Glutamate receptors contribute to excitatory synaptic transmission at sites throughout the brain and the spinal cord. In the late 1970s, the ligand-gated ionotropic glutamate receptors in the vertebrate central nervous system (CNS) were classified，on the basis of their pharmacology，into three types: NMDA, AMPA and kainate receptors. Of the three types, the NMDA receptors have received the most attention in the past several years, since they appear to play a key role in a wide range of physiologic and pathologic processes, such as synaptic plasticity, acquisition of memory, ischaemia, epilepsy and certain neurodegenerative disorders.
The NMDA receptor channels differ in fundamental ways from the non-NMDA receptor channels； these properties directly relate to their physiologic roles. The NMDA receptor complex constitutes an ion channel which is permeable to Na + , K + and Ca 2+ and sensitive to voltage-dependent Mg 2+ block. The NMDA receptors are activated by the endogenous neurotransmitter， glutamate, and activation of the NMDA receptor is modulated by a number of endogenous compounds such as glycine, Mg 2+ , polyamines, Zn 2+ , redox reagents, H + and protein phosphorylation. The contribution of these compounds to the function of NMDA receptors in situ remains to be established. From a pharmacological standpoint, these sites represent attractive targets for agents that might be used therapeutically to positively or negatively modulate the NMDA receptor complex. Molecular studies demonstrate the existence of at least five NMDA receptor subunits, NR1 and NR2A-2D. With the identification of these NMDA receptor subunits, it has become possible to investigate the molecular mechanisms responsible for the functional regulation of NMDA receptors.
MOLECULAR BIOLOGY OF NMDA RECEPTORS
Five NMDA receptor subunits from both rat and mouse brain have now been characterized, and several of their human homologues have been identified. It is believed that in native NMDA receptors, these subunits are combined in various combinations to generate heteromeric assemblies that are predicted to contain five subunits. The primary structures of the five NMDA receptor channel subunits were deduced from cDNA sequence. NR1/ζ1 (920 residues) subunits arise from one gene, with eight functional splice variants, whereas the NR2 subunits are coded by four different genes. The predicted amino acid sequences for the mature peptides in the rat subunits， NR2A (1442 residues), NR2B (1456 residues), NR2C (1218 residues), and NR2D (1333 residues) are identical to those in the mouse homologues, ε1 (1445 amino acids), ε2 (1456 amino acids), ε3 (1220 amino acids) and ε4 (1296 amino acids).
Hydropathy plots of the amino acid sequences of the NMDA receptor subunits reveals 4 stretches of about 20 amino acids with high hydrophobicity. These hydrophobic sequences were initially thought to generate the transmembrane regions TM1-TM4, with a large loop between TM3 and TM4. These structural characteristics seem to be common to neurotransmitter-gated ion channels, such as the nicotinic acetylcholine receptor (nAChR), theγ-aminobutyric] acid receptor (GABAR), and the glycine receptor (GlyR). Recently, a three transmembrane segment model was proposed in which the putative channel-lining segment， TM2， loops into the membrane from the intracellular side without traversing it. Phosphopeptide map analysis suggests that the carboxyl-terminal region of the NR1 and NR2 subunits resides on the cytoplasmic side， and mutational analysis of the glycine-binding and redox modulation sites of the NR1 subunit suggests possible extracellular localization of the region between segments TM3 and TM4. These observations support the three transmembrane segment model.
Overall amino acid sequence identity among the NMDA receptor channel subunits is as high as 40%-50% within the NR2 subfamily, but as low as 18% between the NR2 and NR1 subfamilies. Amino acid sequence identity of the NMDA receptor channel subunits with other GluR channel subunits is 12%-18%. The carboxyl-terminal region is highly diverse between the NR1 and NR2 subfamilies. The NR2 subunits have notably larger carboxyl-terminal regions than the NR1 and other GluR channel subunits. The carboxyl-terminal regions of the NR2A and NR2B subunits are very long and similar in size.
Channel pore region
All subunits of the NMDA receptor channel possess an asparagine in segment TM2. Replacement by glutamine of the asparagine in segment TM2 of the NR1 and NR2B subunits strongly reduced both Ca 2+ permeability and sensitivity to Mg 2+ block of heteromeric NMDA receptors. Since there is strong evidence that Mg 2+ produces voltage-dependent channel blocking by binding to a site deep within the channel pore, these results are consistent with the view that segment TM2 constitutes the ion channel pore of the NMDA receptor. As described above, segment TM2 may loop into the plasma membrane from the intracellular side without traversing it, as proposed for the channel-forming pore of voltage-dependent ion channels.
