Cytoplasmic Retention of Protein Phosphatase 2A Inhibitor 2 (I2PP2A) Induces Alzheimer-like Abnormal Hyperphosphorylation of Tau*

Background: In Alzheimer brain, I2PP2A is translocated from the neuronal nucleus to the cytoplasm and promotes abnormal hyperphosphorylation of Tau. Results: Inactivation of nuclear localization signal (NLS) causes retention of I2PP2A in the cell cytoplasm, where it promotes Tau hyperphosphorylation by affecting PP2A signaling. Conclusion: Retention of I2PP2A in cell cytoplasm results in Tau hyperphosphorylation. Significance: The study provides potential tools for investigating Tau-based therapeutics. Abnormal hyperphosphorylation of Tau leads to the formation of neurofibrillary tangles, a hallmark of Alzheimer disease (AD), and related tauopathies. The phosphorylation of Tau is regulated by protein phosphatase 2A (PP2A), which in turn is modulated by endogenous inhibitor 2 (I2PP2A). In AD brain, I2PP2A is translocated from neuronal nucleus to cytoplasm, where it inhibits PP2A activity and promotes abnormal phosphorylation of Tau. Here we describe the identification of a potential nuclear localization signal (NLS) in the C-terminal region of I2PP2A containing a conserved basic motif, 179RKR181, which is sufficient for directing its nuclear localization. The current study further presents an inducible cell model (Tet-Off system) of AD-type abnormal hyperphosphorylation of Tau by expressing I2PP2A in which the NLS was inactivated by 179RKR181 → AAA along with 168KR169 → AA mutations. In this model, the mutant NLS (mNLS)-I2PP2A (I2PP2AAA-AAA) was retained in the cell cytoplasm, where it physically interacted with PP2A and inhibited its activity. Inhibition of PP2A was associated with the abnormal hyperphosphorylation of Tau, which resulted in microtubule network instability and neurite outgrowth impairment. Expression of mNLS-I2PP2A activated CAMKII and GSK-3β, which are Tau kinases regulated by PP2A. The immunoprecipitation experiments showed the direct interaction of I2PP2A with PP2A and GSK-3β but not with CAMKII. Thus, the cell model provides insights into the nature of the potential NLS and the mechanistic relationship between I2PP2A-induced inhibition of PP2A and hyperphosphorylation of Tau that can be utilized to develop drugs preventing Tau pathology.

Abnormal hyperphosphorylation and aggregation of microtubule-associated protein Tau into paired helical filaments/ neurofibrillary tangles is a hallmark of neurodegenerative tauopathies, including Alzheimer disease (AD), 2 frontotemporal dementias, tangle-only dementia, Pick disease, argyrophilic grain disease, progressive supranuclear palsy, corticobasal degeneration, Guam parkinsonism dementia complex, dementia pugilistica, and traumatic brain injury/chronic traumatic encephalopathy (1,2). Tau is a highly soluble and unfolded protein that stabilizes the assembly of microtubules. However, Tau abnormal hyperphosphorylation negatively regulates its microtubule binding, dissociation of Tau from microtubules, and sequestration of normal Tau and other microtubule-associated proteins, causing its aggregation, breakdown of the microtubule network, and eventually cell death (3)(4)(5)(6)(7). Studies have shown that Tau normally contains 2-3 mol of phosphate/ mol of Tau, but it is 3-4-fold more phosphorylated in AD brain (8). The number of neurofibrillary tangles correlates with progressive neuronal dysfunction, synaptic loss, and functional decline in humans and transgenic mouse models (9 -12). Although the triggering mechanism leading to Tau hyperphosphorylation is yet to be clarified, it is well recognized that an imbalanced regulation in Tau protein kinases and phosphatases can directly cause AD-like Tau hyperphosphorylation (1).

