Tyrosine Dephosphorylation and Ethanol Inhibition of N-Methyl-d-aspartate Receptor Function*

The inhibitory effect of ethanol on N-methyl-d-aspartate receptors (NMDARs) is well documented in several brain regions. However, the molecular mechanisms by which ethanol affects NMDARs are not well understood. In contrast to the inhibitory effect of ethanol, phosphorylation of the NMDAR potentiates channel currents (Lu, W. Y., Xiong, Z. G., Lei, S., Orser, B. A., Dudek, E., Browning, M. D., and MacDonald, J. F. (1999) Nat. Neurosci.2, 331–338). We have previously shown that protein kinase C activators induce tyrosine phosphorylation and potentiation of the NMDAR (Grosshans, D. R., Clayton, D. R., Coultrap, S. J., and Browning, M. D. (2002) Nat. Neurosci. 5, 27–33). We therefore hypothesized that the ethanol inhibition of NMDARs might be due to changes in tyrosine phosphorylation of NMDAR subunits. In support of this hypothesis, we found that tyrosine phosphorylation of both NR2A and NR2B subunits was significantly reduced following in situ exposure of hippocampal slices to 100 mm ethanol. Specifically, phosphorylation of tyrosine 1472 on NR2B was reduced 23.5%. These data suggest a possible mechanism by which ethanol may inhibit the NMDAR via activation of a tyrosine phosphatase. Electrophysiological studies demonstrated that ethanol inhibited NMDAR field excitatory postsynaptic potential slope and amplitude to a similar degree as previously reported by our laboratory and others (Schummers, J., Bentz, S., and Browning, M. D. (1997) Alcohol Clin. Exp. Res. 21, 404–408). Inclusion of bpV(phen), a potent phosphotyrosine phosphatase inhibitor, in the recording chamber prior to and during ethanol exposure significantly reduced the inhibitory effect of ethanol on NMDAR field excitatory postsynaptic potentials. Taken together, these data suggest that phosphatase-mediated dephosphorylation of NMDAR subunits may play an important role in mediating the inhibitory effects of ethanol on the N-methyl-d-aspartate receptor.

NMDA 1 receptors are ionotropic glutamate receptors that mediate calcium and sodium entry into neurons. In the CA1 region of the hippocampus, activation of these receptors is known to be required for induction of long term potentiation (LTP), a cellular process proposed by many as a possible mechanism of memory formation. Aberrant NMDA receptor function can result in cell death and may have implications in pathophysiologic disorders such as Parkinson's disease and epilepsy as well as age-related dementias like Alzheimer's disease (7)(8)(9)(10)(11). NMDA receptors are heteromeric assemblies of NR1 and NR2 subunits. The NR1 family is composed of a single gene, which can be alternatively spliced to generate eight theoretical splice variants (12). The NR2 family includes four genes, each encoding a unique subunit: NR2A, NR2B, NR2C, and NR2D (13). A third gene family, NR3, has been discovered, but the function of these subunits is not well understood. Whereas the NR1 subunit is required for a functional channel, differential incorporation of NR2 subunits regulates receptor function. The NR2 subunits, particularly NR2A and NR2B, have been shown to influence the sensitivity of NMDA receptors to ethanol block in cultured neurons as wells as transfected HEK293 cells and oocytes (14 -17). More recent studies have indicated that NR1 subunits may also play a key role in mediating the effects of ethanol on channel activity (18 -20).
Ethanol is known to have widespread effects on the central nervous system. Ethanol inhibits transmission at the NMDAR subtype of glutamatergic excitatory synapses and enhances inhibitory GABAergic synaptic transmission. In addition, we and others have demonstrated that ethanol blocks LTP in the hippocampus (6,(21)(22)(23). Given that NMDA receptors are both required for induction of LTP and inhibited by ethanol, it is likely that inhibition of LTP by ethanol in the hippocampus is due, at least in part, to inhibition of NMDA receptors. The molecular mechanisms of this inhibition however, remain unknown.
