Mammalian Target of Rapamycin (mTor) Mediates Tau Protein Dyshomeostasis

Background: Perturbations of the mammalian target of rapamycin (mTor) signaling pathway are implicated in Alzheimer disease (AD). Results: The activated mTor alters the activity of major tau kinases contributing to the formation of tau dyshomeostasis. Conclusion: We established a cellular system using genetic activation of mTor that developed authentic AD-like changes. Significance: The study provides potential tools for identifying tau-based therapeutics. Previous evidence from post-mortem Alzheimer disease (AD) brains and drug (especially rapamycin)-oriented in vitro and in vivo models implicated an aberrant accumulation of the mammalian target of rapamycin (mTor) in tangle-bearing neurons in AD brains and its role in the formation of abnormally hyperphosphorylated tau. Compelling evidence indicated that the sequential molecular events such as the synthesis and phosphorylation of tau can be regulated through p70 S6 kinase, the well characterized immediate downstream target of mTor. In the present study, we further identified that the active form of mTor per se accumulates in tangle-bearing neurons, particularly those at early stages in AD brains. By using mass spectrometry and Western blotting, we identified three phosphoepitopes of tau directly phosphorylated by mTor. We have developed a variety of stable cell lines with genetic modification of mTor activity using SH-SY5Y neuroblastoma cells as background. In these cellular systems, we not only confirmed the tau phosphorylation sites found in vitro but also found that mTor mediates the synthesis and aggregation of tau, resulting in compromised microtubule stability. Changes of mTor activity cause fluctuation of the level of a battery of tau kinases such as protein kinase A, v-Akt murine thymoma viral oncogene homolog-1, glycogen synthase kinase 3β, cyclin-dependent kinase 5, and tau protein phosphatase 2A. These results implicate mTor in promoting an imbalance of tau homeostasis, a condition required for neurons to maintain physiological function.

Alzheimer disease (AD), 2 the single major cause of dementia in middle and old aged individuals, is histopathologically characterized by filamentous lesions, such as those composed of the 39 -43-amino acid ␤-amyloid peptide and hyperphosphorylated tau (1). AD is multifactorial and heterogeneous. The vast majority of AD cases represent the so-called sporadic form of the disease, which is not associated with any known genetic mutation (2). The sporadic form of AD itself probably involves several different etiopathogenic mechanisms (2). Aging, neuroinflammation, head trauma, and diabetes have been implicated as risk factors for AD. The neurofibrillary degeneration is a slow and progressive retrograde neuronal degeneration; it is observed as neurofibrillary tangles (NFTs) of paired helical filaments (PHFs)/straight filaments in the cell soma, in dystrophic neurites surrounding the ␤-amyloid plaque core, and in neuropil threads (3). Although associations per se cannot prove cause-effect relationships, the formation of tau inclusions (NFTs) is widely thought to contribute to AD pathogenesis as NFT formation correlates with the duration and progression of AD (4). Both insoluble and soluble forms of abnormally hyperphosphorylated tau exist in AD brains, and they do not interact with tubulin (5,6). Furthermore, when the soluble form of abnormally hyperphosphorylated tau is present, it sequesters normal tau and microtubule-associated proteins 1 and 2 (7), accelerating disruption of the microtubule network.
It was demonstrated in transgenic mouse brains that the abnormal hyperphosphorylation of tau precedes the formation of NFTs and neuronal loss (8,9). The expression of tau pseudophosphorylated in vitro at Thr-212, Thr-231, and Ser-262 triggers apoptosis (10), which is accompanied by tau aggregation and breakdown of the microtubule network (10,11). On the other hand, the expression of wild type tau in vivo leads to synaptic loss, whereas deletion of tau rescues ␤-amyloid peptide-induced toxicity at the synapse (12)(13)(14)(15)(16). This evidence suggests that dysregulated production, phosphorylation, and aggregation of tau might be the key events that trigger neuronal degeneration in AD. However, little is known about the upstream intracellular effectors that account for these molecular events in the process of tau deposition, resulting in changes of neuronal function and cognitive decline, although activation of the crucial integrator of multiple signal pathways, mammalian target of rapamycin (mTor), has been proposed (17)(18)(19)(20)(21).
