Positive Regulation of Raf1-MEK1/2-ERK1/2 Signaling by Protein Serine/Threonine Phosphatase 2A Holoenzymes*

Protein serine/threonine phosphatase 2A (PP2A) regulates a wide variety of cellular signal transduction pathways. The predominant form of PP2A in cells is a heterotrimeric holoenzyme consisting of a scaffolding (A) subunit, a regulatory (B) subunit, and a catalytic (C) subunit. Although PP2A is known to regulate Raf1-MEK1/2-ERK1/2 signaling at multiple steps in this pathway, the specific PP2A holoenzymes involved remain unclear. To address this question, we established tetracycline-inducible human embryonic kidney 293 cell lines for overexpression of FLAG-tagged Bα/δ regulatory subunits by ∼3-fold or knock-down of Bα by greater than 70% compared with endogenous levels. The expression of functional epitope-tagged B subunits was confirmed by the detection of A and C subunits as well as phosphatase activity in FLAG immune complexes from extracts of cells overexpressing the FLAG-Bα/δ subunit. Western analysis of the cell extracts using phosphospecific antibodies for MEK1/2 and ERK1/2 demonstrated that activation of these kinases in response to epidermal growth factor was markedly diminished in Bα knock-down cells but elevated in Bα- and Bδ-overexpressing cells as compared with control cells. In parallel with the activation of MEK1/2 and ERK1/2, the inhibitory phosphorylation site of Raf1 (Ser-259) was dephosphorylated in cells overexpressing Bα or Bδ. Pharmacological inhibitor studies as well as reporter assays for ERK-dependent activation of the transcription factor Elk1 revealed that the PP2A holoenzymes ABαC and ABδC act downstream of Ras and upstream of MEK1 to promote activation of this MAPK signaling cascade. Furthermore both PP2A holoenzymes were found to associate with Raf1 and catalyze dephosphorylation of inhibitory phospho-Ser-259. Together these findings indicate that PP2A ABαC and ABδC holoenzymes function as positive regulators of Raf1-MEK1/2-ERK1/2 signaling by targeting Raf1.

PP2A 3 is a major serine/threonine phosphatase implicated in the control of numerous cellular processes including metabolism, tran-scription and translation, ion transport, development, inflammation, cell growth, differentiation, and apoptosis (for reviews, see Refs. 1 and 2). Heterotrimeric PP2A holoenzymes are composed of a scaffolding/ structural subunit (A), a variable regulatory subunit (B), and a catalytic subunit (C). Thus far, four distinct regulatory subunit families have been identified: B or PR55 (2)(3)(4), BЈ or PR61 (5,6), BЉ or PR72 (7)(8)(9), and Bٞ or PR93/PR110 (10). Although the regulatory subunit families share little amino acid sequence homology, isoforms within each family exhibit significant sequence homology. The variable subunit plays a critical role in the control of PP2A by regulating substrate selectivity and/or directing the localization of the enzyme within the cell (11)(12)(13)(14)(15). Recent studies have also revealed a role for PP2A regulatory subunits in cell growth and apoptosis (16 -18), assembly and function of cytoskeletal proteins (14,19,20), and various cell signal transduction pathways (for a review, see Ref. 21).
Numerous protein kinases have been shown to form stable complexes with PP2A including p70 S6 kinase and p21-activated kinases (22,23), IB kinases (24,25), calcium-calmodulin-dependent protein kinase IV (26,27), protein kinase B/Akt (28), Janus-activated kinase (29), casein kinase II (30), and mitogen-activated protein kinase (MAPK) family members (31)(32)(33). The identification of multiple signaling modules containing both kinases and phosphatases suggests that the control of the respective signal transduction pathway is mediated via the interplay of these opposing enzymes. Because dephosphorylation of the associated protein kinase by PP2A has been linked to both positive and negative regulatory effects on kinase activation (24,26,27,32,33), the regulatory subunit likely plays an important role in controlling the function of these protein kinase-PP2A signaling complexes.
The MAPKs represent a large family of protein kinases that regulate diverse intracellular signal transduction cascades controlling cellular proliferation, differentiation, and apoptosis (34). Thus far, four well characterized families of mammalian MAPKs have been identified: the extracellular signal-regulated kinases (ERK1 and ERK2), c-Jun NH 2terminal/stress-activated protein kinases, p38 MAPKs, and ERK5 (for a review, see Ref. 34). Immunoprecipitation studies have revealed that PP2A associates with ERK2 and the MAPK kinase MEK1 (31) as well as with the kinase suppressor of Ras (33) and Raf1 (also known as c-Raf) (32), two upstream kinases in the prototypical MAPK signal transduction pathway (Raf1-MEK1/2-ERK1/2). Although PP2A has been reported to negatively regulate ERK1/2 activity (16,(35)(36)(37)(38), recent studies have demonstrated that PP2A can also function as a positive regulator of Raf1 (32, 39 -41) and kinase suppressor of Ras (33). The positive regulation of these kinases by PP2A is thought to be mediated via dephosphorylation of phosphorylated serine or threonine residues that inhibit kinase activity. Cell type-or species-specific regulatory effects on MAPK signaling have also been documented for PP2A regulatory subunits. For example, knock-down of the single Drosophila B family subunit results in ERK1/2 activation (16), whereas overexpression of a member of the B family of regulatory subunits, B␥, in a mammalian neuronal cell line leads to ERK1/2 activation (41). Furthermore the Caenorhabditis elegans B␣ homolog SUR-6 has been identified as a positive regulator of Raf1 and MPK-1 (the C. elegans ERK ortholog) (42).
Although the PP2A catalytic subunit has been identified in multiple protein kinase macromolecular complexes, the regulatory subunits in these complexes remain largely unknown. Additional studies are needed to determine the oligomeric composition of these complexes as well as the precise function of individual regulatory subunits. To elucidate the role of specific PP2A holoenzymes in MAPK signaling, we altered the cellular levels of B␣ or B␦ regulatory subunits in target cells by overexpression and RNAi approaches. Western analyses of the stable cell lines using phosphospecific antibodies revealed opposing effects on the phosphorylation state of ERK1/2 that were dependent upon overexpression or repression of the targeted B subunit protein. Our findings indicate that PP2A B␣-or B␦-containing holoenzymes control ERK1/2 activation via their interaction with and dephosphorylation of the upstream kinase Raf1, leading to sequential phosphorylation/activation of MEK1/2 and ERK1/2.
