I PP2A 1 Affects Tau Phosphorylation via Association with the Catalytic Subunit of Protein Phosphatase 2A

Abstract

In Alzheimer disease (AD) brain, the level of I ^PP2A₁, a 249-amino acid long endogenous inhibitor of protein phosphatase 2A (PP2A), is increased, the activity of the phosphatase is decreased, and the microtubule-associated protein Tau is abnormally hyperphosphorylated. However, little is known about the detailed regulatory mechanism by which PP2A activity is inhibited by I ^PP2A₁ and the consequent events in mammalian cells. In this study, we found that both I ^PP2A₁ and its N-terminal half I ^{PP2A(1–120)}₁, but neither I ^{PP2A(1–163)}₁ nor I ^{PP2A(164–249)}₁, inhibited PP2A activity in vitro, suggesting an autoinhibition by amino acid residues 121–163 and its neutralization by the C-terminal region. Furthermore, transfection of NIH3T3 cells produced a dose-dependent inhibition of PP2A activity by I ^PP2A₁. I ^PP2A₁ and PP2A were found to colocalize in PC12 cells. I ^PP2A₁ could only interact with the catalytic subunit of PP2A (PP2Ac) and had no interaction with the regulatory subunits of PP2A (PP2A-A or PP2A-B) using a glutathione S-transferase-pulldown assay. The interaction was further confirmed by coimmunoprecipitation of I ^PP2A₁ and PP2Ac from lysates of transiently transfected NIH3T3 cells. The N-terminal isotype specific region of I ^PP2A₁ was required for its association with PP2Ac as well as PP2A inhibition. In addition, the phosphorylation of Tau was significantly increased in PC12/Tau441 cells transiently transfected with full-length I ^PP2A₁ and with PP2Ac-interacting I ^PP2A₁ deletion mutant 1–120 (I ^PP2A₁ΔC2). Double immunofluorescence staining showed that I ^PP2A₁ and I ^PP2A₁ΔC2 increased Tau phosphorylation and impaired the microtubule network and neurite outgrowth in PC12 cells treated with nerve growth factor.

Reversible protein phosphorylation is an essential regulatory mechanism in many cellular processes. Although in the past much attention has been paid to the regulation of protein kinases, it is now apparent that protein phosphatases like the kinases are highly regulated enzymes that play an equally important role in the control of protein phosphorylation. Four main classes, protein phosphatase (PP)2 1, PP2A, PP2B, and PP2C, comprise the majority of cellular serine/threonine phosphatase activity (1). Among them, PP2A has been most implicated in the etiopathogenesis of Alzheimer disease (AD) (2–5).

In AD brain, the axonal microtubule associated protein (MAP) Tau is heavily phosphorylated due in part to decreased Tau phosphatase activity compared with the control brain (6). PP2A, which accounts for ∼70% of the total Tau phosphatase activity in human brain (5), regulates the phosphorylation of Tau (7–10). A decrease in the activity of the enzyme leads to hyperphosphorylation of Tau which, in turn, suppresses its microtubule binding and assembly activities in adult mammalian brain. Inhibition of PP2A also indirectly promotes the activities of several Tau kinases in brain, such as calmodulin kinase II, protein kinase A, MAP kinase kinase (MEK1/2), extracellular regulated kinase (ERK1/2), and P70S6 kinase (7, 11–13). Thus, the abnormal hyperphosphorylation of Tau that results from the inhibition of PP2A activity is probably due to not only a direct decrease in the dephosphorylation by PP2A but also an increase in the phosphorylation of Tau by Tau protein kinases that are regulated by PP2A. However, the mechanisms participating in the regulation of PP2A activity in AD brain are far from understood.

PP2A minimally contains a well conserved catalytic subunit, the activity of which is highly regulated. Regulation is accomplished mainly by members of a family of regulatory subunits, which determine the substrate specificity, subcellular localization, and catalytic activity of the PP2A holoenzymes (1, 14). Li et al. (15) reported the purification of a ∼30-kDa heat-stable protein, I ^PP2A₁, that specifically inhibits PP2A from bovine kidney. Molecular cloning of I ^PP2A₁ revealed it to be a 249-amino acid full-length protein from bovine kidney (16) and human brain (17), which was previously described as the human PHAPI (putative histocompatibility leukocyte antigen class II associated protein-I), mapmodulin, pp32, and LANP (18–23). I ^PP2A₁ inhibits purified preparations of trimeric PP2A, dimeric PP2A, and isolated catalytic C subunit of the phosphatase in a non-competitive manner (15, 17), and PP2Ac has been found in the immunoprecipitates of PHAPI/pp32 from human colon cancer cells (24). I ^PP2A₁ is involved in important physiological events, which include cell proliferation, apoptosis, mRNA transport, and transcription (25). Subcellular localization studies have shown that I ^PP2A₁ is a cytoplasmic/nuclear protein depending on cell type (17, 19). Our previous study showed a significant increase in the neocortical level of I ^PP2A₁ in AD as compared with control cases by in situ hybridization and immunohistochemical colocalization of this inhibitor with PP2A and with abnormally hyperphosphorylated Tau in the neuronal cytoplasm (26). However, the nature of the interaction of I ^PP2A₁ with PP2A and its role in the abnormal hyperphosphorylation of Tau were not understood.