Gene sequence comparison identified that tau mRNA belongs to the 5Ј top mRNA group. It has been established that mTor activation via downstream S6K increases the translation of tau mRNA (19,36). It is plausible that activated mTor may facilitate tau deposition simply by increasing its translation in AD brains. Supporting this view, it has been shown in vitro that rapamycin suppresses tau translation, whereas constitutively active S6K increases tau translation (42). The levels of soluble p-tau were significantly reduced in the brains of rapamycintreated 3xTg-AD mice compared with non-treated controls (43).
In view of the evidence linking mTor to tau deposition, we reasoned that up-regulation of mTor in AD brains could be a key process of AD pathogenesis. However, the manner in which mTor interacts with tau and consequently leads to tau deposition has yet to be modeled. Here we first report that the active form of mTor aberrantly accumulates in NFT-bearing neurons, especially in those at the early stage, and mTor phosphorylates tau in AD epitopes in vitro. To further study the link between mTor and tau, we have created a series of cell lines, based on human SH-SY5Y neuroblastoma cells, with overexpression of wild type mTor, rapamycin-resistant mTor, mTor kinase-dead, and S6K kinase-dead and with suppression of mTor expression. We found that up-regulated mTor promotes tau dyshomeostasis by mediating the synthesis, phosphorylation, and aggregation of tau.
Immunohistochemistry and Immunofluorescence-Brain tissues from six AD cases and six age-matched non-neurological controls from The Netherlands Brain Bank were used in the study (Table 2). Paraffin sections (6 m) of the hippocampus and adjacent temporal cortex were deparaffinized in xylene, rehydrated, and incubated with primary antibodies. The bound rabbit anti-phosphorylated (p-) mTor Ser(P)-2448 or mTor Ser(P)-2481 antibodies were incubated with secondary antirabbit IgG (1:200), detected using the avidin-biotin system from Vectastain (BioNordika AB, Stockholm, Sweden), and visualized with 3,3Ј-diaminobenzidine (Sigma-Aldrich Sweden AB) as brown color as described previously (28). The sections were subsequently incubated with mouse monoclonal anti-tau PHF-1 or rabbit polyclonal anti-p-tau Ser(P)-422 and antimouse/rabbit IgGs (1:200) and visualized by Vector SG (Bio-Nordika AB) as dark gray/blue color. At least 10 contiguous microscopic fields were examined using a 20ϫ objective, and the enzyme-positive neurons that clearly had the perikaryon were counted (48).
For immunofluorescence staining, we followed the method described previously (21) with minor modification. After dewaxing, the sections were incubated with a mixture of primary antibodies against p-mTor Ser(P)-2448 and anti-p-tau (PHF-1 and AT8) overnight at 4°C. Unbound antibodies were removed by washing. Bound antibodies were detected by incubation for 1 h with DyLight 488-conjugated goat anti-mouse IgGs or DyLight 594-conjugated goat anti-rabbit IgGs (1:200 for both; Jackson ImmunoResearch Laboratories). After staining the nuclei with DAPI, fluorescence signals were assessed using confocal microscopy (Zeiss, Oberkochen, Germany). The number of pretangles and tangles including classic tangles and ghost tangles at a relatively early stage were distinguished following the criteria described previously (32) and examined in contiguous fluorescence microscopic fields using a 63ϫ objective.