FLAG-tagged B Subunit Constructs-The FLAG epitope was added to the carboxyl terminus of B␣ and B␦ using a two-step PCR protocol. In the first reaction, PCR amplification of a 300-bp fragment was accomplished using the sense primer 5Ј-GTC ATG ACT GGA TCC TAC AAT-3Ј (which overlaps the BamHI site ϳ300 bp from the stop codon), the antisense primer 5Ј-GTC GTC CTT GTA GTC ATT AAT TTT GTC CTG-3Ј, and rat B␣/pcDNA5/TO as a template. The resulting PCR product served as a template for the second reaction using the same sense primer and the antisense primer 5Ј-GGC GGC CGC CTA CTT GTC ATC GTC GTC CTT-3Ј. Underlined nucleotides indicate BamHI and NotI restriction endonuclease sites in the sense and antisense primers, respectively. The second PCR product was subcloned into pGEM-T (Promega, Madison, WI), and the 300-bp B␣-FLAG fragment was excised with BamHI and NotI and ligated into B␣/pcDNA5/TO digested with BamHI and NotI. A similar strategy was utilized to epitope tag the B␦ subunit. Using rat B␦/pcDNA5/TO as template, the first PCR was performed with the sense primer 5Ј-ACT  GCC GCG GAG TTC CAC CCA-3Ј and the antisense primer 5Ј-GTC  GTC CTT GTA GTC ATT AAT TTT GTC CTG-3Ј; primers for the  subsequent PCR were the same sense primer and the antisense primer  5Ј-GGC GGC CGC TTA CTT GTC ATC GTC GTC CTT-3Ј. Underlined nucleotides indicate SacII and NotI restriction endonuclease sites in the sense and antisense primers, respectively. The second PCR product was cloned into pGEM-T, and the B␦-FLAG fragment was excised with SacII and NotI and ligated into B␦/pcDNA5/TO digested with the same enzymes. Generation of the FLAG-tagged BЈ␤ construct was as described previously (43). Proper construction of the plasmids encoding the FLAG-tagged B subunits was verified by automated DNA sequencing (Vanderbilt University DNA Core Facility).
B␣ RNAi Plasmid-For tetracycline-inducible RNAi, we used the pcDNA3.1/H1-TO plasmid described previously (18) in which two tetracycline operator (TO) sites flank the TATA box of the H1 promoter. The short hairpin RNA-encoding sequence targeting B␣ was synthesized as two complementary primers (forward, GAT CCC CGT GGC AAG CGA AAG AAA GAT TCA AGA GAT CTT TCT TTC GCT  TGC CAC TTT TTG GAA A; reverse, AGC TTT TCC AAA AAG  TGG CAA GCG AAA GAA AGA TCT CTT GAA TCT TTC TTT  CGC TTG CCA CGG G; B␣ complementary sequence underlined), phosphorylated, and ligated into pcDNA3.1/H1-TO digested with BglII and HindIII.
Inducible Cell Lines for B Subunit Overexpression or Knock-down-Human embryonic kidney (HEK) T-Rex cells (Invitrogen) were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 5 g/ml blasticidin. Monoclonal cell lines stably expressing tetracycline-inducible expression plasmids encoding B␣-FLAG, B␦-FLAG, or FLAG-BЈ␤ were generated by transfecting (FuGENE 6, Roche Applied Science) HEK T-Rex cells with B␣-FLAG/pcDNA5/TO, B␦-FLAG/pcDNA5/TO, or FLAG-BЈ␤/pcDNA5/TO followed by selection of individual clones in medium containing 175 g/ml hygromycin B. Monoclonal cell lines for tetracycline-inducible repression of B␣ were generated by transfecting HEK T-Rex cells with the B␣ RNAi vector followed by selection in medium containing 1 mg/ml G418. Control cell lines stably expressing pcDNA5/TO or pcDNA3.1/H1-TP were selected in medium containing 175 mg/ml hygromycin or 1 mg/ml G418, respectively. Expression of B␣-FLAG and B␦-FLAG and suppression of endogenous B␣ was accomplished by treating the respective cells with 1 g/ml tetracycline for 48 h at 37°C.
Reverse Transcription-PCR-Total RNA was obtained from tetracycline-treated HEK T-Rex cells using TRIzol reagent (Invitrogen) according to the manufacturer's protocol. RNA concentrations were determined by spectrometric analysis and adjusted to equal concentrations with Milli-Q H 2 O prior to cDNA synthesis. Synthesis of cDNA was accomplished using Superscript reverse transcriptase (Invitrogen) and ϳ1 g of total RNA. The resulting cDNA products were subjected to PCR amplification using primers specific for B␣, B␦, and GAPDH. PCR primers for the reactions were as follows: B␣ forward, 5Ј-GGA GGG AAT GAT ATT-3Ј; B␣ reverse, 5Ј-TAG TGT AGT AAC TGT AGT-3Ј; B␦ forward, 5Ј-GCG GGC GGC AAC GAC TTC-3Ј, B␦ reverse, 5Ј-TAG CGC CGT GAT CCT AAA-3Ј; GAPDH forward, 5Ј-GGT CAT CAT CTC TGC CCC-3Ј; and GAPDH reverse, 5Ј-CAC CAC CCT GTT GCT GTA G-3Ј. PCR amplification (30 cycles) was performed in a 50-l reaction containing 2.5 units of Amplitaq, 0.2 mM dNTPs, and 0.5 M primers; each cycle consisted of denaturation at 95°C for 30 s, annealing at 50°C (B␣/B␦) or 55°C (GAPDH) for 30 s, and extension at 72°C for 30 s.
Purification of AB␣C, AB␦C, and ABЈ␤C Holoenzymes-Stable HEK T-Rex cell lines harboring B␣-FLAG/pcDNA5/TO, B␦-FLAG/pcDNA5/TO, or FLAG-BЈ␤/pcDNA5/TO were treated with 1 g/ml tetracycline for 48 h to induce protein expression. Stable HEK T-Rex cell lines harboring the empty vector (pcDNA5/TO; EV) were used as controls. Cells were lysed in buffer containing 20 mM Tris-HCl, pH 7.6, 0.1% Igepal CA-630, 150 mM NaCl, 3 mM EDTA, 3 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 17 g/ml aprotinin, and 5 g/ml leupeptin. Following centrifugation for 15 min at 15,000 ϫ g, the clarified lysates (1 mg of protein) were incubated with 20 l of a 50% slurry of anti-FLAG-agarose for 4 h. Bound proteins were washed twice with PAN buffer (10 mM PIPES, pH 7.0, 100 mM NaCl, and 17 g/ml aprotinin) containing 0.5% Igepal CA-630, once with PAN buffer, and once in phosphatase assay buffer (25 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, and 0.25 mg/ml bovine serum albumin). Bound proteins were eluted by incubation for 30 min at 4°C in phosphatase assay buffer (150 l) containing 100 g/ml FLAG peptide. The amount of PP2A catalytic subunit in the purified PP2A holoenzymes was determined by SDS-PAGE and silver staining using serial dilutions of bovine serum albumin as standards. These values were used to calculate the protein concentration of purified PP2A holoenzymes based upon stoichiometric levels of the A, B, and C subunits in each preparation (see Fig. 1A). Aliquots of the purified holoenzymes were either assayed for phosphatase activity or subjected to SDS-PAGE followed by silver stain or immunoblot analysis.