The present study shows that the PP2A activity is dose-dependently inhibited by I ^PP2A₁ and that this decrease is not due to reduced PP2A expression level in NIH3T3 cells. Furthermore, GST-pulldown assay and coimmunoprecipitation from I ^PP2A₁ transiently transfected NIH3T3 cells show an interaction between I ^PP2A₁ and PP2A catalytic subunit, and no interaction between I ^PP2A₁ and PP2A A or B regulatory subunits. The minimal region required for the association with PP2Ac as well as PP2A inhibition is localized at the N-terminal isotype-specific containing region of I ^PP2A₁. In addition, transfection of Tau stably transfected PC12 cells with I ^PP2A₁ increases the phosphorylation of Tau at M4 (Thr-231/Ser-235), 12E8 (Ser-262/356), and Ser-404 sites, and microtubule network and neurite outgrowth are impaired.

EXPERIMENTAL PROCEDURES

Cloning and Generation of Plasmids—The full-length cDNA and deletion mutants of I ^PP2A₁ were designated as I ^PP2A₁FL (aa 1–249), I ^PP2A₁ΔC1 (aa 1–163), I ^PP2A₁ΔC2 (aa 1–120), I ^PP2A₁ΔC3 (aa 1-90), I ^PP2A₁ΔC4 (aa 1–45), I ^PP2A₁F2 (aa 46–90), I ^PP2A₁F2–4 (aa 46–163), and I ^PP2A₁F5 (aa 164–249) and obtained by PCR using pEGFP-N3/I ^PP2A₁ (wt) as a template (17). I ^PP2A₁FL was generated from primer 1 (5′-GATGGATCCATGGAGATGGGCAGA) and primer 2 (5′-GATCTCGAGTTAGTCATCATCTTC), I ^PP2A₁ΔC1 from primer 1 and primer 3 (5′-GATCTCGAGTTAGTAGCCCTCAGC), I ^PP2₁ΔC2 from primer 1 and primer 4 (5′-GATCTCGAGTTAAAGGTCTAAGCT), I ^PP2A₁ΔC3 from primer 1 and primer 5 (5′-GATCTCGAGTTAGAGGTTCGGACA), I ^PP2A₁ΔC4 from primer 1 and primer 6 (5′-GATCTCGAGTTATTCCAGTTCTTC), I ^PP2A₁F2 from primer 7 (5′-GATGGATCCATGTTCTTAAGTACA) and primer 5, I ^PP2A₁F2–4 from primer 7 and primer 3, I ^PP2A₁F5 from primer 8 (5′-GATGGATCCATGGTGGAGGGCCTG), and primer 2 by PCR separately. The BamHI site underlined in the forward primer and the XhoI site underlined in the reverse primer were used to clone these fragments into pGEX6P-1 vector (GE Healthcare Bio-Sciences Corp., Piscataway, NJ) and pCMV2B vector (Stratagene, La Jolla, CA). For cloning into C-terminal Myc tagged pcDNA3.1 vector (Invitrogen, Carlsbad, CA), the stop codon was removed. PP2A-PR65α was cloned into pGEX6P-1 vector by PCR using PR65α cDNA (a kind gift from Dr. Brian A. Hemmings) as a template and using primer 9 (5′-GATGGATCCATGGCGGCGGCCGAC, BamHI site underlined) and primer 10 (5′-GATCTCGAGTCAGGCGAGAGACAG, XhoI site underlined). The amplified cDNAs including the mutations were subcloned into the corresponding vectors and were verified by DNA sequencing.

Rat Brain Tissue Extract Preparation and GST Pulldown Assay—Rat brain was homogenized in buffer A (50 mm Tris·HCl, pH 7.6, 150 mm NaCl, 10 mm β-mercaptoethanol, 1 mm EDTA, 2 mm benzamidine·HCl, 1 mm phenylmethylsulfonyl fluoride, 100 μm leupeptin, 1 μm pepstatin, 2 μg/ml aprotinin) using a Polytron homogenizer and then centrifuged at 16,000 × g for 15 min at 4 °C. The supernatant was saved as brain extract. GST fusion proteins encoding I ^PP2A₁, its deletion mutants, and PP2A-A were purified according to the manufacturer's instructions (GE Healthcare). Recombinant GST or GST fusion proteins were incubated with glutathione-Sepharose 4B beads overnight at 4 °C and washed three times for 15 min at 4 °C with phosphate-buffered saline. The beads were equilibrated with binding buffer (20 mm Tris·HCl, 8.0, 150 mm NaCl, 1 mm EDTA, 0.5% Nonidet P-40, 2 mm benzamidine·HCl, 0.1 mm 4-(2-aminoethyl)benzenesulfonylfluoride hydrochloride, 100 μm leupeptin, 1 μm pepstatin, 2 μg/ml aprotinin). Rat Brain extract was precleared with glutathione-Sepharose 4B beads for 2 h at 4°C and then incubated with GST or GST fusion proteins bound to glutathione-Sepharose 4B beads. After incubation at 4 °C overnight, the beads were washed five times for 10 min in binding buffer, and proteins bound to the beads were eluted into sample buffer, followed by SDS-PAGE, and Western blots.

