THE ROLE OF INTRACELLULAR SODIUM (Na⁺) IN THE REGULATION OF CALCIUM (Ca²⁺)-MEDIATED SIGNALING AND TOXICITY

Abstract

It is known that activated N-methyl-D-aspartate receptors (NMDARs) are a major route of excessive calcium ion (Ca²⁺) entry in central neurons, which may activate degradative processes and thereby cause cell death. Therefore, NMDARs are now recognized to play a key role in the development of many diseases associated with injuries to the central nervous system (CNS). However, it remains a mystery how NMDAR activity is recruited in the cellular processes leading to excitotoxicity and how NMDAR activity can be controlled at a physiological level. The sodium ion (Na⁺) is the major cation in extracellular space. With its entry into the cell, Na⁺ can act as a critical intracellular second messenger that regulates many cellular functions. Recent data have shown that intracellular Na⁺ can be an important signaling factor underlying the up-regulation of NMDARs. While Ca²⁺ influx during the activation of NMDARs down-regulates NMDAR activity, Na⁺ influx provides an essential positive feedback mechanism to overcome Ca²⁺-induced inhibition and thereby potentiate both NMDAR activity and inward Ca²⁺ flow. Extensive investigations have been conducted to clarify mechanisms underlying Ca²⁺-mediated signaling. This review focuses on the roles of Na⁺ in the regulation of Ca²⁺-mediated NMDAR signaling and toxicity.

I. INTRODUCTION

Cytoplasmic Ca²⁺ is the most common signaling factor in all types of cells. Normal intracellular Ca²⁺ concentration ([Ca²⁺]_i) is approximately 40,000-fold lower than extracellular [Ca²⁺], which ranges from 1 to 2 mM ^14,27. Ca²⁺ ions enter neurons via various pathways including voltage-gated Ca²⁺ channels, ligand-gated Ca²⁺ channels and the Ca²⁺ exchangers ^27,34. It is known that activated NMDAR channels are a major route of excessive Ca²⁺ entry in neurons ^{26,41,62,69,91,96,117}. While excessive intracellular Ca²⁺ may activate degradative processes and thereby cause toxic effects ^20,27,40,117, NMDAR channel activity may be inhibited by intracellular Ca²⁺ through: (i) α-actinin/cytoskeleton dissociation from the NR1 subunit of NMDARs ⁵⁵, (ii) calmodulin activation ^{37,55,104,116}, and (iii) activation of phosphatases, such as calcineurin which dephosphorylates NMDARs ^59,72,94. The Ca²⁺-induced down-regulation of NMDARs is considered an important negative feedback mechanism to control NMDAR activity ^33,56,64,66. Based on these findings we questioned: How do excessive amounts of Ca²⁺ get into neurons through NMDARs if NMDARs are inhibited by Ca²⁺ influx?

Na⁺ is the major cation in the extracellular space, and it can enter cells through a variety of routes including permeation through ligand- (e.g., glutamate) and voltage-gated cation channels, uptake via membrane exchangers and gradient-driven co-transporters ⁷³. NMDAR channels are highly permeable to both Na⁺ and Ca²⁺. Short burst or tetanic stimulation of afferents that induces synaptic LTP increases [Na⁺] up to 40 or 100 mM in spines and adjacent dendrites ^82,83. These increases can essentially be prevented by the blockade of NMDARs, indicating that they are mainly mediated by Na⁺ entry through NMDARs ^82,83.

Our initial studies demonstrated that intracellular Na⁺ is an up-regulator of NMDARs, such that raising [Na⁺]_i or activating Na⁺ permeable channels may increase NMDAR-mediated currents ^110,112,113. We then identified that an increase of 5 ± 1 mM in [Na⁺]_i represents a threshold required to mask the down-regulation of NMDARs induced by Ca²⁺ influx. Further increases in Na⁺ influx not only significantly enhance Ca²⁺ influx induced by the activation of NMDARs, but also overcome the Ca²⁺-dependent inhibition of NMDARs ^107,110. This review focuses on the roles of Na⁺ in the development of tissue injury and in the regulation of Ca²⁺-mediated NMDAR signaling and toxicity.