Preparation of Phosphorylated Substrates and Phosphatase Assays-Radiolabeled substrates (e.g. phosphorylase a and cAMP-dependent protein kinase-phosphorylated histone H1) were prepared as described previously (14,44). Purified FLAG-B␣/␦ PP2A holoenzymes (9 ng) were assayed for phosphatase activity in a 50-l reaction containing phosphatase assay buffer and 32 P-labeled substrate. Following a 15-min incubation at 30°C, the reactions were terminated by the addition of trichloroacetic acid (20% final concentration). The samples were incubated on ice for at least 30 min, and proteins were pelleted by centrifugation at 13,000 ϫ g for 10 min. Supernatants were collected and quantified for 32 P i release by liquid scintillation counting. Less than 20% of total phosphorylated substrate was dephosphorylated during the assays, thus assuring that substrate was not limiting. In some experiments, purified PP2A holoenzymes were pretreated for 15 min with 2 nM okadaic acid, a concentration of inhibitor that selectively inhibits PP2A (45), prior to assay of phosphatase activity.
Kinetworks TM Phosphoprotein Analysis-Detergent-solubilized extracts (ϳ350 g of protein) from stably transfected HEK T-Rex cells were subjected to Kinetworks Phospho-Site Screening (KPSS) 1.3 and 4.1 as described on the Kinexus Bioinformatics Corp. website (www.kinexus.ca) and in Ref. 46. These screens use panels of 31-34 highly validated commercial phosphosite-specific antibodies and 20-lane multichannel blotters. The intensity of the ECL signals for the target protein bands on the Kinetworks immunoblots were quantified with a FluorS Max Imager and Quantity One software (Bio-Rad).
Gel Electrophoresis and Immunoblot Analysis-HEK T-Rex cells were cultured for 24 h in the presence of tetracycline and for an additional 16 h in serum-free medium containing tetracycline. As indicated, some cells were incubated with 50 ng/ml epidermal growth factor (EGF) for 10 min or 10% fetal bovine serum for 15 min prior to cell lysis. Cells (from one confluent 10-cm dish) were quickly washed two times with ice-cold phosphate-buffered saline and lysed in buffer A (20 mM Tris-HCl, pH 7.2, 2 mM EGTA, 5 mM EDTA, 30 mM NaF, 40 mM ␤-glycerophosphate, 20 mM sodium pyrophosphate, 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 3 mM benzamidine, 5 M pepstatin, 10 M leupeptin, and 0.5% Igepal CA-630). Clarified cell lysates were obtained by centrifugation at 12,000 ϫ g for 10 min. Cellular extracts (ϳ15 g of protein) were subjected to 10% SDS-PAGE, transferred to nitrocellulose, and immunoblotted with the indicated primary antibodies and corresponding secondary antibodies. Visualization and quantification of the immunolabeled proteins was accomplished using the Odyssey TM infrared imaging system (Li-Cor, Lincoln, NE) and Odyssey software, which measures integrated pixel intensity. Measurements for p-ERK1/2 or p-MEK1/2 were scaled to reflect the value observed for total ERK1/2 or MEK1/2 in each lane. After scaling, a two-sample t test (two-sided) was performed to determine the significance of the differences in the mean p-ERK1/2 and p-MEK1/2 signals in B␣/B␦-overexpressing or B␣ knock-down cell lines relative to control cells (EV).
Raf1 and FLAG-B Immunoprecipitations-Cells were grown in the presence of tetracycline for 24 h, serum-starved for 16 h, and then treated with or without EGF for 10 min immediately prior to harvest. Cells were lysed in buffer B (20 mM Tris-HCl, pH 8.0, 137 mM NaCl, 10% glycerol, 1% Igepal CA-630, 17 g/ml aprotinin, 10 M leupeptin, 1 mM phenylmethylsulfonyl fluoride, and 1 mM Na 3 VO 4 ) or buffer A. Raf1 was immunoprecipitated from the clarified cell lysates (500 g of protein) by incubation with Raf1 antibody (1 g; sc-133) for 1 h followed by subsequent incubation with protein G-Sepharose (20 l of a 50% slurry) for 90 min at 4°C. FLAG-B complexes were immunoprecipitated from the clarified cell lysates (500 g of protein) by incubation with anti-FLAG M2-agarose beads (20 l of 50% slurry) for 4 h at 4°C. Immune complexes were washed three times with the respective buffer, eluted in SDS sample buffer, resolved by 10% SDS-PAGE, and immunoblotted with the indicated antibodies.
Raf1 and ERK1/2 Dephosphorylation Assays-HEK T-Rex cells stably expressing empty vector were serum-starved for 16 h, treated with or without EGF for 10 min, and then lysed in buffer A. Raf immune complexes were isolated from the clarified cell extracts as described above. After washing with buffer A, the immune complexes were washed three times with phosphatase assay buffer lacking or containing 500 nM microcystin-LR and then incubated for 30 min at 37°C with the purified PP2A holoenzyme (ϳ90 ng) that had been pretreated with (ϩ) or without (Ϫ) 500 nM microcystin-LR. For experiments examining the dephosphorylation of phospho-Raf1 and phospho-ERK1/2 by purified AB␣C, AB␦C, or ABЈ␤C holoenzymes, Raf1 immune complexes were isolated from extracts of HEK cells grown under normal conditions and lysed in buffer A. ERK1/2 immune complexes were also isolated from these extracts by incubation with ERK1/2 antibody (1 g) and 20 l of a 50% slurry of protein G-Sepharose for 90 min at 4°C. After washing with buffer A and phosphatase assay buffer, the immune complexes were incubated with purified AB␣C, AB␦C, or ABЈ␤C holoenzymes (ϳ90 ng) for 30 min at 37°C. The phosphatase reactions were terminated by the addition of SDS sample buffer and subjected to immuno-blot analyses using antibodies recognizing phospho-Raf1(Ser-259), total Raf1, phospho-ERK1/2(Thr-202 ϩ Tyr-204/Thr-185 ϩ Tyr-187), and total ERK1/2. The blots were also immunoblotted with a PP2A catalytic subunit (PP2Ac) antibody to ensure equivalent amounts of phosphatase catalytic subunit were added to the dephosphorylation reactions.

