Alcohol inhibition of the NMDA receptor function, long-term potentiation, and fear learning requires striatal-enriched protein tyrosine phosphatase

Abstract

Alcohol's deleterious effects on memory are well known. Acute alcohol-induced memory loss is thought to occur via inhibition of NMDA receptor (NMDAR)-dependent long-term potentiation in the hippocampus. We reported previously that ethanol inhibition of NMDAR function and long-term potentiation is correlated with a reduction in the phosphorylation of Tyr¹⁴⁷² on the NR2B subunit and ethanol's inhibition of the NMDAR field excitatory postsynaptic potential was attenuated by a broad spectrum tyrosine phosphatase inhibitor. These data suggested that ethanol's inhibitory effect may involve protein tyrosine phosphatases. Here we demonstrate that the loss of striatal-enriched protein tyrosine phosphatase (STEP) renders NMDAR function, phosphorylation, and long-term potentiation, as well as fear conditioning, less sensitive to ethanol inhibition. Moreover, the ethanol inhibition was “rescued” when the active STEP protein was reintroduced into the cells. Taken together, our data suggest that STEP contributes to ethanol inhibition of NMDAR function via dephosphorylation of tyrosine sites on NR2B receptors and lend support to the hypothesis that STEP may be required for ethanol's amnesic effects.

Alcohol-induced memory deficits have been well documented in humans, as well as animal studies (1). Long-term potentiation (LTP) is widely accepted as a cellular mechanism underlying learning and memory in the hippocampus, as well as other brain regions (2), and is dependent upon NMDA receptor (NMDAR) activation (3). Acute ethanol application in adult rodents inhibits NMDAR channel activity (4, 5). Ethanol's inhibitory effect on NMDARs is widely thought to underlie both the blockade of LTP and the acute amnesic effects of ethanol. Moreover, recent work suggests that ethanol's effect on the NMDAR may play a role in the addictive nature of ethanol (6, 7).

NMDARs in the hippocampus consist of mainly NR1/NR2A, NR1/NR2B, and NR1/NR2A/NR2B receptor-subunit complexes (8). Although NR1 subunits are required to form an active ion channel, incorporation of the various NR2 subunits regulate NMDAR channel activity by altering the channel kinetics and mediating the differential effects of pharmacological agents, including alcohol (9–11). It has previously been shown that activation of members of the Src-family of kinases (SFKs) enhances NMDAR currents (12). Previous studies have demonstrated a correlation between ethanol inhibition of NMDAR function and dephosphorylation of tyrosine residues on the NR2A and NR2B subunits (13). In particular, ethanol was found to reduce the phosphorylation of a site on the NR2B subunit Tyr¹⁴⁷² that has been shown to regulate endocytosis and function of NMDARs (14, 15). Inhibition of protein tyrosine phosphatase activity using a nonspecific inhibitor, bpV(phen), significantly reduced ethanol inhibition of NMDAR field excitatory postsynaptic potentials (fEPSPs), suggesting that ethanol may inhibit NMDARs via activation of a tyrosine phosphatase (13). Ferrani-Kile et al. independently provided additional support for this hypothesis by demonstrating that ethanol decreased NMDAR tyrosine phosphorylation in the cortex (16).

Striatal-enriched protein tyrosine phosphatase (STEP₆₁, STEP) is a brain-specific protein tyrosine phosphatase that is highly expressed in the striatum as well as the hippocampus and cortex (17). STEP forms a complex with the NMDAR and regulates NMDAR activity in the hippocampus and amygdala (18, 19). Moreover, enhanced STEP activity is associated with dephosphorylation of Tyr¹⁴⁷² on NR2B, a site on which ethanol also has effects. We hypothesized that STEP may mediate the effects of ethanol on memory via decreasing NMDAR tyrosine phosphorylation in the hippocampus. We investigated this hypothesis using dominant negative-STEP and STEP-knockout mice to determine ethanol effects on the biochemistry and physiology of the NMDAR, LTP, and fear learning. Our data indicate that STEP is an important mediator of ethanol inhibition of NMDARs, LTP, and acquisition of fear learning, thereby implicating STEP in ethanol-induced memory loss.

Results

Blocking STEP Substrates Prevents Ethanol Inhibition of NMDAR Excitatory Postsynaptic Currents.

