Protein phosphatase 2A activity associated with Golgi membranes during the G2/M phase may regulate phosphorylation of ERK2.

The extracellular signal-regulated kinase (ERK) 1 and 2 proteins are mitogen-activated protein kinase (MAPK) members that regulate cell proliferation and differentiation. ERK proteins are activated exclusively by MAPK kinase 1 and 2 phosphorylation of threonine and tyrosine residues located within the conserved TXY MAPK activation motif. Although dual phosphorylation of Thr and Tyr residues confers full activation of ERK, in vitro studies suggest that a single phosphorylation on either Thr or Tyr may yield partial ERK activity. Previously, we have demonstrated that phosphorylation of the tyrosine residue (Tyr(P) ERK) may be involved in regulating the Golgi complex structure during the G2 and M phases of the cell cycle (Cha, H., and Shapiro, P. (2001) J. Cell Biol. 153, 1355-1368). In the present study, we examined mechanisms for generating Tyr(P) ERK by determining cell cycle-dependent changes in localized phosphatase activity. Using fractionated nuclei-free cell lysates, we find increased serine/threonine phosphatase activity associated with Golgi-enriched membranes in cells synchronized in the late G2/early M phase as compared with G1 phase cells. The addition of phosphatase inhibitors in combination with immunodepletion assays identified this activity to be related to protein phosphatase 2A (PP2A). The increased activity was accounted for by elevated PP2A association with mitotic Golgi membranes as well as increased catalytic activity after normalization of PP2A protein levels in the phosphatase assays. These data indicate that localized changes in PP2A activity may be involved in regulating proteins involved in Golgi disassembly as cells enter mitosis.

The extracellular signal-regulated kinase (ERK) 1 1 and 2 proteins are important mediators of extracellular signals that promote cell proliferation and differentiation (1). ERK proteins regulate cell proliferation by targeting key proteins at specific times during the cell cycle. For example, receptor activation in response to mitogens causes ERK proteins to phosphorylate and activate transcription factors such as Elk-1, which promotes the expression of immediate early genes such as Fos and Egr-1, and increase cyclin D1 expression required for G 1 to S phase progression (2)(3)(4)(5). Activation of the ERK pathway during G 1 phase may also promote the synthesis of proteins required for the S phase by stimulating the eukaryotic initiation factor eIF4E during G 1 phase indirectly through the MAPK-interacting kinases (Mnk1 and Mnk2) (6,7).
Recent studies indicate that the ERK pathway is also required for regulating cell cycle progression during the G 2 phase and mitosis (M phase) (8 -15). ERK pathway proteins may regulate organelle disassembly and mitotic structures during G 2 /M phase transitions. For example, active ERK localization to the mitotic kinetochore may regulate proteins involved in chromosome segregation during metaphase to anaphase transitions (8,10). In addition, MKK1 activated by Raf-1 kinase has been reported to be involved in Golgi complex disassembly during early mitosis (16,17). Despite ERK1/2 being the only known substrates for MKK1 (1) and despite the fact that an ERK-like protein has been shown to associate with the Golgi complex in early mitosis (16), a role for ERK1 or ERK2 in regulating mitotic Golgi disassembly has not been clearly established. Because ERK1/2 are the only known substrates to be regulated through phosphorylation by MKK1/2, it is likely that any functions ascribed to MKK1 or MKK2 would require the ERK proteins. In vitro kinase assays show that ERK2 can phosphorylate GRASP55 (Golgi reassembly stacking protein 55) on the same site that gets phosphorylated in cells during mitosis (12). However, ERK2 phosphorylation of GRASP55 during normal progression through mitosis and the physiological consequence of phosphorylation in regulating the Golgi complex have yet to be determined.
Although it is recognized that full activation of ERK1/2 requires MKK1/2 phosphorylation of both threonine (Thr 183 ) and tyrosine (Tyr 185 ) (numbering according to a mouse ERK2 sequence; Swissprot P27703) residues within the TEY tripeptide activation site (18,19), in vitro studies suggest that a single phosphorylation on either the Thr or Tyr residues will confer partial enzyme activity (20). We have previously reported that ERK phosphorylated on a single tyrosine residue within the activation site (Tyr(P) ERK) transiently associates with the Golgi complex in the late G 2 phase and with Golgi vesicles throughout mitosis (11). In these studies, an ERK2 (Tyr 185 mutated to Phe 185 ) mutant inhibited Golgi fragmentation induced by a constitutively active MKK1 mutant (11). Given that a catalytically inactive ERK2 mutant, which could still be tyrosine-phosphorylated, did not block active MKK1-induced Golgi fragmentation, these results suggested that ERK proteins may affect mitotic events, such as Golgi fragmentation, through mechanisms that are dependent on ERK phosphorylation but independent of ERK activity. These findings are consistent with other reports indicating that ERK2 stimulates the activity of MAPK phosphatase-3 or topoisomerase II␣ through kinase-independent mechanisms (21,22).
