Role of Protein Phosphatase 2A in mGluR5-regulated MEK/ERK Phosphorylation in Neurons*

The regulation of protein phosphorylation requires coordinated interaction between protein kinases and protein phosphatases (PPs). Recent evidence has shown that the Gαq-protein-coupled metabotropic glutamate receptor (mGluR) 5 up-regulates phosphorylation of MAPK/ERK1/2. However, signaling mechanisms linking mGluR5 to ERK are poorly understood. In this study, roles of a major serine/threonine PP, PP2A, in this event were evaluated in cultured neurons. We found that the PP1/2A inhibitors okadaic acid and calyculin A mimicked the effect of the mGluR5 agonists (RS)-3,5-dihydroxyphenylglycine and (RS)-2-chloro-5-hydroxyphenylglycine in facilitating phosphorylation of ERK1/2 and its upstream kinase, MEK1/2, in a PP2A-dependent but not PP1-dependent manner. Co-administration of either inhibitor with an mGluR5 agonist produced additive phosphorylation of ERK1/2. Enzymatic assays showed a basal level of phosphatase activity of PP2A under normal conditions, and activation of mGluR5 selectively inhibited PP2A, but not PP1, activity. In addition, a physical association of the cytoplasmic C terminus of mGluR5 with PP2A was observed, and ligand activation of mGluR5 reduced mGluR5-PP2A binding. Additional mechanistic studies revealed that mGluR5 activation increased tyrosine (Tyr307) phosphorylation of PP2A, which was dependent on activation of a p60c-Src family tyrosine kinase, but not the epidermal growth factor receptor tyrosine kinase and resulted in dissociation of PP2A from mGluR5 and reduced PP2A activity. Together, we have identified a novel, mGluR5-triggered signaling mechanism involving use- and Src-dependent inactivation of PP2A, which contributes to mGluR5 activation of MEK1/2 and ERK1/2.

Agonist binding of group I metabotropic glutamate receptors (mGluRs) 1 activates phospholipase C␤1 through G␣q-proteins and increases hydrolysis of membrane phosphoinositide. This yields diacylglycerol to activate protein kinase C and inositol-1,4,5-triphosphate to release intracellular Ca 2ϩ (1,2). Altered protein kinase C and Ca 2ϩ signals could then engage in the modulation of various cellular activities, such as survival, synaptic plasticity, gene expression, and so forth (3,4). A number of morphological studies have shown high levels of group I mGluR expression (both mGluR1 and mGluR5 subtypes) in medium spiny projection neurons of rat striatum (5)(6)(7)(8). Because these two receptors are predominantly postsynaptic in this region of the brain, they are thought to play an active role in the regulation of cellular activities of striatal neurons (9). Extracellular signal-regulated kinases (ERKs) 1 and 2, a major subclass of mitogen-activated protein kinases (MAPKs), are densely expressed in striatal neurons and activated through phosphorylation on their Thr 202 and Tyr 204 sites by mitogen-activated protein kinase/ERK kinase (MEK) 1/2 in response to a wide range of extracellular stimuli (10 -12). The group I mGluR-selective agonists have been found to phosphorylate ERK1/2 in cultured glia (13,14) and Chinese hamster ovary cell lines with transfected mGluR1 or mGluR5 (15,16). However, at present, signaling mechanisms underlying the mGluR5 regulation of ERK1/2 phosphorylation are poorly understood, especially in neurons (17).
Protein phosphatase (PP) 2A is one of the major serine/ threonine phosphatases that are highly expressed in striatal neurons (18,19). As a multisubunit complex, PP2A consists of a 36-kDa catalytic subunit (PP2A-C) and a 65-kDa structural subunit (PP2A-A) forming a core enzyme, which associates with a variable regulatory subunit (PP2A-B) to constitute a heterotrimeric PP2A holoenzyme (20). As a negative regulator of protein phosphorylation, PP2A plays a key role in the dephosphorylation of various phosphoproteins (receptors, enzymes, neuropeptides, structural proteins, and so forth) related to many fundamental cellular activities. There is evidence showing that ERK1/2 is among the signaling phosphoproteins that are readily dephosphorylated by PP2A in cell lines (21)(22)(23)(24). It is therefore possible that PP2A participates in the regulation of ERK1/2 phosphorylation via G-protein-coupled receptors, such as group I mGluRs, in neuronal cells.
In this study, we therefore examined the possible involvement of PP2A in the group I mGluR-dependent ERK1/2 phosphorylation in a well-characterized striatal neuronal culture model. We found that inhibition of PP2A mimicked the effect of the mGluR5 agonists in elevating ERK1/2 phosphorylation. Simultaneous PP2A inhibition and mGluR5 stimulation pro-duced an additive increase in ERK1/2 phosphorylation. Interestingly, mGluR5 activation resulted in transient and p60c-Srcdependent tyrosine phosphorylation of PP2A and dissociation of PP2A from mGluR5, leading to reduced PP2A activity. These results unravel a novel signaling model involving inactivation of PP2A after mGluR5 stimulation. This model substrates a significant component of signaling pathways responsible for mGluR5-sensitive activation of ERK1/2.