In Vitro Phosphorylation and Mass Spectrometry-60 ng of mTor (1362-end) active kinase (having a specific activity of 413 units/mg where 1 unit is defined as 1 nmol of phosphate incorporated into 2 mg/ml substrate/min) or 60 ng of S6K (having a specific activity of 197 units/mg where 1 unit is defined as 1 nmol of phosphate incorporated into 100 M substrate/min) (Millipore AB, Solna, Sweden) was used to phosphorylate 10 g of purified tau (2N4R) protein in a buffer containing 450 mM HEPES, pH 7.5, 9 mM EGTA, and 0.09% Tween 20 in the presence of 0.5 mM ATP at 30°C for 2 h. The reaction was stopped by adding 3% formic acid and monitored by Western blotting with specific polyclonal antibody to phosphoserine. Samples that reacted positively were digested by trypsin (1:20 trypsin/ tau ratio) at 37°C for 20 h, and p-sites on tau were analyzed by strong cation exchange and TiO 2 -based fractionation followed by nano-LC and mass spectrometry analysis (49). The identified phosphorylation sites of tau were further validated by Western blots.
Overexpression of mTor and S6K in SH-SY5Y Cells-SH-SY5Y cells were maintained in Neurobasal medium supplemented with 5% fetal calf serum, 2 mM L-glutamine and incubated in a humidified 5% CO 2 atmosphere at 37°C. For stable  expression, SH-SY5Y cells were grown on 6-well plates until reaching 80 -90% confluence and transfected with 3 g of plasmids with various cDNAs of mTor and empty pcDNA3.0 vector/well using FuGENE HD (1:10 DNA) (Roche Applied Science). To make stable cells, after 48 h, the transfected cells were trypsinized and replated into 10-cm plates at various densities in culture medium containing 0.5 mg/ml G418. For S6K, SH-SY5Y cells were co-transfected with EECMV plasmids and pcDNA3.0 (1:10 ratio) in the presence of 0.5 mg/ml G418 in culture medium. After 7-10 days, the resistant cells were plated on 96-well plates (0.5 cell/well). The single clones that survived after G418 suppression were expanded. The expression efficiency was analyzed by Western blots using antibodies against the expressed proteins of targeted genes.
The resistant cells for S6K-SRs and mTor-SRs were expanded and plated on 96-well plates (0.5 cell/well). The single clones that survived after suppression of G418 (pcDNA3.0) and puromycin (pLko.1) were expanded. The silencing efficiency was checked by Western blots.
Cell Cultures and Sample Preparations-To control the expression levels of the targeted genes (S6K and mTor), after transfection, the cells were washed with ice-cold phosphatebuffered saline (PBS), harvested, and suspended in lysis buffer containing 40 mM HEPES, 120 mM NaCl, 1 mM EDTA, 10 mM pyrophosphate, 10 mM glycerophosphate, 50 mM NaF, 0.5 mM orthovanadate, 1% Triton X-100, and 1% protease inhibitor mixture on ice for 20 min. The supernatant samples were obtained after centrifuging at 13,000 ϫ g for 20 min.
For the majority of the experiments, the stable transfected cells were grown to 70 -80% confluence in 100-mm culture dishes using Dulbecco's modified Eagle's medium (DMEM)/ F-12 medium (1:1) supplemented with 10% fetal bovine serum (FBS). The cells were then cultured in 1% FBS medium for 24 h, and experiments were performed in serum-deprived conditions from 30 min to 8 h before cells were harvested. For experiments involving treatment with physiological (100 M) and pathophysiological (300 M) dosages of zinc that have been characterized in our previous studies (21,44,51), cell lysates were sonicated on ice and centrifuged at 1,000 -12,000 ϫ g at 4°C for 10 -20 min to collect supernatants free of nuclei and large cell debris.
For preparation of differentiated SH-SY5Y cells homogenates, SH-SY5Y cells with selectable pcDNA3.0 empty control vector were plated in 100-mm culture dishes at a density of 10 5 cells/cm 2 . After 24 h, the culture medium was changed to serum-free DMEM/F-12 medium supplemented with 10 M retinoic acid for 5 days and 0.5-5 ng/ml BDNF for 2 days. The cells were deprived of BDNF 1 day before samples were collected. After deprivation of both serum and BDNF, cells were washed with PBS and suspended in Triton lysis buffer (1% Triton X-100 in 50 mM Tris, 150 mM NaCl, pH 7.4) containing protease and phosphatase inhibitors such as 2 mM EGTA, 25 mM NaF, 200 M Na 3 VO 4 , 0.5 mM phenylmethylsulfonyl fluoride (PMSF), 5 mM EDTA, 1 M okadaic acid, and protease inhibitor mixture (1:200).