II. Na⁺ IN THE PROCESS OF TISSUE-INJURY

A significant increase in [Na⁺]_i is a characteristic event associated with tissue injury ^{6–9,42,87,92}. Application of voltage-gated Na⁺ channel blockers reduce both Na⁺ entry and apoptotic neuronal death ⁷ whereas increases of Na⁺ entry by application of the voltage-gated Na⁺ channel activator, veratridine, induce neuronal apoptosis and caspase-3 activation ^7,8. There is a report showing that during anoxia Na⁺ entry can occur through either Gd³⁺-sensitive channels or via Na⁺/K⁺/2Cl⁻ co-transporters in cultured hippocampal neurons ⁸⁸, implying that multiple pathways for Na⁺ entry may be activated during tissue injury.

It is known that Na⁺ influx into the cell is accompanied by chloride ions (Cl⁻) and water, which can lead to acute neuronal swelling and damage ^25,26. Previous studies have shown that Na⁺ entry may cause an increase in cytosolic Ca²⁺ through either Na⁺/Ca²⁺ exchangers or activation of voltage-gated Ca²⁺ channels ^17,53, thereby activating Ca²⁺-dependent signaling mechanisms. Moreover, Na⁺ entry via Na⁺/H⁺ exchange may cause changes in intracellular pH, and thereby regulate many cellular functions including enzyme activity, neuronal growth and death ^{10,18,70,71,88,89}. A recent study showed that Na⁺ influx plays an important role in the onset of anti-Fas-induced apoptosis and that blocking Na⁺ influx may rescue programmed cell death in Jurkat cells ¹⁹. Cox and colleagues reported that the binding of agonists to opioid receptors on guinea pig cortical neuron membranes is significantly reduced by increases in [Na⁺]_i of 10 – 30 mM ¹⁰⁵. Maximal inhibition of μ-, δ- and κ-opioid receptor binding by Na⁺ is approximately 60%, 70% and 20%, respectively ¹⁰⁵. Co-occurrence of Na,K-ATPase dysfunction and Na⁺ influx causes α-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate receptor (AMPAR) proteolysis and a rapid reduction of AMPAR cell-surface expression ¹¹⁵. Na+-mediated K⁺ channels such as Slo gene-encoded K⁺ channels ^{15,16,35,76,114}, are widely distributed throughout the nervous system and are involved in both the regulation of the after-potential following action potentials ^16,63, and the protection of neurons from hypoxic stimulation ^16,35,114.

While the details of the mechanisms remain to be clarified, significant pharmacological data have demonstrated the protective effects of blocking Na⁺ influx during injuries to the nervous tissue. The blockade of voltage-gated Na⁺ channels can prevent neurons from traumatic spinal cord injury ^{1,2,8,39,46,92,93} and the loss of white matter ^1,8,30,93, concurrently reducing the sensitization associated with pain ^4,28,32,103 and preventing seizures during kindling development ⁸¹. The inhibition of Na⁺-H⁺ exchange attenuates ischemia-induced cell death ^68,100. As a result, a major focus of pharmaceutical research has been the search for effective therapeutic approaches that target voltage-gated Na⁺ channels ^8,87.