Cell Culture, Transfections, and Differentiation—NIH 3T3 cells (obtained from ATCC, Rockville, MD) were grown in 25-cm² flasks at 37 °C, containing 5% CO₂ in Dulbecco's modified Eagle's medium supplemented with 10% bovine calf serum. Cells were plated on the 6-well plates or 100-mm plates and transfected with expression plasmids using FuGENE 6 (Roche Applied Science, Indianapolis, IN).

PC12 cells or Tau 441 stably transfected PC12 cells were grown in 25-cm² flasks coated with poly-d -lysine at 37 °C, containing 5% CO₂ in RPMI medium 1640 (Invitrogen) supplemented with 10% bovine calf serum, 5% horse serum, and 1% penicillin (27). Cells were plated in 6-well plates or 8-well Lab-Tek II Chamber Slides coated with poly-d -lysine (Nalge Nunc International, Naperville, IL) and then transfected with expression plasmids using Lipofectamine 2000 according to the manufacturer's instructions (Invitrogen). For determining the induction of hyperphosphorylation of Tau by Western blots, cells were grown to 80–90% confluency, and for studying the number and the length of neurites on NGF differentiation, cells were grown to 30–50% confluency. PEGFP-N1 was employed to monitor the transient transfection efficiency, which was ∼30% when using 80–90% confluent cells and ∼2–5% in the case of 30–50% confluent cells. PC12 cells were differentiated into neurons by adding 100 ng/ml NGF to the culture medium for the indicated times.

Coimmunoprecipitation and Western Blots—For coimmunoprecipitation, NIH3T3 cells, transiently transfected with FLAG-tagged I ^PP2A₁ or vector as a control, were grown for 2 days, and lysed in coimmunoprecipitation lysis buffer (50 mm HEPES, pH 7.5, 150 mm NaCl, 1 mm EGTA, 10% glycerol, 1.5 mm magnesium chloride, 1% Triton X-100, 1 mm phenylmethylsulfonyl fluoride, 1 μg/ml leupeptin, and 50 units/ml aprotinin). After centrifugation at 16,000 × g for 15 min, the lysates were immunoprecipitated with anti-PP2Ac mouse monoclonal 1D6 (Millipore, Billerica, MA) followed by protein G-Sepharose (Pierce). After five final washings with lysis buffer, proteins were resolved by SDS-PAGE on 10% gels and transferred to polyvinylidene difluoride membrane according to the protocol of the manufacturer. Western blots of immunoprecipitates were developed with anti-FLAG (Sigma) or 1D6, followed by True Blue horseradish peroxidase-conjugated secondary antibody (eBioscience, San Diego, CA) detecting only native, undenatured IgG. Western blots of cell lysates were developed with anti-FLAG. Immunoreactive protein bands were visualized with enhanced chemiluminescence (ECL) reagents (Pierce). For detecting the C-terminal Myc-tagged I ^PP2A₁ and its deletion mutants, the samples were resolved by 15% SDS-PAGE and transferred to polyvinylidene difluoride at lower voltage and time. The primary antibodies used were mAb FLAG (Sigma, 1:1000), mAb 1D6 (Millipore, Billerica, MA, to PP2Ac, 1:1000), mAb β-actin (Sigma, 1:2000), pAb PP2A-A (Millipore, 1:1000), mAb 2G9 (Millipore, to PP2A-B, 1:1000), pAb GST (GE Healthcare Bio-Sciences Corp., 1:1000), mAb Myc (Cell Signaling Technology, Inc., Beverly, MA, 1:1000), Tau mAb M4 (to phosphorylated Thr-231/Ser-235, 1:2000 (28)), Tau mAb 12E8 (to phosphorylated Ser-262/356, 1:500 (29)), Tau pAb pS262 (BIOSOURCE, 1:1000), Tau pAb pS404 (BIOSOURCE, 1:1000).

PP2A Activity Assay—PP2A activity in immunoprecipitates of this enzyme was assayed using p-nitrophenyl phosphate as a substrate (phosphatase assay kit, Millipore). Briefly, the cells were lysed in coimmunoprecipitation lysis buffer (see above) at 4 °C. The lysates were centrifuged at 16,000 × g for 15 min at 4 °C, and the supernatants were incubated with 2.5 μg of anti-PP2A for 2 h, followed by incubation with protein G-agarose for 1 h at 4 °C. The immunoprecipitates were washed twice with lysis buffer, once with 50 mm Tris buffer (50 mm Tris, pH 7.5, 0.1 mm CaCl₂), resuspended in assay buffer (50 mm Tris, pH 7.5, 0.1 mm CaCl₂, 2.5 mm NiCl₂, and 1 mg/ml p-nitrophenyl phosphate), and incubated at 37 °C for 30 min. The reaction was stopped by the addition of 13% K₂HPO₄, and the absorbance was read at 405 nm.