Isolation of Soluble and Insoluble Fractions-Cell lysates in Triton lysis buffer were sonicated on ice and centrifuged at 100,000 ϫ g at 4°C for 1 h for separation of the supernatant (cytosolic fraction) from the pellet (insoluble fraction), which was resuspended in Triton lysis buffer containing 3% SDS.
In Situ Microtubule Binding Ability Assay of Tau-The samples were prepared as described previously (52) with minor changes. In brief, 4.5 h after serum removal, cells were rinsed with warm PBS and suspended in warm microtubule-stabilizing buffer (80 mM PIPES/KOH, pH 6.8, 1 mM GTP, 1 mM MgCl 2 , 1 mM EGTA, 0.5% Triton X-100, 30% glycerol) containing 1 mM PMSF; 10 g/ml each aprotinin, leupeptin, and pepstatin; 1 M okadaic acid; and 10 M Taxol. The samples were then centrifuged at 5,000 ϫ g at room temperature for 10 min to remove nuclei. The postnuclear lysates were further centrifuged at 100,000 ϫ g at room temperature for 1 h. The supernatant was collected, and the pellet was rinsed twice, resuspended in microtubule-stabilizing buffer, and briefly sonicated.
Protein Measurement and Western Blotting-After cell culture, protein concentrations of samples prepared from the sta-ble cell lines were determined using a bicinchoninic acid (BCA) kit (Pierce). Equal amounts (20 -80 g/lane) of protein were loaded onto 8 -10% (w/v) SDS-polyacrylamide gels. Separated proteins were blotted onto PVDF membranes (Millipore AB) and blocked in 5% (w/v) nonfat milk diluted in Tris-buffered saline supplemented with 0.1% (v/v) Tween 20 (TBST) for 1 h. The membranes were incubated with primary antibodies (see Table 1) at 4°C overnight and then with secondary peroxidasecoupled anti-mouse or anti-rabbit antibodies (1:2,000; GE Healthcare AB) at room temperature for 1 h. After exposure to Hyperfilm MP (Amersham Biosciences), bound antibody intensity was analyzed using ImageJ software. Probed filters were stripped using stripping buffer (100 mM 2-mercaptoethanol, 2% SDS, 62.5 mM Tris-HCl, pH 6.8) at 50°C for 30 min with occasional agitation.
Statistical Analysis-For data from Western blotting, statistical comparisons between different experimental groups were performed using one-way analysis of variance followed by Bonferroni post hoc test analyses. A value of p Յ 0.05 was considered as significant.

RESULTS
Increase of mTor Immunoreactivity in Tangle-bearing Neurons in AD Brains-By double immunostaining, the mottled immunoaggregates of p-mTor Ser(P)-2448 were observed to be increased in CA1 pyramidal neurons that accumulate hyper-ptau Ser(P)-422 (Fig. 1A) and PHF-1 (Fig. 1B) (classic NFTs are indicated by fat white arrows). An increase of p-mTor Ser(P)-2448 was also found in the pretangle neurons that show positive granular structures or completely negative staining for anti-ptau (indicated by thin white arrows). The co-existence of p-tau labeled by PHF-1 and AT8 with p-mTor Ser(P)-2448 was further studied using immunofluorescence microscopy. A similar pattern of co-existence between p-tau (PHF-1 and AT8) and anti-p-mTor Ser(P)-2448 ( Fig. 1, C1, C2, C3, D1, D2, D3, E1, E2, E3, F1, F2, F3, G1, G2, G3, H1, H2, and H3) was found compared with the immunostainings visualized by 3,3Ј-diaminobenzidine (Fig. 1, A and B). More than 90% of the pretangles and more than 80% of the tangles co-existed to different extents with the p-mTor Ser(P)-2448 aggregates (Table 3). A small portion of pretangle neurons and normal looking neurons were positive only for p-tau (PHF-1 and AT8) and for p-mTor Ser(P)-2448 (not shown in picture), respectively. A relatively larger portion of tangle-bearing neurons only positive for p-tau (PHF-1 and AT8) was found, and no tangle-bearing neurons were only positive for p-mTor Ser(P)-2448. Dystrophic neurites positive for both PHF-tau (PHF-1 and AT8) and p-mTor Ser(P)-2448 were also observed (data not shown). Antibodies against p-mTor Ser(P)-2481 were also used but did not work properly in paraffin-embedded sections even after using different approaches for antigen retrieval such as 0.01 M citric sodium buffer, pH 6.0 at 80°C for 10 min by microwave or water bath or Ͼ100°C for 20 min by autoclave.