III. THE ROLES OF Na⁺ IN THE REGULATION OF Ca²⁺-MEDIATED NMDAR SIGNALING AND TOXICITY

Calcium influx through activated NMDARs is regulated by Na⁺ influx

Activated NMDARs are highly permeable to both Na⁺ and Ca^{2+ 33,64,66}. Prolonged increases of intracellular Ca²⁺ during NMDAR activation may act as a negative feedback mechanism controlling NMDAR activity ^33,64,66. In light of our findings demonstrating that: 1) intracellular Na⁺ up-regulates NMDAR channel gating and 2) multiple types of receptor/channels such as AMPARs, voltage-gated Na⁺ channels, non-selective cation channels and remote NMDARs may regulate NMDAR activity through a Na⁺-dependent mechanism ^110,112, we investigated how NMDARs are regulated when both Ca²⁺ and Na⁺ flow into neurons during the same time period through activated NMDARs ^107,110. Recordings were conducted in the cell-attached single-channel configuration. In this recording model, recorded surface NMDARs are isolated by a recording electrode from the bath environment and therefore cannot be directly stimulated by bath-applied agents. We recorded the activity of surface NMDARs before and after activation of remote NMDARs (outside the patch) induced by bath application of NMDA or L-aspartate ¹⁰⁷. To prevent toxic effects which may be induced by application of NMDA or aspartate, a standard extracellular solution in which NaCl and KCl were replaced by Na₂SO₄ and Cs₂SO₄, was utilized ¹⁰⁷. Consistent with previous findings ^{25,54,59,111,112}, no damage of neurons bathed with this standard solution was observed following NMDA or aspartate application. NMDAR single-channel activity was evoked with 10 µM NMDA and 3 µM glycine included in the standard extracellular solution filling the recording electrodes.

We found that bath application of NMDAR agonists may change NMDAR channel activity recorded in cell-attached patches in a concentration-dependent manner. While a significant increase in NMDAR channel gating occurred during L-aspartate (>100 µM) application to neurons bathed with the standard extracellular solution, the activity of NMDARs was inhibited in neurons when Na⁺ influx was blocked by replacing extracellular Na⁺ with Cs⁺ or N-methyl-D-glutamine (NMDG) ^107,112.

We measured the ratio of fluorescence at 346 nm versus 380nm for the Na⁺-sensitive dye, sodium-binding benzofuran isophthalate (SBFI), and the Ca²⁺-sensitive dye, Fura-2, in the soma region of neurons. When the Na⁺ gradient across the cell membrane was decreased by reducing extracellular Na⁺ concentration ([Na⁺]_e) to 20 mM and the Na⁺ ionophore, monensin (10 µM) was included in the extracellular solution, basal [Ca²⁺]_i and [Na⁺]_i of neurons were approximately 84 nM and 16 mM, respectively. Under this condition bath application of L-aspartate increased [Ca²⁺]_i by 66 nM, decreased [Na⁺]_i by 5.8 mM and inhibited NMDAR activity ¹⁰⁷. On average, the overall channel open probability and mean open time were reduced to 64% and 77% of controls. The burst and cluster lengths were also significantly reduced. These inhibitory effects produced by the bath application of L-aspartate were prevented by either application of APV or removal of Ca²⁺ from extracellular solution, indicating that the activation of remote NMDARs may also down-regulate recorded NMDAR activity through Ca²⁺ influx ¹⁰⁷. Thus, it is demonstrated that NMDARs can be up- and down-regulated by influxes of Na⁺ and Ca²⁺, respectively.

