Aberrant early-phase ERK inactivation impedes neuronal function in fragile X syndrome

Abstract

Fragile X syndrome (FXS) has so far resisted efforts to define the basic cellular defects caused by the absence of a single protein, fragile X mental retardation protein (FMRP), because the patients have a wide variety of symptoms of varying severity. Immature-appearing dendritic spines on neurons found in FXS patients and fmr1-KO mice suggest a role for FMRP in modulating production of synaptic structural proteins. We isolated cortical synaptoneurosomes from WT and KO mice and studied MAPK pathway activation after group I metabotropic glutamate receptor (mGluR) stimulation. Here, we show that ERK in KO synaptoneurosomes is rapidly dephosphorylated upon mGluR1/5 stimulation, whereas it is phosphorylated in WT mice, suggesting that aberrant activation of phosphatases occurs in KO synapses in response to synaptic stimulation. In KO synapses, protein phosphatase 2A (PP2A) is overactivated after mGluR1 stimulation, and tyrosine phosphatase is overactivated after mGluR5 stimulation, causing the rapid deactivation of ERK. ERK activation can be restored in KO by pretreatment with phosphatase blockers; blocking of PP2A by okadaic acid could successfully restore normal ERK activation in KO synaptoneurosomes. We propose that overactivation of phosphatases in synapses may be a key deficit in FXS, which affects synaptic translation, transcription, and synaptic receptor regulation.

Abnormal numbers and/or shapes of dendritic spines on neurons are associated with fragile X syndrome (FXS) and other neurological and psychiatric diseases such as schizophrenia, mood disorder, Alzheimer's dementia, autism, and mental retardation (1). We previously reported that the brains of FXS patients and the fmr-1 KO mouse model have immature-appearing, long and thin dendritic spines and higher dendritic spine density (2), suggesting that production of synaptic structural proteins is not appropriately modulated by neuronal activity. Neurotransmitter-triggered synaptic protein synthesis is regulated by a number of signaling pathways. We have chosen to study phosphorylation of the extracellular-signal-regulated kinase ERK because it is a point of convergence of several signaling cascades; thus misregulated ERK activation can be used as a first indicator for deficient translational regulation in synapses. Blocking of overactivated upstream factors to restore normal ERK activation might be a focus for development of treatments and understanding of FXS.

FXS is an inherited, X-linked disorder, caused by hypermethylation of an extreme expansion (200–1,000 repeats) of a (CGG)n trinucleotide repeat in the 5′ UTR of the FMR1 gene, blocking transcription of the gene. The resulting absence of the fragile X mental retardation protein (FMRP) leads to a wide variety of symptoms, including mental retardation, macroorchidism, and behavioral abnormalities. Mental retardation and behavioral abnormalities may be caused principally by disrupted synaptic transmission as a result of a lack of FMRP's role in maintaining synaptic structure and function. Mice lacking FMRP exhibit generalized dysregulation of total protein translation (3–5), but local (synaptic) translation of specific “cargo” mRNAs is affected as well (6).

We have shown that the rapid initiation of synaptic protein translation, triggered by metabotropic receptor stimulation, is defective in mice lacking a functional fmr-1 gene (7). Synaptic translation of specific mRNAs for AMPA receptor and PSD-95 was also seen to be defective in the absence of FMRP (8). A translational response follows neurotransmitter-initiated phosphorylation and activation of second messengers, prominently ERK 1/2. Here, we investigate changes in the cellular signaling cascade triggered by neurotransmitter receptors. We have examined the early time course of ERK phosphorylation, using cortical synaptoneurosomes from WT and fmr-1 KO mice. We show here that rapid activation of ERK1/2 subsequent to group I metabotropic glutamate receptor (mGluR) stimulation, as seen in WT mice, is dramatically altered in fmr1-KO mice, where there is instead a long-lasting dephosphorylation of ERK after glutamate receptor agonist stimulation. Inhibition of the serine–threonine protein phosphatase 2A (PP2A) by okadaic acid, protein phosphatase 1 (PP1) and PP2A by calyculin A or tyrosine phosphatases by orthovanadate restores agonist-stimulated ERK activation in KO synaptoneurosomes, suggesting that overactivity of phosphatases may be a key deficit in FXS. Fmr-1 KO mice show, in addition, a residual phosphatase activation in the presence of okadaic acid or orthovanadate.