One hypothesis that we favor is that ethanol may inhibit the NMDAR by reducing its phosphorylation. Indeed, ethanol has been shown to enhance protein-tyrosine phosphatase activity and also inhibit receptor tyrosine kinase activity (24,25). Moreover, previous studies suggest that Fyn, a member of the Src family of tyrosine kinases, may be a key player in regulating the sensitivity of NMDARs to ethanol in the hippocampus (26 -28). Electrophysiological studies have demonstrated that Src kinase regulates NMDAR channel gating by increasing mean channel open time and open probability (29). Complementary studies demonstrated that Src activation of NMDARs is tonically opposed by striatal enriched tyrosine phosphatase (30). The striatal enriched tyrosine phosphatase family of tyrosine phosphatases was originally characterized in the striatum, but subsequent studies demonstrated that striatal enriched tyrosine phosphatases are expressed throughout the brain including the hippocampus (31). Taken together, these findings led us to hypothesize that ethanol may inhibit the NMDA receptor by reducing the tyrosine phosphorylation of NMDA receptor subunits.
Utilizing the rat hippocampal slice preparation, electrophysiology, immunoprecipitation, and Western blotting, we examined the role of tyrosine phosphorylation in mediating the inhibitory effect of ethanol on NMDA receptors in area CA1 of the hippocampus. Our results indicate that reduced tyrosine phosphorylation of NMDAR subunits correlates with ethanol inhibition of NMDAR fEPSPs. Furthermore, pretreatment with a tyrosine phosphatase inhibitor significantly reduces ethanol inhibition of NMDAR activity.
Hippocampal Minislice Preparation and Ethanol Treatment-Hippocampal CA1 minislices were prepared with a slight modification of a previously described procedure (5,35). Male Sprague-Dawley rats (6 -9 weeks) were used for all experiments. The experimental protocol was approved by the University of Colorado Health Sciences Center Institutional Animal Use and Care Committee. After sacrifice, the brain was quickly removed and placed in ice-cold oxygenated artificial cerebrospinal fluid (aCSF: 124 mM NaCl, 4 mM KCl, 1 mM MgSO 4 , 2.5 mM CaCl 2 , 10 mM dextrose, 1 mM KH 2 PO 4 , 25.7 mM NaHCO 3 ). Both hippocampi were dissected out and unrolled along the hippocampal fissure. Two cuts were then made to isolate area CA1 of each hippocampus before preparing 400-m slices on a McIlwain tissue chopper. Slices were then incubated at interface in an aCSF-filled chamber humidified with 95% O 2 , 5% CO 2 at 32°C. Because sacrifice can alter the basal state of the tissue, such as the polarization of neurons and the phosphorylation state of proteins, the slices were allowed to recover under these conditions for 90 min, with fresh aCSF being added every 20 -30 min. aCSF was then drained from the incubation chamber, and aCSF containing drug or aCSF alone was added. Slices were harvested at the times indicated; sonicated in buffer containing 10 mM Tris, 1 mM EDTA, and 1% SDS; and frozen until analysis. Protein concentrations were determined using the BCA protein assay kit from Pierce with bovine serum albumin (BSA) as a standard.
Immunoprecipitation-Immunoprecipitations were performed using Pansorbin cells (Calbiochem) as previously described (33,36). Immunoprecipitation under these conditions was previously shown to isolate 90 Ϯ 1% of NR2A and 93 Ϯ 1% NR2B subunits as determined by Western blotting of pellets and final supernatants (33). Previous studies also demonstrated that the antibodies to NR2A and NR2B do not cross-react in immunoprecipitation or Western blotting experiments (32,33). Control experiments were conducted in which the fusion protein used to generate the NR2A and NR2B antibodies was absorbed with the respective antibody for 1 h prior to immunoprecipitation. Preabsorbtion of the antibody with its respective fusion protein completely blocked the immunoprecipitation (data not shown).