Phosphorylation of Tau by mTor in Vitro-To understand whether mTor can directly phosphorylate tau and how the global tau phosphorylation profile is mediated by mTor, we carried out in vitro phosphorylation of tau (2N4R) by active mTor. Mass spectrometry resulted in Ͼ88% coverage of tau protein and uncovered two phosphorylation sites, Ser-214 at the flanking region and Ser-356 at the microtubule binding region (Fig. 2, A-C). These two sites were also confirmed by Western blots (Fig. 2D). Additionally, increased phosphorylation on the Thr-231 site that was not covered by mass spectrometry was identified by Western blots (Fig. 2D). Other antibodies such as anti-p-tau Ser(P)-262 and PHF-1 were also used in Western blot analysis, and no additional phosphorylation sites were recognized by these antibodies.
Tau Synthesis and Phosphorylation in Cells Influenced by Genetic Interference of mTor Activity-To study how mTor is involved in tau synthesis and phosphorylation, we established stable cellular models with SH-SY5Y cells as background in which mTor activity is constitutively genetically modified. We centrifuged the cell lysates from these cell lines at 1,000 ϫ g to remove the nuclei, and the postnuclear samples were resolved by Western blotting. The SH-SY5Y cells that overexpressed m-WT or m-S showed an increased level of p-S6K Thr(P)-389, the well characterized downstream substrate of mTorC1, whereas no difference in the level of p-S6K Thr(P)-389 was found in the SH-SY5Y cells that overexpressed m-SD as compared with SH-SY5Y cells carrying empty pcDNA3.0 vector (V1) (Fig. 3A). The increased levels of p-S6K Thr(P)-389 in both m-WT and m-S cells indicated mTOR activity in these cell lines. The total S6K level was not changed among the four cell lines. In contrast, the SH-SY5Y cells in which mTor was knocked out (partial m-SR1; complete m-SR2) showed a decreased level of p-S6K in a dose-dependent manner as compared with empty pLko.1 vector (V2). The level of total S6K was not changed among the different cell lines (Fig. 3B).
We also analyzed the levels of tau in the postnuclear samples from these cell lines. Up-regulation of mTor activity by overexpression of m-WT increased levels of total (R134d), de-p-(tau-1), and p-tau (TG3 and PHF-1) at ϳ50 kDa as compared with V1 control, m-S, and m-SD cells (Fig. 3C). Down-regulation of mTor (m-SR1 and m-SR2) decreased the levels of all forms of tau (total, de-p-, and p-tau) (Fig. 3D). The increased level of de-p-tau might represent the newly synthesized tau driven by up-regulated mTorC1 activity. The mechanism by which m-S cells did not show increased levels of total and phosphorylated tau compared with V1 control cells remains to be investigated further.