Results

Group I mGluR Stimulation Triggers Rapid Deactivation of ERK in fmr-1 KO Cortical Synaptoneurosomes.

To study synapse-specific second messenger activation, we used the synaptoneurosome preparation, based on brief size-selection of synaptic particles so that energy stores are preserved (9). This process makes it possible to take repeated samples from a homogeneous suspension and analyze changes in levels of substrate phosphorylation at successive time points. The two group I mGluRs, mGluR5 and mGluR1, are known to elicit somewhat different cellular responses (10–12). Therefore, we stimulated each receptor independently to test the responses of cortical synaptoneurosomes from WT and fmr-1 KO mice. Stimulation of mGluR1 [by (S)3,5-dihydroxyphenylglycine (DHPG) in the presence of the mGluR5 blocker 2-methyl-6-(phenylethynyl)pyridine hydrochloride (MPEP)] in WT led to rapid phosphorylation of ERK, peaking at ≈1 min, and persisting for ≈5 min. In KO synaptoneurosomes, in contrast, there was rapidly decreased phosphorylation of ERK1/2 within the first 1–2 min, which persisted for at least 10 min (Fig. 1A). Activated ERK needs to be deactivated at a certain point after synaptic stimulation, otherwise abnormal accumulation of pERK is toxic to cells (13). It is worth noting that this safety mechanism of shutting off ERK activation (likely by phosphatases), which was seen after its peak in WT, also functioned in KO (see the parallel decreasing curves after the time points of 2 min in Fig. 1A) even though ERK had already been rapidly dephosphorylated in the initial response to mGluR1 stimulation in KO.

Fig. 1.

ERK activity after group I mGluR stimulation in WT and fmr-1 KO cortical synaptoneurosomes. ERK activity was measured by Western blots, using an antibody to ERK1/2 phosphorylated on Thr-202/Tyr-204. (A) mGluR1 stimulation triggered ERK phosphorylation in WT but ERK was rapidly dephosphorylated in KO synaptoneurosomes (n = 7 for WT; n = 4 for KO). Cortical synaptoneurosomes were preincubated with 10 μM MPEP (mGluR5 blocker) for 15 min, then stimulated by 100 μM DHPG (group I mGluR agonist) to activate mGluR1s. (B) Stimulation of mGluR5 by specific agonist CHPG (750 μM) induced a more transient activation of ERK in WT but a rapid deactivation in KO (n = 4 each for WT and KO). Control samples were collected from an unstimulated pool at 0 min (C0′) and 10 min (C10′). Total protein level of each lane was measured from an amido black-stained membrane and was used as a loading standard. Data are presented as means ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001. Asterisks between the two curves represent P values for WT versus KO; the rest are for P values of stimulated versus baseline values.

When WT synaptoneurosomes were stimulated by mGluR5 [by the specific agonist (RS)-2-chloro-5-hydroxyphenylglycine (CHPG)] ERK1/2 was rapidly phosphorylated, with somewhat different kinetics than mGluR1, peaking at 1 min, then quickly deactivated. Similarly to mGluR1, we observed dephosphorylation of ERK1/2 in the mGluR5-stimulated preparations from fmr1-KO mice (Fig. 1B). The safety mechanism to avoid excessive pERK accumulation also operated in both WT and KO.

MEK Is also Rapidly Deactivated in KO Synapses upon Group I Metabotropic Receptor Stimulation.