RESULTS
Purification of PP2A Holoenzymes Containing FLAG-tagged B␣/␦ Subunits-A tetracycline-inducible system was utilized for controlled expression of FLAG epitope-tagged PP2A regulatory subunits in HEK T-Rex cells. B␣-FLAG or B␦-FLAG complexes were isolated from the respective extracts using anti-FLAG-agarose affinity resin, eluted from the resin with FLAG peptide (FLAG eluates), resolved by SDS-PAGE, and examined by immunoblot and silver stain analysis. The cellular levels of B␣-FLAG and B␦-FLAG proteins were ϳ2.6and 3.5-fold higher, respectively, than the total levels of endogenous B␣ and B␦ subunits (Fig. 1A). Similar results were obtained with multiple independent clones (data not shown). Consistent with previous reports (41,47), inducible overexpression of B␣ or B␦ did not alter expression of the A and C subunits (Fig. 1A). To determine whether the ectopic B subunits bind endogenous A and C subunits, the FLAG eluates were probed with specific antibodies recognizing the PP2A subunits or the FLAG epitope (Fig. 1A). Both A and C subunits were detected in the FLAG eluates from tetracycline-treated cells, whereas no PP2A subunits were observed in FLAG immune complexes isolated from lysates of control cells.
The relative purity and subunit stoichiometry of the isolated PP2A holoenzymes were addressed by SDS-PAGE and silver stain analysis. Three major proteins migrating at 65, 55, and 36 kDa (the predicted molecular masses of the A, B, and C subunits, respectively) were detected in the FLAG eluates (Fig. 1B). Analysis of these proteins by mass spectrometry verified their identity as the A, B␣ or B␦, and C subunits of PP2A (data not shown). The relative staining intensity of the A, B, and C subunits also indicated that these proteins were present in roughly stoichiometric levels in the FLAG eluates.
To test whether the isolated PP2A holoenzymes exhibited phosphatase activity, aliquots of the FLAG eluates were assayed for catalytic activity toward several phosphorylated substrates. AB␣C and AB␦C holoenzymes efficiently dephosphorylated all substrates and exhibited specific activities of 0.062 and 0.074 mmol/min/mg, respectively, when [ 32 P]histone H1 was used as a substrate. The catalytic activity of the purified holoenzymes was completely inhibited by 2 nM okadaic acid (data not shown), a concentration that selectively inhibits PP2A activity (45). Together with the immunoblot and protein staining analyses, these results demonstrate the utility of this system for the successful isolation of functional PP2A holoenzymes from cells expressing a FLAG-tagged PP2A regulatory subunit.
Knock-down of PP2A-B␣ by Inducible RNAi-As a complementary strategy to study the function of select PP2A holoenzymes, we exploited an RNAi strategy for knock-down of the B␣ regulatory subunit. HEK T-Rex cell lines inducibly expressing short hairpin RNA for RNAi-mediated degradation of B␣ mRNA were generated and analyzed for B␣ mRNA and protein levels following tetracycline induction. Reverse transcription-PCR was used to verify B␣ mRNA knock-down and to ensure experimental effects were specific for the B␣ subunit. Tetracycline-inducible knock-down of B␣ mRNA was observed in multiple clones harboring the B␣ RNAi plasmid but not in control cell lines expressing the empty vector ( Fig. 2A). Immunoblot analyses of cells expressing B␣-targeted RNAi revealed a greater than 75% reduction in   DECEMBER 30, 2005 • VOLUME 280 • NUMBER 52 B␣ subunit protein compared with control cells (Fig. 2B). The inducible knock-down of B␣, like the inducible overexpression of B␣ or B␦, did not alter the expression of the A and C subunits (see Figs. 1A and 7B). In comparison with the immunoblot analysis of cell lines expressing B␣/␦-FLAG (Figs. 1A and 2B), reverse transcription-PCR analysis of these cells revealed similar increases in the corresponding regulatory subunit transcript ( Fig. 2A).

PP2A Holoenzymes Regulate Raf1 Signaling
Effects of Alterations in B␣/␦ Subunit Levels on Cellular Phosphoproteins-To monitor signaling pathways targeted by AB␣C holoenzymes, we utilized two Kinetworks phosphoprotein analyses to examine the phosphorylation status of at least 59 phosphosites in over 50 known proteins (not including alternative splice variants) in control, B␣-overexpressing, and B␣ knock-down cells treated with or without EGF (Fig. 3). These screens apply phosphosite-specific antibodies to monitor the phosphorylation state of select proteins in cellular lysates. With 10-min EGF treatment, control cells exhibited dramatic increases in the phosphorylation status of numerous proteins including Raf1, MEK1/2, and ERK1/2 (Fig. 3, compare A and B). Several prominent changes in the phosphorylation status of these members of the MAPK pathway were observed in B␣ knock-down cells and B␣-overexpressing cells relative to control cells. In the absence of EGF, ERK2 showed a loss of phosphorylation (Ϫ70%) in the B␣ knock-down cells and an increase of phosphorylation (ϩ181%) in cells overexpressing B␣ (Fig.  3A). In the presence of EGF, overexpression of B␣ caused a 214% increase in ERK2 phosphorylation, whereas knock-down of B␣ resulted in a 43% decrease in ERK2 phosphorylation (Fig. 3B). A modest increase (ϩ20%) and decrease (Ϫ36%) in EGF-induced MEK1/2 phosphorylation was observed in B␣-overexpressing cells and B␣ knock-down cells, respectively. In contrast to MEK1/2 and ERK1/2, EGF-induced Raf1 (p70) Ser-259 phosphorylation was decreased (Ϫ32%) in cells overexpressing B␣ and increased (ϩ50%) in cells lacking B␣. The opposing effects on Raf1, MEK1/2, and ERK1/2 phosphorylation observed in cells with elevated B␣ levels versus cells with reduced B␣ levels suggest that these members of the MAPK pathway may be physiologic targets for B␣-containing PP2A holoenzymes.