We investigated whether STEP mediates the inhibitory effect of ethanol on NMDAR excitatory postsynaptic currents (EPSCs) in Sprague-Dawley rat hippocampal CA1 pyramidal neurons by administering TAT-STEP C/S (STEP C/S) intracellularly into postsynaptic neurons via the recording electrode. STEP C/S has a point mutation in its catalytic site that renders it catalytically inactive and prevents dephosphorylation of its substrates, consequently acting as a substrate-trapping protein (19). We tested ethanol effects on NMDAR EPSCs after microinjection of varying concentrations of STEP C/S (10 nM: −22.7 ± 5.7%; 30 nM: 12.6 ± 4.7%; 100 nM: 15.5 ± 4.8%) that, with the exception of 10 nM, all prevented ethanol inhibition of NMDAR EPSCs. We, therefore, used 30 nM STEP C/S for our studies. Fig. 1A shows the time course of the effects of acute ethanol (80 mM) on isolated NMDAR EPSC in hippocampal slice recordings. Ethanol produced a significant inhibition of NMDAR currents in control neurons and in neurons pretreated with the TAT-control peptide TAT-Myc (Fig. 1C). TAT-Myc was used as a control because it was necessary to identify any potential effects of intracellular injection either TAT or the protein portion of a fusion protein on NMDAR currents. The ethanol inhibition in both control and TAT-Myc neurons was similar to that seen in previous reports (13, 20). In contrast, CA1 neurons pretreated intracellularly with 30 nM STEP C/S did not show ethanol inhibition of NMDAR EPSCs (Fig. 1C). Interestingly during ethanol washout, a significant enhancement of NMDAR currents was only observed on STEP C/S-treated neurons.

Fig. 1.

Microinjection of STEP C/S into postsynaptic neurons blocks the inhibitory effects of ethanol on NMDAR EPSCs. (A) Time course of cumulative data shows ethanol effects on NMDAR whole-cell current recordings from male Sprague-Dawley rats for control (Con), TAT-Myc (Myc), and STEPC/S (C/S)-injected neurons. (B) Representative NMDAR whole-cell current traces for control, C-Myc, and C/S-injected neurons. BSL, baseline; EtOH, ethanol; and WASH, ethanol washout period. (C) Composite data show quantification of the percent change from baseline in NMDAR currents during ethanol treatment (E) and ethanol washout (W). One-way ANOVA analyses with post hoc Tukey test show that ethanol significantly inhibited NMDAR currents in Con (P < 0.001) and in Myc-treated (P < 0.001) neurons, but it had no significant effect on C/S treated neurons (P = 0.847). During washout, there were some residual effects of ethanol in Con (P < 0.005) and no effects in Myc-treated neurons (P = 0.150), but C/S-treated neurons showed a significant increase (P < 0.001). ***P < 0.001. Number of cells: Con (n = 10); Myc (n = 5); STEPC/S (n = 5). (Scale bar, 50 ms and 50 pA.)

Ethanol Fails to Inhibit NMDAR EPSCs in STEP KO Mice.

To extend these initial findings, we next examined the effect of various concentrations of ethanol (40, 80, and 120 mM) on NMDAR EPSCs in STEP KO mice (21). These results show that 80 and 120 mM produced significant inhibition of the NMDAR currents from WT neurons, but none of the ethanol concentrations tested inhibited the NMDAR in STEP KO neurons (Fig. 2A). Both 80 and 120 mM ethanol significantly enhanced NMDAR EPSCs in the STEP KO neurons. In all subsequent work we used 80 mM ethanol. Fig. 2B shows a time course from pharmacologically isolated synaptic NMDAR EPSCs from WT and STEP KO hippocampal slices. Bath application of ethanol (80 mM) significantly inhibited synaptic NMDAR EPSCs in slice recordings from WT mice neurons, which was reversed (responses returned to baseline values) following ethanol washout (Fig. 2 B and C). In contrast, ethanol (80 mM) application on STEP KO neurons failed to inhibit NMDAR EPSCs. Rather, NMDAR EPSCs began to increase during the later part of ethanol exposure and this enhancement persisted during the early part of ethanol washout. Cumulative data in Fig. 2C (Left) show that the enhancement seen in STEP KO mice during washout was statistically significant (38.3 ± 17.87%; P < 0.01) from that seen during ethanol exposure (24.21 ± 4.61%, P < 0.02). The potentiation seen following washout of ethanol in STEP KO neurons was similar to that seen in neurons injected with STEP (C/S). It has previously been shown that activation of the tyrosine kinase Src enhances NMDAR currents (12). To explore this possibility, we used PP2 [3-(4-chlorophenyl)1-(1,1-dimethylethyl)-1H-pyrazolol[3,4-d] pyrimidin-4-amine] to inhibit the SFKs activity and then examined the effects of ethanol on WT and STEP KO neurons. As shown in Fig. 2C (Right), the effect of ethanol and washout are not affected by PP2 in WT neurons. However, in STEP KO neurons, PP2 treatment blocks the enhancement of NMDAR currents seen during ethanol washout. We also used an inactive PP3 (2 μM) to control for the nonspecific effects of PP2; PP3 (4-amino-7-phenylpyrazol[3,4] pyrimidine) did not alter the effects of ethanol on NMDAR EPSCs in either WT or STEP KO neurons, and it did not block ethanol enhancement of NMDAR EPSCs in STEP KO neurons (11.9 ± 2.0% enhancement). Acute ethanol produced comparable potentiation of the GABAAR inhibitory postsynaptic currents in both WT (29.8 ± 5.2% enhancement) and STEP KO neurons (35.6 ± 1.8% enhancement) (t = 1.055, P > 0.320), suggesting that STEP does not play a key role in ethanol's effects on the GABAAR (Fig. S1).