PP2A
, also known as SET␣, TAF-1␤, and PHAPII, is a nuclear protein that regulates cell cycle (21), cell proliferation (22), and cell motility (23). Moreover, I 2 PP2A controls gene transcription (24) by regulating histone acetylation (25) and is possibly involved in neuronal apoptotic pathways in AD brain (26). We have shown previously that both the mRNA and protein expressions of I 2 PP2A are up-regulated, and I 2 PP2A is selectively cleaved at Asn 175 into two fragments, N-terminal and C-terminal fragments (I 2NTF and I 2CTF ), by asparaginyl endopeptidase and is translocated from its primary localization in the nucleus to the cytoplasm (27)(28)(29). Because PP2A and Tau are localized in the cytoplasm, the increased neuronal cytoplasmic residing of I 2 PP2A in the AD brain leads to the inhibition of PP2A, Tau hyperphosphorylation, and formation of neurofibrillary tangles (28,30,31). More recently, we have shown that I 2CTF alone sufficiently induces Tau pathology and cognitive impairment in a non-transgenic rat model of AD (32). As a nuclear protein, I 2 PP2A requires a specific sequence called the nuclear localization signal (NLS) to be targeted to the nucleus. Indeed, the NLS of I 2 PP2A has been reported at 168 KRSSQTQNKASRKR 181 , and targeted expression of I 2 PP2A is found to be associated with neuronal death (33). Therefore, cytoplasmic withholding of I 2 PP2A regulated by NLS and its association with PP2A appear to be the key upstream molecular mechanism controlling the abnormal phosphorylation of Tau, which is considered one of the earliest signs of neuronal degeneration that precede Tau aggregation/neurofibrillary tangles in AD and related tauopathies.
In this study, we first show that the functional NLS of I 2 PP2A is localized at 179 RKR 181 , which controls the shuttle of I 2 PP2A between the cellular nucleus and cytoplasm. Second, we report a PC12 stably expressing human Tau 441 cell line that allows the inducible expression of mNLS-I 2 PP2A ( 168 KR 169 3 AA/ 179 RKR 181 3 AAA) based on the Tet-Off system. Employing this cell line, we found that cytoplasmic accumulation of I 2 PP2A is associated with inhibition of PP2A activity and activation of Tau kinases regulated by PP2A, hyperphosphorylation of Tau, and promotion of microtubule instability. The mNLS-I 2 PP2A cell model can be useful in screening of therapeutic drugs for AD and other tauopathies.  (sense  primer 5Ј-ACATCGGATCCATGTCGGCGCCGGCGGCCAA-AGTC-3Ј (BamHI-I 2   PP2A , N-terminal positions 1-24 of human I 2 PP2A and a BamHI site) and antisense primer 5Ј-GGTAAAGA-AGCTCTCTGGTTCCTCATGCTGCGCCGCCGCGCTGGC-TTTATTCTGCGTTTGACTC-3Ј (reverse sequence 510 -573 of human I 2 PP2A mutant); sense primer 5Ј-GAGTCAAACGCAGA-ATAAAGCCAGCGCGGCGGCGCAGCATGAGGAACCAG-AGAGCTTCTTTACC-3Ј (sense sequence 510 -573 of human I 2 PP2A mutant) and antisense primer 5Ј-ATAAGAATGCGGCC-GCCTAAGCGTAATCTGGAACATCGTATGGGTAGTCAT-CTTCTCCTTCATCCTCCTCTCC-3Ј (C-terminal 805-834 of human I 2 PP2A and an HA tag plus NotI site) were individually incubated with pGEX-6P-1-I 2 PP2A as a template in the first PCR. The PCR products were gel-purified, combined, and incubated with BamHI-N-I 2 PP2A and NotI-C-I 2 PP2A primers to synthesize fulllength I 2 PP2A RKR 179 3 AAA cDNA in the second PCR. The resulting product was digested with NotI and BamHI and inserted into NotI and BamHI sites of pTRE2hyg to generate the pTRE2hyg I 2 PP2A RKR 179 3 AAA mutation. pTRE2hyg I 2 PP2A mutants were generated by the same procedure.

Construction of Expression Vectors-To
Cell Culture, Transfection, and Differentiation-COS-7 cells (obtained from ATCC (Manassas, VA) were grown in 25-cm 2 flasks at 37°C, containing 5% CO 2 in Dulbecco's modified Eagle's medium supplemented with 10% bovine calf serum. To investigate the localization I 2 PP2A and its mutants, cells were plated on 4-well Lab-Tek II Chamber Slides coated with poly-L-lysine (Nalge Nunc International, Naperville, IL) and transfected with expression plasmids using FuGENE 6 (Roche Applied Science).
In order to generate a stable cell line of Tau 441 /I 2 PP2A or its mutants, PC12 pheochromocytoma cells were grown at 37°C in Dulbecco's modified Eagle's medium (high glucose) supplemented with 10% heat-inactivated horse serum plus 5% fetal calf serum. Cells were co-transfected with pCI-neoTau 441 , pTet-Off, and pTRE2hyg (vector) or pTRE2hyg containing human I 2 PP2A WT, I 2 PP2A AA, and I 2 PP2A AA-AAA tagged with HA using Lipofectamine 2000 (Invitrogen). The pTet-Off vector contains the neomycin resistance gene as a selectable marker. Drug selection was carried out for 2 weeks by seeding 1 ϫ 10 4 cells/well in 96-well tissue culture plates (Nunc, Roskilde, Denmark) containing the same media supplemented with 400 g/ml G418 (Sigma) and 100 g/ml hygromycin B (EMD Biosciences). Among the G418-hygromycin-resistant clones, 20 clones were selected randomly and transferred into 6-well tissue culture plates (Nunc) for cell expansion in the presence of 200 g/ml G418 and 100 g/ml hygromycin. Overexpression of I 2 PP2A has been reported to be associated with cellular death (26,30). We therefore maintained cells at a low dose of doxycycline (Dox; Sigma) (100 ng/ml) to keep a low level of I 2