Immunocytochemical Staining—To study colocalization of I ^PP2A₁ and PP2A, and Tau staining in PC12 cells, the cells were fixed with 4% paraformaldehyde in 0.1 m phosphate buffered saline. In the case of tubulin staining, the cells were rinsed once with warm (37 °C) phosphate-buffered saline and once with warm PEM (80 mm PIPES, 5 mm EGTA, 1 mm MgCl₂, pH 6.8), and then fixed with 0.3% glutaraldehyde containing 0.5% Non-idet P-40 in PEM for 10 min at 37 °C. After fixation, the cells were treated with 0.5% Triton X in Tris-buffered saline for 10 min, blocked with 4% normal horse serum containing 0.1% Tween 20 in phosphate-buffered saline for 30 min, and immunostained with primary antibodies, followed by incubation with Alexa 488-conjugated goat anti-mouse antibody (1:1000) and Alexa 594-conjugated goat anti-rabbit antibody (1:1000) (Invitrogen). The primary antibodies used were mAb 1D6 (to PP2Ac, 1:100), pAb Myc (1:400), mAb M4 (to phosphorylated Thr-231/Ser-235, 1:1000), mAb DM1A (to alphatubulin, 1:2000) (Sigma), rabbit affinity-purified antibody R-42089 to a synthetic peptide corresponding to amino acid residues 10–23 of I ^PP2A₁ (30). Representative images from several cells were generated using a PCM 2000 Confocal Imaging System (Nikon, Melville, NY) or an Act-2U Digital Imaging System (Nikon, Melville, NY). The length of the neurites borne by randomly selected 30 differentiated PC12 cells were measured by means of an image analysis system (Nikon), using Image J software without knowledge of the transgenes.

RESULTS

Inhibition of PP2A by I ^PP2A₁ in NIH3T3 Cells—I ^PP2A₁ is known to be a multifunctional protein that, besides inhibiting PP2A, binds to hypoacetylated over hyperacetylated histones and plays a signaling role in integrating histone hypoacetylation to gene inactivation (31–33). We investigated whether I ^PP2A₁ decreased PP2A activity directly or through inhibiting PP2A expression level in NIH3T3 cells transiently transfected with different amounts of FLAG-tagged I ^PP2A₁. We found that in FLAG-tagged I ^PP2A₁-transfected cells, the expression level of I ^PP2A₁ as well as the inhibition of PP2A activity increased in a dose-dependent manner, whereas, PP2A expression level was not significantly altered (Fig. 1). These data suggest that in cells PP2A is inhibited by I ^PP2A₁ in a dose-dependent manner, and that this decrease in the phosphatase activity is not due to the reduction of PP2A expression level.

FIGURE 1.

Inhibition of PP2A activity by I ^PP2A₁ in cultured cells. A, NIH3T3 cells were transiently transfected with different amounts of FLAG-tagged I ^PP2A₁ or mock DNA (vector) as a control. After 48-h transfection, cells were lysed and expression of PP2Ac, I ^PP2A₁, and β-actin were determined by Western blots. B, the relative PP2Ac expression level was normalized by the expression of β-actin. PP2Ac expression level did not show any significant difference among different levels of I ^PP2A₁ transfection in NIH3T3 cells. C, inhibition of PP2A activity by I ^PP2A₁. NIH3T3 cells were transiently transfected with different amounts of FLAG-tagged I ^PP2A₁ or with mock DNA (vector) as a control. After 48-h transfection, cells were lysed and PP2A was immunoprecipitated with mAb 1D6 to PP2A, and the phosphatase activity was assayed colorimetrically using p-nitrophenyl phosphate as a substrate. I ^PP2A₁ inhibited the PP2A activity in a dose-dependent fashion. Error bars indicate means ± S.E.; n = 3. *, p < 0.05 compared with control.

Characterization of Interaction of PP2A with I ^PP2A₁—PP2A consists of a heterotrimer that exists in multiple forms. The core components of all trimeric forms are 36-kDa catalytic subunit (PP2Ac) and the 65-kDa regulatory subunit (A subunit, PR65). This core heterodimer is ubiquitous, and it forms complexes with “variable” B subunits (some of which are expressed in a tissue- and/or in development restricted manner) and confers distinct properties to the enzyme (14). I ^PP2A₁ was shown to inhibit PP2A activity in vitro and could interact with the C subunit of PP2A (15, 24). It was of interest to determine whether I ^PP2A₁ can only bind to PP2A C subunit and the association is sufficient for inhibition of PP2A activity. To study the interaction between PP2A and I ^PP2A₁, in vitro pulldown assays were carried out using bacterially expressed GST alone, GSTPR65α (PP2A-A), or GST-I ^PP2A₁. Rat brain extract, which contains PP2A holoenzyme, was incubated with purified GSTPR65α, GST-I ^PP2A₁, or GST alone on glutathione-Sepharose 4B beads, and subsequently PP2A subunits on beads were detected by Western blots. We found that I ^PP2A₁ interacted with PP2A catalytic subunit (PP2Ac). However, no associations between I ^PP2A₁ and PR65α (PP2A A regulatory subunit) or PR55α (PP2A B regulatory subunit) were observed (Fig. 2A). The interaction between I ^PP2A₁ and PP2Ac was further confirmed by coimmunoprecipitation. NIH3T3 cells were transiently transfected with FLAG-tagged I ^PP2A₁ for 48 h, and then the cell lysate was collected, and endogenous PP2Ac was immunoprecipitated with 1D6 antibody to PP2Ac. The immunoprecipitated PP2Ac complex was resolved by SDS-PAGE, and Western blots were developed with anti-FLAG antibody. We found that the PP2Ac coimmunoprecipitated with FLAG-tagged I ^PP2A₁ but not vector (Fig. 2B). In addition, PP2Ac and I ^PP2A₁ were found to colocalize in PC12 cells double labeled with antibodies to PP2A catalytic subunit and anti-I ^PP2A₁ (Fig. 2C). These findings suggest that I ^PP2A₁ interacts with PP2A catalytic subunit and not with PP2A regulatory subunits PP2A-A or PP2A-B.