Phosphopeptide map analysis suggests that the carboxyl-terminal regions of NR1 subunits are phosphorylated in vivo. Furthermore, the carboxyl-terminal regions of NR1 and ε2 subunits are phosphorylated by protein kinase C (PKC). These findings support the view that the carboxyl-terminal region of the NMDA receptor subunits is located on the cytoplasmic side. In studies of heteromeric NMDA receptor subunit combinations, PKC activators potentiate the responses of ζ1/ε1 and ζ1/ε2, but not ζ1/ε3 and ζ1/ε4, suggesting a selective regulation of phosphorylation of only certain subtypes of heteromeric NMDA receptors, controlled probably by the COOH-terminal region of the ε-subunit.
Allosteric modulation region
Full activation of the NMDA receptor channel requires the binding of both glutamate and glycine. Site-directed mutagenesis of the NR1 subunit identified three regions which can alter the required concentration for half-maximum response (EC) for glycine but having little effect on glutamate efficacy. The three regions are phenylalanine at residue 448 in the extracellular amino terminal domain, a glycine binding motif (phenylalanine-X-tyrosine) located around residue 370, and a region between segments TM3 and TM4.
NMDA receptor channels are also modulated by compounds, such as sulfhydryl (redox) reagents with strong oxidizing or reducing potential, ethanol, spermine, nitric oxide and protons. Mutation of two cysteine residues (cysteine 744 and 798), located in the region between segments TM3 and TM4 of the NR1 subunit, eliminated potentiation by dithiothreitol (DTT, a redox reagent) treatment, potentiation by spermine as well as inhibition by protons. It also shifted the EC values for NMDA of the NR1/NR2B channel. Splice variants of the NR1 subunit, with the insertion encoded by exon 5 near the amino terminus, are less sensitive to proton inhibition and relieve proton inhibition in the NR1/NR2A, NR1/NR2B and NR1/NR2D channels, but not in the NR1/NR2C channel.
Distribution in the developing brain
Distribution of the NMDA receptor subunit mRNA in the rodent brain was examined by in situ hybridization analysis. Expression of the NR1/ξ1 subunit mRNA occurs ubiquitously in the brain at all developmental stages. In contrast, expression of the respective NR2 (ε) subunit mRNA is differentially regulated during development. In embryonic brain, only the NR2B (ε2) and NR2D (ε4) subunit mRNA is expressed.In contrast to the wide distribution of the NR2B (ε2) subunit mRNA, the NR2D (ε4) subunit mRNA is found exclusively in the diencephalon and brainstem. During the first two weeks after birth, expression patterns of the NR2 (ε) subunit mRNA change drastically. The NR2A (ε1) subunit appears in the entire brain while the NR2C (ε3) subunit mRNA appears only in the cerebellum. In contrast, the expression of the NR2B (ε2) subunit mRNA becomes restricted in the forebrain and that of the NR2D (ε4) subunit is almost completely abolished.
Distribution in the adult brain
The NR1 subunit mRNA is distributed ubiquitously in the brain; by contrast, the NR2 subunit mRNA shows restricted distributions in the brain. The NR2A subunit distributes widely, but the level of expression is higher in the cerebral cortex, the hippocampus and cerebellar granule cells. The NR2B subunit is expressed selectively in the forebrain. And high levels of expression are observed in the cerebral cortex, the hippocampal formation, the septum, the caudate-putaman, the olfactory bulb and the thalamus. The NR2C subunit is found predominantly in the cerebellum. Strong expression is observed in the granule cell layer of the cerebellum, while weak expression is detected in the olfactory bulb and the thalamus. Low levels of the NR2D subunit are found in the thalamus, the brainstem and the olfactory bulb.
Distribution in the spinal cord
The NR2B, NR2D and NR1 subunit mRNA is expressed widely in the spinal cord at embryonic stages. Restricted expression of the ε1 subunit mRNA has been shown in the developing ventral horn. However, in adult rat spinal cord the expression patterns are controversial. The NR2D and NR1 subunit mRNA is the predominant transcripts detected in the spinal cord, whereas NR2C subunit mRNA is found in the rat lumbar spinal cord and NR2A and NR2B are found in the cervical cord.
As the case for subunits from several other families of ligand-gated ion channels, NR1, expressed as a homomer in oocytes, forms a receptor-channel complex with functional properties characteristic of native NMDA receptors. By contrast, the NR2 family of NMDA receptor subunits is functionally inactive when expressed alone, but in combination with NR1, generates responses 10- to 100-fold larger than for homomeric (NR1) receptors, indicating that NR2 subunits greatly increase the efficiency of receptor assembly or function. Each type of heteromeric assembly displays an individual profile of functional properties such as channel conductance and deactivation kinetics, agonist, antagonist and co-agonist sensitivities, and sensitivity to channel blockade by Mg 2+ and MK-801.