We then measured changes of [Na⁺]_i and [Ca²⁺]_i in neurons bathed with extracellular solution containing a [Na⁺] of 10, 20 or 145 mM before and during the activation of NMDARs induced by bath application of L-aspartate. We found that with an increase in [Na⁺]_e, the activation NMDARs produced increases in [Na⁺]_i as expected, but also increased [Ca²⁺]_i. Excluding the effect of Ca²⁺ influx-induced Ca²⁺ release (CICR) from intracellular stores, the increase in [Ca²⁺]_i of neurons bathed with extracellular solution containing 145 mM Na⁺ was still significantly higher than that found in neurons bathed with extracellular solution containing 10 mM Na^{+ 107,110}. When [Na⁺]_e was reduced to 10 mM, the activation of NMDARs produced increases in [Na]_i and [Ca²⁺]_i by around 0.8 mM and 35 nM, respectively. Under this condition, the activation of remote NMDARs inhibited NMDAR activity recorded in cell-attached patches ¹⁰⁷. When [Na⁺]_e was increased to 20 mM, NMDAR activation produced a 5 mM increase in [Na⁺]_i and a 50 nM increase in [Ca²⁺]_i, but no change in the activity of recorded NMDARs ¹⁰⁷. Similarly, increasing [K⁺]_e by 30 mM in an extracellular solution containing 170 mM Na⁺ and 1 µM TTX produced increases in [Na⁺]_i and [Ca²⁺]_i by around 7 mM and 48 nM, respectively, but again showed no change in the activity of NMDARs recorded in cell-attached patches either ^107,112. Thus, an increase in [Na⁺]_i of approximately 5 mM appeared to be a critical concentration for masking the inhibitory effects induced by Ca²⁺ influx on NMDARs in cultured hippocampal neurons ¹⁰⁷. Since a modest increase of [Ca²⁺]_i by approximately 35 nM inhibited NMDAR activity when [Na⁺]_e was reduced to 10 mM ¹⁰⁷, it was possible that Na⁺ influx not only enhanced Ca²⁺ influx but also masked the inhibitory effects of Ca²⁺.

To confirm this hypothesis, we recorded NMDAR single-channel activity before and during the activation of remote NMDARs in cell-attached patches with pipettes filled with a Ca²⁺-free extracellular solution containing 200 mM Na⁺ from neurons that had been pre-treated with BAPTA-AM (10 µM for 4 hrs) and bathed with the same Ca²⁺-free extracellular solution, or with pipettes filled with extracellular solution containing 0.3 or 1.2 mM Ca²⁺ from neurons bathed with the extracellular solution containing the same amount of Ca²⁺, respectively. We found that the activation of remote NMDARs produced a similar up-regulation of NMDAR channel activity when local and bath [Ca²⁺] was set at 0, 0.3 and 1.2 mM, implying again that the effects of Ca²⁺ influx in the regulation of NMDARs by remote NMDARs are overcome by Na⁺ under normal condition ¹⁰⁷. Furthermore, removal of extracellular Ca²⁺ did not produce any effect on the up-regulation of NMDARs by remote NMDARs in neurons bathed with the standard extracellular solution containing 200 mM Na^{+ 107,112}. Thus, we conclude that Ca²⁺ influx through activated NMDARs is regulated by Na⁺ influx, and that the effect of Na⁺, which overcomes Ca²⁺-induced inhibition, provides an essential positive feedback mechanism enhancing both the NMDAR activity and the inward flow of Ca²⁺.

Depletion of extracellular Ca²⁺ enhances Na⁺ influx and thereby causes NMDAR-mediated toxicity

Based on findings that glutamate concentration may increase in both humans ^5,80 and animals after nervous system injury ¹¹, and that application of NMDAR antagonists may protect neurons from excitotoxic injuries in both humans ^24,61 and animal ^24,25,60, it has been believed that NMDAR-mediated excitoxicity plays a key role in the development of neuronal death associated with stroke/traumatic CNS injury. However, it remained unclear how NMDARs are recruited to cause neurotoxicity.