FIGURE 2.

Interaction between I ^PP2A₁ and PP2A. A, GST pulldown assay. Rat brain extract (Ext), used as a source of PP2A holoenzyme, was incubated with Sepharose 4B beads bearing GST, GST-I ^PP2A₁, or GST-PP2A-A. After washing, bound PP2Ac, PP2A-B, and PP2A-A were detected by Western blots. The GST, GST-I ^PP2A₁, or GST-PP2A-A used in pulldown assay are shown in the lowest panel. GST-I ^PP2A₁ fusion protein pulled down PP2Ac (PP2A catalytic subunit) but not PR55α (PP2A Bα regulatory subunit) or PR65α (PP2A A regulatory subunit). From left to right, lane 2 (GST/Ext.) is a negative control, and lane 4 (Ext.) is input. Positions of protein size markers are indicated on the left of the panel. B, association of PP2Ac with I ^PP2A₁ was confirmed by in vitro coimmunoprecipitation. Immunoprecipitation was carried out using mAb to PP2Ac in lysates of FLAG-tagged I ^PP2A₁ transiently transfected NIH3T3 cells. I ^PP2A₁ was detected using antibodies against FLAG (shown in the top and bottom panels). The middle panel depicts the amount of immunoprecipitated PP2Ac. C, colocalization of PP2A catalytic subunit and I ^PP2A₁ in PC12 cells. PC12 cells were doubly stained with monoclonal anti-PP2Ac and polyclonal I ^PP2A₁ (R42089), followed by, rhodamine-conjugated anti-rabbit IgG (panel a), fluorescein-labeled anti-mouse IgG (panel b), and additional staining of nuclei with 4′,6-diamidino-2-phenylindole (DAPI, panel d). The colocalization of the two is demonstrated by the yellow color in the merged image (panel c). Scale bar = 10 μm.

Domains of I ^PP2A₁ Involved in Its Binding to PP2A Catalytic Subunit—The structural feature of I ^PP2A₁ is characterized by four leucine-rich repeats (including the first 25-amino acid subtype-specific N-terminal region), a putative nuclear localization signal, and a long stretch of acidic amino acids at C-terminal ends (18, 22). To define the binding domain as well as PP2A inhibition activity, the full-length I ^PP2A₁ protein and, based on its structure, a series of deletion mutants I ^PP2A₁ΔC1, I ^PP2A₁ΔC2, I ^PP2A₁ΔC3, I ^PP2A₁ΔC4, I ^PP2A₁F2, I₁^PP2AF2–4, and I₁^PP2AF5 were generated (Fig. 3A). In vitro pulldown assays were carried out with bacterially expressed GST fusion I ^PP2A₁ full-length protein or its deletion mutants. Rat brain extract as a source of PP-2A holoenzyme was added to purified GST fusion proteins or GST alone bound to glutathione Sepharose 4B beads and then PP2A catalytic subunit that associated with beads was detected by Western blots. We found that the full-length I ^PP2A₁ and I ^PP2A₁ deletion mutants containing the N-terminal isotype-specific region (I ^PP2AΔC2, I ^PP2A₁ΔC3, and I ^PP2A₁ΔC4) except I ^PP2A₁ΔC1 I ^PP2A₁ΔC1 had the capability to bind with PP2Ac, whereas, the other I ^PP2A₁ mutants without N-terminal isotype-specific region (I ^PP2A₁F2, I ^PP2A₁F2–4, and I ^PP2A₁F5) lost the PP2Ac binding ability (Fig. 3B). Thus, the minimal region required for the binding of PP2Ac is localized at N-terminal region (amino acid 1–45) in I ^PP2A₁.

FIGURE 3.