Radioligand and electrophysiological studies indicate that NMDA receptor properties vary throughout the central nervous system. Regional variations in agonist and antagonist modulation of［ 3 H］MK-801 binding have led to the description of "agonist-preferring", "antagonist-preferring", "cerebellar" and "medial thalamic" types of NMDA receptor glutamate binding sites. Because functional properties and ligand preferences of ligand-gated ion channels often depend on subunit composition, regional and developmental differences in subunit composition of native NMDA receptors may underlie their functional and pharmacologic heterogeneity. Receptors assembled from NR1 expressed in combination with various NR2 subunits show quantitative differences in affinity for glutamate and glycine and their competitive antagonists. Studies on rat subunit recombinant NMDA receptors expressed in HEK 293 cells show a rank order of affinity for ［ 3 H glutamate as NR1/NR2B > NR1/NR2A ≈ NR1/NR2D > NR1/NR2C > NR1.NMDA had approximately equal affinity for all heteromeric types. Glycine exhibited an affinity order of NR1/NR2C > NR1 = NR1/NR2B = NR1/NR2D > NR1/NR2A. The channel-site ligand ［ 3 H］MK-801 showed the affinity ranking NR1/NR2A = NR1/NR2B > NR1 > NR1/NR2C = NR1/NR2D. However, the NMDA receptor expressed in Xenopus oocytes show an affinity order for glutamate and NMDA as NR1/NR2C > NR1/NR2A, the affinity order for D-APV and 7-CK as NR1/NR2A > NR1/NR2C.Studies on mouse subunit recombinant NMDA receptors expressed in Xenopus oocytes show a rank order of affinity for glutamate and glycine as ζ1/ε4 > ζ1/ε3 ≈ ζ1/ε2 > ζ1/ε1.The sensitivity to D-APV is in the order ζ1/ε1 > ζ1/ε2 >ζ1/ε3] > ζ1/ε4 channels, whereas that to 7-CK is in the ord er ζ1/ε3 > ζ1/ε2 > ζ1/ε1 ≈ ζ1/ε4.Thus the ligand binding affinities of recombinant NMDA receptors depends on their subunit composition.
Dose-response curves of Mg 2+ inhibition show different sensitivities to Mg 2+ blockade in heteromeric NMDA receptors. The NR1/NR2A (ζ1/ε1) and ζ1/ε2 channels are more sensitive to Mg 2+ block, whereas the NR1/NR2C (ζ1/ε3) and ζ1/ε4 channels are resistant.The modulatory effects of extracellular Ca 2+ on these five subunit recombinant NMDA receptors are also different. The steady-state currents of the ζ1/ε1, ζ1/ε3 and ζ1/ε4 recombinant NMDA receptor-channels are suppressed by extracellular Ca 2+ , whereas those of the ζ1/ε2 are potentiated by Ca 2+ . Ca 2+ -dependent inactivation of the recombinant NMDA receptor is NR2 subunit specific. NR1-1a/2A and NR1-1a/2D show reversible inactivation that is very similar to native NMDA receptors. By contrast, NR1-1a/2B and NR1-1a/2C show no significant inactivation. NR1-1a/2A also shows Ca 2+ - and glycine-independent desensitization which is less pronounced in NR1-1a/2B and absent in NR1-1a/2C. The Ca 2+ permeability, estimated by measuring reversal potentials, is similar among the four heteromeric channels.
NMDA receptor-associated proteins
Ion channels may not exist as isolated functional units in vivo, but are probably associated with a variety of other proteins that define and specify their cellular function. In the mammalian central nervous system, synapse-associated proteins (SAPs) are constituents of the cytomatrix at synaptic junctions. Recently, a novel family of structurally related SAPs has been identified.Three characterized members are SAP90 (also known as PSD-95), SAP97 and SAP102. The cytoplasmic tails of NMDA receptor subunits NR2B and two isoforms of NR1-3 and NR1-4 have been shown to interact with SAP90, SAP97 and SAP102.Both the NR2B subunit and PSD-95 were found to co-localize in cultured rat hippocampal neurons. SAP102 could be immunoprecipitated with NMDA receptors from rat brain synaptosomes. These results suggest that in vivo SAPs may form a protein complex with postsynaptic NMDA receptors.
Recently, a novel class of NMDA receptor-like (NMDAR-L or χ-1) subunits was identified,[10,11] which do not form functional receptors when expressed as monomers but markedly decrease the current responses when coexpressed with NR1 and NR2B subunits. They exhibit an average identity of 27% to NMDA subunits and 23% to non-NMDA subunits, with the highest levels present in the spinal cord, brainstem, hypothalamus, thalamus, CA1 field of the hippocampus, and amygdala. The spatial distribution of χ-1 expression is similar from postnatal day 1 to adulthood. However, transcript levels decline sharply between postnatal day 7 and postnatal day 14 and remain attenuated in adulthood.