EXPERIMENTAL PROCEDURES
Primary Rat Striatal Neuronal Cultures-Standardized procedures were employed to prepare primary striatal neuronal cultures from day 18 rat embryos or neonatal 1-day-old rat pups (Charles River, New York, NY) (25)(26)(27). Predominant neuronal cells were obtained using the procedures, as evidenced by the fact that Ͼ90% of total cells were immunoreactive to the specific marker for neurons (microtubule-associated protein-2aϩ2b), but not for glia (glial fibrillary acidic protein). Cells were cultured for 15-18 days before use.
Single and Double Immunofluorescent Labeling-The immunofluorescent labeling on 8-chamber glass slides was performed as described previously (25,26). Briefly, cultures were fixed in cold 4% paraformaldehyde (10 min), followed by incubation in 4% normal donkey serum and 1% bovine serum album (20 min) to block nonspecific staining. The cells were treated with primary rabbit anti-pERK1/2 or anti-ERK1/2 antibodies (Cell Signaling; 1:100) overnight at 4°C. Sections were then incubated for 1 h with donkey anti-rabbit secondary antibodies (1:200) conjugated to fluorescein isothiocyanate (Jackson ImmunoResearch Laboratories). For double labeling, the cells were treated with a mixture of primary antibodies containing rabbit anti-mGluR5 antibodies (1:200; Upstate, Charlottesville, VA) and mouse anti-PP2A (catalytic subunit) antibodies (1:200; Upstate) overnight at 4°C. Sections were then incubated for 1 h with donkey anti-rabbit secondary antibodies conjugated to fluorescein isothiocyanate and donkey anti-mouse secondary antibodies conjugated to TRITC (Jackson ImmunoResearch Laboratories) at 1:200. The immunofluorescent images were analyzed using confocal microscopy (Nikon C1 laser scanning confocal microscope).
Co-immunoprecipitation-Rat striatal cell proteins were prepared under weakly denaturing conditions to permit the interaction of mGluR5 with other intracellular proteins (26). Briefly, striatal cultures were scraped into a microtube containing ice-cold sample buffer (10 mM Tris-HCl, pH 7.4, 5 mM NaF, 1 mM Na 3 VO 4 , 1 mM EDTA, and 1 mM EGTA) and homogenized by sonication. The homogenate was centrifuged at 800 ϫ g (10 min) at 4°C. The supernatant was again centrifuged at 11,000 ϫ g at 4°C for 30 min. The pellet was resuspended in sample buffer and solubilized in 1% sodium deoxycholate. After incubation at 37°C for 30 min, Triton X-100 was added to a final concentration of 0.1%. Insoluble proteins were sedimented at 100,000 ϫ g at 4°C for 20 min. The supernatants were used for co-immunoprecipitation. The mGluR5 or PP2A was then precipitated using rabbit anti-mGluR5 antibodies (Upstate) or mouse antibodies against the catalytic subunit of PP2A (Upstate), respectively, and 50% protein A-agarose/ Sepharose bead slurry (Amersham Biosciences). In some experiments, a 5-fold (by weight) excess of an immunizing peptide, KSSPKYDTLI-IRDYTNSSSSL (mGluR5) or CRGEPHVTRRTPDYFL (PP2A), was preincubated with the respective antibodies before co-immunoprecipitation. The c-Src was precipitated using rabbit anti-c-Src antibodies (Santa Cruz Biotechnology). Proteins were separated on Novex 4 -12% gels and probed with rabbit anti-mGluR5 antibodies (Upstate), rabbit anti-c-Src antibodies (Santa Cruz Biotechnology), or mouse anti-PP2A antibodies (Upstate). To detect tyrosine phosphorylation of PP2A catalytic subunit, PP2A catalytic subunits were immunoprecipitated with rabbit PP2A catalytic subunit antibodies (Upstate). PP2A immunoprecipitates were then blotted with rabbit anti-phosphotyrosine antibodies (Upstate). Horseradish peroxidase-conjugated secondary antibodies and enhanced chemiluminescence were used to detect proteins. Negative controls with antigen pre-absorption were carried out for antibodies used in immunoprecipitation.
PP1 and PP2A Phosphatase Activity Assay-The activity was assessed by dephosphorylation of the synthetic PP1/PP2A-specific phosphopeptide (K-R-pT-I-R-R) using a serine/threonine phosphatase assay kit from Upstate (catalogue no. . Following the manufacturer's instructions, cultures were scraped with 0.3 ml of phosphatase extraction buffer containing 20 mM imidazole-HCl; 2 mM EGTA, pH 7.0; 10 g/ml each of aprotinin, leupeptin, antipain, and soybean trypsin inhibitor; 1 mM benzamidine; and 1 mM phenylmethylsulfonyl fluoride. Cells were sonicated (10 s) and centrifuged at 2000 ϫ g for 5 min. Supernatants were used for phosphatase activity assays. To immunoprecipitate PP1 or PP2A, rabbit antibodies against PP1␦ (4 g) or mouse antibodies against the catalytic subunit of PP2A (4 g) were added, respectively, to a total of 150 g/200 l lysate, followed by addition of 50% protein A-agarose/Sepharose bead slurry (Amersham Biosciences) and incubation for 1-2 h at 4°C. Beads were washed three times with PBS, followed by one wash with assay buffer, before the phosphopeptide was added to a final concentration of 0.75 mM for incubation for 10 min at 30°C. The samples were centrifuged briefly, and 25 l of sample was transferred into a well of a 96-well microtiter plate. A detection solution containing malachite green (100 l) was added into each well and incubated for 10 -15 min at room temperature for color development. A colorimetric assay was made by reading at 650 nm in a microtiter plate reader. Absorbance values of samples were compared with negative controls containing no enzyme and a freshly prepared phosphate standard. Specific activity was determined as the amount of pciomoles of phosphate released per minute per microgram of protein.