Interaction between PP2Ac and various domains of I ^PP2A₁. A, schematic diagram of the constructs of functional domains of I ^PP2A₁ employed for the interaction studies. B, rat brain extract as a source of PP-2A holoenzyme was incubated with Sepharose 4B beads bearing GST, GST-I ^PP2A₁, or GST-I ^PP2A₁ deletion mutants. The pulldown assay was carried out using 0.5 μg of GST fusion protein per mg of brain extract, except in cases of GST-I ^PP2A₁F2, and GST alone, double the amount, i.e. 1 μg was employed. After washing, bound PP2Ac was detected by Western blots (upper panel). The GST, GST-I ^PP2A₁, or GSTI ^PP2A₁ deletion mutants used in the pulldown assay are shown in the lower panel. The I ^PP2A₁ constructs, which included the N-terminal (aa 1–45) isotype-specific region, except construct I ^PP2A₁ΔC1, interacted with PP2Ac. The right panel shows that, even in overloaded blots, GST-I ^PP2A₁F2 does not show interaction with PP2Ac. Positions of protein size markers are indicated on the left for the left panel and on the right for the right panel. C, NIH3T3 cells were transiently transfected with C-terminal Myc-tagged I ^PP2A₁, its deletion mutants, or empty vector as a control. After 48-h transfection, transfected Myc-tagged I ^PP2A₁ and its mutants were detected using antibodies against Myc from same amounts of lysates. Positions of protein size markers are indicated on the left. D, relative PP2A activity in transiently transfected NIH3T3 cells with C-terminal Myc-tagged constructs was detected as described in Fig. 1C. I ^PP2A₁FL reduced PP2A activity to ∼70%, whereas I ^PP2A₁ΔC2, I ^PP2A₁ΔC3, and I ^PP2A₁ΔC4 inhibited PP2A activity to ∼50%, 67%, and 76% in NIH3T3 cells, respectively, and I ^PP2A₁F5 had no detectable effect on the activity. *, p < 0.05 compared with the vector transfection control.

To measure the inhibition of PP2A by I ^PP2A₁ mutants, I ^PP2A₁FL, I ^PP2A₁ΔC2, I ^PP2A₁ΔC3, I ^PP2A₁ΔC4, and I ^PP2A₁F5 were chosen to characterize the PP2A activity. NIH3T3 cells were transiently transfected with Myc-tagged-I ^PP2A₁FL, I ^PP2A₁ΔC2, I ^PP2A₁ΔC3, I ^PP2A₁ΔC4, or I ^PP2A₁F5 for 48 h, followed by PP2A activity assays as described above in Fig. 1. We found that compared with vector alone, I ^PP2A₁FL, I ^PP2A₁ΔC2, I ^PP2A₁ΔC3, and I ^PP2A₁ΔC4 reduced PP2A activity to ∼70%, ∼50%, 67, and 76%, respectively. However, I ^PP2A₁F5 did not inhibit PP2A activity in this assay (Fig. 3, C and D). Therefore, in subsequent studies, only I ^PP2A₁FL and I ^PP2A₁ΔC2 were employed.

Effect of I ^PP2A₁ on Tau Phosphorylation, Microtubule Assembly, and Neurite Outgrowth—There is accumulating evidence that reduced PP2A activity is sufficient to induce Tau hyperphosphorylation in cultured cells and in brain (3, 4, 10, 34–36). We investigated the functional consequences of the inhibition of PP2A activity by I ^PP2A₁ in cultured cells by Western blots and by immunocytochemistry. To assess whether Tau phosphorylation is altered in cultured cells, the Tau441 stably transfected PC12 cells (PC12/Tau441) were transiently transfected with Myc-tagged I ^PP2AFL, I ^PP2A₁ΔC2, or I ^PP2A₁ F5, and vector as a control for 48 h, and Tau phosphorylation was studied at several sites by Western blots, and the level of Tau phosphorylation was normalized by total Tau expression. Phosphorylation of Tau at M4 (Thr-231/Ser-235), 12E8 (Ser-262/356), and at Ser-404 sites was significantly increased in cells that overexpressed I ^PP2A₁ΔC2 and in the case of 12E8 site I ^PP2A₁FL but not I ^PP2A₁F5 (data not shown) (Fig. 4A).

FIGURE 4.

The effects of I ^PP2A₁ and its deletion mutants on Tau phosphorylation and neuronal morphology of PC12 cells. A, Tau441 stably transfected PC12 cells were transiently transfected with vector, Myc-tagged I ^PP2A₁FL, or I ^PP2A₁ΔC2. After 48-h transfection, Tau phosphorylation at M4 (pT231/pS235), 12E8, (pS262/356) and pS404, sites, total Tau (R134d), as well as transfected I ^PP2A₁ were detected by Western blots from same amounts of lysates. The level of Tau phosphorylation was quantified by densitometry and normalized to total Tau levels. *, p < 0.05 compared with the vector transfection control. B and C, immunocytochemical staining of PC12 cells transiently transfected with Myc-tagged I ^PP2A₁FL, or I ^PP2A₁ΔC2. At 6 h post-transfection, cells were differentiated with 100 ng/ml NFG for further 90 h and then processed by double immunofluorescence using rabbit polyclonal antibody to Myc and mouse monoclonal antibody M4 to phosphotau (pThr-231/pSer-235) as primary antibodies (B), followed by fluorescein-labeled anti-mouse IgG (panels c and d), or rhodamine-conjugated anti-rabbit IgG (panels a and b) and additional staining of nuclei with 4′,6-diamidino-2-phenylindole (DAPI, panels e and f). M4 signal in cells that overexpressed I ^PP2A₁FL or I ^PP2A₁ΔC2 was increased as compared with non-transfected cells. In C, the cells were processed by double immunofluorescence using rabbit polyclonal antibody to Myc and mouse monoclonal antibody DMIA to tubulin as primary antibodies, followed by rhodamine-conjugated anti-rabbit IgG or fluorescein-labeled anti-mouse IgG, respectively. The morphology of the cells expressing I ^PP2A₁FL or I ^PP2A₁ΔC2 was significantly altered with marked decreases in the number and the length of neurites as compared with mock transfected cells. Scale bar = 10 μm. D, effect of I ^PP2A₁ and its deletion mutant on neurite outgrowth from PC12 cells. The total number of cells expressing Myc-tagged proteins was counted, and cells with processes were regarded as neurite-positive. The number of cells with such outgrowths was then expressed as a percentage of the total number of cells. The percentage of cells bearing neurites in cells transfected with I ^PP2A₁FL and I ^PP2A₁ΔC2 was significantly decreased compared with non-transfected cells used as a control. All cells on any individual dish were counted up to a total of 250 cells, and these results show the average from three independent experiments. E, neurite outgrowth was analyzed quantitatively by measuring the average length of randomly selected 30 neurons in three fields without knowledge of the transgenes under the same conditions described in B and C. The length of neurites in cells transfected with I ^PP2A₁FL and I₁PP2AΔC2 was remarkably decreased compared with non-transfected cells used as a control. Error bars indicate means ± S.E.; n = 30. *, p < 0.05; **, p < 0.001.