To understand what causes this activity-dependent rapid dephosphorylation of ERK in KO, we first examined the activity level of its upstream kinase. ERK is activated by a MAPK/ERK kinase (MEK)-catalyzed dual phosphorylation. MEK phosphorylation has been shown to be necessary to the ERK activation observed in hippocampal slices (12, 14). Stimulation of mGluR1 led to a modest and sustained MEK phosphorylation in WT synaptoneurosomes, with slower kinetics than ERK phosphorylation (Fig. 2A). Interestingly, although the pERK response rises rapidly, peaking at 1 min, the pMEK response is slower and more prolonged, suggesting a role for an additional enzyme pathway. MEK, again like ERK, is rapidly dephosphorylated in KO samples. Similarly, upon mGluR5 stimulation, MEK is phosphorylated in WT but dephosphorylated in KO samples (Fig. 2B). The kinetics of this reaction differ as well: phosphorylation of MEK is stronger and more prolonged than the phosphorylation of ERK. These differences in kinetics suggest involvement of another regulatory factor, such as deactivation of phosphatases.

Fig. 2.

MEK and PP2A activity after group I mGluR stimulation in WT and KO cortical synaptoneurosomes. MEK and PP2A activity was measured by Western blots, using antibodies to phosphorylated MEK1/2 and phosphorylated PP2A. The ratios of activated (phosphorylated) MEK to total MEK and the ratios of deactivated (phosphorylated) PP2A to total PP2A C subunit were used. (A) mGluR1 stimulation led to a modest and sustained activation of MEK in WT synaptoneurosomes but it was rapidly dephosphorylated in KO (n = 4 each for WT and KO). (B) mGluR5 activation triggered a stronger activation of MEK in WT than in A and the opposite reaction in KO (n = 4 each for WT and KO). (C) PP2A was strongly deactivated in WT but its activity level in KO was sustained at 1 and 2 min and increased at 5 and 10 min in response to mGluR1 stimulation (n = 6 each for WT and KO). (D) mGluR5 stimulation maintained the basal activity level of PP2A in both WT and KO (n = 4 for WT; n = 6 for KO). Control samples were collected from an unstimulated pool at 0 min (C0′) and 10 min (C10′). Total protein level of each lane was measured from an amido black-stained membrane and was used as a loading standard. Data are presented as means ± SEM. *, P < 0.05; **, P < 0.01. Asterisks between the two curves represent P values for WT versus KO and the rest for versus baselines.

PP2A Is Overactivated in KO Synapses After mGluR1 but Not mGluR5 Stimulation.

The level of ERK activation is controlled not only by its upstream kinase MEK but also by phosphatases such as PP2A. PP2A is a serine–threonine phosphatase, long identified as the major Ser/Thr phosphatase involved in ERK inactivation (14). By staining samples with antibody to phosphorylated (deactivated) PP2A, we were able to demonstrate that stimulation of mGluR1s led to transiently deactivated PP2A in WT, but not KO, synaptoneurosomes (Fig. 2C). This transient deactivation corresponds kinetically to the increase in ERK phosphorylation seen in WT samples. Sustained basal activity level of PP2A in KO samples together with rapidly deactivated MEK can account for the rapid deactivation of ERK upon mGluR1 stimulation. There was no significant difference in PP2A activation of WT and KO after mGluR5 stimulation (Fig. 2D), which suggests that the deactivation of ERK and MEK in KO upon mGluR5 stimulation was caused by overactivation of a different phosphatase than PP2A.

Thus, the aberrant initial responses to mGluR stimulation in KO synaptoneurosomes, rapidly dephosphorylated ERK and MEK, are caused mainly by inappropriate activation of phosphatases, which does not happen in WT. The data so far strongly indicate that mGluR stimulation triggers overactivation of phosphatases in KO synapses, which leads to the deficient early-phase ERK activation, and thus a lack of activity-dependent modulation of synaptic protein synthesis. To test this hypothesis, we applied phosphatase blockers to KO synaptoneurosomes to see whether normal ERK activation in response to mGluR stimulation could be restored.

Inhibition of PP2A Restores Normal ERK Activation in KO Synapses.

