found to greatly enhance the kainate-induced 22Na+ ion influx. We propose that the Bergmann glia kainate receptors modulate the efficacy of the glutamatergic synapses between the parallel fibers and Purkinje cell spines and form part of a glial machinery responsible for plastic changes in synaptic transmission.

[3H]kainate localization in goldfish brain: receptor autoradiography and membrane binding

The anatomical distribution of specific [3H]kainate binding in goldfish brain was investigated by membrane binding and autoradiographical techniques. Saturation binding of the radioligand was determined in 8 anatomically defined regions and demonstrated a single class of high affinity sites with Kd values ranging from 290 to 650 nM. Kainate receptor densities, however, varied significantly. The cerebellum contained the highest concentration of binding sites (964 pmol/mg prot.), while the optic tectum had the lowest (96 pmol/mg prot.). Binding site distributions determined by autoradiographic studies demonstrated the same regional variation and allowed more specific localization of the binding sites. Within the cerebellum, the molecular layers of the corpus, valvula and lobus caudalis displayed a uniform and highly intense image while the granule cell layers (except for the medial granule cell mass of the lobus caudalis) did not. Other areas of intense binding were the posterior tubercle of the diencephalon, inferior lobes of the hypothalamus and layers 1 and 2 of the optic tectum (deep to the periventricular granule cells).

Autoradiographic distribution of binding sites for the non-NMDA receptor antagonist CNQX in chick brain

The anatomical distribution of binding sites for the non-N-methyl-D-aspartate (non-NMDA) glutamate receptor antagonist [3H]6-cyano-7-nitroquinoxaline-2,3-dione ([3H]CNQX) in 1-day-old chick brain was investigated. Specific [3H]CNQX binding sites were widely distributed but were particularly densely localised in the molecular layer of the cerebellum. In the midbrain, the binding was comparatively low, except for relatively high levels in the nucleus isthmi of the optic lobe. The distribution of [3H]CNQX binding was markedly different to that of the NMDA receptor ligand [3H]MK-801. Overall, the localisation of [3H]CNQX binding was similar to the combined distributions of [3H]AMPA and [3H]kainate binding sites. Kainate inhibited [3H]CNQX binding throughout the brain. AMPA also inhibited [3H]CNQX binding in the fore- and midbrain, but the dense binding in the molecular layer of the cerebellum was notable in being AMPA-insensitive.

Subcellular localization of a putative kainate receptor in Bergmann glial cells using a monoclonal antibody in the chick and fish cerebellar cortex

A monoclonal antibody, IX-50, that was raised against a kainate binding protein (Mr = 49,000) from chicken cerebellum, was used in light and electron microscopic immunocytochemical studies to localize putative kainate receptors. Pre- and postembedding immunoperoxidase and immunogold methods were used in the cerebellar cortices of one to 26-day old chickens and adult rainbow trout. Immunoreactivity was detected only in association with Golgi epithelial/Bergmann glial cells. Intracellular immunoreactivity was present in the granular and agranular endoplasmic reticulum, Golgi apparatus and in lysosomes, representing the sites of synthesis, glycosylation and degradation of the protein. In the fish the granular endoplasmic reticulum was not immunoreactive. Extracellular immunoreactivity was associated with the plasma membrane. In the fish it was established that the epitope is on the outer surface of the membrane. The protein seems to be uniformly distributed along the membrane including the somata, the radial stem processes and the leafy lamellae surrounding Purkinje cell dendrites. Areas of the glial membrane in contact with other glial cells were also immunopositive. High-resolution light microscopy demonstrated all the Bergmann glial plasma membrane in the cortex, providing a "negative" image of Purkinje cell dendrites. It is apparent that Bergmann glial processes selectively outline the dendrites of the Purkinje cells by surrounding the parallel fibre terminal/Purkinje cell spine synaptic complexes. The parallel fiber terminals were highly immunoreactive for glutamate, as shown by an immunogold procedure. The association of Bergmann glial processes, carrying the Mr = 49,000 kainate binding protein, with the Purkinje cell dendrites and spine synapses could provide a basis for neuronal signalling to the Bergmann glia, possibly by glutamate.

Solubilization and partial purification of alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid binding sites from rat brain

alpha-Amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) binding sites were solubilized from rat brain membranes using 1% Triton X-100 in 0.5 M potassium phosphate buffer containing 20% glycerol. The solubilized binding sites were stable, permitting biochemical and pharmacological characterization as well as partial purification. Pharmacological and binding analyses indicated that the solubilized binding sites were similar to the membrane-bound sites. Both the solubilized and the membrane-bound preparations contained high- and low-affinity AMPA binding sites in the presence of potassium thiocyanate. A similar rank order for inhibition of [3H]AMPA binding by several excitatory amino acid analogs was obtained for the soluble and membrane-bound preparations. [3H]AMPA binding to both soluble and membrane-bound preparations was increased in the presence of potassium thiocyanate. The solubilized AMPA binding sites migrated as a single peak with gel filtration chromatography, with an Mr of 425,000. Beginning with the solubilized preparation, AMPA binding sites were purified 54-fold with ion-exchange chromatography and gel filtration. The characterization and purification of these soluble binding sites is potentially useful for the molecular characterization of this putative excitatory amino acid receptor subtype.

Molecular structure of the chick cerebellar kainate-binding subunit of a putative glutamate receptor

Kainate receptors mediate some of the excitatory transactions carried out in the central nervous system by the neurotransmitter glutamate. They are involved in neurotoxicity, possibly in neurodegenerative disorders and it has been suggested that they have a role in long-term potentiation. Kainate receptors are present both on neuronal and glial cell membranes where they regulate the gating of a voltage-independent ion channel. Nothing is known about their molecular structure. Taking advantage of the unusually high abundance of 3H-kainate binding sites in the chick cerebellum, we have isolated an oligomeric protein that displays a pharmacological profile similar to that of a kainate receptor, and have demonstrated, using the monoclonal antibody IX-50, that this protein is composed of a single polypeptide of Mr 49,000 which harbours the specific kainate recognition site. The structure of this kainate binding protein (KBP) is also of interest because of its exclusive cerebellar localization on Bergmann glial membrane in close proximity to established glutamatergic synapses. We now report the isolation of the complementary DNA containing the complete coding region of the kainate binding protein. The predicted structure of the mature protein has four putative transmembrane domains with a topology analogous to that found in the superfamily of ligand-gated ion channels. This raises the possibility, that kainate binding protein may form part of an ion channel and may be a subunit of a kainate subtype of glutamate receptor.

Cloning by functional expression of a member of the glutamate receptor family

We have isolated a complementary DNA clone by screening a rat brain cDNA library for expression of kainate-gated ion channels in Xenopus oocytes. The cDNA encodes a single protein of relative molecular mass (Mr) 99,800 which on expression in oocytes forms a functional ion channel possessing the electrophysiological and pharmacological properties of the kainate subtype of the glutamate receptor family in the mammalian central nervous system.