Cell Viability Assay-Cell viability was measured using a double fluorescein diacetate/propidium iodide staining procedure (26). Fluorescein diacetate is membrane-permeable and freely enters intact cells, in which it is hydrolyzed by cytosolic esterase and converted to membraneimpermeable fluorescein with a green fluorescence, exhibited only by live cells. Propidium iodide is non-permeable to live cells but penetrates the membranes of dying/dead cells, showing red fluorescence. Cells were rinsed twice with 1ϫ PBS and incubated at 37°C for 5 min with 1ϫ PBS (0.5 ml/per well) containing 10 g/ml fluorescein diacetate (Sigma) and 5 g/ml propidium iodide (Sigma). Cultures were washed once with PBS and examined under fluorescent light microscopy. The total numbers of viable cells stained by green fluorescein and dead cells stained by red propidium iodide were determined by counting cells in five random fields. Positive control was produced by treating cultures with kainic acid (500 -1000 M, 24 h).
Statistics-The results are presented as means Ϯ S.E. and were evaluated using a one-or two-way analysis of variance, as appropriate, followed by a Bonferroni (Dunn) comparison of groups using least squares-adjusted means. Probability levels of Ͻ0.05 were considered statistically significant.

Inhibition of PP2A Increases ERK1/2 Phosphorylation-Al-
though PP2A dephosphorylates ERK1/2 in heterologous systems (21-24), it is not known whether PP2A could exert the same effect in neurons. We then set out to test whether an inhibitory pathway exists from PP2A to ERK1/2 in our cultured striatal neurons for controlling ERK1/2 phosphorylation. OA was used to selectively inhibit PP2A action on the ERK pathway. As shown in Fig. 1, A and B, the inhibition of PP2A with OA increased ERK1/2 phosphorylation in concentration-and time-dependent manners. At a concentration of 0.1 nM, OA did not alter basal levels of pERK1/2. Starting at a fairly low concentration (1 nM), OA induced a reliable increase in pERK1/2 levels. Greater increases in pERK1/2 were induced by higher concentrations of the inhibitor. No significant changes were seen in cellular levels of ERK1/2 after OA application at all concentrations surveyed. In a time-course study in which OA was applied at 0.5-1 M for different durations, a rapid and prolonged increase in ERK1/2 phosphorylation occurred with no changes in ERK1/2 levels. Reliable increases in ERK1/2 phosphorylation were seen as early as 2 min after OA application and peaked around 5-10 min after OA application. The increases remained at 30 min and declined at 60 min. A different time-course evaluation was also performed, in which OA (1 M) was incubated for 5 min, and cultures were then lysed 5, 25, or 60 min after termination of the drug incubation. The data from this study (data not shown) were similar to those obtained by incubating OA at different durations. To define the subcellular distribution of pERK1/2 activated by the PP2A inhibition, immunofluorescent labeling was performed after OA treatment (1 M, 5 min). A high level of pERK1/2 immunostaining was present within the nucleus, and weak to moderate immunostaining was seen in the cytoplasm and neural processes (Fig. 1C). There was no significant difference in cell viability as detected by the double fluorescein diacetate-propidium iodide staining between control cultures and cultures treated with OA (1 M, 5 or 60 min).
Another commonly used PP2A inhibitor, calyculin A, was also tested for its effects on pERK1/2 and ERK1/2 in the same protocol for OA. Calyculin A at a dose range from 0.1 to 1000 nM (for 2, 5, 10, or 30 min) increased ERK2 phosphorylation indistinguishable from that seen after OA application ( Fig. 2A). Calyculin A (0.1-1000 nM, 10 min) had no effects on ERK1/2 levels (data not shown). In contrast, tautomycin, an inhibitor relatively selective for PP1 (28), did not alter basal levels of pERK2 ( Fig. 2A) and ERK1/2 (data not shown) at doses ranging from 10 nM to 3 M. To test the efficacy and selectivity of the PP1/PP2A inhibitors in this culture system, the PP1 and PP2A activity was assayed in cultures after drug treatments at different concentrations for 10 min. From Fig. 2B, OA and calyculin A at 0.01 M, a dose that significantly increased pERK1/2 levels, reduced PP2A but not PP1 activity. At 1 M, the two inhibitors reduced PP1 and, to a greater extent, PP2A activity. In contrast to these two drugs, tautomycin at 1 M substantially reduced PP1 activity, sparing PP2A activity. Together, the data obtained above support the existence of a PP2Asensitive, but not PP1-sensitive, dephosphorylation of pERK1/2 in cultured striatal neurons.