Western Blotting-Samples from immunoprecipitation experiments and CA1 homogenates were prepared for SDS-polyacrylamide gel electrophoresis on 7.5% gels. Separated proteins were transferred to Poly-Screen polyvinylidene difluoride membrane (PerkinElmer Life Sciences). Blots were blocked in 3% BSA for 1 h at room temperature. Blots were incubated overnight with general anti-phosphotyrosine antibody (1:5000) or the site-specific anti-phospho-Tyr-1472 (1:500) in 1% BSA at 4°C. Alternatively, blots were blocked in 3% BSA overnight at 4°C and then incubated in primary antibody for 2 h at room temperature. Blots were then subjected to three 10-min washes in Tris-buffered saline (140 mM NaCl, 20 mM Tris, pH 7.6) plus 0.1% Tween 20 (TBST) before incubating with horseradish peroxidase-conjugated goat anti-mouse (general anti-phosphotyrosine) or goat anti-rabbit (site-specific antiphospho-Tyr-1472) secondary antibody (1:10,000) in 1% BSA for 1 h at room temperature. Finally, three additional 10-min TBST washes were performed, and immunodetection was accomplished using the Pierce SuperSignal chemiluminescence kit and the Alpha Innotech imaging system. Each blot was stripped with Restore Western blot stripping buffer (Pierce) and reprobed with anti-NR2A (1:1000) or anti-NR2B (1:3000). Quantitation was performed using AlphaEase software (Alpha Innotech) and Excel (Microsoft). Standard curves were included on each Western blot, and analyzed bands were always within the linear range of detection for each antibody. Where indicated, statistical analysis was evaluated by Student's paired t test.
Electrophysiology-Slices were prepared as described above, except the hippocampus remained intact such that "whole" hippocampal slices were generated. The slices were placed in a perfusion chamber containing aCSF flowing at Ͼ3 ml/min. After a Ͼ90-min recovery, fEPSPs were recorded in area CA1. fEPSPs in CA1 were elicited by stimulation of Schaeffer collateral-commissural fibers with a bipolar nichrome-stimulating electrode. Extracellular recordings were obtained with a silver electrode contained within a drawn glass capillary containing aCSF, placed in the dendritic layer. NMDAR responses were isolated by simultaneous application of 2 M NBQX and 50 M picrotoxin for at least 60 min. Electrophysiological data were analyzed by analysis of variance with Fisher's protected least significant difference test post hoc tests to determine the p value of each measurement.

RESULTS
The effect of 100 mM ethanol on the level of tyrosine phosphorylation in hippocampal CA1 minislices was examined. We chose this particular concentration of ethanol as well as the exposure time of 10 min based on previous experiments in CA1 minislices, which demonstrated that 100 mM ethanol inhibited NMDAR fEPSP slope by ϳ23% and nearly completely (Ͼ95%) blocked LTP (6). To facilitate direct examination of the effect of ethanol on the level of tyrosine phosphorylation of NR2A and NR2B, these subunits were first immunoprecipitated from hippocampal homogenates. The immunoprecipitates were then probed with the anti-phosphotyrosine antibody (4G10) to assess the extent of tyrosine phosphorylation of the NR2A (Fig.  1a) and NR2B (Fig. 1b) proteins. The experimental design employed was identical to that used to measure the effects of ethanol on NMDAR function presented in Fig. 5. Pretreatment with the AMPAR and GABA A receptor antagonists NBQX and picrotoxin, respectively, had no effect on tyrosine phosphorylation of NR2 subunits. Subsequent ethanol treatment resulted in a significant reduction of phosphotyrosine on both NR2A (37.4%) and NR2B (36.4%) subunits (p Ͻ 0.01) (Fig. 1, c and d). Treatment with ethanol alone in the absence of NBQX and picrotoxin resulted in a similar reduction of phosphotyrosine on both NR2A and NR2B (data not shown). In order to confirm that these results were not due to changes in the levels of the NR2A or NR2B protein, each Western blot was stripped and reprobed with the respective antibody ( Fig. 1, a and b). No significant difference in NR2A or NR2B levels were observed in control versus ethanol-treated samples (Fig. 1, c and d).
It has been suggested that Fyn tyrosine kinase may be important for determining an animal's sensitivity to ethanol and that this effect may be due to direct phosphorylation of the NMDAR by Fyn (26). There are several tyrosine residues in the NR2B subunit that are phosphorylated by Fyn in vitro (34). It has been shown that phosphorylation of one of these tyrosines, residue 1472, is increased following induction of LTP in area CA1 of the hippocampus (34). Using a previously characterized antibody to Tyr-1472 (34), we next attempted to establish whether this particular tyrosine was also subject to reduced phosphorylation following ethanol treatment. The amino acids surrounding Tyr-1472 of NR2B share greater than 50% identity with the homologous region of the NR2A subunit. Of the remaining residues, several are conservative substitutions between the two subunits. Like NR2B, NR2A is a substrate of Fyn kinase (37). Therefore, we first confirmed the specificity of the Tyr-1472 antibody for NR2B in our immunoprecipitates.