We examined the effects of extracellular Ca²⁺ depletion and reperfusion, which may occur in stroke patients, on cultured hippocampal neurons ^108,110. Neurons were bathed initially with an extracellular solution containing: 140 mM NaCl, 5 mM CsCl, 1.8 mM CaCl₂, 33 mM glucose, 25 mM HEPES; pH: 7.35; osmolarity: 310–320 mOsm. The reduction of [Ca²⁺]_e from 1.8 mM to 0.5 or 0 mM caused a significant increase in Caspase-3 activity and morphological changes in neurons such as swelling, beading, and/or process disintegration. Significantly less formazan was observed in 3-(4,5-dimethylthiazol-2-yl)2,5-diphenyl-tetrazolium bromide (MTT) assays in which neurons were treated with the extracellular solution containing 0.5 or 0 mM Ca²⁺, indicating a change in mitochondrial function associated with neuronal injury ^43,78,86,101. Unexpectedly, application of NMDAR antagonists APV (100 µM) and MK801 (2 µM) significantly prevented the above mentioned changes in neurons only when the drugs were applied concomitant to the reduction of [Ca²⁺]_e from 1.8 to 0 mM. No protective effects of the drugs could be found when they were applied during Ca²⁺ reperfusion or when [Ca²⁺]_e was reduced from 1.8 to 0.5 mM ¹⁰⁸. These findings suggest that the depletion of extracellular Ca²⁺ may evoke NMDAR-mediated neurotoxicity ¹⁰⁸, and also raised the questions of how and when NMDAR activity is recruited to induce neuronal injury following the removal of extracellular Ca²⁺.

To address this question, we recorded NMDAR single-channel activity before and during a depletion of extracellular Ca²⁺ from 1.8 to 1.3, 0.5 or 0 mM in cell-attached patches from cultured hippocampal neurons. To prevent cell damage during the reduction of extracellular Ca²⁺, the Cl⁻ in the standard extracellular solution was replaced by SO₄^{2–25,107,108,112}. Bath application of a low [Ca²⁺] solution to neurons caused a parallel shift of the current-voltage (I/V) relationship in NMDAR single-channels recorded in cell-attached patches, which indicates that there is a cell-depolarization, but no change in single-channel conductance ^107,108. In order to account for this, the holding potential was re-adjusted to maintain a 70 mV patch-potential from the reversal potential of recorded channels. We found that a depletion of extracellular Ca²⁺ from 1.8 to 1.3 or 0.5 did not induce any significant change in the activity of recorded channels until [Ca²⁺]_e was reduced from 1.8 to 0 mM ^107,108,112. The channel activity could be subsequently abolished with application of the NMDAR antagonist, MK801, confirming that a [Ca²⁺]_e reduction from 1.8 to 0 mM produces increases in NMDAR activity ^107,108,112. Since the concomitant blockade of NMDARs to the reduction of [Ca²⁺]_e from 1.8 to 0.5 mM may actually increase the number of injured neurons ¹⁰⁸, the up-regulation of NMDARs appears to be essential in the triggering of toxicity mediated by NMDARs and the application of NMDAR antagonists in the Ca²⁺ reperfusion model may be protective only when NMDARs are recruited.

To identify the mechanisms by which the removal of extracellular Ca²⁺ results in the up-regulation of NMDARs, we measured [Ca²⁺]_i and [Na⁺]_i in cultured hippocampal neurons before and during reductions of extracellular Ca²⁺. A [Ca²⁺]_e reduction-dependent decrease in [Ca²⁺]_i and increase in [Na⁺]_i were observed ¹⁰⁸. A depletion of extracellular Ca²⁺ from 1.8 to 0 mM produced sufficient increases in [Na⁺]_i capable of enhancing NMDAR activity ^107,108,112. Furthermore, we found that the up-regulation of NMDAR activity induced by extracellular Ca²⁺ depletion was prevented by the blockade of Na⁺ influx ¹⁰⁸.

Previous studies showed that the removal of extracellular Ca²⁺ to 0 mM may increase NMDAR single-channel conductance ^33,66,106, and that reducing intracellular Ca²⁺ may reduce the Ca²⁺-dependent inhibition of NMDARs and thereby enhance NMDAR channel activity ^{33,37,55,59,66,72,94,104,116}. Therefore, it is possible that NMDAR gating may be enhanced by the removal of extracellular Ca²⁺ through Na⁺ and/or Ca²⁺-dependent mechanisms.

The ensemble currents produced by the summation of consecutive super-clusters were compared before (1.8 mM) and after reducing [Ca²⁺]_e to 0 mM. We found that the removal of extracellular Ca²⁺ may significantly increase the decay time of ensemble currents and that this effect can be abolished by blocking Na⁺ influx. This suggests that the removal of extracellular Ca²⁺ may affect NMDAR-mediated whole-cell responses through the action of Na^{+ 108}.