OA or calyculin A may change cell firing, which, in turn, could cause a release of an assortment of transmitters and modulators in the culture, leading to activation of receptors and/or kinases other than PP2A to account for the observed changes in ERK1/2 phosphorylation. Thus, the effect of OA and calyculin A (1 M, 10 min) was tested in the presence of a Na ϩ channel blocker, tetrodotoxin (1 M, 30 min before OA or calyculin A). We found that the ERK1/2 phosphorylation induced by the two drugs was unaffected by tetrodotoxin (data not shown). This suggests that the effect of the two inhibitors was independent of cell firing and transmitter release.
The immediate upstream kinase that phosphorylates ERK1/2 is MEK1/2, which is activated through phosphorylation at its serine 217/221 sites by the MEK kinase, i.e. the serine kinase Raf-1 (or c-Raf) (10 -12). We examined the effect of OA on MEK1/2 phosphorylation and the effect of the MEK1/ Representative immunoblots are shown to the left of the quantified data (mean Ϯ S.E., n ϭ 9). C, confocal immunofluorescent images illustrating subcellular distributions of pERK1/2 induced by okadaic acid (1 M, 5 min). Strong pERK1/2 immunostaining was seen in the nucleus, and weak to moderate staining was seen in the cytoplasm and neural processes. * (pERK1) and ϩ (pERK2), p Ͻ 0.05 versus basal levels.
2-selective inhibitors (U0126 and PD98059) on OA-induced phosphorylation of ERK1/2. OA produced a concentration-dependent increase in pMEK1/2 without a change in the amount of MEK1/2 proteins (Fig. 2C). These sites of MEK1/2 phosphorylation are involved in kinase activation, indicating that OA produced activation of MEK1/2. The phosphorylation of MEK1/2 induced by OA also corresponded well with the kinetics of OA-induced ERK1/2 phosphorylation in a time-course study (data not shown). Pretreatment of cultures with U0126 (10 M) abolished OA-induced ERK1/2 phosphorylation, without changing basal ERK1/2 levels (Fig. 2D). Similar results were also obtained after application of PD98059 at 100 M (data not shown). In addition, OA (1 M, 5 min) increased basal levels of phosphorylated Raf-1, and the elevation of Raf-1 phosphorylation started to decline 1 h after OA application (data not shown). These results support a model in which OA application results in a sequential phosphorylation of Raf-1, MEK1/2, and ERK1/2 because they are hierarchically organized as a three-tiered MAPK module. Moreover, the effects of OA on ERK1/2 were probably indirect and mediated by activation of the upstream kinase MEK1/2.
Activation of mGluR5 Reduces PP2A Activity-Activation of group I mGluRs with the selective agonist DHPG increased ERK1/2 phosphorylation in striatal neurons and other cell lines (13)(14)(15)(16). It is likely that DHPG may inhibit basal protein phosphatase activity, which contributes, at least in part, to the elevation of ERK1/2 phosphorylation. To investigate this likelihood, effects of DHPG on PP1 and PP2A activity were examined in this study. We found that DHPG concentration-dependently reduced basal levels of PP2A but not PP1 activity (Fig.  3A). At 100 M, DHPG reduced basal PP2A activity by ϳ23% of control (Fig. 3A). The inhibition of PP2A activity induced by 100 M DHPG appeared to be a rapid and transient event because it became evident at 2 min and peaked at 5 min (Fig.  3B), which corresponds well with the kinetics of ERK1/2 phosphorylation in response to DHPG stimulation (27). A series of pharmacological studies were next performed to identify the receptor mechanism mediating the DHPG inhibition of PP2A activity. Pretreatment of cultured neurons with the noncompetitive mGluR5-selective antagonist MPEP at a low concentration (0.2 M) blocked the DHPG inhibition of PP2A activity (Fig. 3C). In contrast, the noncompetitive mGluR1-selective antagonist CPCCOet at a high concentration (20 M) did not significantly alter the DHPG effect (Fig. 3C). The mGluR5selective agonist CHPG at 1-2 mM (5 min) reduced PP2A activity to an extent comparable with that induced by DHPG (100 M, 5 min), and the CHPG effect was sensitive to MPEP (0.2 M), but not CPCCOet (20 M; data not shown). These data suggest that selective activation of mGluR5 rather than mGluR1 mediates the DHPG-sensitive PP2A inhibition. Additionally, because MPEP alone did not affect basal PP2A activity (Fig. 3D), mGluR5 is less likely to be involved in maintaining basal PP2A activity but exerts a ligand (use)-dependent inhibition of basal PP2A activity.