Activation of MEK1/2 and ERK1/2 by AB␣C and AB␦C Holoenzymes-PP2A exhibits diverse regulatory roles in the control of various MAPK signal transduction pathways (16, 41, 48 -50). Based upon the Kinetworks analysis of our cell lines, we chose to further address the role of PP2A AB␣C and AB␦C holoenzymes in the regulation of Raf1-MEK1/2-ERK1/2 signaling. Stable cell lines with increased or decreased expression of B family subunits were serum-starved (basal phosphorylation) or serum-starved and then stimulated with fetal bovine serum. The phosphorylation states of MEK1/2 and ERK1/2 were assessed by immunoblot analysis of the resulting cell lysates using phosphosite-specific antibodies recognizing the active forms of these MAPK kinase and MAPK members (supplemental Fig. 1). In comparison to control cells, a significant increase in ERK1/2 (Thr-202 ϩ Tyr-204/Thr-185 ϩ Tyr-187) phosphorylation was observed in the cell lines overexpressing B␣ or B␦ under both basal and serum-stimulated conditions. In contrast, the phosphorylation of ERK1/2 was significantly decreased in B␣ knockdown cells, relative to control cells, under both conditions. Similar changes in the phosphorylation of MEK1/2 (Ser-217 ϩ Ser-221) were observed in these cell lines (data not shown).
Additional studies of the cell lines were conducted using EGF. In all three cell lines, EGF induced a transient phosphorylation of MEK1/2 and ERK1/2 peaking at 2 and 5 min, respectively (Fig. 4). However, in response to EGF (2, 5, 10, and 15 min), the cells with elevated levels of B␣ or B␦ exhibited increased phosphorylation of MEK1/2 and ERK1/2 as compared with control cells at the respective time points. This effect was most pronounced at 10 and 15 min of EGF treatment (Fig. 4, compare lane 10 with lanes 11 and 12 or lane 13 with lanes 14 and 15) at which times control cells showed near basal levels of MEK1/2 and ERK1/2 phosphorylation, whereas B␣and B␦-overexpressing cell lines still exhibited increased phosphorylation of these kinases. Quantitation of the 10-min EGF time points revealed a significant increase in ERK1/2 phosphorylation in the cell lines overexpressing B␣ or B␦ (2.97-or 2.83fold, respectively; p Ͻ 0.05; n ϭ 13) compared with control cells. Likewise MEK1/2 phosphorylation was significantly increased in the cell lines overexpressing B␣ or B␦ (4.12-or 3.67-fold respectively; p Ͻ 0.05; n ϭ 4) relative to control cells. In contrast, ERK1/2 phosphorylation was reduced by 30% in B␣ knock-down cells relative to control cells following a 10-min EGF treatment (Fig. 5, compare lane 6 with lane 5). Together these findings demonstrate increased phosphorylation of MEK1/2 and ERK1/2 in cells with elevated levels of B␣ or B␦ and decreased phosphorylation of these kinases in cells lacking B␣. Thus, PP2A AB␣C and AB␦C holoenzymes appear to positively regulate the activation of these MAPK kinases and MAPKs.
PP2A Holoenzymes Indirectly Influence the Phosphorylation State of MEK and ERK-To address the level at which B␣ and B␦ subunits exert their effects on the MAPK activation cascade, we first tested whether pretreatment of our cell lines with a MEK1/2 inhibitor (PD98059) influenced EGF-induced phosphorylation of MEK1/2 and ERK1/2. PD98059 reduced EGF-induced ERK1/2 phosphorylation in all cell lines (Fig. 5, FIGURE 5. AB␣C and AB␦C holoenzymes act upstream of MEK1/2 and ERK1/2. HEK T-Rex cells stably expressing EV, B␣-FLAG, B␦-FLAG, or B␣ siRNA (B␣-KD) were treated with tetracycline for 24 h and then serum-starved overnight in medium containing fresh tetracycline. Prior to harvest, the cells were incubated in serum-free medium lacking (Basal) or containing EGF (50 ng/ml) alone or EGF plus PD98059 (50 M) for 10 min; PD98059 was added 30 min prior to EGF treatment. Cell lysates were prepared and subjected to Western analysis using antibodies recognizing the indicated proteins. The data are representative of experiments that were performed at least three times. Cell lysates (ϳ15 g of protein) were resolved by SDS-PAGE and immunoblotted with antibodies recognizing B␣/␦, p-ERK1/2, total ERK1/2, p-MEK1/2, and total MEK1/2. The data are representative of experiments that were performed at least three times. B, graphical representation of the data shown in A. The p-ERK1/2 and p-MEK1/2 signals were normalized to total ERK1/2 and MEK1/2 levels, respectively; the -fold activations are relative to normalized values of unstimulated cells expressing EV, which were set as 1. DECEMBER 30, 2005 • VOLUME 280 • NUMBER 52 lanes 9 -12). Consistent with a previous report (51), we found that PD98059 did not completely abolish ERK phosphorylation in response to 50 ng/ml EGF. However, somewhat surprisingly, increased MEK1/2 phosphorylation was observed in all cells treated with the MEK inhibitor and EGF (Fig. 6, lanes 9 -12). The EGF/PD98059-induced hyperphosphorylation of MEK is most likely due to blockade of an inhibitory feedback loop, which requires ERK-dependent phosphorylation and inactivation of Raf1 (Ref. 52; also see "Discussion"). The presence of elevated, yet comparable, levels of MEK1/2 phosphorylation in EGF/ PD98059-treated cell lines overexpressing either the B␣ or B␦ subunit and the lack of any increases in MEK1/2 and ERK1/2 phosphorylation in the EGF/PD98059-treated cell lines lacking B␣ suggest that these MAPK family members are not themselves direct substrates for AB␣C and AB␦C holoenzymes. Instead these findings support the idea that the PP2A holoenzymes act upstream of MEK1/2 and ERK1/2 to activate this signal transduction pathway.

