3. Properties of Glutamate Receptors

The excitatory amino acid receptors can be grouped into ionotropic receptors (i.e. those where receptor activation is directly coupled to a membrane ion channel) and metabotropic receptors (i.e. those where receptor activation is coupled to an intracellular biochemical cascade: this may eventually lead to opening or closing of membrane ion channels, amongst other effects). The ionotropic receptors were the first to be classified pharmacologically, largely due to the efforts of Watkins and his colleagues (Watkins & Evans, 1981), and the broad scheme of NMDA receptors and non-NMDA (AMPA/kainate) ionotropic receptors, based on responses evoked by the selective agonists NMDA, AMPA and kainate is still in use (Watkins, 1991). Subsequently, metabolic responses to excitatory amino acid agonists were discovered (Foster & Roberts, 1981; Manzoni et al., 1991; Nicoletti et al., 1986) and this ultimately led to the characterisation of the metabotropic glutamate receptors (mGluRs) (Pin & Duvoisin, 1995; Watkins & Collingridge, 1994).

3.1 AMPA/Kainate receptors

These receptors were originally classified by their activation by the agonists quisqualate and kainate, but not NMDA. The use of quisqualate as an agonist for these receptors has now been abandoned in favour of the more selective agonist AMPA, and these receptors are thus referred to either as 'non-NMDA ionotropic receptors' or 'AMPA/kainate receptors'. It is noteworthy that kainate itself can activate AMPA receptors, and AMPA can activate most of the kainate receptors (Watkins et al., 1990). Molecular biological techniques have so far revealed the existence of four glutamate receptor subunits (GluR1-GluR4) Footnote. An alternative nomenclature refers to GluR1-GluR4 as GluRA-GluRD. which can be regarded as AMPA receptor subunits, and five receptor subunits which can be regarded as kainate receptor subunits (GluR5-GluR7 and KA1, KA2). Both of these subunit groups can form homomeric and heteromeric channel assemblies with other members of their groups (Gasic & Hollmann, 1992; Hollmann & Heinemann, 1994). Furthermore, immunoprecipitation experiments demonstrate that GluR6 and/or GluR7 subunits assemble with KA2, but not with GluR2-GluR4 (Puchalski et al., 1994). All of the AMPA and kainate receptors appear to be blocked by the competitive AMPA/Kainate antagonists CNQX and NBQX, although these antagonists do appear to show selectivity to native AMPA receptors and some novel non-competitive antagonists (e.g. GYKI52466) also appear to have selectivity for AMPA receptors compared to kainate receptors (Lerma et al., 1993; Paternain et al., 1995). The dye Evans Blue has been shown to block all combinations of GluR1-GluR4 apart from homomeric GluR3 and GluR6 (Keller et al., 1993). More recently, it has been shown that GluR6 but not GluR2/4 can be antagonised by the novel compound NS-102 (Verdoorn et al., 1994).

Apart from GluR2, the cloned AMPA receptors have a non-linear voltage relationship and are relatively Ca2+ permeable. However, in heteromeric AMPA receptors the linear voltage properties and Ca2+-impermeability of GluR2 are dominant. In most CNS neurones (but see references in Jonas et al., 1994) AMPA/kainate responses show little Ca2+ permeability and this is in accordance with the widespread expression of GluR2 throughout the CNS. The peculiar property of the GluR2 subunit appears to be due to RNA editing at one site (Gasic & Hollmann, 1992; Hollmann & Heinemann, 1994). Each of the AMPA receptor subunits can exist in two forms due to alternative splicing ('flip' and 'flop' forms), the efficacy of L-glutamate being higher at the 'flip' form (Sommer et al., 1990).

GluR5-GluR7 are thought to correspond to the low-affinity kainate receptors, whereas KA1 and KA2 correspond to the so-called high-affinity kainate receptors (Bettler et al., 1992; Lomeli et al., 1992). Homomeric GluR7, KA1 or KA2 receptors do not appear to give agonist responses, but this may (for example) be due to a very rapid desensitisation which might obscure responses. However, heteromeric complexes of KA2 with GluR5 or GluR6 do form functional receptors, and it is noteworthy that KA2/GluR6 shows a substantial response to AMPA. It is thus feasible that, in vivo, there are few receptors amongst the AMPA/Kainate receptors which are insensitive to AMPA, and this may explain the difficulty in segregating AMPA and kainate receptors functionally in most CNS areas (Hollmann & Heinemann 1994). The use of more selective agonists and antagonists will help to resolve some of these matters in the future.

3.2 NMDA receptors

The NMDA-receptor-channel-complex has been extensively studied, and it is known that they have a relatively higher Ca2+ permeability than the non-NMDA ionotropic receptors, they are blocked by Mg2+ in a voltage-dependent manner, they have a requirement for glycine (or similar ligand) as a co-agonist, and they have modulatory sites for polyamines, reducing agents, Zn2+ and protons (McBain & Mayer, 1994). These receptors can be antagonised in a competitive manner by a growing number of substituted five-carbon and seven-carbon chain glutamate analogues such as D-2-amino-phosphonopentanoic acid (AP5) and 3-((+)-2-carboxypiperazin-4-yl)-propyl-1-phosphonate (CPP), and in a non-competitive manner by phencyclidine and the dissociative anaesthetic ketamine (for review see McBain & Mayer 1994).