Inhibition of PP2A and Stimulation of mGluR5 Produce an Additive Increase in ERK1/2 Phosphorylation-If inhibition of PP2A contributes to mGluR5-mediated ERK1/2 phosphorylation, then the PP2A inhibitor might augment the level of ERK1/2 phosphorylation induced by mGluR5 stimulation. To evaluate this possibility, separate dose-response curves for OA and DHPG in their effects of elevating ERK1/2 phosphorylation were constructed again in cultured striatal neurons in this FIG. 2. ERK1/2 or MEK1/2 phosphorylation induced by the inhibition of PP2A with calyculin A or OA in cultured rat striatal neurons. A, calyculin A (0.1-1000 nM, 10 min) concentrationdependently increased pERK2 levels (mean Ϯ S.E., n ϭ 6). In contrast, tautomycin, an inhibitor relatively selective for PP1, did not alter basal pERK2 levels after incubation at 0.01-3 M for 10 min (mean Ϯ S.E., n ϭ 6). B, effects of OA, calyculin A (CA), and tautomycin (Tauto) on PP1 and PP2A activity in cultured rat striatal neurons measured by the protein phosphatase activity assays. The inhibitors were incubated for 10 min before the activity assay (mean Ϯ S.E., n ϭ 5-7). OA and CA reduced PP2A but not PP1 activity at 0.01 M, whereas tautomycin reduced PP1 but not PP2A activity at 1 M. C, OA (0.1-1000 nM, 5-10 min) concentration-dependently increased pMEK1/2 but not MEK1/2 levels. Representative immunoblots are shown to the left of the quantified data of pMEK/MEK analyzed from separate experiments (mean Ϯ S.E., n ϭ 5). D, the MEK1/2 inhibitor U0126 blocked OA-induced ERK1/2 phosphorylation. Representative immunoblots are shown to the left of the quantified data of pERK2/ERK2 analyzed from separate experiments (mean Ϯ S.E., n ϭ 4 -6). U0126 (10 M) was incubated for 30 min before and during a 10-min incubation of OA (1 M). *, p Ͻ 0.05 versus basal levels. ϩ, p Ͻ 0.05 versus OA alone. study. The two drugs were then co-applied at different concentrations to test the effect of co-application of both of them on ERK1/2 phosphorylation. As shown in Fig. 4, the two drugs at a low concentration (3 M for DHPG and 0.1 nM for OA) did not alter basal levels of pERK1/2. At the middle concentrations (10 or 30 M DHPG in combination with 1 or 10 nM OA, respectively), the two drugs produced an additive increase in ERK1/2 phosphorylation. At the two highest concentrations of each drug, no greater ERK1/2 phosphorylation was seen after coincubation of the two drugs as compared with that induced by either drug alone. Co-incubation of the two drugs at all concentrations had no effect on cellular levels of ERK1/2 (data not shown). Apparently, the two drugs, when co-applied, produced additive ERK1/2 phosphorylation mostly when they were administered at a middle dose range. Once a maximal (ceiling) increase in ERK1/2 phosphorylation was reached by either drug alone at a high dose range, no additive increase could be induced by the two drugs together. This pattern of additive effects supports a signaling model in which mGluR5 increases ERK1/2 phosphorylation partially through PP2A inhibition.
The mGluR5 Is Physically Associated with PP2A-Recent studies with protein-protein interaction assays in vitro suggest physical association of mGluRs with certain protein phosphatases. For instance, mGluR3 specifically forms a complex with PP2C, whereas mGluR7 specifically forms a complex with PP1C (29 -31). In this study, a co-immunoprecipitation approach was used to examine the physical association of mGluR5 with PP2A in cultured striatal neurons. Under normal conditions, a clear and specific PP2A immunoreactive band was detected in mGluR5 immunoprecipitates (Fig. 5A). This band, as well as the mGluR5 band, was not shown in the immunoblots if the mGluR5 immunizing peptide was added to absorb the mGluR5 antibodies in immunoprecipitation procedures (Fig. 5A). In a reverse co-immunoprecipitation, specific mGluR5 immunoreactivity was also detected in PP2A precipitates (Fig. 5B). Again, the PP2A immunizing peptide was able to block the mGluR5 and PP2A immunoblots in PP2A precipitates (Fig. 5B). In addition, an association of mGluR1␣ with PP2A was tested in cultured striatal neurons. We found that no detectable PP2A immunoreactivity was displayed in the mGluR1␣ precipitates (data not shown). These results indicate a selective physical association of mGluR5 with PP2A in cultured striatal neurons under normal conditions. To detect cel-lular co-localization of mGluR5 and PP2A, double immunofluorescent labeling was performed on cultured striatal neurons. As seen in Fig. 5D, strong mGluR5 immunostaining was distributed in the soma (membrane and cytoplasm) and neurites, whereas strong PP2A staining was shown in the soma and neurites with moderate staining in the nuclear envelope. A merged image in Fig. 5D exhibits clear co-localization of the two proteins in the same neurons. The co-localization occurred in Ͼ90% of neurons.