Fig. 2.

Ethanol fails to inhibit NMDAR EPSCs but reintroduction of STEP can restore ethanol's inhibition in STEP KO mice. (A) Ethanol (40, 80, and 120 mM) show differential effects on NMDAR EPSCs. Among WT mice, one-way ANOVA indicated that ethanol dose-dependently inhibited NMDAR currents [F(3,21) = 24.976, P < 0.001]. Post hoc Tukey test analyses showed that 80 and 120 mM ethanol significantly affected NMDAR responses (P < 0.001). In KO mice, ethanol significantly enhanced NMDAR currents [F(3,28) = 3.017, P < 0.05]. Post hoc Tukey test analyses showed that 80 and 120 mM ethanol significantly increased NMDAR currents (P < 0.05). (B) Time course of ethanol (80 mM) effects on NMDAR EPSC amplitudes in WT (n = 8) and KO (n = 6) animals. Representative traces show whole-cell currents of a hippocampal CA1 neuron from a WT and KO mice. Dashes indicate baselines. (C) (Left) Composite data show that ethanol application (WT, E, n = 8) inhibited the NMDAR EPSCs in slices from WT mice but had no effect during the washout (WT, W, n = 8). In contrast, ethanol potentiated the NMDAR EPSCs both during ethanol application (KO, E, n = 10) and washout (KO, W, n = 8). (Right) Composite data show that, in neurons from WT mice, PP2 had no effect either during ethanol application (WT-PP2, E, n = 13) or during washout (WT-PP2, W, n = 10). However, in neurons from KO mice, PP2 blocked the NMDAR enhancement seen with ethanol application (KO-PP2, E, n = 10) and during washout (KO-PP2, W, n = 9). (D) Intracellular microinjection of the control peptide, TAT-Myc, (WT, Con, C, Clear bars) and TAT-wtSTEP (STEP, cross-hatched bar) do not alter ethanol effects on NMDAR EPSCs in neurons from WT mice. Intracellular microinjection of TAT-Myc (KO, Con, C, striped bar) does not alter the enhancing effects of ethanol on NMDAR EPSCs, but microinjection of TAT-wtSTEP (KO, STEP, cross-hatched bar) restores ethanol inhibition of NMDAR EPSCs. TAT-wtSTEP has no effect in WT neurons [Con (C), n = 8]; [wtSTEP (STEP), n = 9]. TAT-STEP restores ethanol inhibition in STEP KO neurons [Con (C), n = 6]; [wtSTEP (STEP), n = 5]. One-way ANOVA with a post hoc Tukey test analysis determined significance, *P < 0.05, **P < 0.01, ***P < 0.001. (Scale bar, 150 ms and 150 pA.)

Genetic alteration of mice can result in unintended compensatory mechanisms (22), which can convolute interpretation of results. We wanted to determine whether the absence of an ethanol effect was specifically because of lack of STEP activity, and not developmental anomaly. To this end we introduced a STEP TAT-fusion protein (TAT-wtSTEP; 30 nM) into CA1 pyramidal neurons from WT and STEP KO hippocampal slices. Ethanol (80 mM) inhibited NMDAR EPSCs in WT neurons by −25.2 ± 5.70% and the addition of TAT-wtSTEP did not significantly alter this effect (−30.2 ± 3.9%) (Fig. 2D). However, in neurons from STEP KO mice, intracellular administration of TAT-wtSTEP rescued the inhibitory effect of ethanol on NMDAR EPSCs (−31.1 ± 4.38%) (Fig. 2D). These results indicate that STEP is required for ethanol inhibition of NMDAR EPSCs.

Ethanol Does Not Inhibit Tetanus-Induced LTP and Fear Learning in STEP KO Mice.

A number of previous studies have demonstrated that ethanol blocks induction of NMDAR-dependent LTP (23–26). We predicted that because STEP was necessary for ethanol inhibition of NMDAR function, ethanol effects on LTP would also be abolished in STEP KO mice. To test this hypothesis, we measured high-frequency stimulation (HFS)-induced LTP in the CA1 stratum radiatum region of hippocampal slices obtained from WT and STEP KO mice. In slice recordings from WT mice, HFS elicited robust LTP, as measured by the slope of the fEPSP (Fig. 3C), and pretreatment with ethanol (80 mM) before HFS blocked the induction of LTP (Fig. 3 A and C). In contrast, ethanol was unable to block the induction of LTP in slices from STEP KO mice (Fig. 3 B and C). These data suggest that STEP is required for ethanol inhibition of LTP.