PP2A
expression. For the present study, cells were grown without Dox for 72-96 h to allow the expression of I 2 PP2A . The Tau 441 / I 2 PP2A cell lines were screened by Western blot analysis of HA expression in the presence or absence of 1 g/ml Dox. For studying the effect of mNLS-I2PP2A-induced hyperphosphorylation of tau on the number and the length of neurites, stably transfected PC12 cells were differentiated into neurons by adding 100 ng/ml NGF to the culture medium for 5 days.
Subcellular Fractionation-Cytoplasmic and nuclear fractions were prepared from cells as described previously with minor modifications (29). Briefly, the cells were scraped and washed with cold PBS, and centrifuged at 200 ϫ g for 7 min. The pellet was resuspended in lysis buffer comprising 50 mM Tris-HCl, pH 7.4, 0.32 M sucrose, 150 mM NaCl, 10% glycerol, 1 mM sodium vanadate, 50 mM sodium fluoride, 5 mM EDTA, 1 mM PMSF, 10 g/ml aprotinin, 10 g/ml leupeptin and homogenized for 1 min on ice using a Potter-Elvehjem homogenizer set to 600 rpm with 5-10 gentle strokes. The lysate was then inspected, and if intact cells were still evident, the homogenization was repeated. The cell lysate was centrifuged at 750 ϫ g for 10 min in a swinging bucket rotor. The supernatant was kept, and the pellet was resuspended with half of the initial volume of the lysis buffer with 3-4 strokes and centrifuged as above. The second supernatant was combined with the previous one, and the pellet was resuspended with the same buffer and saved as the nuclear fraction. The pooled supernatant was centrifuged at 100,000 ϫ g for 1 h, and the resulting supernatant was saved as the cytosolic fraction. After protein measurement, the samples were boiled in Laemmli's buffer and used for Western blots.
PP2A Activity Assay-PP2A activity was assayed in cell lysates using the phosphatase ELISA described previously (29). Briefly, 96-well plates were precoated with a 17-aa phosphopeptide corresponding to Tau aa 194 -207 in which Ser 199 was replaced with phosphoserine and to which KKK were added at the C terminus (coating buffer, 35 mM NaHCO 3 ). After overnight blocking with a protein-free blocking solution (Pierce), the enzymatic reaction was performed by adding 2 g of cell lysate protein in 60 l of reaction buffer (50 mM Tris-HCl, pH 7.0, 2 mM MnCl 2 , 0.1 mg/ml BSA, 20 mM ␤-mercaptoethanol) at 30°C for 30 min (in the presence or absence of 15 nM of okadaic acid) in a moist chamber. The reaction was stopped by adding 150 l of Pierce blocking solution containing 50 mM NaF, followed by an overnight incubation with monoclonal antibody Tau-1, which recognizes Tau unphosphorylated at Ser 198 , Ser 199 , or Ser 202 (30). HRP-conjugated secondary antibody treatment was for 60 min at room temperature. Finally 75 l of tetramethylbenzidine substrate reagent (Sigma) was added, and the colorimetric development was measured using a microtiter plate reader at 650-nm wavelength.
Methylation Sensitivity Assay of PP2A Antibodies-In order to characterize the methylation sensitivity of PP2Ac antibodies used in the present study, PC12/Tau cells were incubated with 0.5 M NaOH for 5 min on ice, followed by neutralization with an equal volume of HCl and 0.5 volume of 1 M Tris-HCl, pH 6.8 (34). Control was treated with preneutralized base solution followed by 0.5 volume of Tris-HCl, pH 6.8. The samples were then analyzed by Western blots using three anti-PP2Ac antibodies: 1D6, R123d, and BD.
Immunofluorescent Staining-Histopathologically confirmed AD and age-matched control brain paraffin sections (Sun Health Research Institute, Brain Donation Program, Sun City, AZ) were treated with antigen retrieval buffer containing 7 mM citric acid and 10 mM sodium citrate in microwave for 2 min followed by primary antibody treatment. On the other hand, PC12 cells were grown on 4-chamber glass slides, fixed in cold 4% paraformaldehyde for 15 min, and permeabilized in 0.1% Triton X-100 for 15 min, followed by incubation in 5% goat serum to block the nonspecific staining for 1 h. The primary antibodies used at the indicated dilutions were as follows: anti-HA (1:3000, Sigma-Aldrich), anti-PP2Ac (R123d, 1:300), anti-I 2 PP2A (1483, 1:200) (29), anti-␣-tubulin (DM1A, 1:2000; Sigma), and anti-GSK-3␤ (1:300; Cell Signaling) overnight at 4°C. An Alexa-488/555-conjugated appropriate combination of anti-mouse/rabbit secondary antibody (1:500; Jackson Laboratory) was used for fluorescent labeling. The cells on slides were rinsed with PBS, mounted with Fluorogel (Electron Microscopy Sciences, Hartfield, PA), and examined under a laser confocal microscope (Nikon Eclipse 90i, Melville, NY). For quantitative analysis, the images were captured using the ϫ40 objective, and the antibody staining was quantified by measuring mean pixel intensity with the software Image Pro-Plus version 5.0 (Media Cybernetics, Silver Spring, MD), as described previously (39,40). Each experimental condition was calculated using the average number of 5 fields/well for 3-4 wells. The numbers of cells bearing neurites and the length of the neurites were analyzed by 25-30 randomly selected NGFdifferentiated PC12 cells using ImageJ software, as described previously (41).
Statistical Analysis-The data were analyzed with Student's t test when comparing means between two groups or with oneway analysis of variance plus Bonferroni post hoc multiple comparison test to compare mean values across multiple treatment groups. In all cases, a probability statistic Ͻ0.05 was taken to indicate significance. All data are expressed throughout as means Ϯ S.E.