PHYSIOLOGIC AND PHARMACOLOGIC PROPERTIES OF NMDA RECEPTORS
Modulation of channel activity
NMDA receptors are ligand-gated ion channels. In addition to the glutamate recognition site, they contain separate recognition sites for glycine, Mg 2+ , Zn 2+ , polyamine and H + , and can be modulated by a number of endogenous and exogenous compounds.
Modulation by glutamate and glycine
The NMDA receptor is activated by the endogenous neurotransmitters glutamate and glycine. A unique feature of NMDA receptors is their requirement for both glutamate and the co-agonist glycine for efficient gating. Glycine potentiates the NMDA-induced response. The activation of NMDA receptors can be detected as an increase in the binding of noncompetitive NMDA antagonists, such as tritiated (+)-5-methyl-10,11-dihydro-5H-dibenzo［a,d］cyclohepten, 5,10-imine maleate (［[3H］ MK-801), which selectively binds to the open state of the NMDA channel. Pharmacologic and electrophysiologic studies indicate that glutamate and glycine bind to distinct sites on the NMDA receptor. However, the sites for glutamate and glycine appear to be allosterically coupled. From analysis of the kinetics of NMDA receptor activation by glutamate and glycine, it is generally agreed that two molecules of glutamate and two molecules of glycine must bind to native NMDA receptors on hippocampal neurons to activate ion channel gating.
Modulation by cations
High Ca 2+ permeability and voltage-sensitive blockade by extracellular Mg 2+ are thought to be important for NMDA receptor function in mediating long-term changes in synaptic efficacy. At a resting membrane potential, the NMDA-receptor ion-channel is strictly controlled by Mg 2+ ions which, at physiological concentrations, totally block the channel. The channels containing NR2A or NR2B are more sensitive to Mg 2+ block compared with NR2C- or NR2D-containing channels. The prevention of NMDA receptor response by Mg 2+ ions may be viewed as an intrinsic protective mechanism by which excessive entry of Ca 2+ into a target cell and subsequent neuronal toxicity can be prevented.
In addition to this voltage-sensitive block by Mg 2+ , the modulatory effects of Mg 2+ on NMDA receptor function are very complicated. It has been reported that at least four effects of Mg 2+ on NMDA responses can be distinguished: the voltage-dependent block by intracellular Mg 2+ ; a potentiating effect of Mg 2+ on NMDA-induced responses at positive membrane potentials; a reduction of the elementary current by Mg 2+ measured at positive membrane potentials; and a voltage-independent and subunit specific potentiation of NMDA responses by Mg 2+ . This subunit specificity involves both the NR2 subunit (the presence of NR2B is required) and the NR1 subunit (the presence of exon 5 is suppressive). Receptor binding studies in rat brain membranes show that low concentrations of Mg 2+ increase ［[3H］MK-801 binding, and high concentrations of Mg 2+ inhibit ［[3H］MK-801.
Ca 2+ not only readily passes through the NMDA receptor channel but also modulates the activity of NMDA receptors. Ca 2+ has at least three different effects on NMDA receptor function: to potentiate the response of the NMDA receptor at nonsaturating concentrations of glycine; to increase the desensitization or inactivation of the NMDA receptor in an NR2 subunit-specific fashion at high concentration; to block ionic flux of NMDA receptors in a voltage-independent fashion. Receptor binding studies also show that Ca 2+ increases ［ 3 H］MK-801 binding at low concentrations and inhibits ［ 3 H］MK-801 at high concentrations. Both Ca 2+ and Mg 2+ probably act at a single site in the extracellular domain of the receptor to increase its affinity for glycine, for Ca 2+ and Mg 2+ have been shown to enhance strychnine-insensitive ［ 3 H］ glycine binding and increase the apparent affinity of the NMDA receptor for glycine by a voltage-independent mechanism. Some monovalent cations such as K + and Na + have been show to increase ［ 3 H］MK-801 binding at low concentrations and inhibit ［ 3 H］MK-801 at high concentrations.
Zn 2+ blocks the NMDA current in a noncompetitive and voltage-independent manner. In addition, Zn 2+ potentiates agonist-induced currents in certain splice variants of the NMDA receptor. Zn 2+ may act as a neuromodulator of NMDA receptor-mediated responses, at least in some specific brain areas, for activation of glutamatergic neurons lead to the release of Zn 2+ .
The heavy metal Pb 2+ , which is a ubiquitous environmental pollutant, has been shown to impair learning and memory. The distribution of Pb 2+ and Zn 2+ in brain and their selective accumulation in hippocampal formation are similar. In the presence of nonsaturating concentrations of glycine, currents activated by NMDA were potentiated by Pb 2+ and this potentiation was antagonized by kynurenic acid, suggesting that it may increase affinity for glycine. In the presence of saturating concentrations of glycine, Pb 2+ reduced the current of NMDA in a voltage-independent manner, and this inhibitory effect was observed at low micromolar concentrations and could be competitively antagonized by Ca 2+ .