Activation of mGluR5 Reduces mGluR5 and PP2A Association-Agonist binding of glutamate receptors may result in dissociation of submembranous signaling proteins from the membrane receptors. In this study, we tested the effects of mGluR5 stimulation on mGluR5 association with PP2A. DHPG (100 M, 5 min) markedly reduced the amount of PP2A in mGluR5 immunoprecipitates (Fig. 6A) and the amount of mGluR5 in PP2A immunoprecipitates (data not shown). The reduction of PP2A levels in mGluR5 immunoprecipitates induced by DHPG (100 M for different durations) was a rapid and transient event (Fig. 6B), which kinetically corresponded with dynamic decreases in PP2A activity (Fig. 3B). In addition, the mGluR5 antagonist MPEP (0.2 M) prevented DHPG (100 M, 5 min) from reducing PP2A in mGluR5 immunoprecipitates (Fig. 6C). The results here reveal a ligand-driven dissociation of mGluR5 with PP2A, which may contribute to the reduction of PP2A activity and the elevation of ERK1/2 phosphorylation after mGluR5 stimulation.
Activation of mGluR5 Increases Tyrosine Phosphorylation of PP2A-The catalytic subunit of PP2A can be inactivated through phosphorylation of Tyr 307 in its conserved C terminus in vitro and in living cells (32)(33)(34)(35). We then set out to determine whether activation of the G-protein-coupled receptors (mGluR5) could inactivate PP2A via a tyrosine phosphorylation mechanism in living neurons. The tyrosine phosphorylation of the PP2A catalytic subunit was tested in PP2A immunoprecipitates using immunoblots with anti-phosphotyrosine antibodies. Activation of mGluR5 with DHPG (100 M, 5 min) elevated the level of tyrosine-phosphorylated PP2A (pPP2A), without affecting basal levels of PP2A (Fig. 7). The elevated phosphorylation was blocked by pretreatment of cultures with 0.2 M MPEP, but not 20 M CPCCOet (Fig. 7). Both MPEP and CPCCOet had no effect on basal levels of PP2A tyrosine phosphorylation (Fig. 7). Moreover, DHPG (100 M incubated for 2, 5, 10, or 30 min) induced a rapid and transient tyrosine phosphorylation of PP2A (data not shown), similar to the ERK1/2 phosphorylation (Fig. 1B), reduction of PP2A activity (Fig. 3B), and PP2A dissociation from mGluR5 (Fig. 6B) in response to DHPG stimulation under the same conditions. These results indicate that selective activation of mGluR5 elevates the tyrosine phosphorylation of the PP2A catalytic subunit without changing cellular levels of PP2A.
Non-receptor Tyrosine Kinase (p60c-Src) Mediates the PP2A Phosphorylation by mGluR5-To further examine the mechanism by which mGluR5 activation increases tyrosine phosphorylation of PP2A, we tested the role of a p60c-Src family nonreceptor tyrosine kinase in this event. The two Src inhibitors (PP2 and genistein) and their negative controls (PP3 and genistin, respectively) were used to define the importance of Src in cultured striatal neurons. We found that PP2 at 5 M blocked the DHPG-induced tyrosine phosphorylation of PP2A (Fig. 8A). PP2 also prevented the DHPG-induced reduction of PP2A association with mGluR5 (Fig. 8B) and the reduction of PP2A activity (Fig. 8C). Finally, PP2 partially reduced the DHPGstimulated ERK1/2 phosphorylation (Fig. 8D). In contrast to PP2, PP3 at the same concentration did not achieve any of the above-mentioned PP2 effects (Fig. 8). When genistein versus genistin was used at 50 M, results similar to those observed after PP2 versus PP3 application were obtained (data not shown). To examine the possible role of receptor tyrosine kinases, such as EGF receptors (ErbB1 receptors), we tested the effects of the ErbB1 inhibitor tyrphostin AG1478 on DHPGinduced tyrosine phosphorylation of PP2A. AG1478 at 1 M, a dose that blocked the ERK1/2 phosphorylation induced by hu- man EGF (30 ng/ml, 5-10 min), did not inhibit increases in tyrosine phosphorylation of PP2A (data not shown), nor did AG825, a tyrphostin that selectively inhibits the EGF receptor tyrosine kinase ErbB2 (data not shown). These data suggest a Src-dependent and EGF receptor-independent tyrosine phosphorylation of PP2A in response to mGluR5 stimulation in striatal neurons. To determine whether mGluR5 has a physical association with Src proteins, Src immunoreactivity was detected in the mGluR5 precipitates. From Fig. 9, specific bands of Src proteins were displayed in the mGluR5 precipitates. Likewise, mGluR5 bands could be shown in the Src precipitates in a reverse co-immunoprecipitation (data not shown). Thus, Src proteins form a complex with mGluR5 for processing mGluR5 signals to the downstream effectors.

DISCUSSION
The present study investigated signaling mechanisms involved in linking group I mGluRs to ERK in cultured neurons. We found that selective activation of mGluR5 activated ERK1/2. Inhibition of PP2A, but not PP1, mimicked the effect of the mGluR5 agonists. The PP2A inhibitor also increased phosphorylation of MEK1/2. Co-application of the PP2A inhibitor and the mGluR5 agonist produced an additive increase in ERK1/2 phosphorylation. Additional studies showed that activation of mGluR5 increased tyrosine phosphorylation of PP2A through a Src family non-receptor tyrosine kinase, but not the receptor kinase (EGF receptors). Increased tyrosine phosphorylation may result in dissociation of PP2A with mGluR5 and reduction of PP2A activity, leading to increased MEK1/2 and ERK1/2 phosphorylation. These results reveal a PP2A-dependent mechanism imperative for transmitting mGluR5 signals to the MEK/ERK pathway in neurons.