Mutant NLS-I 2 PP2A
Translocates to the Cell Cytoplasm from the Nucleus and Inhibits PP2A Activity-We previously reported that I 2 PP2A is translocated from neuronal nucleus to the cytoplasm, co-localizes with PP2A, and promotes abnormal hyperphosphorylation of Tau by inhibiting PP2A in AD brain (27,28), suggesting the critical role of I 2 PP2A translocation and its association with PP2A in the pathogenesis of AD. Because the NLS of proteins plays an important role in facilitating the nuclear translocation, we focused on the 168 KRSSQTQNKAS-RKR 181 region of I 2 PP2A , as described by Qu et al. (33). We further carried out multiple alignments of I 2 PP2A sequence among different species and found that 178 RKR 181 emerged as the most conserved sequence, suggesting its important functional implication (Fig. 1A). To identify the critical region of I 2 PP2A NLS, we generated various mutants of this protein (Fig.  1B) and overexpressed them in COS7 cells transiently. Confocal microscopy analysis of their subcellular localization revealed that the mutants I 2 PP2A AAA and I 2 PP2A AA-AAA were FIGURE 1. Identification of nuclear localization signal of I 2 PP2A . A, the structural similarities between human, chimpanzee, pig, bovine, rat, mouse, and Drosophila were analyzed to determine the conservation and homology of potential I 2 PP2A NLS among various species. B, schematic diagram of the constructs of I 2 PP2A and its NLS mutants employed to study their intracellular localization. C, photomicrographs of COS7 cells transiently transfected with vector (vec), HA-tagged human I 2 PP2A (I 2 PP2A WT), and its NLS mutants (I 2 PP2A AA, I 2 PP2A AAA, and I 2 PP2A AA-AAA) for the identification of critical regions of I 2 PP2A required for its translocation from the cell nucleus to the cytoplasm. Following 48 h of transfection with HA-tagged I 2 PP2A (WT and NLS mutants), COS7 cells were subjected to immunocytochemical staining with anti-HA, and TOPRO-3 was used for nuclear staining. Mutations at amino acid residues 178 -181 markedly increased the cytoplasmic retention of I 2 PP2A . D, COS7 cells transiently transfected with R179A-I 2 PP2A and K180A-I 2 PP2A , followed by immunostaining with anti-HA (I 2 PP2A ). Immunofluorescence images indicate that lysine 180 is required for I 2 PP2A nuclear localization (i.e. NLS). Scale bar, 50 m. present in a diffuse manner throughout the cells, whereas the wild type I 2 PP2A and I 2 PP2A 168 AA 169 were localized mainly in the nuclei (Fig. 1C), but low cytoplasmic staining can also be observed at higher microscopic gain (data not shown), suggesting that the 178 RKR 181 motif is required for the nuclear localization of I 2

PP2A
. We further observed that mutation of Lys 180 to Ala alone confined the I 2 PP2A into the cytoplasm of COS7 cells (Fig. 1D).
Next we evaluated whether translocation of I 2 PP2A is associated with inhibition of PP2A activity. As expected, total cell lysates from all transfections of I 2 PP2A constructs inhibited PP2A activity significantly, independent of their cellular localization ( Fig. 2A). We found similar expressions of endogenous and various constructs of HA-tagged exogenous I 2 PP2A (Fig. 2, B and C). In order to investigate the effect of the cytoplasmic localization of I 2 PP2A on PP2A activity, we fractionated the total cell lysate into cytosolic and nuclear fractions where histone-3 and ␤-actin were used as markers for nuclear and cytosolic fractions, respectively (Fig. 2, D--G). When we assayed PP2A activity with cytosolic fractions, only NLS mutants (I 2 PP2A AAA and I 2 PP2A AA-AAA) significantly inhibited the activity (Fig.  2H). These results suggest that the translocation of I 2 PP2A from neuronal nucleus to cytosol is the critical step in inhibition of PP2A activity.
Direct Interaction of PP2Ac with Translocated mNLS-I 2

PP2A
Leads to Inhibition of PP2A Activity-Abnormal hyperphosphorylation of Tau is considered a key event leading to neurofibrillary tangle formation, which correlates with dementia in AD (42,43). Based on the present finding that translocated mutant NLS-I 2 PP2A inhibits PP2A activity, the major phosphatase in the brain, we developed a Tet-Off cellular model of ADtype abnormal hyperphosphorylation of Tau by stably overexpressing HA-tagged mNLS-I 2 PP2A (I 2 PP2A AA-AAA) along with human Tau 441 in PC12 cells. We also developed stably transfected cell lines for vector and I 2 PP2A WT in an identical manner. To validate this model, we first determined the PP2A activity in the cytosolic fraction of the cell lines. PP2A activity was significantly decreased in mNLS-I 2 PP2A cells as compared with controls (vec) (Fig. 3A). PP2A activity was also decreased by I 2 PP2A WT but did not reach statistical significance (Fig. 3A). We then characterized the methylation sensitivity of PP2Ac antibodies used in the present study by treating PC12 cell lysates with NaOH, which causes the demethylation of the PP2Ac pool (34). We found that R123d and BD PP2Ac antibodies recognize total PP2Ac, whereas 1D6 recognizes the demethylated PP2Ac pool (Fig. 3B). Next, we examined whether translocated mNLS-I 2 PP2A can bind to PP2Ac and thus inhibit its activity. To study the interaction between PP2Ac and mNLS-I 2

PP2A
, the cytosolic fraction of PC12 cell lysates was also used to immunoprecipitate endogenous PP2Ac, using rabbit polyclonal antibody R123d to PP2Ac. Western blots developed with HA antibody clearly showed significantly higher interaction between I 2 PP2A AA-AAA and PP2Ac as compared with I 2 PP2A WT (Fig. 3, C and D). It is also evident that I 2 PP2A WT also co-immunoprecipitated with cytosolic PP2Ac, possibly caused by the overexpression of the transgene. To further confirm these findings, we carried out immunofluorescence analysis with stably transfected cell lines using antibodies to PP2Ac (R123d) and I 2 PP2A (HA). A diffuse pattern of I 2