Fig. 3.

Ethanol fails to inhibit LTP induction and fear learning in STEP KO mice. Cumulative time course shows ethanol significantly blocks LTP slope in WT (A) but not in STEP KO (B) hippocampal slices. Control responses (○) and ethanol- (●) (80 mM) treated responses. (C) Composite data show ethanol inhibits LTP slope from WT slices, control (Con, C, n = 11); ethanol (E, n = 5) and not in STEP KO slices (Con, C, n = 10); EtOH (E, n = 4). Solid bars: WT; striped Bars: KO. In addition, the amplitude of LTP did not differ between WT and STEP KO slices (t = 1.158, P > 0.261). (D) Quantification of freezing behavior during a 3-min interval testing cued fear conditioning. WT animals froze 80.6 ± 3.8% of the time with saline and 43.1 ± 8.6% with ethanol (P < 0.001). STEP KO animals froze 69.1 ± 5.3% and froze 56.2 ± 7.8% with saline and ethanol, respectively (P > 0.19). WT (n = 8); KO (n = 9). (E) Quantification of freezing behavior during a 5-min testing interval for contextual learning. WT treated with saline froze 50.4 ± 7.4% of the time- and ethanol-treated froze 15.8 ± 5.3% of the time (P < 0.002). Saline-treated STEP KO mice froze 38.9 ± 8.7%, and ethanol-treated STEP KO animals froze 22.2 ± 4.9% (P > 0.10). (F) Composite data for BEC (mg%) for WT and STEP KO animals at 15 and 60 min post ethanol administration. Two-way ANOVA determined significant time-dependent decrease in BEC levels but no genotypic differences of BEC. Solid bars, WT animals; striped bars, KO animals; S, saline (white); E, 1.5 g/kg ethanol (gray). *P < 0.05, **P < 0.002.

To further determine ethanol's effects on memory of STEP KO mice, we studied the effects of ethanol on fear learning. Freezing behavior was defined as the absence of any movement except for respiration, and is expressed as the percent time the mouse was frozen. During the cued-conditioning testing period, WT mice treated with ethanol showed a significant reduction in freezing (Fig. 3D) (P < 0.001). In contrast, ethanol treatment had no significant effect on freezing in response to the tone in the STEP KO mice (Fig. 3D) (P > 0.19). The results of the contextual-conditioning test were similar (Fig. 3E). Saline-injected WT mice showed a robust freezing response upon introduction into the chamber, which was significantly reduced in the ethanol treated WT mice (P < 0.002) (Fig. 3E). In contrast, ethanol caused no significant reduction in freezing in the STEP KO mice (P > 0.11). We also determined the effects of novelty and within-session memory extinction on cued conditioning in WT and KO mice. Our data indicated that these factors did not contribute significantly to the difference of freezing in these mice (Fig. S2). In addition, we investigated the anxiolytic and sedative/hypnotic effects of ethanol on these mice: the results showed that ethanol did not produce differential anxiolytic or sedative/hypnotic effects (Fig. S3). Therefore, these data suggest that STEP is required for ethanol attenuation of fear-conditioned learning. To see whether WT and STEP KO animals might differ in ethanol metabolism, we measured blood ethanol concentrations (BEC) 15 and 60 min after ethanol administration to these animals. A two-way ANOVA determined a significant time-dependent effect [F(1,24) = 43.2; P < 0.0001], but BEC levels were not significantly different between genotypes at either time point (Fig. 3F).

STEP Is Necessary for Ethanol Effects on Tyrosine Phosphorylation of NR2B.