The generation of a single phosphorylation within the TEY activation site of ERK, such as Tyr(P) ERK during the G 2 /M phase of the cell cycle, could occur by a single phosphorylation of unphosphorylated ERK by MKK1/2 or by threonine dephosphorylation of dually phosphorylated ERK. Given that there is no evidence that MKK1/2 will selectively phosphorylate ERK proteins on only one of the activation sites, an alternative possibility is that dually phosphorylated ERK is targeted by a specific phosphatase activity that results in a single phosphorylation within the TEY sequence. Tyrosine-phosphorylated ERK may be generated through the actions of protein phosphatase 2A (PP2A), which is a major regulator of ERK inactivation through threonine dephosphorylation (23). The possibility that PP2A mediates Tyr(P) ERK generation at the Golgi is further supported by evidence demonstrating a role for PP2A in regulating the phosphorylation of GM130, a protein whose phosphorylation status may regulate Golgi complex disassembly and reassembly as cells enter and exit mitosis, respectively (24).
In this study, we examined mechanisms involved in generating the tyrosine-phosphorylated form of ERK during G 2 /M phase transitions. Using dually phosphorylated ERK2 as a substrate, biochemical assays showed an increase in serine/ threonine phosphatase activity associated with Golgi-enriched membranes isolated from cells synchronized in the late G 2 / early M phase compared with G 1 phase cells. Using varying concentrations of the serine/threonine phosphatase inhibitors okadaic acid and fostriecin, our findings suggest that PP2A is the most likely phosphatase that associates with Golgi-enriched membranes isolated from the G 2 /M phase. These data suggest a potential mechanism for the observed localization of Tyr(P) ERK to the Golgi during mitotic transitions and further suggest that the phosphorylation status of ERK proteins depends on their subcellular location.

MATERIALS AND METHODS
Cell Culture-HeLa and Ptk 1 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and penicillin (100 units/ml)/streptomycin (100 g/ml) purchased from Invitrogen. HeLa cells were synchronized at the G 1 /S phase boundary by a double thymidine block. Briefly, cells that were ϳ50% confluent were treated with 2 mM thymidine for 16 h, washed with Hanks' buffered saline solution (Invitrogen), and released into complete medium for 8 h. The cells were then treated a second time for 16 h with 2 mM thymidine to generate cells that are synchronized at the G 1 /S phase boundary. G 1 /S phase synchronized cells were then washed with Hanks' buffered saline solution and incubated in complete medium for 7-8 h to obtain late G 2 /early M phase synchronized cells. We have previously characterized the synchronization of HeLa cells at the G 1 or G 2 /M phase using this methodology by flow cytometry (fluorescence-activated cell sorter) and biochemical analysis (13,25).
Golgi Membrane Enrichment-Synchronized Golgi-enriched fractions were prepared with the method of Misumi et al. (26) with some modifications. HeLa cells were grown in three 150-mm tissue culture dishes, synchronized in the G 1 /S or G 2 /M phases, and harvested with a cell scraper in 10 ml of phosphate-buffered saline, pH 7.2 (Invitrogen). The cells were collected by centrifugation at 5000 ϫ g for 5 min, resuspended in homogenization buffer (0.25 M sucrose, 10 mM Tris-HCl, pH 7.5, and protease inhibitors), and lysed with 20 passages through a 25-gauge needle. The homogenate was centrifuged at 600 ϫ g for 5 min to remove the nuclei and loaded into the bottom of an ultracentrifuge tube (UltraTube, 13 ϫ 51 mm; Nalgene, Rochester, NY) with an equal volume of 2.3 M sucrose in 10 mM Tris-HCl, pH 7.5. The sample was overlaid with 1.2 and 0.8 M sucrose solutions (1 ml each), and the sucrose gradient was centrifuged at 73,000 ϫ g for 2.5 h. Following centrifugation, the aliquots (0.4 ml) were taken sequentially from the top of the tube. As an alternative method, the cell lysates were loaded on top of a 5-25% (v/v) glycerol gradient and centrifuged at 150,000 ϫ g for 30 min, and the fractions were collected sequentially from the top of the gradient (27). Aliquots were removed from each fraction and mixed with an equal volume of 2ϫ SDS-PAGE sample buffer, and the proteins were separated and analyzed using SDS-PAGE and immunoblotting, respectively. The remaining samples were immediately frozen at Ϫ80°C for further analysis. Golgi enrichment was assessed by im-munoblotting fractions for the Golgi protein markers GM130 (BD Biosciences, San Jose, CA), and Golgin-97 (Molecular Probes, Eugene, OR). The presence of PP2A was determined by immunoblotting with antibodies directed against the 36-kDa PP2A catalytic subunit (Sigma; Santa Cruz Biotechnology, Inc., Santa Cruz, CA; Upstate, Waltham, MA). The protein concentrations were determined using Bradford Reagent (Sigma).