High concentrations of Cu 2+ were shown to reduce specific ［[3H］ glutamate binding. Studies in cultured neurons show that Cu 2+ strongly antagonizes responses induced by NMDA. This voltage-independent inhibition is reduced by increasing the concentration of NMDA. Cu 2+ may directly interact with the NMDA recognition site of the NMDA receptor. This is in contrast to Zn 2+ inhibition, which is shown to be independent of agonist concentration.
Modulation by polyamines
Another well-established example of the allosteric regulation of NMDA receptors involves the modulatory action of polyamines. Polyamines are ubiquitous compounds, present at high concentrations in brain and other tissues, and may exert an important facilitatory influence on NMDA-receptor mediated transmission under pathological conditions such as brain ischemia or trauma, during which their production is dramatically enhanced. Multiple recognition sites for polyamines are proposed for NMDA receptors to account for their complex modulatory activity in vitro. It has been demonstrated that polyamines do not directly activate NMDA receptors, but instead act to potentiate or inhibit glutamate-mediated responses. The potentiating effect of polyamines can be subdivided into glycine-dependent stimulation which results in an increased affinity of the receptor for subsaturating concentrations of glycine, and glycine-independent stimulation which increases glutamate-induced currents in the presence of saturating concentrations of glycine. The potentiation effects of polyamines may also be due to a decrease in NMDA receptor desensitization. The inhibitory effects of polyamines can also be subdivided into voltage-dependent channel inhibition and a decrease in the affinity of the receptor for glutamate. Voltage-dependent inhibition involves an action within the pore of the NMDA receptor channel, probably the binding site for Mg 2+ or other channel blockers. Receptor binding studies show that low concentrations of spermine and spermidine enhance the binding of MK-801, and higher concentration of spermine inhibits the binding of MK-801. For recombinant NMDA receptors, the effects of spermine are controlled by subunit composition in the receptor complex, which may explain the confusing, and apparently contradictory effects of spermine on native NMDA receptors where stimulation (or inhibition) varies widely between cultured neurons.
Modulation by H +
The pH of extracellular space in the vertebrate CNS is surprisingly plastic and can undergo stereotypic alkaline and acid shifts following neuronal activity. Neuronal excitability has also been shown to be enhanced following extracellular alkaline shifts and to be depressed following acidification of the extracellular medium. The activity of NMDA receptors is strikingly sensitive to changes in H + concentration, and partially inhibited by the ambient concentration of H + under physiological conditions. Further acidification will produce strong suppression of NMDA receptor activity. Receptor binding studies show that high concentrations of H + decrease the［ 3 H］MK-801 binding and this decrease is due to a decrease in the observed association rate of［ 3 H］MK-801 binding. Similar to native NMDA receptors, homomeric NR1 function is strongly inhibited by protons. Examination of the eight alternatively spliced NR1 complementary DNA indicated that exon 5 controls H + inhibition. Coexpression of NR1 and NR2 revealed that exon 5 relieved pH inhibition for heteromeric receptors comprised of NR1 plus NR2A, NR2B or NR2D but not NR2C.
The modulation of NMDA receptors by H + may provide an important local feedback mechanism to limit calcium influx through the NMDA channel and to protect neurons in hypoxic/ischemic conditions during which substantial acidification of the extracellular space by lactic acid takes place.
Modulation by redox agents
The NMDA receptor has been suggested to be modulated by the redox state. Sulfhydryl reagents with strong oxidizing or reducing potentials, including those endogenous to brain, have been found to modulate the amplitude of NMDA-mediated responses in a variety of neuronal preparations. This modulation may be effected via a redox site consisting of thiol groups associated with the NMDA receptor complex. Reduction of this site by sulfhydryl reducing agents such as dithiothreitol (DTT) enhances the activity of the NMDA receptor; by contrast, oxidizing agents such as 5.5'-dithiobis,2-nitrobenzoic acid (DTNB) decrease NMDA receptor function. Although DTNB and DTT are experimental agents, a growing body of evidence indicates that there exist endogenous oxidizing and reducing agents capable of modulating the activity of NMDA receptor in vivo. One of the most interesting proposals concerns the action of nitric oxide (NO). There is general agreement that NO inhibits NMDA receptor responses, but the mechanism of action remains controversial. Molecular studies found that 2 cysteines in the NR1 subunit (Cys-744 and Cys-798) are required for redox modulation of the NMDA-gated currents in oocytes expressing NR1/NR2B, NR1/NR2C, or NR1/NR2D receptors. Redox modulation of NR1/NR2A receptors appeares to be different from that of the other heteromeric receptors, indicating the presence of one or more unique redox modulatory sites on NR1/NR2A receptor. Mutation of these 2 cysteines in the NR1 subunit also reduces the sensitivity of potentiation by spermine, the sensitivity for protons and the affinity of NMDA. These results suggest that changes in the redox state of the cysteines lead to conformational changes in the NMDA receptor protein that share certain features with changes induced by at least two other allosteric modulators, H + and spermine.