As reported by other groups (53,54), we also noticed multiple bands of tau in SH-SY5Y cell homogenates when loading 30 g/lane (data not shown). The molecular weights of these bands were judged by both the positions of the molecular markers used in every electrophoresis and the existence of the positive bands (masses) in different preparations such as in differentiated SH-SY5Y cell lysates, purified soluble AD p-tau, insoluble PHF-tau, and homogenates prepared from murine neuroblastoma N2a cells, medial temporal cortex from a case of AD, and mouse brain cortex. When the experimental conditions were optimized by titrating different concentrations of primary antibodies or loading different amount of proteins (not shown), only the 50-(55) and/or 36-kDa bands remained when loading 40 g/lane (data not shown). Only the 50-kDa band remained for R134d, tau-1, TG3 (existing only in high protein amount; not shown), and PHF-1 (existing only in high protein amount; not shown), and only the 36-kDa band remained for anti-p-tau Ser(P)-262 and anti-p-tau Ser(P)-214. Both 50-and 36-kDa bands with stronger intensity at 36 kDa remained for anti-p-tau Ser(P)-356, and only 110-kDa band remained for anti-p-tau Thr(P)-212. The faint 50-kDa band for TG3 and PHF-1 was observed when more than 80 g/lane postnuclear supernatant from V1 control cells was loaded (not shown).
Two Faces of S6K in Tau Synthesis-When the kinase-dead form of S6K was stably overexpressed in SH-SY5Y cells (Fig.  4A), a dramatic increase of total S6K and a dramatic decrease of p-S6K Thr(P)-389 were observed as compared with V3 empty control vectors. Correspondingly, in a manner similar to mTor knock-out shown in Fig. 3D, all forms of tau (total, de-p-, and p-tau) were decreased. In an attempt to clarify the role of S6K in tau synthesis, we also tried to establish stable cell lines with S6K depletion. We managed to achieve the clones but failed to passage the cell lines as the growth rate of both cell lines (S6K-SR1 and S6K-SR2) was very slow. To our surprise, the growth rate of S6K-silenced cells (S6K-SR1 and S6K-SR2) was slower than mTor-silenced m-SR1 and m-SR2 cells. From the limited amount of cell lysates, we detected decreased S6K and increased total tau in S6K-SR1 cells as compared with V4 control vectors; S6K-SR2 cells did not show a significant change (Fig. 4B).
Rapamycin Decreases Tau Synthesis-To further explore the role of mTor in tau synthesis, V1 control cells were cultured and treated with 100 or 300 M zinc sulfate for 4 h in the presence of 20 ng/ml rapamycin. Both 100 and 300 M zinc induced increased levels of total tau (Fig. 5A), de-p-tau (tau-1) (Fig. 5B), and p-tau Ser(P)-214 (Fig. 5C). However, a consistent decreased effect for all forms of tau was observed in V1 control cells treated with both 100 and 300 M zinc in the presence of 20 ng/ml rapamycin as compared with only 100 M zinc-treated controls with the exception of total tau in the condition treated with 300 M zinc in the presence of rapamycin.
Tau Aggregation in Cells Influenced by Genetic Interference of mTor Activity-To evaluate tau aggregation mediated by mTor, the soluble and insoluble forms of p-tau were extracted with 1% Triton X-100 (Fig. 6, A and B). Tau in the insoluble fraction represents aggregated tau. Regarding the different p-tau sites in Western blots, PHF-1 mainly recognized p-tau mass at ϳ50 kDa. Both anti-p-tau Ser(P)-214 and anti-p-tau Ser(P)-262 recognized p-tau mass at ϳ36 kDa, whereas anti-ptau Ser(P)-356 recognized two p-tau masses at 50 and 36 kDa. p-tau recognized by PHF-1 was predominantly observed in the soluble fraction and was increased in m-WT cells as compared with V1 control cells (Fig. 6A, the histogram for PHF-1). In contrast, the increased level of p-tau Ser(P)-214 was predominantly located in the insoluble fraction of m-WT cells as compared with V1 control cells (Fig. 6A, the histogram for p-tau Ser(P)-214). Levels of p-tau Ser(P)-356 at ϳ50 kDa showed a