Molecular biological techniques have revealed that the NMDA-receptor-channel-complex comprises two subunits (NR1 and NR2). There are eight splice variants of NR1, and it is thought that NR1 is a component of all native NMDA receptors, although NR1 subunits can be assembled into homomeric NR1 channels. There are four NR2 subunit types (NR2A-NR2D), which when co-expressed with NR1 are thought to form native NMDA-receptor-channel complexes (Gasic & Hollmann, 1992; Hollmann & Heinemann, 1994; McBain & Mayer,1994) Footnote. In the mouse, the NR1 subunit is referred to as the z1 subunit, and the NR2A-NR2D subunits are referred to as the e1-e4 subunits (e.g., Watanabe et al., 1992,1993).. The different NR2 subunits appear to confer different physiological and pharmacological properties on the receptors: for example, NR1-NR2C channels are more sensitive to Mg2+ blockade and display the highest affinity sites for glycine binding compared to other heteromeric channels, whereas the NR1-NR2A channel differs from the others in its response to reducing agents. The NR1 subunit is ubiquitous throughout the CNS, whereas there is a differential distribution of NR2 subunits: for example NR2C expression levels are high in the cerebellum, but low elsewhere. NR2A and NR2B are found in the thalamus, although NR2A is distributed more prominently in the lateral thalamic nuclei, especially the ventrobasal complex (Buller et al., 1994; Monyer et al., 1994; Watanabe et al., 1993) and NR2D is expressed early during development rather than in the adult (Monyer et al., 1994; Watanabe et al., 1992). It is thought that the NR1-NR2A complex in particular displays a higher affinity for competitive NMDA antagonists than for agonists, and NR1-NR2A has the fastest offset decay time following pulsatile L-glutamate application (Laurie & Seeburg, 1994; Monyer et al., 1994).

3.3 Metabotropic glutamate receptors

At the time of writing, there are known to be at least eight metabotropic glutamate receptors (mGluR1-8), which can be placed into three groups on the basis of sequence homology, agonist pharmacology, and coupling to intracellular transduction mechanisms (Nakanishi, 1992; Pin & Duvoison 1995; Watkins & Collingridge 1994). Group I comprises mGluR1 and mGluR5 (In addition, there are splice variants of mGluR1 and mGluR5), and these receptors appear to be coupled to postsynaptic inositol phosphate metabolism. Group II comprises mGluR2 and mGluR3, and Group III comprises mGluR4, mGluR6, mGluR7 and mGluR8. Groups II and III can couple to an inhibitory cAMP cascade in many expression systems, but may also couple to other transduction mechanisms under physiological conditions. The Group II and III receptors have been suggested to mediate presynaptic actions of glutamate in several brain areas (Baskys & Malenka, 1991; Calabresi et al., 1992; Forsythe & Clements, 1990; Hayashi et al., 1993; Ishida et al., 1993; Jane et al., 1994; Salt & Eaton 1995a), although this does not preclude the possibility that these receptors may also mediate postsynaptic effects in some locations. Similarly, there is evidence that Group I receptors may be found in what appear to be presynaptic locations in some cases (Romano et al., 1995). Within the thalamus of adults, in situ hybridisation studies have shown particularly prominent expression of mGluR1 (Fotuhi et al., 1993; Masu et al., 1991; Shigemoto et al., 1992), mGluR4 (Tanabe et al., 1993) and mGluR7 (Okamoto et al., 1994; Saugstad et al., 1994). mGluR3 mRNA is highly expressed in neurones of the thalamic reticular nucleus (Tanabe et al., 1993). It is noteworthy that metabotropic effects of excitatory amino acid agonists have been described which do not fit into currently known scheme of receptors: it is likely that this is due to further metabotropic receptors which remain to be discovered (Boss et al., 1994; Boss & Boaten 1995).

Use of the agonist (1S,3R)-ACPD, which acts at most of the known metabotropic receptors with a varying degree of potency, has made it very difficult to draw conclusions about the physiological role(s) of the various metabotropic receptors when agonists are applied to complex neural systems which almost certainly contain a variety of receptors at different pre- and postsynaptic loci (Nakanishi, 1992; Watkins & Collingridge, 1994). These difficulties have been exacerbated by the lack of selective competitive antagonists to the metabotropic receptors. Several compounds (e.g. L-AP3, L-AP4, L-aspartic acid-ß-hydroxamate) had been suggested as antagonists on the basis of neurochemical studies, but electrophysiological experiments with these compounds indicate that they cannot be regarded as antagonists and are in some cases full agonists (e.g. L-AP4). Thus, functional studies where these compounds have been used must be considered with caution. The situation is however likely to improve as more selective agonists (e.g. 3,5-dihydroxyphenylglycine, CCG-I, cis-MCG-I) and antagonists (e.g. 4-CPG, MCPG, MAP4, MCCG, MPPG) become available for use in studies of synaptic function (Pin & Duvoisin, 1995; Roberts 1995; Watkins & Collingridge, 1994).

NOTE Please note: this document is part of the HTML version of a paper originally published in print in Progress in Neurobiology.

contents Contents back Previous Section Forward Next Section References Reference List

Return to Tom Salt's Home Page