For some experiments, the Golgi-enriched fractions were salt-washed as described by Dirac-Svejstrup et al. (28) with some modifications. Briefly, equal aliquots of each fraction were incubated in sucrose buffer (0.2 M sucrose, 50 mM potassium phosphate, pH 7.0, 5 mM MgCl 2 , and protease inhibitors) in the presence or absence of 1 M KCl. The membranes were then pelleted by centrifugation at 100,000 ϫ g and resuspended in lysis buffer (20 mM Tris, pH 7.4, 137 mM NaCl, 2 mM EDTA, 10% glycerol, 1% Triton X-100, 0.1% SDS, and protease inhibitors) and stored at Ϫ80°C until use in phosphatase assays and immunoblotting.
Immunofluorescence-Cells were grown on glass coverslips and fixed in either 4% paraformaldehyde or cold methanol (Ϫ20°C) as described previously (11). The cells were then incubated with antibodies against Tyr(P) ERK, GM130, and PP2A followed by fluoroscein or Texas Redconjugated secondary antibodies. The cells were counterstained with 4Ј,6-diamidino-2-phenylindole to identify chromosomes. The digital images were captured using a Nikon E800 epifluorescence microscope (Image Systems Inc., Columbia, MD) and processed with IPLab Scientific Imaging Software (Scanalytics Inc., Fairfax, VA).
Phosphatase Assays-Approximately 15 g each of G 1 /S phase or G 2 /M phase Golgi fractions were incubated in the presence of dually phosphorylated active ERK2 (1 ng; New England Biolabs, Beverly, MA) in the kinase buffer (50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 1 mM EGTA, 2 mM dithiothreitol, and 0.01% Brij 35) at 30°C for 1, 15, 30, or 60 min. Aliquots were taken at each time point and analyzed by SDS-PAGE and immunoblotting. 10 M microcystin (Sigma) was used as a control to block serine/threonine phosphatase activity. The phosphatase inhibitors okadaic acid (0 -100 nM; EMD Biosciences, San Diego, CA) and fostriecin (50 nM; LC Laboratories, Woburn, MA) were preincubated with Golgi-enriched fractions for 20 min at 4°C, and in vitro phosphatase assays were performed for 45 min at 30°C with dually phosphorylated ERK2 as a substrate to assess inhibition of the serine/ threonine phosphatase activity. As a control, dually phosphorylated ERK2 was incubated with purified PP2A (Promega) for the times indicated above. The phosphatase reactions were stopped by adding an equal volume of 2ϫ SDS-PAGE sample buffer.
Phosphatase activity was assessed by immunoblotting for the presence of Tyr(P) ERK2 (E-4 antibody; Santa Cruz Biotechnology), ppERK2 (M8159 antibody; Sigma), and ERK2 (C-14 antibody; Santa Cruz Biotechnology). Densitometry software (NIH Imager; National Institutes of Health, Bethesda, MD) was used to quantify the amount of Tyr(P) ERK2 immunoreactive signal generated over time. A densitometry calibration curve was generated using bovine serum albumin of known concentrations as the standard, and the intensity of the Tyr(P) ERK2 signal was measured within the linear range and normalized to total ERK2 intensity.
PP2A Immunoprecipitations and Immunodepletions-PP2A was immunoprecipitated from 100 l of the Golgi-enriched fraction by incubation with 20 g of PP2A monoclonal antibody raised against the catalytic subunit (Upstate) for 2 h on ice. Forty l of a 50% slurry of protein G-Sepharose beads (Sigma) in phosphate-buffered saline was added to the mixture and incubated overnight at 4°C on a bidirectional rotator. The mixture was centrifuged to pellet the immune complexes and then washed three times with phosphate-buffered saline. To immunodeplete PP2A, the Golgi-enriched fractions were subjected to two additional rounds of incubation with PP2A antibody and protein G-Sepharose beads. The PP2A immunoprecipitates, PP2A immunodepleted fractions, and appropriate nondepleted PP2A controls were then used in the phosphatase assays and immunoblotting analysis.