Large reductions in [Ca²⁺]_e have been found during instances of high neuronal activity ^48,84,97, the development of seizures ⁵⁰, hypoglycemic coma ³⁸, and periods of hypoxia and ischemia ^49,74. Ca²⁺-depletion has also been reported to induce cell injury and death ⁹⁰. Thus, the Na⁺-dependent enhancement of NMDAR activity induced by depletion of extracellular Ca²⁺ may be an important mechanism underlying the development of neurotoxicity in the CNS.

Na⁺ regulation of Ca²⁺ homeostasis

Under resting conditions [Ca²⁺]_i in neurons is normally maintained at 10 – 100 nM, and is tightly regulated by both Ca²⁺ influx and efflux across the membrane. [Ca²⁺]_i can be increased by Ca²⁺ entry through Ca²⁺ channels (including ligand- and voltage-gated Ca²⁺ channels and non-selective cation channels) located on the plasma membrane and by CICR from the endoplasmic reticulum (ER) upon binding of inositol trisphosphate (IP3) to the inositol trisphosphate receptor (IP3R). Ca²⁺-mediated injury is usually acute and rapid ⁹⁵. Disturbances of Ca²⁺ homeostasis in the cytoplasm, ER, or mitochondria can be harmful to cells ³⁴. Since Ca²⁺ stores are closely connected within the cells and interact with each other, dysregulation of one compartment is usually followed by responses from the others. Together, they may overwhelm the cell’s capacity to maintain overall homeostasis and kill the cell ³⁴.

In the plasma membrane Ca²⁺-ATPase and the Na⁺/Ca²⁺ exchanger act to transport cytosolic Ca²⁺ to the extracellular space. The Na⁺/Ca²⁺ exchanger has a low affinity for Ca²⁺ but a high velocity; as such, it removes Ca²⁺ only when cytosolic concentrations are high. The Ca²⁺- ATPase, has a high affinity for Ca²⁺ and pumps out Ca²⁺ even at low cytosolic concentrations ^13,22,34. In resting cells, [Ca²⁺] in the mitochondrial matrix is around 100 nM. When cytosolic [Ca²⁺] rises, Ca²⁺ can enter the mitochondria through a uniporter and thereby regulate Ca²⁺ signals ³⁶. In mitochondria the Na⁺/Ca²⁺ exchanger extrudes Ca^{2+ 34,36,45}. However, the activity of the Na⁺/Ca²⁺ exchanger may be reversed on the influx of Na^{+ 51}. This reversal in Na⁺/Ca² exchange is observed under pathological conditions ³⁴. If the mitochondrial Na⁺/Ca²⁺ exchanger is overwhelmed by Ca²⁺ entry, the Ca²⁺ levels in the mitochondrial matrix may increase enough to trigger a mitochondrial permeability transition. The sustained transitions may cause mitochondrial depolarization, inhibition of ATP production, and cell death ^12,29,57. In the nucleus Ca²⁺ is involved in the gene transcription and DNA metabolism ^47,67,79. Unlike in the mitochondria, nuclear Ca²⁺ is found to be rapidly equilibrated with cytosolic Ca²⁺. This may occur by diffusion across nuclear pores ³ and/or Ca²⁺ channels in the nuclear envelope ⁶⁵.