Okadaic acid and calyculin A are both blockers of PP2A. Fig. 3 shows that in the presence of okadaic acid, both WT and KO synaptoneurosomes respond to metabotropic receptor agonist stimulation with a strong phosphorylation of ERK lasting ≈5 min. Calyculin A blocks PP1 and PP2A; when it is used on KO synaptoneurosomes, the deblocked pERK response is even stronger (Fig. 3B; P < 0.05 at t_min = 1, KO + okadaic acid versus KO + calyculin) than when blocking only PP2A. These results suggest that there is a reservoir of ERK-phosphorylating activity that emerges when phosphatase activity is blocked; the dephosphorylation of ERK observed in KO synaptoneurosomes is abolished under these conditions. Interestingly, the release from negative phosphatase feedback seems to be complete in both WT and KO when stimulated by mGluR1 (Fig. 3A; WT + okadaic acid versus KO + okadaic acid, not statistically different); but after stimulation by mGluR5, KO synaptoneurosomes do not reach the same levels of ERK phosphorylation as do WT (Fig. 3B; P < 0.05 at t_min = 5, 10, WT + okadaic acid versus KO + okadaic acid). This finding suggests the presence in KO synapses of residual blocking activity by another okadaic acid-resistant phosphatase (that is, not PP2A) after mGluR5 stimulation, which agrees with our finding in Fig. 2D. Okadaic acid has been reported to inhibit PP1 as well at higher concentration and after longer incubation (15); however, okadaic acid entry into cells is known to be slow (16). We postulate that the amount of okadaic acid that penetrates into synaptoneurosomes under our condition is enough to inhibit PP2A, but probably not PP1, because pretreatment by calyculin A, inhibiting both, induced dramatically stronger ERK activation than okadaic acid.

Fig. 3.

ERK activity after group I mGluR activation in WT and KO cortical synaptoneurosomes with serine/threonine phosphatase blockers. (A) mGluR1 stimulation in the presence of PP2A blocker okadaic acid restored ERK activation in KO to the level that was not statistically significant from WT. However, blocking both PP1 and PP2A with calyculin A caused overactivation of ERK in KO, which was even stronger than in WT (KO + calyculin versus WT, P < 0.01 at t_min = 1, 2 and P < 0.05 at t_min = 5, 10) (n = 4 each for KO + okadaic acid, KO + calyculin, and WT + okadaic acid). (B) mGluR 5 stimulation together with the serine/threonine inhibitors in KO synaptoneurosomes showed similar restoration of ERK activation to A except that KO with okadaic acid showed a peak of ERK phosphorylation at 2 min after stimulation, which was the only time point at which activated ERK level was significantly higher in KO + okadaic acid than in WT (P < 0.05). ERK activity levels in KO + calyculin were significantly higher than WT at the time points of 1 min (P < 0.01) and 2 and 5 min (P < 0.05) (n = 4 each for KO + okadaic acid, KO + calyculin; n = 3 for WT + okadaic acid). Control samples were collected from an unstimulated pool at 0 min (C0′) and 10 min (C10′). Total protein level of each lane was measured from an amido black-stained membrane and was used as a loading standard. Data are presented as means ± SEM.

Stimulation of mGluR5 Induces Overactivation of Tyrosine Phosphatase in KO Synaptoneurosomes.

Orthovanadate is a blocker of tyrosine phosphatases (17). We incubated synaptoneurosome suspensions for 20 min in 1 mM orthovanadate before stimulation with mGluR1 or mGluR5. Vanadate likewise caused the phosphorylation of ERK to rise dramatically upon metabotropic agonist stimulation in both KO and WT samples (Fig. 4). The rise in pERK is faster, larger, and more prolonged than when okadaic acid is used. In the presence of vanadate, ERK activity never goes back to the basal level in WT and KO as it does when blocking PP2A and/or PP1, suggesting activation of tyrosine phosphatases is responsible for the safety mechanism seen in both WT and KO to avoid toxic excessive accumulation of pERK after mGluR stimulation. When stimulation is by mGluR5, vanadate completely releases both KO and WT strains to high, prolonged ERK phosphorylation levels (Fig. 4B). Because the residual okadaic acid-resistant phosphatase activity seen in KO (Fig. 3B) is abolished in the presence of vanadate (Fig. 4B), tyrosine phosphatase is likely to be overactivated in response to mGluR5 stimulation in KO synapses. When stimulation is by mGluR1, there is less than complete release in KO synaptoneurosomes, suggesting a residual vanadate-resistant phosphatase activity (Fig. 4A), which supports our finding in Fig. 2C that PP2A (vanadate-resistant) is overactivated in KO synapses upon mGluR1 stimulation. Interestingly, the residual phosphatase activity is different for stimulation by mGluR1 (vanadate-resistant) or mGluR5 (okadaic acid-resistant).