Besides detecting Tau phosphorylation in PC12/Tau441 cells, wild-type PC12 cells were transiently transfected with Myc tagged I ^PP2AFL, I ^PP2A₁ΔC2, or I ^PP2A₁F5 as well. At 6 h post-transfection, cells were differentiated with 100 ng/ml NGF for another 90 h to stimulate cell neuronal differentiation as well as increasing endogenous Tau expression level. The cells were then processed to detect endogenous Tau phosphorylation at M4 site and to assess microtubule network by immunohistochemistry using antibody M4 to Tau, and DM1A to tubulin, respectively. Furthermore, the neurite length was measured in randomly selected 30 transfected or non-transfected cells in three fields. I ^PP2A₁FL and I ^PP2A₁ΔC2 but not I ^PP2A₁F5 (data not shown) transfected cells showed Tau phosphorylation at M4 site and as well as a reduction in microtubule network/staining (Fig. 4, B and C). Meanwhile, in 96-h culture of mock transfected cells, ∼80% of PC12 cells possessed neurites, and the mean length of neurites was ∼20 μm (Fig. 4, D and E). For the proportion of PC12 cells expressing I ^PP2A₁FL or I ^PP2A₁ΔC2 the number of neurite bearing cells was reduced to ∼80 and ∼40%, respectively, of the mock transfected cells (Fig. 4D). The mean neurite length of cells overexpressing I ^PP2A₁FL and I ^PP2A₁ΔC2 was reduced significantly from ∼20 μm to ∼11.4 μm and ∼1.5 μm, respectively (Fig. 4E), suggesting that probably the dissociation of Tau from microtubules by I ^PP2A₁FL and I ^PP2A₁ΔC2 impairs neurite outgrowth and neuronal differentiation.

DISCUSSION

Neurofibrillary degeneration of abnormally hyperphosphorylated Tau is a hallmark of AD and a family of related neurodegenerative diseases called tauopathies (37–39). These tauopathies include frontotemporal dementia with Parkinsonism linked to chromosome 17, Pick disease, corticobasal degeneration, progressive supranuclear palsy, Guam Parkinsonism dementia complex, and dementia pugilistica (40). The degree of neurofibrillary degeneration is well known to correlate directly with the degree of dementia in AD patients (41–43). The abnormal hyperphosphorylation of Tau is the key step that, through sequestration of normal MAPs, results in the breakdown of the microtubule network and consequent neurodegeneration (44–50). PP2A, which accounts for >70% of total phosphoseryl/phosphothreonyl protein phosphatase activity in human brain (5), is a major regulator of the phosphorylation of Tau (7–9, 51). The activity of PP2A is compromised in AD brain (4, 34, 52–54). I ^PP2A₁, which is homologous to PHAP-1, pp32, mapmodulin, and LANP (15, 18–20, 22, 23, 55) regulates PP2A activity by functioning as a non-competitive inhibitor of this enzyme (15). In addition, several studies have shown that pp32 is involved in “histone masking,” which can lead to inhibition of histone acetyltransferase-dependent transcriptional activation (19, 33, 56). We have previously shown that the I ^PP2A₁ colocalizes with neurofibrillary pathology and that the transcription of I ^PP2A₁ is up-regulated in the affected areas of AD brain, and I ^PP2A₁ is, thus, a primary suspect in Alzheimer neurofibrillary degeneration (26). However, neither the nature of the interaction between I ^PP2A₁ and PP2A nor the sequence of events by which this interaction could lead to the abnormal hyperphosphorylation of Tau and neurodegeneration were understood.