Modulation by protein phosphorylation
Recent studies suggest that protein phosphorylation is a major mechanism in the modulation of neurotransmitter receptor function. PKC has been shown to potentiate the NMDA-induced current. PKC phosphorylation occurs on several distinct sites of the NR1 subunit, and most of these sites are contained within a single alternatively spliced exon in the carboxyl-terminal domain. Protein tyrosine kinases (PTKs) have been found to regulate NMDA receptors, and these kinases exert their effects on NMDA currents by phosphorylation of tyrosine in the carboxyl-terminal domain of NR2A subunit. Protein tyrosine phosphatases (PTPs) have also been reported to reduce NMDA channel activity. This regulation by PTP is Ca 2+ -independent. NMDA channels may also be regulated by a phosphorylation that is Ca 2+ -dependent. For example, NMDA channel activity has been found to be decreased by calcineurin, a serine/threonine phosphatase, which requires activation by Ca 2+ /calmodulin. The modulation of NMDA channels by protein phosphorylation provides a potentially important way for the nervous system to control the excitability of neurons and the efficacy of synaptic transmission.
Modulation by ATP and Ca 2+ -dependent depolarization of actin
Rosenmund and Westbrook demonstrated that filamentous actin, a cytoskeletal component under dynamic control by intracellular Ca 2+ and ATP, regulate NMDA receptor activity. Increased ［Ca 2+ ］i levels were shown to induce actin depolymerization with subsequent instability of the NMDA receptor channel complex, resulting in the rundown of responses to NMDA. This was mimicked by cytochalasins and prevented by phalloidin, agents which respectively promote and prevent actin depolymerization. Such Ca 2+ -dependent inactivation could act as a safety mechanism to prevent excessive influx of Ca 2+ through NMDA receptors in energy-depleted cells.
Dopamine and histamine also regulate the NMDA receptors. The binding of ［ 3 H］MK-801 is inhibited by dopamine, perhaps by a competitive inhibition. The dose-response curve for modulation of NMDA receptor activity by histamine is biphasic, with potentiation at low concentrations and inhibition at high concentrations.
Desensitization of NMDA receptors
Desensitization is a common property of all membrane receptors and is the process by which receptors are inactivated in the prolonged presence of agonist. This phenomenon is universal, but little is known about its mechanism. So far, three types of desensitization have been described for NMDA responses: Ca 2+ -dependent desensitization or inactivation, glycine-sensitive desensitization and Ca 2+ - and glycine-independent desensitization. The Ca 2+ -dependent desensitization has a relatively slow time course of decay (τ ≈ 2 s), the magnitude of which is dependent on the extracellular Ca 2+ concentration. Glycine dependent desensitization has a much faster time course (τ < 500ms) and is most readily observed in the absence of Ca 2+ -dependent desensitization.
Inactivation is a form of desensitization in which the current is reduced during agonist application. However, inactivation of NMDA receptor does not require ligand binding or channel opening, thus distinguishing it from NMDA receptor desensitization. Although inactivation does not seem to directly involve kinases or phosphatases, its molecular mechanism is not understood.
The physiologic role of NMDA desensitization remains uncertain. However, the process is regulated by physiologic concentrations of glycine and Ca 2+ and could serve as a mechanism to limit excitation during periods of prolonged or repeated transmitter exposure.