Enrichment of Golgi Membrane Proteins-Previously
, we reported that ERK proteins phosphorylated only on the tyrosine residue within the TEY activation site transiently co-localized with Golgi complex proteins during the G 2 phase and mitosis (11). To assay enzymatic activities that may be involved in generating a tyrosine-phosphorylated form of ERK (Tyr(P) ERK) during this time of the cell cycle, we used gradient fractionation methods to isolate Golgi membrane-enriched fractions from cells synchronized at the G 1 /S boundary (referred to as the G 1 phase) or late G 2 /early M phase (referred to as the G 2 /M phase). The first method employed a discontinuous sucrose gradient as previously described (26) (see "Materials and Methods"). Immunoblotting for the Golgi proteins GM130 and Golgin-97 was used to identify the fractions that were Golgi-enriched. The highest levels of GM130 and Golgin-97 were evident in fraction 3 of G 1 /S phase synchronized cells (Fig.  1A), which was in agreement with previous reports using this method to enrich Golgi membranes (26). In the G 2 /M phase synchronized cells, the GM130 and Golgin-97 expression levels were enriched in fraction 3 (Fig. 1B). However, GM130 and Golgin-97 staining in G 2 /M phase cells also showed a somewhat more diffuse staining pattern across a broader range of the gradient (Fig. 1B). The diffuse association of these Golgi proteins with multiple gradient fractions is likely due to changes that are beginning to occur in the Golgi membrane density during preparation for mitosis as previously suggested (27).
Thus, an alternative isolation method was employed where Golgi membranes from the G 2 /M phase synchronized cells were enriched using a 5-25% continuous glycerol gradient as previously described (27). Immunoblotting for GM130 and Golgin-97 in the G 2 /M phase-specific Golgi-enriched fraction showed the highest expression levels in fraction 2 ( Fig. 1C), which is in agreement with previous studies using this methodology (27). A high level of Golgi protein expression was observed at the bottom of the gradient (fraction 7, Fig. 1C), which was previously attributed to intact Golgi membranes from interphase cells (27). Because these cells were harvested at a time corresponding to the late G 2 phase, when tyrosine-phosphorylated ERK is first observed (11), and early mitosis, it is expected that some cells contain Golgi membranes that are still largely intact.
Flow cytometry analysis of DNA content following staining with propidium iodide confirmed that the cells were highly synchronized in the G 1 (2 N DNA content) or the G 2 /M (4 N DNA content) phase (Fig. 1D). The fractions containing low or high levels of Golgi proteins from either the sucrose or glycerol gradient method for enriching Golgi membranes from both G 1 and G 2 /M synchronized cells were used in subsequent biochemical assays.
Increased Serine/Threonine Phosphatase Activity on Golgi Membranes Isolated during G 2 /M Phase-We next tested whether cell cycle-dependent changes in serine/threonine phosphatase activity could be responsible for accumulation of Tyr(P) ERK on the Golgi complex during the G 2 /M phase (11). Samples containing equal amounts of total protein from Golgienriched fraction 3 from G 1 phase synchronized cells or fraction 2 from G 2 /M phase synchronized cells (Fig. 1, A and C) were incubated with purified ERK2 that was dually phosphorylated on the threonine and tyrosine activation site residues (ppERK2). ERK2 phosphorylation status was determined over time by immunoblotting with phospho-ERK1/2 antibodies specific for the monophosphorylated tyrosine (Tyr(P) ERK2) or dually phosphorylated threonine and tyrosine residues (ppERK2) within the TEY activation site. As shown in Fig. 2A, significant Tyr(P) ERK2 immunoreactivity was observed over time after ppERK2 was incubated with Golgi-enriched fractions from G 2 /M phase as compared with G 1 phase synchronized cells. Correspondingly, the level of ppERK2 immunoreactivity decreased following incubation with Golgi-enriched membranes from the G 2 /M phase ( Fig. 2A). Similarly, incubations of ppERK2 with fraction 3 from G 2 /M-synchronized Golgienriched membranes isolated by the sucrose gradient ( Fig. 1B) also showed enhanced Tyr(P) ERK2 reactivity compared with fraction 3 of G 1 phase synchronized cells (data not shown). Given that Tyr(P) ERK2 immunoreactivity was blocked when microcystin was present in the phosphatase reaction ( Fig. 2A), these data indicate that serine/threonine phosphatase activity associated with Golgi membranes is increased during the G 2 /M phase of the cell cycle and can dephosphorylate the threonine residue on ppERK2.
We next determined whether the increased phosphatase activity associated with Golgi membranes toward ppERK2 was due to increases in the total cellular phosphatase activity at this time of the cell cycle. In these experiments, whole cell lysates containing equal amounts of total protein generated from G 1 or G 2 /M phase cells were incubated with ppERK2 for varying times, and ERK2 phosphorylation was determined by immunoblotting. As shown, the increase in Tyr(P) ERK2 immunoreactivity over time and the corresponding decrease in ppERK2 were similar in both G 1 and G 2 /M synchronized cells (Fig. 2B). These data indicated that the total serine/threonine phosphatase activity in the cell does not change significantly during the cell cycle, which is in agreement with previous studies (29). However, our findings suggested that localized serine/threonine phosphatase activity associated with the Golgi membranes is elevated during the G 2 /M phase as compared with the G 1 phase. These data are also consistent with a previous study indicating that the serine/threonine phosphatase PP2A is involved in regulating Golgi complex structure during mitosis (24).