The present study shows that the PP2A activity is inhibited by I ^PP2A₁ in a dose-dependent manner and that this decrease is not due to any reduction in its expression level of PP2Ac in NIH3T3 cells. Furthermore, I ^PP2A₁ interacts with PP2A catalytic subunit and not with PP2A-A or -B regulatory subunits, and the minimal region required for the association with PP2Ac as well as PP2A inhibition is localized at the N-terminal isotype-specific containing region of the inhibitor. This study also shows that I ^PP2A₁ significantly increases the phosphorylation of Tau at M4 (Thr-231/Ser-135), 12E8 (Ser-262/356) and at Ser-404 sites and impairs microtubule network and neurite outgrowth in PC12 cells during differentiation by NGF.

Unlike the variable B regulatory subunit of the trimeric PP2A holoenzyme, which regulates the phosphatase activity by determining its intracellular localization and substrate specificity, we found that I ^PP2A₁ regulated the PP2A activity by directly interacting with and inhibiting the activity of its catalytic subunit. We found that I ^PP2A₁ interacted with PP2Ac and no such association between the inhibitor and PP2A-A subunit or -B subunit was detected, suggesting that I ^PP2A₁ inhibits PP2A activity by directly interacting with its catalytic subunit. PP2Ac bound to the full-length I ^PP2A₁ and its deletion mutants containing the N-terminal isotype-specific region (I ^PP2A₁ΔC2, I ^PP2A₁ΔC3, and I ^PP2A₁ΔC4) except I ^PP2A₁ΔC1, and not to deletion mutants without the N-terminal isotype-specific region (I ^PP2A₁F2, I ^PP2A₁F2–4, and I ^PP2A₁F5), which suggests that the N-terminal isotype-specific region is required for I ^PP2A₁ binding to PP2Ac as well as the PP2A inhibitor activity. The deletion mutant I ^PP2A₁ΔC1 (aa 1–163) with the N-terminal region but without the C-terminal acidic region and the mutant I ^PP2A₁F5 (aa 164–249) with the C-terminal acidic region alone lost the ability to bind to PP2Ac. These findings suggest 1) that the region between aa 121–163 probably functions as an autoinhibitory domain that negatively affects the interaction of I ^PP2A₁ with PP2Ac and its ability to modulate the PP2A activity, and 2) that the C-terminal acidic region itself cannot bind to PP2Ac but might be involved in neutralizing the aa 121–163 autoinhibitory domain. In cells the inhibitory role of the aa 121–163 domain of I ^PP2A₁ might be regulated through phosphorylation of Ser-158 or both Ser-158 and Ser-204, which have been shown by Hong et al. (57) to be phosphorylated by casein kinase II in the cell.

The present study strongly suggests a link among I ^PP2A₁, PP2A, and abnormally hyperphosphorylated tau. We found that I ^PP2A₁ deletion mutants with the binding capacity to PP2Ac as well as inhibition of PP2A activity, when overexpressed in Tau441 stably transfected PC12 cells, increased the phosphorylation of Tau at M4 (Thr-231/Ser-235), 12E8 (Ser-262/356) and at Ser-404 sites. Phosphorylation of Tau at these sites is critically involved in converting it from microtubule assembly promoting to microtubule assembly inhibitory protein and promoting its self-assembly into paired helical filaments (39, 44, 45, 58–60). This increase in the abnormal hyperphosphorylation of Tau is probably both due to an inhibition of PP2A activity as well as an increase in the activity of one or more Tau kinases. Calmodulin kinase II, protein kinase A, extracellular regulated kinases ERK1 and ERK2, glycogen synthase kinase-3, p70S kinase, and stress-activated kinase p38 are among the Tau kinases activities that are regulated by PP2A (11, 13, 61–63).

In the present study we found that overexpression of I ^PP2A₁ full-length protein and its deletion mutant I ^PP2A₁ΔC2 (aa 1–120) in PC12 cells decreased the microtubule stability as evidenced by a marked reduction in the size of the microtubule network and cell size and reduction in the number and the length of neurites during NGF-induced neuronal differentiation. These deleterious effects of I ^PP2A₁ could occur through the abnormal hyperphosphorylation of Tau, which has been previously demonstrated to disrupt microtubules by sequestering normal MAPs (44, 45, 49, 58, 64). Mapmodulin has been shown to bind to normal MAPs, especially as the free proteins, and in this way destabilize microtubules (21, 55). Thus, I ^PP2A₁ could also contribute to destabilization of microtubules through essentially competing with tubulin.

In short, the direct binding of I ^PP2A₁ to PP2Ac, causing inhibition of the phosphatase activity, provides a new mechanism of the regulation of the cellular PP2A activity. I ^PP2A₁ appears to interact through its N-terminal isotype region (aa 1–45) and have an autoinhibitory domain, which lies within amino acid residues 121–163. This autoinhibitory domain appears to be at least partly neutralized by the C-terminal acidic region (aa 164–249) of I ^PP2A₁, because both the full-length protein and the aa 1–120 N-terminal domains have PP2A inhibitory activities, whereas neither the aa 1–163 nor 46–163 can inhibit PP2A. Inhibition of PP2A by I ^PP2A₁ results in the abnormal hyperphosphorylation of Tau at several sites, which are known to be critically involved in converting normal Tau to an inhibitory and pathological state. Thus, inhibition of I ^PP2A₁ activity offers a very promising therapeutic target for inhibition of neurofibrillary degeneration of the abnormally hyperphosphorylated Tau through the restoration of the PP2A activity.