PP2A Holoenzymes Regulate Raf1 Signaling
AB␣C and AB␦C Holoenzymes Activate the MAPK Cascade Downstream of Ras-To further explore the target(s) of AB␣C or AB␦C holoenzymes within the MAPK cascade, we utilized a reporter assay to monitor MAPK-dependent transcriptional activation of Elk1 in our cell lines after transfection of oncogenic Ras V12 or constitutively active MEK1 (CA-MEK1). Both Ras V12 and CA-MEK1 stimulated MAPK reporter activity severalfold in all cell lines (Fig. 6). However, in comparison with the uninduced cell lines (ϪDox), the inducible expression of B␣ or B␦ (ϩDox) significantly augmented the Ras V12-induced MAPK activation at three different doses of Ras V12. In contrast, the inducible expression of B␣ or B␦ did not have any significant effect on CA-MEK1-induced MAPK activation. Furthermore the addition of doxycycline did not influence Ras V12-or CA-MEK1-induced reporter activity in the control cells, thus demonstrating that the augmentation of Ras V12-induced MAPK activation was specific to cells with elevated levels of B regulatory subunits. Together with results obtained with the MEK inhibitor (Fig. 5), these findings indicate that the AB␣C and AB␦C holoenzymes act downstream of Ras and upstream of MEK1/2 to promote MAPK activation.
Regulation of Raf1 Phosphorylation by PP2A Holoenzymes-It seemed counterintuitive to us that phosphatase overexpression would promote increased phosphorylation of target proteins. However, the upstream kinase of MEK1/2, Raf1, is tightly regulated by its phosphorylation state, and PP2A-dependent dephosphorylation of a Raf1 inhibitory residue (phospho-Ser-259) is critical for kinase activation (32,39,40). Thus, given the importance of PP2A activity in the control of Raf1 activity, we examined Raf1 Ser-259 phosphorylation in our cell lines with altered B family regulatory subunit levels (Fig. 7). Upon EGF stimulation, significant decreases in Raf1 Ser-259 phosphorylation were observed in cells with elevated B␣ or B␦ levels relative to control cells (Ϫ35 or Ϫ50%, respectively; p Ͻ 0.05; n ϭ 5); in contrast, increased phosphorylation of this site (ϩ17%) was observed in the B␣ knockdown cell line relative to control cells (Fig. 7A). A similar trend in Raf1 Ser-259 phosphorylation status was observed in the unstimulated cell lines. To confirm these findings, Raf1 was immunoprecipitated from the corresponding cell lysates and analyzed for phospho-Raf1 Ser-259. Similar to the analyses of the cell lysates, Ser-259 phosphorylation was decreased ϳ38 and 78% in the Raf1 immune complexes from EGFstimulated cells with elevated B␣ and B␦ levels, respectively (Fig. 7B). These results are consistent with our Kinetworks phosphoprotein analyses of EGF-stimulated cells (Fig. 3B) that revealed decreased (Ϫ32%) phosphorylation of Ser-259 in the B␣-overexpressing cell line and increased (ϩ50%) phosphorylation of this residue in the B␣ knock- HEK T-Rex cells were treated as described in the legend to Fig. 5. A, cell lysates (ϳ15 g of protein) were immunoblotted for phospho-Raf1 (Ser-259) and GAPDH. p-Raf1(Ser-259) phosphorylation was normalized to GAPDH levels. B, Raf1 immune complexes (Raf1 IPs) from lysates of cells prepared in buffer A were subjected to Western analysis using phospho-Raf1 (Ser-259) and total Raf1 antibodies. HEK T-Rex cells inducibly overexpressing B␣ or B␦ or cells selected for integration of the EV were cotransfected with MAPK reporter plasmids and the indicated amounts of expression plasmids (ng) for oncogenic Ras (Ras V12) or CA-MEK1. Following treatment with or without Dox for 2 days, MAPK activation was determined by dual luciferase assays. Activities are normalized to basal conditions (no activator plasmid) and plotted as means Ϯ S.D. of triplicate wells from a representative experiment. Ras V12-induced MAPK activation was significantly different in cells with elevated B␣ or B␦ levels (ϩDox) as compared with uninduced cells (ϪDox) when analyzed by Student's t test. *, p Ͻ 0.002. down cell line relative to control cells. Dephosphorylation of the Raf1 inhibitory site (phospho-Ser-259) and subsequent activation of this kinase would account for the phosphorylation/activation of the downstream kinases MEK1/2 and ERK1/2 in cell lines overexpressing B␣ or B␦. Likewise increased phosphorylation of Ser-259 in B␣ knock-down cell lines would explain the inactivation of MEK1/2 and ERK1/2 in these cell lines. Together these findings support the idea that PP2A holoenzymes dephosphorylate Raf1 Ser-259, leading to the phosphorylation and activation of the downstream kinases MEK1/2 and ERK1/2.
PP2A AB␣C and AB␦C Holoenzymes Associate with Raf1-To determine whether PP2A regulatory subunits modulate Raf1-MEK1/2-ERK1/2 signaling by targeting PP2A holoenzymes to Raf1, we performed a series of co-immunoprecipitation experiments. Western analysis of FLAG immune complexes from cells expressing B␣-FLAG or B␦-FLAG revealed that endogenous Raf1 together with the endogenous PP2Ac specifically co-precipitated with FLAG-tagged B subunits (Fig.  8A). In reciprocal immunoprecipitation experiments, FLAG-B␣ and FLAG-B␦ were found to co-precipitate with Raf1 in both the presence and absence of EGF stimulation; however, the associations of B␣ and B␦ with Raf1 were enhanced by 47 and 38%, respectively, following EGF treatment (Fig. 8B). PP2Ac also co-immunopurified with Raf1, and this interaction was markedly increased as well with EGF treatment. Furthermore in the presence of EGF, the amount of phosphatase catalytic subunit co-immunoprecipitating with Raf1 from lysates of B␣ knockdown cells was 40% less than the corresponding immunoprecipitation from control cell lysates (Fig. 8B, top panel, compare lane 8 with lane 2). In contrast, the amount of PP2Ac co-immunoprecipitating with Raf1 from lysates of cells overexpressing B␣/B␦ was increased ϳ35% as compared with the immunoprecipitation from control cell lysates (Fig. 8B,  top panel, compare lanes 4 and 6 with lane 2). Interestingly EGF stimulation of cells expressing empty vector or FLAG-B␣/␦ subunits produced a subtle upward shift in the migration of Raf1 on SDS-polyacrylamide gels (Fig. 8B, top panel). Because PP2A AB␣C and AB␦C appear to be recruited to Raf1 following EGF treatment, the slower migrating form of Raf1 is likely due to PP2A-mediated dephosphorylation of phospho-Ser-259 and subsequent phosphorylation of several activating residues in the kinase (e.g. Ser-338 and Ser-621). Consistent with this idea, no EGF-induced shift in the migration of Raf1 was observed in the B␣ knock-down clones, indicating that Raf1 Ser-259 dephosphorylation (and subsequent activating site phosphorylations) is prevented by the reduced levels of B␣ subunit. Together these data provide compelling evidence that B␣ and B␦ subunits target the PP2A holoenzyme to Raf1.