PP2A
staining predominantly localized to the nucleus was observed in I 2 PP2A WT cells (Fig. 3E). In contrast, PP2Ac and I 2 PP2A were found to co-localize in the cytoplasmic compartment in mNLS-I 2 PP2A (I 2 PP2A AA-AAA) cells (Fig. 3E). These findings suggest that the direct physical interaction of NLS mutant-I 2 PP2A with PP2A catalytic subunit could be involved in PP2A inhibition.
Effect of mNLS-I 2 PP2A on Tau Phosphorylation and Neurite Outgrowth-PP2A is the most implicated protein phosphatase in regulating Tau phosphorylation because it accounts for more than ϳ70% of Tau dephosphorylation activity (13). Inhibition of PP2A activity has been reported to be associated with abnormal hyperphosphorylation of Tau in cultured cells and in the brain (16, 29, 32, 40, 44 -46). In the current study, we therefore investigated the functional consequences of the inhibition of PP2A activity due to direct interaction of PP2Ac with mNLS-I 2 PP2A in the I 2 PP2A AA-AAA cells as compared with I 2 PP2A WT cells. Because both I 2 PP2A WT and I 2 PP2A AA-AAA were Tet-Off cell lines, we determined abnormal hyperphosphorylation of Tau at 12E8 (Ser 262 /Ser 356 ) and PHF1 (Ser 396 /Ser 404 ) sites in the presence or absence of 1 g/ml Dox for 72 h. The level of the abnormal hyperphosphorylation of Tau at 12E8 and PHF1 sites, which are considered major sites in AD-type neurofibrillary degeneration, was increased significantly both in I 2 PP2A WT (p Ͻ 0.05) and mNLS-I 2 PP2A (p Ͻ 0.01) as compared with the vector cells (Fig. 4). However, the extent of Tau hyperphosphorylation at 12E8 (p ϭ 0.054) and PHF1 (p ϭ 0.046) sites in mNLS-I 2 PP2A was higher than in I 2 PP2A WT cells. Treatment with Dox normalized the level of I 2 PP2A transgene expression and the associated increase in abnormal hyperphosphorylation of Tau (Fig. 4). No significant change was detected in total Tau level as determined by human-specific anti-Tau, 43D, among the cell lines (Fig. 4A).
The functional aftermath of abnormal hyperphosphorylation of Tau is the disruption of microtubule stability, which could reduce the number and the length of neurite outgrowth (41). To investigate this possibility, we differentiated the stable cell lines of I 2 PP2A WT and I 2 PP2A AA-AAA with 100 ng/ml NGF for 5 days and then carried out immunofluorescence using rabbit polyclonal antibody to HA and mouse monoclonal antibody DM1A to tubulin. The morphology of cells expressing I 2 PP2A AA-AAA was significantly altered, with marked decreases in the number and length of neurites. The neurite outgrowth was also reduced in I 2

PP2A
WT cells but was not statistically significant when compared with controls (vec) (Fig. 5, A-C). The number of neurite-bearing cells was reduced to ϳ22 and ϳ39%, whereas the mean neurite length was reduced to ϳ17 and ϳ49% in PC12 cells stably expressing I 2

PP2A
, respectively (Fig.  5). These data imply that the abnormal hyperphosphorylation of Tau by mNLS-I 2 PP2A suppresses its microtubule binding and assembly, which eventually impairs the neurite outgrowth and neuronal morphology.
Effect of mNLS-I 2 PP2A on GSK-3␤ and CAMKII Signaling Pathways-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 phosphorylation of Tau by Tau kinases that are regulated by PP2A. Among the Tau kinases in the brain, OCTOBER 3, 2014 • VOLUME 289 • NUMBER 40 GSK-3␤, CAMKII, ERK1/2, MEK1/2, and P70S6 kinase are regulated by PP2A (47). To explore the involvement of PP2A-regulated Tau kinases, we measured the levels of total and activated/inactivated forms of these kinases by Western blots. The levels of phosphorylated/activated CaMKII increased, whereas phosphorylated/inactivated GSK-3␤ at Ser 9 decreased noticeably in mNLS-I 2 PP2A cells (Fig. 6, A and B). However, we detected no significant change in the levels of phosphorylated WT and its NLS mutants (I 2 PP2A AA, I 2 PP2A AAA, and I 2 PP2A AA-AAA), COS7 cells were lysed, and PP2A activity was measured by ELISA using total cell lysate. I 2 PP2A WT along with all of its NLS mutants inhibited PP2A activity. B, the expression of endogenous and exogenous I 2 PP2A -HA was determined by Western blots. C, the quantitation of blots in B. The exogenous I 2 PP2A expression did not show any significant difference among the groups. D and F, nuclear and cytoplasmic fractions were prepared from COS7 cells transiently transfected (48 h), and the expression of I 2 PP2A in each fraction was analyzed by Western blots. Histone-3 and ␤-actin were used as markers of nuclear and cytoplasmic fractions, respectively. E and G, quantitative analysis of blots in D and F, respectively. The cytoplasmic retention of I 2 PP2A is associated with inactivation of NLS. H, PP2A activity assayed in the cytosolic fraction showed a significant decrease in NLS-I 2 PP2A mutants (I 2 PP2A AAA and I 2 PP2A AA-AAA  (Fig. 6, C and D). On the contrary, LiCl treatment significantly (p Ͻ 0.05) prevented the Tau phosphorylation at PHF1 but not at the 12E8 site induced by mNLS-I 2 PP2A expression (Fig. 6, C and D). Combined treatment of both inhibitors (KN-93 and LiCl) showed a synergistic effect in reversing the hyperphosphorylation of Tau in mNLS-I 2 PP2A cells. These results suggested that mNLS-I 2 PP2A promotes abnormal hyperphosphorylation of Tau both by directly inhibiting PP2A and by up-regulating the activities of Tau kinases regulated by PP2A.