Although group I mGluRs increase ERK phosphorylation in cultured glia (13,14) and Chinese hamster ovary cell lines transfected with mGluR1 or mGluR5 (15,16), the similar role of group I mGluRs in neurons and the detailed signaling mechanisms bridging mGluR1/5 signals to ERK are poorly understood. In the Chinese hamster ovary cells expressing transfected mGluR5, the mGluR5 agonist induced ERK1/2 phosphorylation in the absence of intracellular or extracellular Ca 2ϩ ions (16), indicating a Ca 2ϩ -independent mechanism involved in the event. In neurons, we found that extracellular Ca 2ϩ ions are not important because removal of extracellular Ca 2ϩ ions and blockade of major Ca 2ϩ channels (N-methyl-Daspartate receptors, ␣-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors, and voltage-operated Ca 2ϩ channels) did not affect DHPG-stimulated ERK1/2 phosphorylation (27). In contrast to the Chinese hamster ovary cells, a Ca 2ϩ -dependent mechanism derived from intracellular Ca 2ϩ release via activation of the conventional phospholipase C␤1/inositol-1,4,5-triphosphate pathway contributes to the mGluR5-regulated ERK1/2 phosphorylation, even though this pathway appears to mediate only a small portion of signals (27). The participation of Ca 2ϩ signals by Ca 2ϩ release in neurons provides a common signaling network node (or hub) for multiple signaling pathways to actively interact with the ERK pathway through Ca 2ϩ . A recent interesting finding in this laboratory is the identification of Homer proteins in forming a signaling pathway to link mGluR5 to ERK1/2 (27). Constitutively expressed Homer1b/c proteins (36,37) are preferentially co-localized with mGluR5 (38). The C-terminal coiled-coil domain of Representative immunoblots (IB) are shown above the quantified data of tyrosine-phosphorylated PP2A (pPP2A) and PP2A immunoreactivity analyzed from separate experiments (mean Ϯ S.E., n ϭ 9). Note that DHPG increased the level of tyrosine-phosphorylated PP2A, which was sensitive to MPEP, but not CPCCOet. PP2A in samples was precipitated through immunoprecipitation (IP) with rabbit antibodies against the PP2A catalytic subunit, and levels of tyrosinephosphorylated PP2A in PP2A precipitates were then detected through immunoblots with anti-phosphotyrosine antibodies. *, p Ͻ 0.05 versus basal levels. ϩ, p Ͻ 0.05 versus DHPG alone. the proteins promotes the association of mGluR5 with various signaling proteins (38 -41). Some of these linkages participate in the signaling function to transmit mGluR5 signals to ERK. Disruption of Homer1b/c-mGluR5 binding with a cell-permeable Tat peptide and reduction of cellular levels of Homer1b/c through a small interfering RNA approach reduced DHPGstimulated ERK1/2 phosphorylation in striatal neurons (27).
In addition to intracellular Ca 2ϩ and synaptic Homer proteins, the current study unraveled a novel PP2A-dependent mechanism involved in transmitting mGluR5 signals to MEK/ERK. Several lines of evidence support the role of PP2A. First, the PP2A inhibitors produced the same effect as the mGluR5 agonist in elevating basal levels of MEK/ERK phosphorylation. Moreover, the subcellular distribution pattern of increased pERK proteins induced by the PP2A inhibitor parallels that seen after mGluR5 stimulation (27). These results indicate a possibility that the mGluR5 activation may facilitate MEK/ERK phosphorylation via PP2A inhibition. Second, co-administration of the PP2A inhibitors with the mGluR5 agonist produced an additive increase in ERK phosphorylation. This indicates the participation of a signaling mechanism involving PP2A inhibition in mGluR5 activation of MEK/ERK. Third, a basal level of PP2A activity exists in striatal neurons, and PP2A activity was reduced in response to mGluR5 activation. This provides direct evidence for the inhibition of PP2A after mGluR5 activation. Fourth, selective physical association of PP2A with mGluR5 exists in cultured neurons. This is mainly supported by the results from the co-immunoprecipitation studies. Co-localization of PP2A and mGluR5 in most neurons detected with double immunofluorescent labeling seems to provide a morphological support for the occurrence of PP2A/ mGluR5 association at a cellular level. However, due to the limited spatial resolution of confocal microscopy, the association of PP2A with mGluR5 has to rely largely on the biochemical data from the co-immunoprecipitation studies. Fifth, in the efforts in additional mechanistic studies, we found that mGluR5 inactivated PP2A via enhancement of tyrosine phosphorylation of PP2A. The tyrosine phosphorylation was mediated by activation of a p60c-Src family of non-receptor tyrosine kinases, independent of the EGF receptor tyrosine kinase. Kinetically, the mGluR5 agonist produced a series of closely corresponded biochemical events, including tyrosine phosphorylation of PP2A, dissociation of PP2A from mGluR5, reduction of PP2A activity, and ERK phosphorylation. These results reveal a signaling model in which mGluR5 activation activates Src to facilitate tyrosine phosphorylation of PP2A, followed by the dissociation of PP2A from mGluR5 and the reduction of PP2A activity, leading to activation of MEK/ERK. Finally, it should be pointed out that PP2A is believed to regulate ERK indirectly via a mechanism involving MEK. Based on the findings that the PP2A inhibitor increased MEK1/2 phosphorylation and that the MEK1/2 inhibitor abolished ERK phosphorylation induced by the PP2A inhibitor, the upstream kinase MEK may serve as a control point for ERK signaling in response to PP2A action (42).