Western blots of our NR2A and NR2B immunoprecipitates demonstrated that only immunoprecipitates of NR2B reacted with the Tyr(P)-1472 antibody (data not shown).
Following treatment with 100 mM ethanol for 10 min, the level of Tyr(P)-1472 in the NR2B immunoprecipitates was evaluated by Western blotting with the site-specific anti-Tyr(P) antibody. A representative Western blot (Fig. 2a) shows that ethanol treatment resulted in a significant reduction of Tyr(P)-1472 (23.5%, p Ͻ 0.01; see Fig. 2b). No difference in the total amount of NR2B reactivity was seen between treatments (Fig. 2a).
We next used PP2, a selective inhibitor of the Src family of tyrosine kinases, as a positive control to assess the relative magnitude of the dephosphorylation effect. When slices were exposed to 10 M PP2 for 10 min, tyrosine phosphorylation of NR2B subunits was reduced ϳ35% as compared with control slices (p Ͻ 0.05) (Fig. 3, a and b). After exposure to PP2 for 30 min, an even more robust effect was observed (ϳ60% reduction of tyrosine phosphorylation).
The concentration of ethanol used in these studies was chosen based on previous studies in our laboratory, which demonstrated that 100 mM ethanol reduces NMDAR fEPSP slope and completely blocks LTP in CA1 minislices (6). Because this concentration of ethanol is usually associated with severe intoxication, we examined the effect of lower doses of ethanol on tyrosine phosphorylation of NR2A and NR2B subunits. The effects of 10 min exposure to 25, 50, 75, or 100 mM ethanol on tyrosine phosphorylation of NR2A and NR2B subunits are summarized in Fig. 4. When tyrosine phosphorylation of NR2A subunits was measured, only 75 mM ethanol produced a significant effect similar in magnitude to 100 mM ethanol (Fig. 4a). Cumulative data demonstrate that 50 and 75 mM ethanol significantly reduced tyrosine phosphorylation of NR2B subunits (Fig.  4b). Overall, 100 mM ethanol produced the most robust effect on tyrosine phosphorylation of NR2A and NR2B subunits.
If ethanol affects on NMDAR function are indeed due to enhanced dephosphorylation of the receptor, we should be able to reduce ethanol's effect by blocking tyrosine phosphatase activity. To test this possibility, we examined the effect of ethanol on NMDAR activity in the presence of the tyrosine FIG. 1. 100 mM ethanol reduces tyrosine phosphorylation of immunoprecipitated NMDAR subunits. a, representative Western blot demonstrating reduced tyrosine phosphorylation of the NR2A subunit after exposure to 100 mM ethanol for 10 min. Pretreatment with NBQX and picrotoxin had no effect on the level of tyrosine phosphorylation. There was no difference in the amount of NR2A reactivity between treatments. b, representative Western blot demonstrating reduced tyrosine phosphorylation of the NR2B subunit after exposure to 100 mM ethanol for 10 min. Pretreatment with NBQX and picrotoxin had no effect on the level of tyrosine phosphorylation. There was no difference in the amount of NR2B reactivity between treatments. c, quantitation of tyrosine phosphorylation and subunit protein reactivity of NR2A immunoprecipitates. Controls were normalized to 100%. phosphatase inhibitor bpV(phen). Pharmacologically isolated NMDAR fEPSPs were recorded in area CA1 of hippocampal slices (Fig. 5a). Application of 100 M DL-2-amino-5-phosphonovaleric acid confirmed that observed responses were NMDARmediated. The slope and amplitude of the NMDAR fEPSPs was measured during a 30-min exposure to 100 mM ethanol. The graph of NMDAR fEPSP amplitude as a function of time shown in Fig. 5b demonstrates that ethanol inhibition reaches a maximum after 10 min of ethanol exposure. Pretreatment with bpV(phen) significantly reduced the effect of ethanol on NMDAR function. Cumulative data shown in Fig. 5c demonstrates that the percentage inhibition of NMDAR fEPSP slope and amplitude by ethanol alone was consistent with previous reports. Perfusion of 10 M bpV(phen), a selective phosphotyrosine phosphatase inhibitor, for 25 min before and during ethanol exposure significantly reduced the inhibitory effect of ethanol on NMDAR function (Fig. 5c). Specifically, bpV(phen) almost completely blocked the effect of ethanol on NMDAR fEPSP slope (5.7 Ϯ 3.0% inhibition, p Ͻ 0.01). However, bpV-(phen) reduced the effect of ethanol on NMDAR fEPSP amplitude by only 50%. These latter results may indicate that additional mechanisms may contribute to the