Fig. 4.

ERK activity after group I mGluR activation in WT and KO cortical synaptoneurosomes with tyrosine phosphatase blocker. (A) mGluR1 stimulation in the presence of orthovanadate led to a sustained phosphorylation level of ERK in KO, which was significantly higher than in WT at 10 min after stimulation (n = 4 each for KO + orthovanadate and WT + orthovanadate). KO with orthovanadate did not achieve the level of ERK activation in WT + orthovanadate. Asterisks between the two curves represent P values for either WT + vanadate versus KO + vanadate or KO + vanadate versus WT. (B) Phosphorylation of ERK in KO synaptoneurosomes rose dramatically in response to mGluR5 stimulation together with orthovanadate, to a peak level that was not statistically different from WT + orthovanadate (n = 3 for KO + orthovanadate; n = 4 for WT + orthovanadate). Asterisks represent P values for either WT versus KO + orthovanadate or WT + orthovanadate versus KO + orthovanadate. Control samples were collected from an unstimulated pool at 0 min (C0′) and 10 min (C10′). Total protein level of each lane was measured from an amido black-stained membrane and was used as a loading standard. Data are presented as means ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Discussion

FMRP is an RNA-binding protein, increasingly considered to be important for localization and modulation of translation of a subset of mRNAs, perhaps as a conditional repressor protein that becomes permissive for the translation of specific mRNA cargoes upon appropriate stimulation (7, 18, 19). We have suggested that the dendritic spine abnormalities seen in patients with FXS, as well as fragile X KO mice, are caused by a deficit in pruning of immature spines (20), reflecting changes in local protein synthesis.

The varying range of symptoms exhibited by patients could be explained by the effect of FMRP on targeted mRNA transport and modulation of localized translation of a set of proteins, among which are proteins important for cell signaling and structural modification of synapses. Antibody-positioned RNA amplification (APRA) (6) was previously used to identify ≈100 mRNAs bound to FMRP in cultured neurons in situ. In addition to these “APRA cargo” mRNAs, a number of additional FMRP-bound cargo mRNAs have been identified in other laboratories (5, 21, 22).

Interestingly, some mRNAs identified as FMRP cargos by the APRA technique code for enzymes in the ERK pathway, upstream of translation initiation (Fig. 5). One APRA cargo phosphatase is the ERK-specific tyrosine phosphatase VHR (also termed DUSP3, dual specificity phosphatase, initially identified on the array as MKP3, another ERK-specific tyrosine phosphatase; ref. 6); it has previously been found to temper and shorten ERK activation (23). In HeLa cells in which VHR has been down-regulated by siRNA, serum activation of ERK and Jnk is elevated and prolonged (24). Additional signaling enzymes may well be identified as cargos when improved arrays are used. Castets et al. (25) pointed out that PP2A has an N-terminal G-quartet in its mRNA that binds to FMRP, and thus levels of this protein may be subject to translational regulation by FMRP. They observed higher basal levels of the phosphatase PP2Ac in cultured fibroblasts from fmr1-KO mice, suggesting a potential effect on actin-regulated spine shape change; spine shape differs from normal in FXS humans and fmr1 KO mice. It should be noted that activation level, not the amount of PP2A, will determine its net effectiveness. In cultured rat striatal cells, PP2A deactivation by phosphorylation was linked to a sharp, temporary increase in ERK activity by Mao et al. (26). On the basis of coimmunoprecipitation from whole cells, they concluded that active PP2A binds to mGluR5, and that tyrosine phosphorylation of PP2A after mGluR5 stimulation corresponds kinetically with dissociation and temporary deactivation of the phosphatase.