PP2A-mediated Dephosphorylation of Raf1
Ser-259-Because previous reports have established that phosphatase inhibitors block dephosphorylation of Raf1 Ser-259 and prevent maximal activation of the kinase (32,33,39,40), we tested whether phospho-Ser-259 is a direct target of PP2A. Raf1 immune complexes were isolated from lysates of target cells prepared in the presence of phosphatase inhibitors (i.e. to retain the phosphorylation state of Raf1 prior to dephosphorylation assays). After washing, the immune complexes were incubated with purified AB␣C or AB␦C holoenzymes that had been pretreated with or without the PP2A inhibitor microcystin-LR. Immunoblot analysis of the reaction mixtures revealed that both holoenzymes completely dephosphorylated phospho-Ser-259 in a microcystin-sensitive fashion (Fig.  9A). Prior cell treatment with EGF did not influence the ability of the purified holoenzymes to dephosphorylate phospho-Ser-259 in these in vitro dephosphorylation assays.
To determine the specificity of Raf1 Ser-259 dephosphorylation, a BЈ␤-containing PP2A holoenzyme was also tested for activity toward phospho-Raf1. In contrast to AB␣C and AB␦C, ABЈ␤C failed to dephosphorylate Ser-259 (Fig. 9B). Furthermore we tested whether these purified PP2A holoenzymes could dephosphorylate activating phosphosites in ERK1/2 (Thr-202 ϩ Tyr-204/Thr-185 ϩ Tyr-187); no detectable dephosphorylation of ERK1/2 was observed for AB␣C, AB␦C, or ABЈ␤C (Fig. 9C). Presumably the dephosphorylation of ERK1/2-activating sites in cells is catalyzed by one of the ERK-specific dual specificity phosphatases (35). The results of our phosphatase assays reinforce our conclusion that Raf1 is the preferred AB␣C and AB␦C substrate within the Raf1-MEK1/2-ERK1/2 pathway and also indicate that dephosphorylation of the Raf1 inhibitory site (phospho-Ser-259) is specifically mediated by these two holoenzyme forms of PP2A.

DISCUSSION
PP2A regulatory subunits play key roles in dictating the localization and substrate specificity of PP2A, thereby controlling PP2A activity toward numerous target proteins involved in cellular signaling events (11)(12)(13)(14)(15). However, the precise identity of these variable regulatory subunits has been determined in relatively few signal transduction pathways. Although genetic strategies to investigate the physiologic role of PP2A have been hampered by the inability to overexpress or knock down the phosphatase catalytic subunit in mammalian cells (53,54), our recent studies have revealed that PP2A regulatory subunits can be successfully overexpressed or repressed in mammalian cell lines (18,41,43,47). In this report, we used overexpression and RNAi approaches to experimentally alter the expression of select PP2A regulatory subunits in HEK cells. To identify the cellular pathways targeted by AB␣C and AB␦C holoenzymes, we initially exploited the Kinetworks screen for analysis of a wide range of phosphoproteins in our cell lines treated with or without EGF. Among the few proteins that exhibited opposing changes in their phosphorylation state in cells overexpressing versus cells lacking the B␣ subunit were the MAPKs (Fig. 3).
Numerous studies have implicated PP2A as a key regulator of the prototypical MAPK signal transduction cascade (Raf1-MEK1/2-ERK1/ 2). Although several reports have linked PP2A to the negative control of MAPK signaling (16,(35)(36)(37)(38), recent studies indicate that PP2A functions as a positive regulator of this pathway (32, 33, 39 -41, 52). To address these apparent contradictory findings and to elucidate the role of specific PP2A holoenzymes in the control of this pathway, we monitored the activation of Raf1-MEK1/2-ERK1/2 in cell lines with increased or decreased PP2A regulatory subunit levels. Our studies demonstrate that in response to cell stimuli (e.g. EGF and fetal bovine serum), there were increased phosphorylations of MEK1/2(Ser-217 ϩ Ser-221), ERK1(Thr-202 ϩ Tyr-204), and ERK2(Thr-185 ϩ Tyr-187) in cells with elevated B␣ or B␦ levels and reduced phosphorylation of these kinases in B␣ knock-down cells relative to control cells (Figs. [3][4][5]. In contrast to MEK1/2 and ERK1/2, the phosphorylation of Raf1 Ser-259 was decreased in cells with elevated B family regulatory subunit levels and increased in B␣ knock-down cells relative to control cells (Figs. 3 and 7). Because phosphorylation of Ser-259 is inhibitory with respect to Raf1 activity, our data lend additional support to the idea that PP2A functions as a positive regulator of Raf1-MEK1/2-ERK1/2 signaling by dephosphorylating phospho-Ser-259 resulting in Raf1 activation and consequential phosphorylation and activation of MEK1/2 and, in turn, ERK1/2 (33,41,52). Consistent with this hypothesis, our immunoprecipitation experiments indicate that the B␣ and B␦ regulatory subunits play an important role in targeting PP2A to Raf1 (Fig. 8). Although the B␣ subunit has been shown to associate with Raf1 in a growth factordependent manner (33), our results provide evidence that both B␣ and B␦ subunits interact with Raf1 even in serum-starved cells and that these interactions increase following growth factor treatment. Of particular importance, we now demonstrate that the novel B family subunit member B␦ also functions as a positive regulator of Raf1-MEK1/2-ERK1/2 signaling.
Multiple mechanisms control Raf1 activity including intra-and intermolecular protein interactions as well as inhibitory and activating phosphorylations (for a review, see Ref. 55). In support of the idea that Raf1 is subject to feedback inhibitory phosphorylations, recent studies have revealed that Raf1 activation is prolonged following pharmacological inhibition of MEK1/2 (51,52). Consistent with this hypothesis, we show hyperphosphorylation of MEK1/2 in our cell lines treated with EGF and the MEK1/2 inhibitor PD98059 (Fig. 5). Importantly the presence of elevated levels of MEK1/2 phosphorylations in EGF/PD98059-treated cell lines with increased B family regulatory subunit expression and the lack of any increases in MEK1/2 and ERK1/2 phosphorylations in EGF/ PD98059-treated B␣ knock-down cell lines indicate that these MAPK family members are not themselves direct substrates for AB␣C and AB␦C holoenzymes. If MEK1/2 and ERK1/2 were direct substrates for the PP2A holoenzymes, we would have expected to see decreased phosphorylation of these proteins in the cell lines with elevated B subunit levels and increased phosphorylation of these proteins in cells lacking B␣; however, this was not observed in our model system. Furthermore our studies of RasV12-and constitutively active MEK1-dependent MAPK activation indicate that AB␣C and AB␦C holoenzymes act downstream of Ras and upstream of MEK1 (Fig. 6). These observations together with our findings showing that both PP2A holoenzymes form stable interactions with Raf1 and directly dephosphorylate Raf1 Ser-259 (Figs. 8 and 9) demonstrate that the major target of AB␣C and AB␦C holoenzymes in the Raf1-MEK1/2-ERK1/2 pathway is in fact Raf1.