full effect of ethanol on NMDAR function. There was no significant effect of bpV-(phen) alone on NMDAR function. Taken together, these data suggest that reduced tyrosine phosphorylation, possibly as a result of ethanol-stimulated tyrosine phosphatase activity (25), may account, at least in part, for the inhibitory effect of ethanol on the NMDA receptor in area CA1 of the hippocampus. DISCUSSION A number of previous studies have demonstrated that pharmacologically relevant concentrations of ethanol inhibit NMDAR function in the hippocampus (6,22,38). Conversely, tyrosine phosphorylation of the NMDAR has been shown by a number of laboratories to produce potentiation of receptor function (39,40). We and others have hypothesized, therefore, that ethanol may inhibit NMDARs by inhibiting tyrosine phosphorylation of the NMDAR. However, there has been no direct evidence that ethanol can produce such an effect. We tested the effect of several doses of ethanol on tyrosine phosphorylation of NR2A and NR2B subunits in area CA1 of rat hippocampus. We demonstrated that ethanol doses as low as 75 mM produce a significant reduction in tyrosine phosphorylation of both NR2A and NR2B. Given that tyrosine phosphorylation enhances NMDAR function, this reduction in the tyrosine phosphorylation of NR2A and NR2B may be responsible, at least in part, for the inhibitory effect of ethanol on NMDA receptor function in area CA1 of the rat hippocampus.
A previous study suggested that ethanol exposure in vivo results in increased tyrosine phosphorylation of NR2B subunits assayed in vitro (26). One possible interpretation of these results is that ethanol enhances the tyrosine phosphorylation of the NMDAR. Given that ethanol inhibits NMDAR function, such an interpretation conflicts with data from a number of laboratories showing that phosphorylation enhances NMDAR function. A possible explanation of these results may be that ethanol treatment was delivered in vivo and phosphorylation FIG. 5. Tyrosine phosphatase inhibition significantly reduces ethanol inhibition of the NMDA receptor. a, pharmacological isolation of NMDAR fEPSPs. AMPAR and GABA A receptor responses were blocked by application of NBQX and picrotoxin, respectively. The resulting NMDAR fEPSP was amplified by increasing the stimulus intensity 100 -300%. Perfusion of the NMDAR antagonist DL-2-amino-5phosphonovaleric acid verified that the isolated responses were indeed NMDAR-mediated. b, data for a series of recordings where 100 mM ethanol was applied for 30 min. Each point represents the mean amplitude Ϯ S.E. of five fEPSPs measured every 30 s. Open circles, control data (n ϭ 8 slices). Closed circles, experiments where 10 M bpV(phen) was included in the perfusion solution for 25 min before and during ethanol treatment (n ϭ 11 slices). c, bar graph showing the average percentage inhibition (mean Ϯ S.E.) of NMDAR fEPSP slope and amplitude by 100 mM ethanol alone versus 100 mM ethanol following pretreatment with 10 M bpV(phen) for 25 min. Percentage inhibition was calculated based on averages of 21 consecutive sweeps taken from the base-line period before ethanol application as compared with 21 consecutive sweeps taken from 5 to 15 min after ethanol was applied. Ethanol alone significantly inhibited NMDAR fEPSP slope (19.5 Ϯ 3.2%, n ϭ 8 slices, p Ͻ 0.01) and amplitude (44.0 Ϯ 8.4%, n ϭ 8 slices, p Ͻ 0.01). Pretreatment with bpV(phen) almost completely blocked the effect of ethanol on NMDAR fEPSP slope (5.7 Ϯ 3.0% inhibition, n ϭ 11 slices) and significantly reduced the effect of ethanol on NMDAR fEPSP amplitude (21.7 Ϯ 4.6%, n ϭ 11 slices). *, p Ͻ 0.01. There was no significant effect of bpV(phen) alone. was assayed in vitro after sacrifice of the animals. Sacrifice is known to produce extremely rapid and profound alterations in phosphorylation state. It is possible that the effects of sacrifice may have obscured the effects of ethanol on tyrosine phosphorylation. To minimize sacrifice effects, the data presented in this paper were generated using an in situ hippocampal slice preparation. We assayed the effects of ethanol on the NMDAR in hippocampal slices that had recovered from the sacrifice. Following treatment, the slices were immediately sonicated in SDS to rapidly halt phosphorylation and dephosphorylation processes. Using such an assay, our data clearly show that ethanol reduces tyrosine phosphorylation of the NMDAR.