Fig. 5.

Products of FMRP cargo mRNAs are members of second messenger cascades converging on ERK. FMRP cargo proteins are indicated by stars (6, 25). Phospho-ERK1/2 can trigger translational activities (via Mnk1 and S6kinase) and stimulating transcriptional factors (Elk1, CREB) by phosphorylation. ERK phosphorylation is also regulated by phosphatases such as PP2A (in turn regulated by PI3-kinase) (41) and by VHR [which is activated by receptor tyrosine kinase (RTK) and Zap70] (23). PKC and RGS 5 form a feedback loop to Gq (42, 43). We have omitted, for clarity, other important signaling pathways, such as NMDA/Ca²⁺/CaMKII, cAMP/PKA, and DARPP32/ STEP (44).

The formation of long-term memories is widely considered to require localized protein synthesis at or near synapses (see review by Sutton and Schuman in ref. 27). In normal mice, FMRP is translated near synapses in response to group I mGluR activation (7, 28). We now show that fmr-1 KO mice exhibit in addition a striking deficit in the rapid phosphorylation of ERK, the kinase on which several activation pathways converge, and whose activation leads via phosphorylation of Mnk1 and eIF4E to translation initiation (29). Using timed sampling of stimulated synaptoneurosome suspensions we have seen that in fmr-1 KO mice ERK is not phosphorylated, but rather dephosphorylated in the first 10 min after stimulation by metabotropic receptor agonists.

Although there is interest in the mGluR theory to describe the FXS (30), and evidence for relief of audiogenic seizures by using MPEP (31), an mGluR5 blocker, there are as yet no direct measurements (e.g., ref. 32) demonstrating excessive activity of mGluR5 in fragile X mammalian systems. A more complex effect, caused by dysregulated levels of a series of FMRP cargo enzymes, may cause an imbalance in second messenger systems, an explanation amenable to experimental approaches. Because the resting cellular level of pERK is higher in hippocampus (33) and cortical synaptoneurosomes (data not shown) of fmr-1 KO mice, the key to signaling efficacy likely lies in the rate of change in phosphorylation. Either a defective kinase pathway, an altered phosphatase activation level, or both can be expected to alter the rate of change. If activation of mGluR in WT cells leads simultaneously to rapid but transient phosphorylation of ERK and activation of phosphatase, the result will be a quick activation of protein translation accompanied by endocytosis of dephosphorylated GluR2 subunits in AMPA receptors, leading to long-term depression (LTD) (34). In cells lacking FMRP, phosphatase activation prevails, so that quick initiation of protein translation is missing, but AMPA-R dephosphorylation and LTD are not only intact in the absence of protein translation but even accentuated in the hippocampus of fmr1-KO mice (35). Moreover, withdrawal of dephosphorylated GluR2 subunits from AMPA receptors because of excessive phosphatase activation can explain why protein translation is not necessary for LTD in fmr-1 KO mice (36). Huang and Hsu (37) demonstrated a reduction of tyrosine phosphorylation and surface expression of GluR2 in hippocampal slices subjected to DHPG-LTD. Volk et al. (38) likewise observed decreased GluR1 surface expression after DHPG-LTD and showed that mGluR1, but not mGluR5, was required for this shift. Aberrant activation of phosphatases by mGluR stimulation in KO cortical synapses may also cause excessive dephosphorylation of AMPA and NMDA receptors, which leads to reduced or absent cortical long-term potentiation seen in fragile X KO mice (39, 40). Dysregulated levels of cargo enzymes in the absence of FMRP might cause an imbalance in the enzyme cascade such that protein translation initiation is impaired, while nevertheless dephosphorylation of key receptor subunits leads to long-lasting synaptic reactivity changes.