OA and calyculin A are the selective PP2A inhibitors, but they inhibit PP1 as well at the relatively high doses. To differentiate the relative importance of PP2A and PP1 for the OA and calyculin A activation of ERK, the effects of the two inhibitors were tested at a dose (0.01 M) that significantly inhibited PP2A, but not PP1. We found that the inhibitors strongly promoted MEK and ERK phosphorylation at this dose. Thus, the two inhibitors act by inhibiting PP2A rather than PP1 to protect pMEK and pERK from dephosphorylation. In support of this, we found the following: 1) the relative PP1 inhibitor tautomycin, at a dose that inhibited PP1 but not PP2A, did not alter the level of ERK phosphorylation, and 2) mGluR5 stimulation with DHPG selectively reduced PP2A but not PP1 activity. Furthermore, selective inhibition of PP1 with a novel PP1 inhibitor, tautomycetin, resulted in no ERK phosphorylation or even suppression of ERK phosphorylation in COS-7 cells (43). The mGluR5 agonist DHPG induced a rapid and transient ERK phosphorylation in cultured striatal neurons (27): ERK phosphorylation was seen at 2 min and returned to normal levels 30 -60 min after DHPG application, a dynamic event similar to that observed in other cell lines (15,16). As compared with DHPG, the PP2A inhibitor OA induced a relatively prolonged MEK/ERK phosphorylation. However, at 60 min, the magnitude of MEK/ERK phosphorylation was decreased. This seems to indicate an early cessation of the OA effect, despite the fact that OA usually produces a longer period of action (19). The early decline of the OA effect may suggest an activation of compensatory mechanisms involving some other OA-insensitive phosphatases, such as mitogen-activated protein kinase phosphatases (44,45). Further studies need to elucidate the interaction between PP2A and mitogen-activated protein kinase phosphatases in the late phase of this event.
As an important regulatory mechanism, the catalytic subunit of PP2A can be phosphorylated on Tyr 307 in vivo and in vitro by the p60c-Src non-receptor tyrosine kinase, the receptor tyrosine kinase (the EGF or insulin receptor), or the cytokine (interleukin-1 or tumor necrosis factor ␣) (32-35, 46 -48). The phosphorylation results in inactivation of the enzyme. Thus, concomitant transient inactivation of PP2A can be an accelerating factor for a given signaling pathway/cascade during the transmission of signals. Indeed, inactivation of PP2A as a result of enhanced tyrosine phosphorylation of the catalytic subunit of PP2A prolonged c-Jun phosphorylation in Rat-1 fibroblasts (49). Similarly, in this study, transient inactivation of PP2A was seen after mGluR5 activation as a result of increased tyrosine phosphorylation. The inactivation of PP2A contributes to the mGluR5-mediated MEK/ERK activation. Apparently, mGluR5 could activate MEK/ERK via coordinated activation of protein kinase and inactivation of protein phosphatase. Because the inactivation of PP2A was blocked by inhibitors selective for Src but not EGF receptors, a signaling mechanism involving Src activation is responsible for the PP2A inactivation. This is further supported by evidence for the interaction of Src with PP2A in a number of reports (50 -53). With regard to the relationship between the tyrosine phosphorylation of PP2A and the dissociation of PP2A from mGluR5, the two events were highly correlated to each other in their dynamic responses to mGluR5 stimulation. It is likely that the increased tyrosine phosphorylation via a Src-dependent mechanism may result in the dissociation of PP2A from mGluR5 because 1) Src formed a complex with mGluR5, and 2) blockade of Src activation prevented the PP2A dissociation from mGluR5.
Tyrosine phosphorylation of the catalytic subunit of PP2A is enhanced in the presence of the inhibitor OA (19), suggesting the following: 1) PP2A can reactivate itself in an auto-dephosphorylation manner under normal conditions, and 2) PP2A can act as a phosphotyrosine phosphatase (19,54) in addition to its apparent phosphoserine/threonine phosphatase. This study revealed a high level of basal PP2A activity in striatal neurons, which was also sensitive to the two inhibitors OA and calyculin A. Thus, the basal level of PP2A may be maintained by an active auto-dephosphorylation mechanism. This auto-dephosphorylation mechanism appears to operate independently from mGluR5 under normal conditions because the mGluR5 antagonist did not affect tyrosine phosphorylation of PP2A and basal PP2A activity (this study).