Ethanol inhibition of the NMDAR is associated with a reduction in the phosphorylation of Tyr¹⁴⁷² in the NR2B subunit of the NMDAR (13), a site that has been shown to regulate endocytosis and function of NMDARs (14, 15). We investigated whether dephosphorylation of NMDARs still occurs in STEP KO mice following ethanol treatment. In agreement with previous studies in hippocampal slices from rats (13, 20), ethanol produced a significant decrease in overall NR2B tyrosine phosphorylation in hippocampal slices from WT mice (Fig. 4A) (−25.98 ± 12.14%; P < 0.01). In contrast, ethanol failed to decrease the tyrosine phosphorylation of the NR2B subunit in STEP KO slices. Western blot analysis did not detect any significant ethanol-induced decrease in tyrosine phosphorylation on the NR2A subunit (Fig. 4B). Next, we examined ethanol's effects on specific phosphorylation sites of the NR2B subunit in the STEP KO and WT mice. We assayed immunoprecipitated NR2B from hippocampal homogenates with phospho-antibodies specific for Tyr¹⁴⁷², Tyr¹³³⁶, and Tyr¹²⁵². In WT slices, Western blot analysis revealed a significant decrease in the phosphorylation of Tyr¹⁴⁷² (Fig. 4C) (−15.54 ± 8.48%; P < 0.03). In contrast, ethanol had no significant effect on Tyr¹⁴⁷² phosphorylation in STEP KO brain slices (−6.50 ± 9.38%; P > 0.100). Neither the Tyr¹³³⁶ nor Tyr¹²⁵² was altered by ethanol. In addition, we measured the protein levels of NR2B and found that no protein level differences between WT and STEP KO brain slices (Fig. 4D). Therefore, these data indicate that in the absence of STEP, ethanol does not decrease phosphorylation of NR2B Tyr¹⁴⁷² or total NR2B tyrosine phosphorylation.

Fig. 4.

Ethanol treatment does not reduce NR2B tyrosine phosphorylation in STEP KO mice. (A) Representative Western blot and quantification of decreased tyrosine phosphorylation on immunoprecipitated NR2B subunits after 10-min 80 mM ethanol treatment. Tyr (P): WT (n = 13); KO (n = 10). (B) Tyrosine phosphorylation on NR2A subunits after ethanol treatment. No significant change on NR2A tyrosine phosphorylation was detected in either WT or STEP KO mice Tyr (P): WT (n = 15); KO (n = 11). (C) Significant decrease in Tyr¹⁴⁷² phosphorylation of NR2B subunits after 80-mM ethanol treatment in slices from WT (n = 13) (P < 0.05), but not KO (n = 10) mice. (D) NR2B from immunoprecipitated hippocampal homogenates WT (n = 15); KO (n = 12). *P < 0.05, **P < 0.01. Solid bars, WT; striped bars, KO; C, control aCSF-treated minislices; E, ethanol-treated minislices.

Discussion

Two theories have been put forth to explain ethanol inhibitory effects on NMDARs. Some groups have proposed that ethanol directly interacts with the receptor, possibly via a binding pocket for ethanol within the transmembrane domain (27). An alternate model proposes that ethanol affects NMDARs by modulating the phosphorylation of the receptor (13, 16, 28–30). Importantly, these two models are not necessarily mutually exclusive, as it is possible for phosphorylation to influence an ethanol-binding pocket. Our research has focused on ethanol's effects on NMDAR phosphorylation after we and another group independently observed that ethanol decreased tyrosine phosphorylation of the NMDAR in both the hippocampus and cortex (13, 16). Moreover, we previously demonstrated that a broad spectrum tyrosine phosphatase inhibitor bpV(phen) significantly reduced ethanol effects on the NMDAR, suggesting that ethanol may inhibit the NMDARs by promoting tyrosine dephosphorylation of the receptor. In the present report, we provide evidence that the tyrosine phosphatase STEP plays a key role in mediating ethanol's inhibitory effects on NMDAR function and phosphorylation, as well as the inhibitory effects of ethanol on LTP and fear conditioning.

Our study focused on STEP because STEP and NMDARs have been shown to be associated in a complex. Moreover, STEP reduces NMDAR activity, negatively regulates LTP, and is required for internalization of both NMDARs and AMPARs from neuronal surfaces (31, 32). Based on these results, we hypothesized that STEP may mediate the inhibitory effects of ethanol on NMDAR activity in the brain. To address the role of STEP in ethanol action, we first examined the effects of ethanol on NMDAR function after microinjections of the substrate-trapping STEP (C/S) peptide (19) into CA1 neurons in rat hippocampal slices. STEP C/S completely attenuated the effects of ethanol on NMDAR EPSCs, suggesting STEP activity was required for ethanol inhibition of NMDAR activity. STEP C/S binds to several STEP substrates and some of these substrates may be used by other cell-signaling pathways to regulate NMDAR activity (19, 21). Therefore, we next tested the effects of ethanol on hippocampal synaptic NMDAR EPSCs in WT and STEP KO mice. In these experiments, synaptically evoked NMDAR EPSCs were resistant to the inhibitory effects of ethanol in STEP KO mice, providing additional evidence that STEP is necessary for ethanol's inhibition of NMDAR EPSCs.

An important consideration in studies with KO animals is that compensatory mechanisms may contribute to the effects seen. For example, previous studies in mice created with other gene deletions report that compensatory mechanisms develop and contribute to the behavioral effects of ethanol as a means of homeostasis (22). To address this issue, we attempted a rescue experiment by microinjecting wtSTEP into CA1 neurons from STEP KO slices and found that ethanol inhibition of NMDAR EPSCs was restored. These findings strongly support the hypothesis that STEP is directly involved in mediating the ethanol inhibition of NMDAR EPSCs. This role of STEP in ethanol action seems relatively specific for the NMDAR, as we saw no alteration in ethanol's effects on the GABAAR in STEP KO mice.