Materials and Methods

Synaptoneurosomes.

Synaptoneurosomes were prepared from the cortices of single WT or KO mice, P10–P14, of the strain FVB.129P2-FMR1^tm1Cgr. Briefly, mice were quickly decapitated, brains were removed and dissected, and cortices were homogenized in a glass-Teflon homogenizer in 1 ml of homogenizing buffer (50 mM Hepes, pH 7.5, 125 mM NaCl, 100 mM sucrose, 2 mM potassium acetate), filtered through a series of nylon mesh filters (149, 62, and 30 microns; Small Parts) and finally through a 10-μm polypropylene filter (Gelman Sciences). Filters were washed at each step with the homogenizing buffer (used total 1 ml). The final filtrate was spun briefly (4,000 × g, 1 min); final supernatant volume was ≈1 ml. Before stimulation paradigms, this suspension was incubated, stirring, with 1 μM tetrodotoxin (Tocris), on ice for 5 min then at room temperature for another 5 min. Reactions proceeded at room temperature.

Stimulation.

Each synaptoneurosome preparation was divided into smaller pools, which were either not stimulated or were stimulated with one of the following: 750 μM CHPG (mGluR5-specific agonist; Tocris) or blocked with 10 μM MPEP (mGluR5 blocker; Tocris) for 15 min before addition of 100 μM DHPG (agonist for group 1 mGluR; Tocris), to stimulate specifically mGluR1s. To block phosphatases, synaptoneurosomes were preincubated with 50 nM Calyculin A (Cell Signaling Technology) for 10 min or 200 nM okadaic acid (Tocris) for 10 min or 1 mM sodium orthovanadate (Sigma) for 20 min before stimulation. Samples were removed and instantly lysed at 1, 2, 5, and 10 min after stimulation, in a lysis buffer [final concentration 50 mM Tris (pH 8) 50 mM NaCl, 1% Nonidet P-40, 20 nM okadaic acid, 20 μM sodium orthovanadate, phosphatase inhibitor mixture (Pierce), and protease inhibitor mixture (Pierce)].

Western Blots.

For each sample, 30 μg of lysed protein samples was run on 12% polyacrylamide gels, blotted to nitrocellulose membranes, and stained with rabbit mAbs specific for phospho-ERK1/2 (Thr-202/Tyr-204) (Cell Signaling), phospho-MEK 1/2 (Ser-217/221) (Cell Signaling), phospho-PP2A (Y307) (Epitomics), MEK 1/2 (Cell Signaling), or PP2A C subunit (Cell Signaling). HRP-labeled secondary anti-rabbit antibody (Cell Signaling) was detected by enhanced chemiluminescence (Pierce). To quantify and standardize protein levels, total protein was stained with Amido Black (J. T. Baker Chemical). Chemiluminescence was scanned in an FluorChem 8900 Imager (Alpha Innotech), and relative optical densities were determined by using AlphaEaseFC software, version 4.0.1 (Alpha Innotech), normalized to total protein loaded.

Statistical Methods.

ANOVAs of 2 × 5 repeated measures were conducted to determine the influence of genotype (WT, KO) and time [0 (baseline) and 1, 2, 5, and 10 min] on MEK andPP2A activity measures after mGluR1 or mGluR5 stimulation (Fig. 2). ANOVAs of 5 × 5 (five groups, five time points) were conducted on ERK activity measures after mGluR1 or mGluR5 stimulation (Fig. 3); the significant post hocs for the groups of WT and KO are shown separately in Fig. 1. ANOVAs of 2 × 2 × 5 to determine the influence of genotype (WT, KO), blocker (vanadate, control), and time (five time points) were conducted on ERK activity after mGluR1 and mGluR5 stimulation (Fig. 4). When appropriate, post hoc analyses were conducted by using the least significant differences method. The SPSS statistical package (version 15.0) was used for all analyses, and P < 0.05 was considered significant.