Cell Model of Tau Hyperphosphorylation
However, the question remains to be answered as to whether mNLS-I 2 PP2A has any direct interaction with CAMKII or GSK- 3␤, which might also contribute to the activation of these kinases. To address this issue, we performed co-immunoprecipitation studies with cell lysates from COS7 cells transiently transfected with a combination of I 2 PP2A AA-AAA-HA/ CAMKII-GFP or I 2 PP2A AA-AAA-Myc/GSK-3␤-HA. I 2 PP2A coimmunoprecipitated with GSK-3␤ along with PP2Ac, whereas no in vivo association of CAMKII with I 2 PP2A was detected (Fig.  7, A and B). Furthermore, protein phosphatase 1 (PP1) is reported to dephosphorylate GSK-3␤ at Ser 9 and activate the kinase activity (49). Because mNLS-I 2 PP2A is found to be associated with GSK-3␤ with undefined functional implications, it is reasonable to speculate that I 2 PP2A might affect GSK-3␤ indirectly through PP1. Thus, we also evaluated the direct interaction between I 2 PP2A and PP1 by co-immunoprecipitation experiments. We did not find any association of I 2 PP2A with PP1 in our cell model (Fig. 7A). Furthermore, Western blot analysis of PP1 did not show any significant change between I 2 PP2A WT and mNLS-I 2 PP2A cells used in the study (Fig. 7, C and D). In addition, GSK-3␤ and I 2 PP2A were found to co-localize in cytoplasmic compartments of COS7 cells double-labeled with anti-GSK-3␤ and anti-HA (I 2 PP2A AA-AAA) (Fig. 7E). Next, we asked the question whether or not similar cross-talk between I 2 PP2A and GSK-3␤ is also evident in AD brain. Consistent with previous reports (27,50), a significantly enhanced I 2 PP2A staining was detected in the cytoplasm of hippocampus of AD brains as compared with age-matched control brains, and the increased I 2 PP2A was co-localized with GSK-3␤ (Fig. 7F).

DISCUSSION
Tauopathies are age-associated neurodegenerative diseases characterized by neurofibrillary pathology made up of abnormally hyperphosphorylated Tau, and the severity of these lesions directly correlates with dementia. Given the magnitude and the impact of tauopathies, including AD, on public health, there is increasing interest in the field to decipher the molecular mechanism underlying the Tau pathology as well as identifying the therapeutic interventions where disease-relevant cellular models are of crucial importance. We have previously shown that the activity of PP2A, the major Tau phosphatase regulating abnormally hyperphosphorylated Tau, is decreased in AD brains due to the cleavage and translocation of I 2 PP2A from the neuronal nucleus to the cytoplasm by asparagine endopeptidase, possibly as a consequence of brain acidosis (27,29). As observed in AD brain, we recently reported a selective increase in the cleavage of I 2 PP2A and inhibition of PP2A activity in the lateral sclerosis cases (46), suggesting I 2 PP2A -PP2A signaling as a common molecular target in different neurodegenerative diseases. To clarify whether cytoplasmic localization of I 2 PP2A is vital for its association and inhibition of PP2A and thus neurofibrillary degeneration, we identified the potential NLS of I 2 PP2A , and, employing the mutated (inactivated) NLS, we generated a cell model of the hyperphosphorylated Tau that exhibits the following characteristic features. 1) The system is developed on PC12 cells based on the mutation of NLS of I 2 PP2A , which restricts it to cytoplasmic localization where both PP2A and expressed human Tau 441 are present.
2) The expression of mNLS-I 2 PP2A can be switched on and off in response to doxycycline (Tet-Off system) in order to control the inhibition of PP2A activity and resultant hyperphosphorylation of Tau. 3) The expression of mNLS-I 2 PP2A is associated with the downregulation of PP2A activity along with up-regulation of PP2Aregulated Tau kinases, possibly synergistically increasing the phosphorylation of Tau and impairment in microtubule network and neurite outgrowth. I 2 PP2A is predominantly a nuclear protein; hence, the NLS of I 2 PP2A is supposed to be the central element controlling its nuclear localization. Previous studies revealed two NLSs for I 2 PP2A ; one is close to the N terminus, 6 AKVSKK 11 (51), and other one is in the area of the I 2 PP2A cleavage site at Asn 175 , 168 KRSSQTQNKASRKR 181 (33). Based on the sequence comparison analysis among different species, we found 168 KRSSQTQNKASRKR 181 as an apparent conserved region of I 2 PP2A and inferred that this region is a potential NLS. By using site-directed mutagenesis in combination with immunofluorescence analysis, we confirmed the NLS in the I 2CTF in the current study and further narrowed it down to 179 RKR 181 because mutations of these three amino acid residues cause the I 2 PP2A to be diffusely distributed throughout the cytoplasm. The present study clearly suggests that NLS at amino acids 179 -181 is sufficient to determine its subcellular localization. The most likely reason is that NLS at amino acids 6 -11 is neutralized by phosphorylation at Ser 9 in contrast to the NLS at 179 -181 because the phosphorylation of Ser 9 is considered to be critical for its interaction and thus inhibition of PP2A (23). However, one cannot rule out the existence of double NLS in a protein similar to that found in 5-lipoxygenase (52). In the present study, cell fractionation and subsequent analysis with subcellular markers followed by PP2A activity further confirmed