It is important to discuss the fact that although ethanol did not inhibit the NMDAR in STEP KO animals, it actually produced a potentiation of NMDAR currents during both the ethanol application and drug washout period. A similar increase of NMDAR currents after ethanol treatment had been previously observed in the hippocampus (33) and in dorsal striatum (34). It has been reported that STEP inactivates Fyn (35), a SFK that phosphorylates tyrosine residues on the NMDAR and enhances its function (12). Therefore, it seemed plausible that ethanol activation of Fyn is unopposed by STEP in hippocampal neurons of STEP KO mice (and also possibly in the dorsal striatum), leading to activation of NMDARs. We tested this possibility by blocking SFK activity before ethanol treatment using the SFK inhibitor PP2. Our data demonstrate that PP2 blocks or prevents the ethanol-induced increase in NMDAR EPSCs during both ethanol treatment and the drug washout period. All experiments with kinase or phosphatase inhibitors are fraught with difficulties in interpretation. Nevertheless, these data suggest the possibility that ethanol's potentiating effects on the NMDAR might be regulated by the SFK Fyn, as has been suggested previously (28, 33). However, the potentiating effect of ethanol that we have seen in the STEP KO animals is not likely to be caused by an increase in NMDAR phosphorylation, as ethanol produces no increase, but rather a modest decrease, in NMDAR phosphorylation in STEP KO animals.

To further examine the role of STEP in ethanol's inhibitory action, we measured the effect of ethanol on tyrosine phosphorylation of the NMDAR in the STEP KO mice, because the NMDAR activity is regulated by protein phosphorylation (36). Recent work has shown that the functional expression of NMDARs is regulated by phosphorylation of NR2B Tyr¹⁴⁷². Activation of STEP has been shown to dephosphorylate this site (31), and we previously showed that ethanol treatment also leads to dephosphorylation of Tyr¹⁴⁷² in rat hippocampal slices (13, 20). Therefore, we tested whether ethanol reduced the tyrosine phosphorylation of the NMDARs in WT and STEP KO mice. Our data show that although ethanol reduced phosphorylation of Tyr¹⁴⁷² in WT animals, it did not significantly alter phosphorylation of Tyr¹⁴⁷² in STEP KO animals. These data suggest that STEP is required for ethanol to stimulate the dephosphorylation of NR2B Tyr¹⁴⁷². In addition, these data suggest that ethanol may inhibit the NMDAR by activating STEP to dephosphorylate the NMDAR. This model is supported by previous work, which demonstrated that ethanol accelerates the catalytic activity of tyrosine phosphatases in vitro (37). However, we have been unable to demonstrate any direct effect of ethanol on purified recombinant STEP using in vitro assays.

Given that ethanol had no inhibitory effect on the function and phosphorylation of NMDAR in STEP KO slices, we next investigated the effects of ethanol on NMDAR-dependent LTP induction in STEP KO and WT mice. Our results show that STEP KO animals also appear resistant to the inhibitory effects of ethanol on LTP induction. Taken together, our data support a model in which ethanol inhibition of the NMDAR and of LTP requires STEP. Given that our data suggest that STEP plays a key role in ethanol-induced inhibition of NMDAR function and promoted NR2B dephosphorylation, as well as inhibiting LTP, we also examined the potential role of STEP in mediating ethanol inhibition of fear-conditioned learning. Fear conditioning is a behavioral learning task that measures freezing, a natural defense mechanism in mice (38). Importantly, NMDAR antagonists block fear conditioning, suggesting that NMDAR activation is necessary for fear conditioning (39). Previous studies demonstrated that ethanol, administered just before training, significantly impairs both contextual and cued fear conditioning (40, 41). Our study supports these previous findings, demonstrating that ethanol significantly reduces freezing responses during both the cued and contextual fear-conditioning testing in WT mice. However, in STEP KO animals, ethanol failed to significantly inhibit fear conditioning. There are a number of possible explanations for this effect of the STEP KO on ethanol blockade of fear conditioning. However, we found no difference in anxiolytic and sedative effects of ethanol, in responses to novel environment, and in the extinction of fear memory between WT and STEP KO animals (Figs. S2 and S3). Our data suggest the key role played by STEP in ethanol's inhibitory action on the function and phosphorylation of the NMDAR and on LTP. Therefore, we favor the hypothesis that ethanol fails to block fear conditioning in STEP KO mice because ethanol requires STEP to inhibit the NMDAR and LTP, both of which underpin fear-conditioned learning. However, it is important to also note that ethanol potentiates the NMDAR responses in the STEP KO animals and thus the failure of ethanol to block learning may reflect a consequence of this effect rather than simply the loss of the STEP molecule.