We next tested the possibility that sucrose or glycerol used in the fractionation methods could cause an artificial inhibition of phosphatase activity in the G 1 phase fraction or phosphatase activation in the G 2 /M phase fraction. Glycerol (5% final concentration in the G 2 /M phase fractions) was added to the G 1 phase fraction containing high levels of Golgi proteins. Similarly, sucrose (final concentration, 200 mM in the G 1 fractions) was added to the G 2 /M phase fraction containing the highest level of Golgi proteins. The final concentrations for sucrose and glycerol were the same used in the fractionation and phosphatase experiments shown in Figs. 1 (A and C) and 2A. The adjusted fractions were incubated with ppERK2 for 60 min, and ERK2 phosphorylation was determined by immunoblotting. The addition of glycerol or sucrose to the fractions had no effect on the Tyr(P) ERK2 immunoreactivity observed after incubating ppERK2 with Golgi fractions isolated from G 1 or G 2 /M phase cells (Fig. 2C).
The phosphorylation and activity of ERK proteins is also regulated by tyrosine phosphatases (30,31), and complete dephosphorylation of ERK proteins may be through dual specificity phosphatases (31) or cooperative complexes consisting of serine/threonine and tyrosine phosphatases (32). Thus, the observed Tyr(P) ERK2 generation in the phosphatase assays could be an intermediate toward the complete dephosphorylation of ERK2. Therefore, the next experiments tested whether tyrosine phosphatase activity was also present in Golgi-enriched membranes during the G 2 /M phase. G 2 /M phase Golgienriched fractions were incubated in the absence or presence of the general tyrosine phosphatase inhibitor sodium orthovanadate (NaV), and phosphatase assays were performed using ppERK2 as a substrate. We predicted that if tyrosine phosphatase activity was present in the Golgi-enriched membranes, then the addition of NaV should cause a further increase the Tyr(P) ERK2 generation. However, as shown in Fig. 3A, the addition of NaV had no effect on Tyr(P) ERK2 immunoreactivity in G 2 /M phase samples compared with samples treated in the absence of NaV, which indicated that tyrosine phosphatase activity toward ERK2 was not increased on the Golgi membranes during the G 2 /M phase.
To verify that tyrosine phosphatase activity was present in these cells and could be measured by our methods, the postnuclear supernatant of G 2 /M phase cells was assayed for Tyr(P) ERK2 immunoreactivity treated with or without phosphatase inhibitors. Compared with untreated cells, the presence of NaV enhanced the Tyr(P) ERK2 immunoreactivity (Fig. 3A). As

FIG. 2. Increased serine/threonine phosphatase activity in Golgi-enriched fractions isolated from G 2 /M phase cells.
A, Golgi membrane-enriched fractions from G 1 or G 2 /M phase synchronized cells were incubated with purified ppERK2 (2 ng) for the indicated time points. B, the total post-nuclear supernatant from G 1 and G 2 /M synchronized cells was incubated with ppERK2 as in A. Following incubations, samples from A and B were immunoblotted for Tyr(P) ERK2, ppERK2, and total ERK2. The graphs below the immunoblots in A and B show the relative amounts of Tyr(P) ERK2 as determined by densitometry versus time. C, G 1 or G 2 /M phase Golgi-enriched fractions were incubated with 5% glycerol or 200 mM sucrose, respectively, prior to incubation with ppERK2 for 60 min. Immunoblotting shows the levels of Tyr(P) ERK2, ppERK2, and total ERK2 in the presence or absence of 10 M microcystin (MC). The results are representative of at least three separate experiments.
shown previously, Tyr(P) ERK2 generation was abolished by microcystin irrespective of the presence of NaV (Fig. 3A). These results suggest that although tyrosine phosphatases are active in the G 2 /M phase cells, this activity does not affect the Tyr(P) ERK2 generation.
PP2A Activity Is Increased in Golgi Membranes Isolated from G 2 /M Phase-Recent biochemical studies have characterized the in vitro activities of a number of serine/threonine phosphatases with respect to their activity toward phosphorylated ERK2 (30). Of these, PP2A shows significant activity toward ERK2 in vitro and has also been shown to regulate the ERK pathway in vivo (30,33). PP2A regulatory subunits have been shown to localize to the Golgi, and PP2A activity may be critical for Golgi complex reassembly late in mitosis (24,34). Therefore, the role of PP2A or a PP2A-related activity in generating increased Tyr(P) ERK2 immunoreactivity from dually phosphorylated ERK2 following incubation with Golgi-enriched fractions isolated from G 2 /M phase cells was examined.