PHYSIOLOGIC AND PATHOLOGIC ROLE OF NMDA RECEPTORS
Neural plasticity and learning
NMDA receptors have been shown to have many important physiologic roles in the vertebrate CNS. They are mediators of synaptic transmission in many pathways in the brain, including the generation of rhythmic motor activity and the regulation of neuronal development in the embryonic nervous system. It is now widely accepted that NMDA receptors play a crucial role in both synaptic plasticity and formation of the neural network during development. Tetanic stimulation-induced LTP (long-term potentiation) of synaptic transmission in the hippocampus has been extensively studied as a model of activity-dependent change in synaptic efficacy, which is thought to provide the physiologic basis for information storage in the brain. During tetanic stimulation, there is a greater and longer-lasting depolarization, which alleviates the Mg 2+ block of NMDA channels. The relationship between the NMDA receptor channel-dependent synaptic plasticity and learning has been examined by different approaches. Chronic intraventricular infusion of 2-amino-5-phosphonovaleric acid (APV), a competitive NMDA receptor antagonist, impaired both hippocampal LTP and spatial learning in rats. Mutant mice defective in the ε1 subunit of the NMDA receptor channel as a result of gene targeting, and mutant mice with a CA1 pyramidal cell-specific knockout of the NR1 gene, had impaired spatial learning and LTP in the hippocampal CA1 region. These results support the notion that NMDA receptor channel-dependent synaptic plasticity is the cellular basis of certain forms of learning. Although detailed mechanisms for both synaptic plasticity and formation of the neural circuit remain to be further elucidated, there is consensus that Ca 2+ entry into the postsynaptic neuron through NMDA receptors triggers a biochemical cascade leading to persistent changes in the CNS. It has been reported recently that the intracellular domain of NR2 subunits is very important for NMDA receptor function. Gene-targeted mice expressing NMDA receptors without the large intracellular C-terminal domain of any one of three NR2 subunits were made deficient in that particular subunit. Mice expressing the NR2B subunit in a C-terminally truncated form die perinatally. NR2A mice are viable but exhibit impaired synaptic plasticity and contextual memory. NR2C mice display deficits in motor coordination.
Neurological disorders and clinical implications
Free glutamate is present in most neural cells at an extracellular (synaptic cleft) concentration of 10 mmol/L, much higher than any other excitatory agent. In addition to their importance to normal CNS function, excessive activation of glutamate receptors may mediate neuronal injury or death termed "excitotoxicity", which may constitute a final common pathway for neuronal injury due to disease with diverse pathophysiologic processes. This form of injury appears to be predominantly mediated by excessive influx of Ca 2+ into neurons through ionic channels, triggered by activation of glutamate receptors. A new form of excitotoxicity of NMDA receptors was described recently, in which glutamate also increased the intracellular Mg 2+ concentration in Na + /Ca 2+ -free solution and produced delayed neuronal cell death. The neurotoxicity was correlated with the extracellular Mg 2+ concentration and could be blocked by the addition of NMDA receptor antagonists. It was reported currently that NMDA receptor-mediated K + efflux may contribute to neuronal apoptosis after brain ischemia. Cortical neurons exposed to NMDA in medium containing reduced Na + and Ca 2+ lost substantial amounts of intracellular K + and underwent apoptosis. Both K + loss and apoptosis were attenuated by increasing the extracellular K + .
In many neuropathologic conditions (such as ischaemia, trauma and epilepsy) and in chronic neurodegenerative states (such as Parkinson's disease), injury to neurons may be caused, at least in part, by over-stimulation of NMDA receptors. Researchers have proposed therapeutic applications for drugs that modulate NMDA receptors. These drugs include competitive NMDA receptor antagonists such as 2-amino-7-phosphonoheptanoate (AP7) and 3-［(±)-2-carboxypiperazin-4-yl］-propyl-1-phosphonate (CPP), and noncompetitive NMDA receptor antagonists such as dizocilpine (MK-801), phencyclidine (PCP) and ketamine.
Potential clinical applications for NMDA antagonists were initially defined on the basis of animal models, and epilepsy was the first target syndrome identified. Interestingly, antagonists of the NMDA receptor, including the competitive types (such as AP7 and CPP) and NMDA channel blockers (such as MK-801), all displayed anticonvulsant properties in animal models of epilepsy. MK-801 was found to be the most potent compound in protecting against NMDA-induced damage and global ischaemia. Today stroke and traumatic injury of the spinal cord and brain are the most important clinical targets. A protective action for CPP in the spinal cord and brain trauma has been reported. Parkinson's disease may also be a target, since NMDA antagonists have been shown to provoke the release of dopamine in basal ganglia and to cause monoamine-like locomotor stimulation in monoamine-depleted rodents. There is strong evidence from rodent models of Parkinsonism that motor improvement induced by L-dopa can be greatly enhanced by the co-administration of NMDA antagonists.
NMDA receptor antagonists such as MK-801 also have a number of other proprieties, such as delaying or preventing the development of morphine tolerance and dependence, blocking the development of behavioral sensitization to cocaine and amphetamine, inhibiting tolerance to alcohol and alcohol withdrawal-induced seizures and blocking the neurotoxic effects of the HIV viral coat protein gp120 in vitro.
It is expected that the clinical use of drugs modulating excitatory amino acid systems will be limited, because NMDA receptors are involved in basic physiologic processes and development. Cloning and expression of specific genes for NMDA receptor proteins and evidence that subunit composition varies in different parts of the brain open up a large new field for pharmacologic research. It may be possible to identify NMDA receptor antagonists that are specific for particular heteromeric forms of the receptor which may be responsible for specific normal or abnormal functions. Furthermore, identifying the modulatory effects of endogenous agents on NMDA receptors may provide evidence for novel modulatory binding sites, thus forming a new basis for the design of new drugs with high specificity and fewer side effects.