JOURNAL OF BIOLOGICAL CHEMISTRY 27685
that cytoplasmic localization of I 2 PP2A is associated with inhibition of PP2A activity, which is considered to be the major regulator of Tau hyperphosphorylation as seen in tauopathies. Thus, utilizing this observation, we developed an inducible cellular model of Tau hyperphosphorylation by overexpressing  mNLS-I 2  PP2A (I 2   PP2A AA-AAA) along with human Tau 441 in PC12 cells.
PP2A has been reported to interact physically with its endogenous inhibitors I 1 PP2A and I 2 PP2A and its cleavage fragments and thus inhibits its activity (31,41). In the present study, we validated the newly developed cellular model in which cytoplasmic localization of I 2 PP2A leads to increased physical association with PP2Ac and inhibits its activity. Co-localization studies with PP2Ac and I 2 PP2A clearly demonstrated a significant cross-talk in the cytoplasmic compartment, which is consistent with the reduced PP2A activity. Co-immunoprecipitation experiments further confirmed that the increased physical interaction between PP2A and I 2

PP2A
facilitates the inhibition of PP2A activity in our cell model.
Next, we addressed the question of whether or not PP2A inhibition promotes the abnormal hyperphosphorylation of Tau and microtubule instability as a functional consequence of cytoplasmic retention of I 2 PP2A in this cell model. We found a marked increase in the abnormal hyperphosphorylation of Tau in cells at Ser 262/356 (12E8) and Ser 396/404 (PHF1) sites, two prominent abnormal hyperphosphorylated sites found in PHF-Tau in AD brain (53). Ser 262/356 is located in the microtubule (MT) binding domain of Tau, and phosphorylation at this site has been shown to induce MT disassembly, whereby the new unbound pool of Tau is susceptible to self-assembly into PHFs (14,43,54,55). Phosphorylation of Tau at Ser 262 in combination with Thr 212 and Thr 231 results not only in loss of its normal function but also gain of a toxic activity that causes disruption of microtubule networks and cell death (3,14). With respect to Ser 262 , there is evidence that Tau phosphorylation at this site plays a critical role in A␤42-induced Tau toxicity because Tauinduced toxicity was prevented by using a transgenic fly expressing a nonphosphorylatable form of Tau at Ser 262 (56). The phosphorylation of Tau at Ser 262 and Ser 356 has another functional consequence: priming the phosphorylation at other sites, which increases further its microtubule inhibitory activity and self-aggregation into filaments, because we previously showed that a prior phosphorylation of Tau at Ser 262/356 containing the Ser/Thr-X motif by CAMKII, a non-proline-dependent protein kinase (PDPK) was found to stimulate a subse- quent phosphorylation catalyzed by GSK-3␤, a PDPK (57). In contrast to Ser 262 , Ser 396 and Ser 404 are present in the C-terminal domain and have only a moderate influence on Tau-MT interaction but are up-regulated in AD and other tauopathies (58,59). Moreover, phosphorylation of Tau at the PHF1 (Ser 396/404 ) site in combination with the AT8 (Ser 199 /Ser 202 /Ser 205 ) site has been shown to induce a compaction of the paper clip folding of Tau that generates a pathological conformation, which aggregates somewhat more readily (60). Thus, it appears that the missorting of I 2 PP2A into the cytoplasm probably contributes to the deficit in PP2A activity, which in turn promotes the hyperphosphorylation and loss of biological function of Tau in our cellular model.
There is significant evidence that abnormal hyperphosphorylation of Tau results in Tau dysfunction due to the disturbances in microtubule dynamics, axonal transport, and neurite outgrowth, which synergistically contribute to the pathogenic processes. In the current study, we found that mNLS-I 2