Using an antibody against the catalytic subunit of PP2A, we first examined the pattern of PP2A expression in the cellular fractions isolated by the sucrose or glycerol gradient from G 1 or G 2 /M phase cells, respectively. The expression pattern for the Golgi protein marker GM130 in Fig. 4A was similar to that observed in Fig. 1 (A and C). In G 1 phase cells, some PP2A co-localized to fraction 3 containing GM130; however, most PP2A was found in fractions where GM130 was absent (Fig.  4A). In contrast, high levels of PP2A were found in the fractions that also contained the highest levels of GM130 when examining G 2 /M phase cells (Fig. 4A). The examination of Tyr(P) ERK2 generation from ppERK2 using these fractions demonstrated that the highest phosphatase activity associated with the fractions containing the highest levels of PP2A (Fig. 4, B and C). Because most of the phosphatase activity in G 2 /M phase cells localized to the fractions containing GM130 (Fig.  4C), whereas most of the phosphatase activity in G 1 phase cells associated with fractions lacking GM130 (Fig. 4B), these data indicated that a transient increase in PP2A association with Golgi membranes occurs during mitotic transitions.
To further demonstrate PP2A interactions with the Golgi membranes during mitosis, Golgi fractions were isolated from asynchronous and nocodazole-arrested mitotic cells using the glycerol gradient method. We postulated that if PP2A was associating with the Golgi membranes, PP2A phosphatase activity could still be measured after pelleting Golgi membranes, and high salt concentration could remove phosphatase activity. To test this, fraction 2, which contained the highest level of Golgi markers as shown in Fig. 1C, was diluted with or without 1 M KCl and the Golgi membranes were pelleted by ultracentrifugation. The Golgi membrane-enriched pellets were then subject to phosphatase assays and immunoblotting. As shown in Fig. 5A, increased ERK phosphatase activity was observed in mitotic cells probably as a result of increased PP2A expression that was still associated with the Golgi membranes after pelleting (Ϫ salt wash). The phosphatase activity could be block by the presence of fostriecin (FST) (Fig. 5A). In the salt-washed samples, PP2A levels and the corresponding phosphatase activities were reduced (Fig. 5A). The phosphatase activity in the asynchronous sample is likely due to the expected presence of some cells in G 2 /M phase. As expected, the expression of GM130 and ␣-tubulin could also be removed by salt washing because these proteins are not integral Golgi membrane proteins (Fig. 5B). The increased expression of GM130 in the fraction from mitotic cells (Fig. 5B) supports previous studies indicating that the glycerol gradient is the preferred method for isolating mitotic Golgi membranes (27). Thus, these data support the association of PP2A with mitotic Golgi membranes.
Next we compared the cellular locations of PP2A, Tyr(P) ERK, and Golgi proteins by immunofluorescence. Ptk 1 cells were used in these experiments because they retain a relatively large flat morphology during mitosis and are amenable for visualizing intracellular structures. Prophase cells, identified by the appearance of condensed chromosomes using 4Ј,6-diamidino-2-phenylindole staining, showed typical nuclear and perinuclear Golgi staining with the Tyr(P) ERK antibody (Fig.  6A) as we have previously described (11). Interphase cells only show Tyr(P) ERK staining in the nucleus (Fig. 6A). PP2A staining, although primarily cytoplasmic, showed enhanced staining at the perinuclear Golgi region during prophase, which partially overlapped with Tyr(P) ERK and the Golgi marker GM130 (Fig. 6). Thus, a portion of PP2A is in the area of the Golgi complex in early mitotic cells, consistent with its potential role in regulating protein phosphorylation and function at this intracellular location.
Pharmacological Inhibition of PP2A-A number of serine/ threonine phosphatase inhibitors have been used to differentiate between the activities of different phosphatases in vitro (35) as well as to examine potential PP2A phosphatase activity associated with the Golgi complex (24). Okadaic acid (OA) is a broad range inhibitor of serine/threonine phosphatases causing complete inhibition of PP2A and PP4 at ϳ1 nM, whereas the IC 50 of OA for PP1 is 10 -15 nM (35,36). Golgi-enriched membranes from the G 2 /M phase were incubated with ppERK2 in the presence of varying concentrations of OA, and threonine phosphatase activity was determined by immunoblotting for Tyr(P) ERK2. OA at concentrations between 1 and 2.5 nM completely eliminated the Tyr(P) ERK2 immunoreactivity observed in the untreated samples (Fig. 7A). Next, we tested the phosphatase inhibitor FST, which is a specific inhibitor of PP2A and PP4 at an IC 50 of 5 nM (36,37). In contrast, PP1 inhibition by FST has a reported IC 50 of 131 M and is therefore much less sensitive to inhibition (37). The presence of FST (50 nM) in G 2 /M phase Golgi-enriched membranes blocked phosphatase activity toward ppERK2 as demonstrated by the inability to detect Tyr(P) ERK2 immunoreactivity (Fig. 7B).