Cued and contextual testing conditions have been reported to be largely mediated by two distinct brain regions, the amygdala and the hippocampus, respectively (42, 43), and also by a compensatory fear neurocircuitry involving the stria terminalis (44). NMDAR-dependent LTP has been demonstrated in the amygdala and has been reported to be necessary for cued fear conditioning (38). Therefore, given our data, one intriguing hypothesis is that STEP also plays a role mediating ethanol's inhibitory effects on NMDAR phosphorylation, function, and LTP in the amygdala and perhaps other brain areas. It is also important to keep in mind that there is strong evidence that ethanol may have direct effects on the NMDAR (45). Moreover, it has been shown that ethanol effects on the NMDAR can be very rapid and may alter NMDAR channel kinetics (46). Thus, if STEP is involved in such effects, it may not be via dephosphorylation of sites involved in trafficking but rather in gating of the NMDAR. Lastly, it has also been reported (47) that ethanol effects do not require the C terminus of the NMDAR in CHO-K1 cells. Future studies are required to reconcile this result with our work on STEP in hippocampal slices. In sum then, it is likely that several processes are working together to mediate the underlying acute inhibition by ethanol.

In addition to its well-known role in memory, LTP and its molecular underpinnings have also come under investigation as a cellular mechanism underlying addiction (6, 7). Ethanol's effects on NMDARs have also been implicated in ethanol tolerance and dependence (6). NMDAR antagonists have been shown to block tolerance to ethanol (48) and decrease ethanol intake in operant free-choice, self-administration paradigms, as well as after periods of abstinence (49). Given STEP's ability to modulate NMDARs, LTP, and ethanol effects, this tyrosine phosphatase may serve as a key target for research and development of pharmaceutical interventions for alcoholism and addiction.

Materials and Methods

Animals.

All procedures were approved by the Institutional Animal Care and Use Committee of the University of Colorado, Denver. The rats and mice were cared for in an accredited facility where the animals were kept in 12-light/12-dark cycle and were on food and water ad libitum. Six- to 9-wk-old Sprague-Dawley rats or STEP WT and KO mice were used in this study. STEP KO mice were generated by classic homologous recombination and back-crossed for 7 to 12 generations onto C57/B6 mice (21).

Hippocampal Slice Recordings.

After killing, the brains were rapidly removed and immersed in ice-cold, sucrose containing cutting buffer (in mM: 87 NaCl, 2.5, KCl, 7 MgCl₂, 0.5 CaCl₂, 1.25 NaH₂PO₄, 25 d-glucose, 35 sucrose, and 25 NaHCO₃) for 40 to 60 s to cool the interior of the brain. Transverse slices (400-μm thickness) were made using a tissue chopper/slicer (TC-2 Tissue Sectioner; Sorvall) and the slices were transferred to individual compartments in a storage system for recovery (50). For whole-cell electrophysiological recording of NMDAR EPSCs, a single slice was transferred to a recording chamber and superfused with artificial cerebrospinal fluid (aCSF) at a bulk flow rate of 2.0 mL/min at 33.5 °C. The aCSF contained the following (in mM): 126 NaCl, 3.0 KCl, 1.5 MgCl₂, 2.4 CaCl₂, 1.2 NaH₂PO₄, 11 d-glucose, and 25.9 NaHCO₃. The whole-cell microelectrodes have resistances of 6 to 9 MΩ when filled with a K⁺-gluconate internal solution, which contained (in mM): 130 K⁺-gluconate, 1 EGTA, 2 MgCl₂, 0.5 CaCl₂, 2.54 disodium-ATP, and 10 Hepes; adjusted to pH 7.3 with KOH and to 280 mOsm. CA1 pyramidal neurons were voltage-clamped at −60 mV (corrected for the liquid-junction potential) from the normal resting membrane potential (−65 to −70 mV). The NMDAR EPSCs were isolated pharmacologically using CNQX (20 μM), bicuculline methiodide (30 μM), and CGP-52432 (1.0 μM) to block AMPA, GABA_A, and GABA_B receptor-mediated currents, respectively. NMDAR EPSC responses were evoked at proximal positions (i.e., a depolarizing pulse that produced a depolarizing current of 1–3 nA for a duration of 10–15 ms). The stimulus intensity was sufficient to trigger an NMDAR current of ∼100 to 160 pA (average 121.6 ± 0.6 pA, n = 101). Acute drug treatments were measured as the average of the 10-min period starting 2 min after the drug application until 2 min after start of drug washout. The average of the amplitude of responses was then compared with that of the baseline period. In some experiments, hippocampal slices were incubated with PP2 or PP3 (2 μM) for 3 h before recordings.