PP2A
decreased not only the mean neurite length but also the average number of NGF-induced differentiated cells bearing neurites. Our data are consistent with our previous observation that I 1 PP2A , another endogenous inhibitor of PP2A, impairs MT network and neurite outgrowth via hyperphosphorylation of Tau (41). Although our data do not imply a direct role of I 2 PP2A in microtubule instability, it is well recognized that once Tau is abnormally hyperphosphorylated, it can disrupt microtubules by sequestering normal microtubule-associated proteins (61). There is indirect evidence that Tau hyperphosphorylation might facilitate neurite retraction (62). Pharmacological inhibition, dominant negative down-regulation, and siRNA knockdown of PP2Ac lead to neurite retraction and inhibition of axiogenesis (63). Thus, I 2 PP2A could contribute to the destabilization of microtubules indirectly through PP2A-dependent phosphorylation states of Tau in our cell model. However, we cannot rule out the possibility that the effects of PP2A observed in the present study on neurite outgrowth and microtubule stability are partially Tau-independent because PP2A is not only the major Tau phosphatase but also the major brain phosphatase, which plays a crucial role in regulating most cellular functions, including neurite outgrowth (64,65).
In addition to the direct action of PP2A on Tau, the activation of Tau protein kinases can also induce Tau hyperphosphorylation. For instance, the proline-independent 12E8 site is phosphorylated by CaMKII, PKA, and MARK (66,67). Ser 396 is almost exclusively phosphorylated by GSK-3␤, whereas both GSK-3␤ and Cdk5 phosphorylate Tau at Ser 404 . We found a significant activation of CAMKII, which is consistent with the report that the treatment of metabolically active rat brain slices by a PP2A inhibitor, okadaic acid, activated CAMKII activity and promoted Tau hyperphosphorylation at the 12E8 site (68). However, the levels of phosphorylated ERK1/2, MEK1/2, and p70 S6 kinase did not change among the cells. Overexpression of I 2 PP2A has been shown to increase the levels of ERK1 and ERK2 in HeLa cells (69), whereas Fukukawa (70) reported the opposite; overexpression of I 2 PP2A results in suppression of EGF-stimulated ERK activation, and knocking down I 2 PP2A by siRNA resulted in enhancement of the MEK/ERK pathway in HeLa cells. Phosphorylated ERK1/2 was reported to be reduced in head and neck squamous cell carcinoma stably expressing shRNA against I 2 PP2A (71). The exact reason for this discordance is not clear, but it is apparent that the regulation of the MEK/ERK pathway by I 2 PP2A is possibly a cell type-specific phenomenon; however, it demands further investigation. Additionally, we noticed a significant activation of GSK-3␤, as detected by a decrease in inhibitory phosphorylation at Ser 9 . Our current finding is consistent with previous reports that showed that accumulation of I 2 PP2A or overexpression of the C-terminal cleaved fragment of I 2 PP2A , I 2CTF , results in an increase in GSK-3␤ activity (32,72). Co-treatment with LiCl (GSK-3␤ inhibitor) and KN-93 (CAMKII inhibitor) significantly reduced the I 2 PP2A -induced Tau hyperphosphorylation. The exact reason underlying the GSK-3␤ activation in our model is not clear. However, the direct physical interaction between I 2 PP2A and GSK-3␤ could possibly explain the observed phenomenon. Another possible reason could be that PP1 activation could lead to the dephosphorylation of Ser(P) 9 -GSK-3␤ associated with its activation. Consistent with this possibility, it has been reported that I 1 PP2A and I 2 PP2A can markedly increase PP1c activity in the presence of a nearly physiological concentration of Mn 2ϩ in a substrate-specific manner (73). Although we did not measure PP1 activity, we could not detect any change in total PP1 protein level. On the other hand, coimmunoprecipitation experiments clearly suggest a physical association of I 2 PP2A with GSK-3␤ in contrast with CAMKII. Consistent with our finding, Liu et al. (72) showed that overactivation of GSK-3␤ inhibits PP2A through up-regulation of I 2 PP2A , indicating a positive correlation between I 2 PP2A and GSK-3␤ activity. We further confirmed the in vivo association of GSK-3␤ with I 2 PP2A in AD brain, where both are co-localized in the neuronal cytoplasm. Collectively, these results (Fig. 8) suggest 1) that inhibition of PP2A caused by cytoplasmic translocation of I 2 PP2A mimics the abnormal hyperphosphorylation of Tau, as observed in AD brain; 2) that the effect of mNLS-I 2 PP2A is modulated both directly by PP2A and indirectly by the Tau kinases that are regulated by PP2A; and 3) that the in vivo FIGURE 8. Proposed mechanism of I 2 PP2A -mediated Alzheimer-like abnormal hyperphosphorylation of Tau. Translocation of I 2 PP2A from the cell nucleus to the cytoplasm enables it to interact with PP2Ac and GSK-3␤, leading to the inhibition of the phosphatase and activation of the kinase. Activation of CaMKII activity is secondary to the I 2 PP2A -dependent PP2A inactivation. Collectively, all of these events contribute to the hyperphosphorylation of Tau. association of I 2 PP2A with GSK-3␤ could imply a direct role of I 2 PP2A in activation of GSK-3␤.
In conclusion, our study demonstrates a potential NLS at amino acids 179 -181, which regulates the translocation of I 2 PP2A from neuronal cytoplasm to nucleus, and offers the development of a Tet-Off-inducible cell model of abnormal hyperphosphorylation of Tau in which the NLS of I 2 PP2A is mutated, and it is translocated to the cell cytoplasm. Cytoplasmic retention of I 2 PP2A results in inhibition of PP2A directly and indirectly by PP2A-regulated Tau kinases, all of which eventually lead to the hyperphosphorylation of Tau. The restoration of PP2A activity through the inhibition of I 2 PP2A -PP2A interaction offers a promising therapeutic target for preventing the neurofibrillary degeneration of the abnormally phosphorylated Tau. Our cellular model of tauopathies is suitable for screening drugs/compounds that can attenuate the abnormal hyperphosphorylation of Tau.