PP2A Immunoprecipitations and Tyr(P) ERK2 Generation-To test whether PP2A phosphatase activity could be isolated, the catalytic domain of PP2A was immunoprecipitated from G 2 /M phase Golgi-enriched fractions and incubated with ppERK2 in the presence or absence of FST. As predicted, Tyr(P) ERK2 generation was observed after incubating ppERK2 with the PP2A immunoprecipitates, and this effect was blocked by the presence of FST (Fig. 7C). Incubation of purified PP2A with ppERK2 gave a similar pattern of Tyr(P) ERK2 generation and indicated that a corresponding kinase activity was not involved in regulating ERK phosphorylation in these assays (Fig. 7D).
Because PP2A appeared to be the primary phosphatase activity associated with G 2 /M phase Golgi fractions and because this increased activity could be explained by increased PP2A protein levels associated with the Golgi (Fig. 4A), we next determined whether PP2A also had increased catalytic activity in G 2 /M phase cells. Golgi-enriched fractions from G 1 and G 2 /M phase cells were normalized to contain the same amount of PP2A, and the ability to generate Tyr(P) ERK2 following incubation with ppERK2 was examined by immunoblotting. As shown, Tyr(P) ERK2 generation was higher in the G 2 /M phase fraction, indicating that PP2A catalytic activity associated with the Golgi membrane is increased at this time of the cell cycle compared with G 1 phase (Fig. 8).
Immunodepletion of PP2A Inhibits Tyr(P) ERK2 Generation-To provide further evidence that PP2A is involved in dephosphorylating ppERK2 following incubation with G 2 /M phase Golgi-enriched fractions, PP2A was immunodepleted from Golgi fractions prior to incubation with ppERK2 in the phosphatase assay. Although complete immunodepletion of PP2A in our hands was difficult, ϳ50% reduction of PP2A could be achieved using multiple rounds of incubation with the PP2A antibody (Fig. 9A). As predicted, PP2A depleted Golgi-membrane fractions incubated with ppERK2 showed a corresponding decrease in the ability to generate Tyr(P) ERK2 as compared with non-PP2A-depleted control samples (Fig. 9B). These data suggest that increased PP2A targeting to the Golgi and activity could be responsible for generating Tyr(P) ERK during progression through mitosis. DISCUSSION In these studies we provide biochemical evidence that PP2A association with Golgi membranes is elevated during the late G 2 and early M phase of the cell cycle and may be involved in regulating localized ERK phosphorylation. Although PP2A has been implicated in regulating protein sorting through targeting of vesicles associated with the trans-Golgi network/endosome system (34,38,39), few studies have examined cell cycle-regulated functions for PP2A at the Golgi complex. One example for PP2A function during mitosis suggests that PP2A activity is required during telophase to facilitate dephosphorylation of the Golgi protein GM130 and promote reassembly of Golgi membranes in each of the daughter cells (24). Previously it was suggested that phosphorylation of GM130 inhibits interactions with the p115 vesicle tethering protein, which may contribute to Golgi complex disassembly during the late G 2 phase or early mitosis (40). Thus, these studies suggest that PP2A regulation of Golgi proteins is important during the later stages of mitosis. However, our data also suggested that PP2A associated with the Golgi membranes may be activated as cells are preparing to enter mitosis and regulate proteins such as ERK, which also associates with Golgi membranes during G 2 to M phase transitions (11). It is possible that localized PP2A is regulated in such a way to target certain substrates, such as ppERK, during mitotic entry and other substrates, such as GM130, as cells exit mitosis. PP2A substrate specificity and subcellular localization could be achieved through the interactions of its different regulatory subunits (41).
We showed by immunodepletion that PP2A is the likely phosphatase in G 2 /M phase Golgi-enriched fractions that is responsible for generating Tyr(P) ERK2 from dually phosphorylated ERK2 (Fig. 9). The inability to completely eliminate the phosphatase activity is most likely due to our inability to immunodeplete PP2A by much more than 50%. Nevertheless, the potential for other phosphatases with PP2A-like activity being involved in generating Tyr(P) ERK2 should be considered. A likely candidate for the involvement of another phosphatase would be PP4, formerly referred to as PPX, which shares 65% amino acid identity with PP2A and is found in the cytoplasm, in the nucleus, and in association with centrosomes (42). PP4 and PP2A are inhibited by similar concentrations of okadaic acid and fostriecin (35)(36)(37). However, several reasons indicate that PP4 is not responsible for the ERK phosphatase activity associated with the G 2 /M phase Golgi membranes. First, whereas PP4 may function in regulating microtubule formation and centrosome maturation in Drosophila and Caenorhabditis elegans during mitotic progression (43,44), this function for PP4 in mammalian cells has not been reported. PP4 reportedly localizes to the centrosome of mammalian cells during interphase and all phases of mitosis, except telophase (42). In contrast, PP4 localization to the centrosome in C. elegans has only been observed only during mitosis (43). Secondly, PP4 is probably not a directly regulator of ERK protein phosphorylation because previous studies indicate an inverse correlation between the two proteins. For example suppression of ERK activity correlates with an increase in PP4 expression (45).