Measurement of LTP.

Synaptic fEPSP responses were evoked with bipolar tungsten electrodes placed in the CA1 pyramidal cell layer. Test stimuli were delivered once every 30 s with the stimulus intensity set to 40 to 50% of the maximum synaptic response. Tetanic stimulation consisted of two trains of 100-Hz stimuli lasting 1 s each, with an intertrain interval of 15 s, at the control stimulus intensity. Field EPSP recordings were made with a glass micropipette filled with aCSF and placed in the stratum radiatum ∼200 to 300 μm from the cell body layer. This stimulation produced a LTP that persisted for more than 30 min. The slopes of fEPSPs were calculated as the slope measured between 10 and 30% from the origin of the initial negative deflection. Each time point is shown as an average of five 30-s interval measurements. The 10 min of LTP data (25–35 min after stimulation) were used in analyses as reported by Schummers et al. (26).

Behavioral Procedures.

Fear conditioning was performed as described in Wehner et al. (51) and Gould (41). The procedure took place over 2 d and consisted of a training session on the first day and a two-phase testing session on the second day. Fifteen minutes before the start of training, mice were administered a single dose of saline (0.01 mL/g, i.p.) or ethanol (1.5 g/kg, i.p.; 15% vol/vol in saline). Mice were trained in a conditioning chamber (26-cm width × 20-cm length × 30-cm height) with a grid floor composed of stainless steel rods (3.2 mm) placed on 7.9-mm centers (Med Associates). The chamber was housed inside a sound-attenuating cubicle, which was ventilated with a built-in fan and illuminated by white light. Following a 2-min exploratory period, the mouse was presented with a conditioning stimulus (CS), which in this case was a broadband clicker (80 dB) that sounds for 30 s. At the culmination of the clicker period, the mouse received a mild foot shock (0.5 mA, 2 s) through the grid floor. This cycle was repeated for a total time in the chamber of 5.5 min (the mouse remained in the chamber for 30 s after the second foot shock). FreezeFrame software was used to control the apparatus (ActiMetrics Software). The mice were tested 24 h later in two phases. The first phase tests the mouse for its ability to remember the context in which it had previously received a foot shock; this is known as contextual conditioning and is considered to be a simple test of spatial learning. The test consisted of returning the mouse for a total of 5 min to the same chamber to which it was trained. During this time, an observer scored the mouse for freezing behavior, defined as the absence of any movement except for respiration, using a previously described sampling method (52). No foot shock was administered during the test. Following the context test, the mouse was returned to its home cage. One hour later, the mouse was tested for its conditioning to the CS by being placed into a distinct test chamber, which consisted of a circular clear Plexiglas compartment (30 cm diameter) with a smooth floor. After 3 min of exploration, the mouse was presented with the CS for an additional 3 min; an observer scored the mouse for freezing behavior throughout the entire 6-min period. For analysis of BEC, two blood samples were drawn from WT and STEP KO animals at 15 and 60 min after a single dose of ethanol (1.5 g/kg, i.p.; 15% vol/vol in saline) was administered. BECs were determined using an enzymatic assay (53).

Immunoprecipitation and Semiquantitative Western Blot Analysis.

The hippocampal minislices (54) were prepared, transferred to an oxygenated incubation chamber at 32 °C for a 90 min recovery. Following 10 min of ethanol exposure, slices were then harvested and homogenized in a STE buffer (1% SDS, 10 mM Tris, and 1 mM EDTA), boiled for 5 min, and immediately stored in the −80 °C until use. Protein concentrations were determined using BCA protein assay kit (Pierce). Immunoprecipitation and immunoblotting were performed as described in Goebel-Goody et al. (55). A five-point standard curve of a hippocampal homogenate was included on each gel to establish the linear response range for each blot. All of the immunoreactivity data were normalized to an independent control on each blot and expressed as percent of control to reduce blot-to-blot variations. Immunodetection of the blot was performed using the Pierce SuperSignal Chemiluminescence kits and the Alpha Innotech ChemiImager 4400 imaging system. Immunoblotting was performed using anti–NR2B-Tyr¹⁴⁷², NR2B-Tyr¹³³⁶, NR2B-Tyr¹²⁵², NR2B, NR2A, and NR1 antibodies obtained from PhosphoSolutions and the 4G10 general antiphospho-tyrosine antibody from Upstate/Millipore.

Data Analysis and Statistics.

All data were expressed as the mean ± SEM. BEC data were expressed as milligram alcohol per deciliter blood (mg %). Differences between two groups were determined by unpaired t tests, and differences between strain and ethanol treatment were determined by two-way ANOVA with post hoc Tukey test pairwise comparisons using SigmaStat software (Systat Software Inc.).