Dougherty et al. (52) recently described a feedback mechanism for Raf1 inactivation following PDGF stimulation that involves ERK1/2-dependent phosphorylation of six additional inhibitory sites in Raf1. Results from cellular experiments performed with the phosphatase inhibitor okadaic acid implicated PP2A in the dephosphorylation of these inhibitory residues and suggested that the hyperphosphorylated/ desensitized Raf1 is returned to a signaling-competent state via a PP2Adependent process. In support of a role for PP2A in the direct dephosphorylation of these newly identified inhibitory phosphorylation sites, we noticed a pronounced shift in the migration of Raf1 on SDS-polyacrylamide gels to a faster migrating species following incubation of phospho-Ser-259 Raf1 (ϪEGF) or hyperphosphorylated Raf1 (ϩEGF) with purified PP2A holoenzymes (Fig. 9A). Thus, in addition to phospho-Ser-259, PP2A appears to target other inhibitory phosphorylation sites and thereby plays a key role in the activation of Raf1.
The regulation of MAPK signaling by individual PP2A holoenzymes may be cell type-specific. For example, overexpression of the neuronspecific B family regulatory subunit B␥, but not B␣, leads to ERK1/2 activation in PC6-3 cells (a neuronal PC12 subline) (41). Furthermore we have recently reported a role for B␣/B␦ and BЈ family members in the modulation of ERK and Akt activation, respectively, in PC6-3 cells (43). Of note, PC12 cells predominantly express the neuronal B-Raf isoform, which may be regulated by phosphorylation/dephosphorylation in a manner distinct from Raf1 (56,57). Our present studies exploit the powerful combination of complementary overexpression and knockdown strategies to delineate the precise role of AB␣C and AB␦C holoenzymes in the control of Raf1-MEK1/2-ERK1/2 signaling in HEK cells. Our findings provide compelling evidence that B␣-and B␦-containing holoenzymes control activation of this pathway via their interaction with Raf1 and dephosphorylation of the inhibitory phospho-Ser-259 FIGURE 9. PP2A-mediated dephosphorylation of Raf1. A, Raf1 immune complexes from cells treated with (ϩ) or without (Ϫ) EGF were prepared in buffer A and then washed three times in phosphatase assay buffer containing or lacking 500 nM microcystin-LR (MC-LR). Dephosphorylation reactions were initiated by the addition of purified PP2A AB␣C or AB␦C holoenzymes (ϳ90 ng) that had been pretreated with (ϩ) or without (Ϫ) 500 nM microcystin-LR. B and C, Raf1 or ERK1/2 immune complexes isolated from extracts of cells grown under normal conditions were incubated with AB␣C, AB␦C, or ABЈ␤C holoenzymes (ϳ90 ng). FLAG eluate from EV-expressing cells was used as a control. Following a 30-min incubation at 37°C, the phosphatase reactions were terminated by the addition of SDS sample buffer and subjected to Western analysis using antibodies recognizing p-Raf1(Ser-259), total Raf1, p-ERK1/2, total ERK1/2, and PP2Ac. site, resulting in consequential activation of MEK1/2 and ERK1/2. Taken together with other recent reports (33,41,43), our studies indicate that specific PP2A holoenzymes target multiple steps in this important signal transduction cascade and appear to do so in a cell typespecific manner.
Additional physiologic targets for AB␣C and AB␦C holoenzymes likely exist in vivo. In addition to ERK1 and ERK2, the phosphorylations of numerous other proteins (e.g. CDK1, GSK3␣, GSK3␤, MKK3/6, protein kinase B, Rb, and cAMP-response element-binding protein) were altered by B␣ overexpression or knock-down (Fig. 3). However, in most cases, the phosphorylation of these proteins did not display dramatic opposing changes in cells with increased versus decreased PP2A regulatory subunit levels. Whether or not opposing changes in the phosphorylation of any protein are observed in these cells likely depends on the relative contributions of PP2A and the counteracting kinase on the steady state phosphorylation levels under a given stimulus condition. A more comprehensive analysis of the phosphoprotein levels in these cell lines over an extended EGF time course will allow us to test whether any of these proteins are true physiologic targets for AB␣C and/or AB␦C holoenzymes.
The tetracycline-inducible systems for overexpression or knockdown of select PP2A regulatory subunits in mammalian cells provide a means to identify physiologic substrates of closely related PP2A holoenzymes. Although B␣ and B␦ display ϳ90% amino sequence identity, these two subunits may differentially target the PP2A heterotrimer to specific substrates within the cell. Consistent with this idea, we previously demonstrated that B␣ and B␦ exhibit differences in their subcellular localization (13). However, our current findings indicate that these two PP2A regulatory subunits likely exhibit redundant functions in the control of Raf1-MEK1/2-ERK1/2 signaling because no major differences in the phosphorylation of these proteins were observed in B␣versus B␦-overexpressing cells. Nonetheless specific cellular functions for these B subunit family members may exist and might be revealed as we examine additional signal transduction pathways in cell lines with altered B␣ or B␦ levels.
The large number of protein kinases regulated by PP2A and the molecular mechanisms controlling protein kinase-PP2A signaling complexes are subjects of intense investigation. Given the presence of 14 genes coding for regulatory subunits that bind to the AC core dimer, more than 70 different PP2A heterotrimers are predicted to exist in mammalian cells. Our present studies define the role AB␣C and AB␦C holoenzymes play in activation of the Raf1-MEK1/2-ERK1/2 signal transduction cascade. Furthermore as reported herein, the development of cell lines that can be induced to overexpress or knock down specific regulatory subunits provides an elegant strategy to elucidate the functional role of specific PP2A holoenzymes in control of various biological processes.