substantial increase in both the soluble and insoluble fractions of m-WT cells, whereas levels of p-tau Ser(P)-356 at ϳ36 kDa were not changed (Fig. 6A, the histograms for p-tau Ser(P)-356). Levels of p-tau Ser(P)-262 were reduced in the soluble fraction but not changed in the insoluble fraction in m-WT cells compared with V1 control cells (Fig. 6A, the histograms for p-tau Ser(P)-262). , which is typical of phosphopeptide high energy collision-induced dissociation fragmentation spectra. The observed fragment ions are consistent with a phosphorylation located at serine 3 (i.e. serine 356 in the protein sequence). In this respect, of particular importance is the presence of the y9 ion (866.4479), which proves that threonine 8 is unmodified, leaving the side chain of serine 3 as the only possible phosphorylation site (B). The high energy collision-induced dissociation spectrum (and respective annotation) of the observed doubly charged precursor ion 573.7858 m/z matched to the phosphopeptide TPsLPTPPTR is shown. Observed fragment masses are shown in black, whereas theoretical fragment masses are shown in red. Mass measurement errors are well under 0.01 m/z in accordance with the high resolution used (7,500 at 400 m/z). Ions are denoted b ions (if extending from the N terminus) or y ions (from the C terminus), and some ions have incurred a loss of the phospho group (H 3 PO 4 ), which is typical of phosphopeptide high energy collision-induced dissociation fragmentation spectra. The observed fragment ions are consistent with a phosphorylation located at serine 3 (i.e. serine 214 in the protein sequence). In this respect, of particular importance is the presence of the ions y3, y4, y5, y6, y7, and b2 ions, which prove that all threonines in the peptide are unmodified, leaving the side chain of serine 3 as the only possible phosphorylation site (C). Western blot analyses (D) are highlighted in black. Increased phosphorylation on the Thr-231 site, which was not covered by mass spectrometry, was revealed by Western blot.
A large variation of p-tau level was observed in soluble and insoluble fractions among V2, m-SR1, and m-SR2 cells (Fig.  6B). In the soluble fraction, PHF-1 was not changed in m-SR1 but increased in m-SR2 as compared with V2. In the insoluble fraction, the PHF-1 signal was not changed among V2, m-SR1, and m-SR2 (Fig. 6B, the histogram for PHF-1). A decrease of p-tau Ser(P)-214 was found in insoluble fractions of m-SR1 cells, whereas an increase of p-tau Ser(P)-356 at ϳ50 kDa was found in both soluble and insoluble fractions of m-SR2 cells (Fig. 6B, the histograms for both p-tau Ser(P)-214 and p-tau Ser(P)-356). Levels of p-tau Ser(P)-262 were decreased in m-SR1 in the soluble fraction but not changed in the insoluble fraction of both m-SR1 and m-SR2 (Fig. 6B, the histogram for p-tau Ser(P)-262).
Microtubule Binding Capacity of tau in Cells Influenced by Genetic Interference of mTor Activity-As both mTor and S6K can phosphorylate the flanking and repeat regions of tau, the effect of mTor on the microtubule binding capacity of tau was analyzed. We found that m-WT cells showed a significant increase for total tau and p-tau Ser(P)-214 in the microtubuleindependent fraction (Fig. 7A, the histograms for both T tau and p-tau Ser(P)-214). No change was found for p-tau Ser(P)-356 in the microtubule-independent fraction. A significant increase of total tau and p-tau Ser(P)-356 in the microtubuledependent fraction was also found in m-WT cells (Fig. 7B, the histograms for T tau (R134d) and p-tau Ser(P)-356) as compared with V1 control cells. Genetic inactivation of mTor activity in m-SR1 and S6K-KD cells reduced total tau in both microtubule-independent and -dependent fractions as well as p-tau  Ser(P)-214 in the microtubule-independent fraction. m-SR1 but not S6K-KD reduced p-tau Ser(P)-214 in the microtubuledependent fraction. No change was found for p-tau Ser(P)-356 in m-SR1 and S6K-KD cells compared with control cells.