An additional issue of significant import is which tyrosine residue(s) on the NMDA receptor subunits is/are responsible for the observed change in phosphorylation. There are many tyrosine residues on the NR2 subunits that could be subject to phosphorylation and dephosphorylation. Moreover, both NR2A and NR2B subunits are known to be substrates of Fyn kinase (37,41). Several of the C-terminal tyrosines on the NR2B subunit, including residue 1472, are phosphorylated by Fyn (34). Furthermore, phosphorylation of residue 1472 is increased following induction of LTP in area CA1 of the hippocampus (34). We used a site-specific anti-phosphotyrosine antibody to assay for changes in phosphorylation of this residue after ethanol treatment. Our results show that phosphorylation of tyrosine 1472 on NR2B is reduced following treatment with 100 mM ethanol. Thus, ethanol may be exerting its effects on the NMDAR and LTP by reducing Fyn phosphorylation of the receptor. In support of this finding, basal tyrosine phosphorylation of NR2A and NR2B subunits was reduced in fyn mutant mice (41), and inhibition of recombinant NMDA receptors by 100 mM ethanol was reduced under conditions of enhanced Fyn-mediated tyrosine phosphorylation (28). Last, LTP is inhibited in mice lacking the gene for Fyn (42), and Fyn-deficient mice are more sensitive to ethanol than wild type animals (26). Taken together, these studies and the work we have presented here suggest that Fyn kinase may be a key player in the inhibition of LTP by ethanol as well as in regulating the sensitivity of NMDA receptors to ethanol in the hippocampus.
Interestingly, the ethanol-mediated reduction in phosphorylation of tyrosine 1472 is nearly as great in magnitude as the reduction of total phosphotyrosine on NR2B. Due to differences in antibody affinity and other technical constraints, we cannot draw any conclusions as to whether the effect on residue 1472 can account for the entire effect seen on total tyrosine phosphorylation of NR2B. Site-directed mutagenesis studies would be a potential method of addressing this issue.
There may be multiple mechanisms that mediate ethanol effects on the NMDARs. Certainly, one mechanism occurs very rapidly (10 -100 ms), since single channel studies of NMDARs indicated that ethanol is an allosteric modulator of channel gating (43). It has been argued by some that such an effect is too rapid to be mediated by phosphorylation effects. However, the turnover numbers for kinases and phosphatases clearly show that these enzymes can mediate events in the low millisecond time scale. Moreover, there is clear precedent for phosphorylation regulation of NMDAR channel gating (29,30). Unfortunately, it is simply not possible for us to examine biochemical changes due to alcohol in the millisecond time scale.