1. Traynelis SF, Hartley M, Heinemann SF. Control of proton sensitivity of the NMDA receptor by RNA splicing and polyamines. Science 1995;268:873-876.
2. Watanabe M, Inoue Y, Sakimura K, et al. Developmental changes in distribution of NMDA receptor channel subunit mRNAs. Neu Rep 1992;3:1138-1140.
3. Watanabe M, Mishina M,Inoue Y. Distinct spatiotemporal distributions of the N-methyl-D-aspartate receptor channel subunit mRNAs in mouse cervical cord. J Comp Neurol 1994;345:314-319.
4. Laurie DJ, Seeburg PH. Ligand affinities at recombinant N-methyl-D-aspartate receptors depend on subunit composition. Eur J Pharmacol 1994;268:335-345.
5. Ishii T, Moriyoshi K, Sugihara H, et al. Molecular characterization of the family of the N-methyl-D-aspartate receptor subunits. J Biol Chem 1993;268:2836-2843.
6. Kutsuwada T, Kashiwabushi N, Mori H, et al. Molecular diversity of the NMDA receptor channel. Nature 1992;358:36-41.
7. Mishina M, Mori H, Araki K, et al. Molecular and functional diversity of the NMDA receptor channel. Annu NY Acad Sci 1993;707:136-152.
8. Gomperts SN. Clustering membrane proteins: it's all coming together with the PSD-95/SAP90 protein family. Cell 1996;84:659-662.
9. Muller BM, Kistner U, Kindler S, et al. SAP102, a novel postsynaptic protein that interacts with the cytoplasmic tail of the NMDA receptor subunit NR2B. Neuron 1996;17:255-265.
10. Ciabarra AM, Sullivan JM, Gahn LG, et al. Cloning and characterization of χ-1: a developmentally regulated member of a novel class of the ionotropic glutamate receptor family. J Neurosci 1995;15:6498-6508.
11. Sucher NJ, Akbarian S, Chi CL, et al. Developmental and regional expression pattern of a novel NMDA receptor-like subunit (NMDAR-L) in the rodent brain. J Neurosci 1995;15:6509-6520.
12. Liu Y, von Euler G. Ca 2+ and H[+ antagonizes the decrease of ［ 3 H］MK-801 binding induced by glutamate and glycine in the presence of Mg 2+ . Neurochem Int 1996;28:401-415.
13. Hollmann M, Boulter J, Maron C, et al. Zinc potentiates agonist-induced currents at certain splice variants of the NMDA receptor. Neuron 1993;10:943-954.
14. Johnson TD. Modulation of channel function by polyamines. Trends Pharmacol Sci 1996;17:22-27.
15. Sullivan JM, Traynelis SF, Chen H-SV, et al. Identification of two cysteine residues that are required for redox modulation of the NMDA subtype of glutamate receptor. Neuron 1994;13:929-936.
` 16. Raymond LA, Tingley WG, Blackstone CD, et al. Glutamate receptor modulation by protein phosphyorylation. J Physiol 1994;88:181-192.
17. Khr G, Seeburg PH. Subtype-specific regulation of recombinant NMDA receptor-channels by protein tyrosine kinases of the src family. J Physiol 1996;492:445-452.
18. Rosenmund C, Westbrook GL. Calcium-induced actin depolymerization reduces NMDA channel acitivity. Neuron 1993;10:805-814.
19. von Euler G, Liu Y. Dopamine inhibits ［ 3 H］MK-801 binding in membrane preparations from rat cerebral cortex and caudate putamen. Acta Physiol Scand 1992;146:547-548.
20. Tsien Jz, Huerta PT, Tonegawa S. The essential role of hippocampal CA1 NMDA receptor-dependent synaptic plasticity in spatial memory. Cell 1996;87:1327-1338.
21. Sprengel R, Suchanek B, Amico C, et al. Importance of the intracellular domain of NR2 subunits for NMDA receptor function in vivo. Cell 1998;92:279-289.
22. Hartnett KA, Stout AK, Rajdev S, et al. NMDA receptor-mediated neurotoxicity: a paradoxical requirement for extracellular Mg 2+ in Na + /Ca 2+ -free solutions in rat cortical neurons in vitro. J Neurochem 1997;68:1836-1845.
23. Yu SP, Yeh C, Strasser U, et al. NMDA receptor-mediated K + efflux and neuronal apoptosis. Science 1999;284:336-339.
24. Lipton SA, Sucher NJ, Kaiser PK, et al. Synergistic effets of HIV coat protein and NMDA receptor-mediated neurotoxicity. Neuron 1991;7:111-118.