Finally, PP4 appears to be more specific for other MAPKs, such as the c-Jun N-terminal kinases, or other substrates, such as NF-B-associated proteins, in response to inflammatory signals (46,47).
Our cell staining data suggested that PP2A is elevated in mitotic cells (Fig. 6), which could explain the increased phosphatase activity toward ERK at this time. However, we have observed using immunoblot analysis that total PP2A levels do not appear to change during the cell cycle (data not shown). This is in agreement with previous studies that examined PP2A levels in synchronized cells (29). One possibility to explain the cell staining data in Fig. 6 is that PP2A may undergo a conformational change during G 2 /M phase transitions that makes the antigenic epitope more accessible to the PP2A antibody. Antibody accessibility caused by changes in PP2A conformation would not be expected to interfere with PP2A detection using immunoblot analysis following SDS-PAGE. Moreover, Golgi-enriched fractions from G 2 /M phase synchronized cells retain higher phosphatase activity toward ppERK2 compared with G 1 phase cells containing equal amounts of PP2A (Fig. 8). Given that total cellular phosphatase activity toward ppERK2 is similar in G 1 and G 2 /M phase cells (Fig. 2B), our data indicated that PP2A redistributes to the Golgi membranes and increases in catalytic activity during G 2 /M phase transitions.
If PP2A activity on the Golgi complex generates tyrosine-phosphorylated ERK during the G 2 /M phase as suggested in these studies, then dually phosphorylated ERK must be used as the substrate. The source of dually phosphorylated ERK may be the nuclei of cells that are in early mitosis (8,10). However, future studies will need to demonstrate whether dually phosphorylated ERK proteins can be transported out of the nucleus. Alternatively, active MKK1, which has been shown to regulate mitotic Golgi fragmentation (16), may generate dually phosphorylated ERK, which is then is dephosphorylated on the threonine residue by PP2A. This process would have to occur rapidly because, to our knowledge, there is no evidence that dually phosphorylated ERK associates with the Golgi complex during G 2 /M phase transitions. Another possibility, which has yet to be demonstrated, is that active MKK1 at the Golgi during mitosis is only capable of phosphorylating the tyrosine residue on the ERK active site. It has been suggested that MKK1 may undergo as conformational change or even partial proteolysis during mitosis, which may affect its interactions with ERK (48,49). Future studies will be needed to verify that a conformational change in mitotic MKK1 occurs and affects ERK phosphorylation. In a previous study, we have shown that MKK1 does not likely undergo partial proteolysis during mitosis and that the reported proteolyzed lower molecular weight MKK1 previously reported (48) could be due to phospho-specific MKK1/2 antibody cross-reactivity with phosphorylated nucleophosmin (50), a multifunctional nucleolar protein involved in centrosome duplication and ribosome assembly (51,52). Lastly, tyrosine-phosphorylated ERK has been shown to localize in the nucleus of cells (53), where it could be targeted to the Golgi complex during G 2 /M phase transitions. The higher PP2A activity at this time of the cell cycle could maintain ERK proteins in the mono-tyrosine-phosphorylated form.
Recent studies have implicated tyrosine-phosphorylated ERK and ERK activators (MKK1) in the regulation of Golgi complex fragmentation during G 2 /M phase transitions (11,16,17,49). As cells enter mitosis, the interphase Golgi network, which is comprised of flattened stacks of cisternae, fragments into smaller stacks and eventually forms clusters of vesicles and tubules (28,29). These Golgi fragments organize around each of the spindle poles and reassemble into stacks in the daughter cells as cells exit mitosis during telophase (30). However, the mechanisms involved in regulating mitotic Golgi fragmentation remain largely unknown and somewhat controversial. In contrast to ERK and MKK1, the mitotic kinase Cdc2 has been reported to be the major regulator of mitotic Golgi fragmentation (40). Future studies will be needed to resolve this controversy; however, it is likely that MKK1, tyrosinephosphorylated ERK, and Cdc2 may all serve some function in regulating mitotic Golgi fragmentation.
Although the current studies provided a mechanism for generating a tyrosine-phosphorylated form of ERK at a specific cellular location during the cell cycle, the physiological function for this phosphorylated form of ERK remains unclear. A single phosphorylation within the active site of ERK may be used to adjust the magnitude and duration of enzyme catalytic activity in response to an extracellular (20,53) or, as suggested in the current study, intrinsic (e.g. cell cycle) stimuli. Phosphorylated ERK may also function in a kinase-independent manner regulating protein-protein interactions involved in maintaining intracellular organelles, such as the Golgi complex.