Studies of C-terminally truncated NR2A and NR2B subunits in human embryonic kidney 293 cells demonstrated that a site of action of ethanol is probably at a locus exposed to the extracellular environment (44,45). However, these studies also demonstrated that the C termini of NR2 subunits are important for regulating the ethanol sensitivity of NMDARs. In addition to the direct action of ethanol on NMDARs, it is likely that NMDAR response to ethanol is regulated by indirect mechanisms. Because ethanol can rapidly diffuse across the cell membrane, kinases and phosphatases are likely intracellular targets of ethanol. Indeed, ethanol has been shown to accelerate the kinetics of protein dephosphorylation by protein-tyrosine phosphatases (25). Consistent with this report, our results demonstrate that pretreatment with a tyrosine phosphatase inhibitor occludes ethanol inhibition of NMDAR activity. These data indicate that ethanol may be exerting its effect on the NMDAR by enhancing the activity of protein-tyrosine phosphatases. This model has been supported by recent data showing that activation of an endogenous protein phosphatase inhibitor, DARPP-32, regulates the ethanol sensitivity of NMDARs in the nucleus accumbens and may promote ethanol reinforcement (20). Thus, dopamine was shown to activate this phosphatase and thereby reduce the effect of ethanol on the NMDAR. In addition, dopamine failed to reduce ethanol effects on the NMDAR in DARPP-32 knockout mice. Although DARPP-32 expression appears to be largely absent in the hippocampus (46,47), these data provide support for the very provocative hypothesis that it may be possible to diminish the effects of ethanol on NMDARs by activating endogenous phosphatase inhibitors. It is important to note that DARPP-32 is a serine phosphatase inhibitor and that the focus of the aforementioned report was on serine phosphorylation of the NR1 subunit of the NMDAR.
We are also very interested in the mechanism by which reduced phosphorylation, possibly of tyrosine 1472, might be transduced into reduced NMDAR responses. Potential mechanisms downstream of a change in phosphorylation include modulation of channel properties, regulation of protein-protein interactions, and altered receptor localization. Recent studies of glutamate receptors and LTP have focused on the idea of dynamic regulation of surface expression as a mechanism of plasticity (48,49). For example, it is believed that AMPARs move rapidly between the cell's surface and interior (50 -55) very early in development (5,56,57). It has also been suggested that surface expression of NMDA receptors is a dynamic process regulated by phosphorylation events (58,59). Moreover, we have recently shown that LTP stimulation promotes membrane insertion of the NMDAR and that this effect is blocked by tyrosine kinase inhibition (5). Thus, reduced tyrosine phosphorylation may be a signal for endocytosis of NMDARs. Consistent with this hypothesis, tyrosine 1472 on NR2B is part of a consensus motif (YppØ) for AP-2 binding, a protein that mediates endocytosis via clathrin-coated pits (60). Studies in immune cells have demonstrated that T cell receptors preferentially interact with AP-2 when a tyrosine residue in the cytoplasmic domain of the receptor is nonphosphorylated (1, 61). Roche et al. (58) have shown that NMDAR internalization in dissociated hippocampal neurons is clathrin-mediated. Thus, it is possible that enhanced endocytosis of the NMDA receptor due to a decrease in phosphorylation of tyrosine 1472 could serve as a mechanism of receptor down-regulation in response to ethanol. The viability of this hypothesis is limited by a number of facts including tethering of NMDAR subunits to the postsynaptic density by scaffolding proteins such as PSD-95. However, it is possible that changes in phosphorylation also modulate NMDAR-PSD protein-protein interactions. For example, it has been reported that binding of the AMPAR-targeting protein stargazing to PSD-95 is negatively regulated by protein kinase A phosphorylation (2,3). Alternative to the hypothesis of endocytosis, NMDARs may also diffuse between synaptic and extrasynaptic sites in response to ethanol.
In summary, we have demonstrated that pharmacologically relevant doses of ethanol significantly reduce tyrosine phosphorylation of NR2A and NR2B subunits of the NMDA receptor. Specifically, we find that tyrosine phosphorylation of residue 1472 on NR2B, a known Fyn kinase site, is specifically reduced with ethanol treatment. Moreover, pretreatment with a tyrosine phosphatase inhibitor occludes ethanol inhibition of NMDAR activity. Taken together, these data suggest that reduced tyrosine phosphorylation of NR2A and NR2B subunits, possibly via activation of a tyrosine phosphatase, may underlie the inhibitory effect of ethanol on the NMDA receptor. There are still a number of questions that must be answered including the mechanism by which ethanol reduces tyrosine phosphorylation of NMDAR subunits, the time course of ethanol-mediated changes in tyrosine phosphorylation of NMDAR subunits, and how the change in phosphorylation level is transduced into inhibition of receptor function.