CICR during NMDAR activation has been reported ^33,85,98. We observed that when extracellular solution contained more Na⁺, NMDAR activation produced greater increases in both [Na⁺]_i and [Ca²⁺]_i ¹⁰⁷. Furthermore, in neurons bathed with extracellular solution containing 145 mM Na⁺, NMDAR activation-induced increases in [Ca²⁺]_i were significantly reduced from 100 ± 30 nM (n = 5) to 62 ± 8 nM (n = 8) with thapsigargin (0.1 µM) treatment ¹⁰⁷, which depletes intracellular stores of Ca²⁺ by blocking Ca²⁺ re-uptake. In the absence of thapsigargin, NMDAR activation only produced a 35 ± 8 nM (n = 8) increase in [Ca²⁺]_i in neurons bathed with extracellular solution containing 10 mM Na^{+ 107}. The blockade of Ca²⁺ influx by removal of extracellular Ca²⁺ abolished the NMDAR activation-induced increase in [Ca²⁺]_i, (data not shown) ^27,99. The increase in [Ca²⁺]_i induced by Ca²⁺ release from intracellular stores during NMDAR activation in neurons bathed with extracellular solution containing 10 mM Na⁺ was significantly reduced when compared with that in neurons bathed with extracellular solution containing 145 mM Na^{+ 107}. These data suggest that CICR from intracellular stores during NMDAR activation may be regulated by intracellular Na⁺. Stys and colleagues provided direct evidence showing that intra-axonal Ca²⁺ release during ischemia in rat optic nerves is mainly dependent on Na⁺ influx. This Na⁺ accumulation stimulates three distinct intra-axonal sources of Ca²⁺: (1) the mitochondrial Na⁺/Ca²⁺ exchanger driven in the Na⁺ import/Ca²⁺ export mode, (2) positive modulation of ryanodine receptors, and (3) promotion of IP3 generation by phospholipase C ⁷⁵.

IV. QUESTIONS AND FUTURE STUDIES

Na⁺ entry is a key factor that initiates fast action potentials and shapes sub-threshold electrical properties to thereby regulate neuronal excitability and neuronal discharge activity ^{21,23,44,52,102}. Present data have shown that: 1) intracellular Na⁺ up-regulates NMDARs; 2) via increasing intracellular Na⁺, multiple types of receptor/channels such as AMPA receptors, voltage-gated Na⁺ channels and non-selective cation channels, may regulate NMDAR activity; 3) Na⁺ influx may enhance Ca²⁺ influx, mask the Ca²⁺-dependent inhibition of NMDARs and significantly alter Ca²⁺ homeostasis.

Based on combined investigations of protein crystal structures in-vitro and functions in cells, Na⁺ binding motifs have been characterized in a number of proteins such as thrombin, Na⁺/K⁺-ATPase and various neurotransmitter transporters. Thrombin is a serine protease, the activity of which is regulated by Na⁺ binding. The sequence, CDRDGKYG, in the Na⁺ binding loop is highly conserved in thrombin from 11 different species ³¹. Investigations into the crystal structure of a bacterial homologue of the Na⁺/Cl⁻ dependent transporters from Aquifex aeolicus revealed that there are two Na⁺ binding sites, named Na1 and Na2 ¹⁰⁹. Na⁺/K⁺-ATPase is found to have three Na⁺ biding sites. Na1 is formed entirely by the side chain oxygen atoms of residues on three helices in the transmembrane regions (TM) 5, 6 and 8. Na2 is formed almost “on” the TM4 helix with three main chain carbonyls plus four side chain oxygen atoms (Asp 811 and Asp 815 on TM6 and Glu 334 on TM4). The Na3 binding site is contiguous to Na1. The carbonyls of Gly 813 and Thr 814 (TM6), the hydroxyl of Tyr 778 (TM5), and the carboxyl of Glu 961 (TM9) contribute to the Na3 binding site ^58,77.

To date there is no evidence of a similar amino acid sequence corresponding to a Na⁺ binding site, as seen in these Na⁺ binding proteins, present in NMDAR subunit proteins (Yu, unpublished data). Molecular mechanisms underlying the regulation of NMDARs and Ca²⁺ signaling by intracellular Na⁺ remain unclear. Investigations aiming to identify critical Na⁺ targeting site(s) in the regulation of NMDARs and Ca²⁺ homeostasis are essentially needed for understanding activity-dependent neuroplasticity in the CNS.