Immunoreactivities of PKA, Akt, GSK-3␤, Cdk5, and PP2A in Cells Influenced by Genetic Interference of mTor Activity-The immunoreactivities of antibodies to these enzymes were measured in the postnuclear supernatants (Fig. 8A). Although V1 control cells were negative for p-Akt Ser(P)-473, the well characterized downstream substrate of mTorC2, the m-WT, m-S, and m-SD cells showed a dramatic increase for p-Akt Ser(P)-473 (Fig. 8B). A stepwise decrease of p-Akt Ser(P)-473 was observed from m-WT to m-S to m-SD, and mTor silencing (m-SR1 and m-SR2) showed a nearly negative level of p-Akt Ser(P)-473 compared with V2 control. The S6K-KD cells showed a dramatic increase of p-Akt Ser(P)-473 compared with V3 control.
The level of p-GSK-3␤ Ser(P)-9 was increased in m-WT cells but was not changed in either m-S or m-SD cells as compared with V1 control cells (Fig. 8D). The mTor partial deletion (m-SR1) showed a reduction of p-GSK-3␤ Ser(P)-9, whereas the mTor complete deletion (m-SR2) showed no change for p-GSK-3␤ Ser(P)-9, suggesting the existence of a compensatory activation of other kinases regulating GSK-3␤ phosphorylation at the Ser-9 site. The level of total GSK-3␤ was highly increased in m-WT, m-S, and m-SD cells as compared with V1 control cells (Fig. 8E). Relatively higher and lower levels of total GSK-3␤ were observed in m-S and m-SD, respectively, as compared with m-WT. mTor depletion did not influence the level of total GSK-3␤. The S6K-KD cells showed a slight decrease of p-GSK-3␤ Ser(P)-9 as compared withV3 control, although the total level of GSK-3␤ was dramatically increased (Fig. 8, D  and E).
Both PKA␣ and PKA␤ were increased in m-WT, m-S, and m-SD cells as compared with V1 control (Fig. 8, F and G). PKA␣ decreased in a stepwise manner from m-SR1 to m-SR2, whereas PKA␤ showed no change in either m-SR1 or m-SR2 as compared with V2 control. Both PKA␣ and PKA␤ were increased in S6K-KD cells compared with V3 control.
The level of Cdk5 seemed to be suppressed in m-WT, m-S, and m-SD but was not changed in cells with mTor depletion (m-SR1 and m-SR2) or in cells that overexpressed S6K-KD (Fig.  8H). The m-WT, m-S, and m-SD cells showed an increase for both p-PP2A Tyr(P)-307 and total PP2A as compared with V1 control cells (Fig. 8, I and J). Silencing of mTor and overexpression of S6K-KD resulted in no reduction in the levels of both total and phosphorylated PP2A catalytic subunit (Fig. 8, I and J) with the exception that complete silencing of mTor (m-SR2) caused a dramatic reduction of PP2A catalytic subunit.

DISCUSSION
Previously we have extensively studied the role of mTor signaling and tau hyperphosphorylation in SH-SY5Y cells, murine N2a neuroblastoma cells, rat primary neurons, and metabolically active rat brain slices that were treated with a physiological or pathological dosage of zinc (21,31,44,51,56,57). In the present study, the role of mTor in biochemical changes (translation, phosphorylation, and aggregation) of tau was studied in AD brains, by in vitro phosphorylation plus mass spectrometry, and in SH-SY5Y cells with different genetic modifications of mTor activity. mTor, a key sensor of cellular stress, maintains protein homeostasis (23,58). Its phosphorylation at Ser-2448 by Akt activates mTor, whereas phosphorylation at the Thr-2446 site by 5Ј-adenosine monophosphate-activated protein kinase inactivates mTor (23,58). Using dot blots, we previously found an ϳ3-fold increase in p-mTor Ser(P)-2481 in the homogenates of AD brains but no change in p-mTor Ser(P)-2448 (20), a site that signals upstream of Akt activity. However, a relatively more focalized change for p-mTor Ser(P)-2448 was found in special groups of neurons in the present study (Fig. 1). The active forms of both S6K and Akt were reported to be aberrantly accumulated in tangle-bearing neurons (21,32). Taken together, it suggests that both mTorC1 and mTorC2 are up-regulated in AD brains.
The increase of total tau in AD brains (59,60) is not likely caused by up-regulated transcription of tau gene but by an upregulated translation of tau mRNA because tau mRNA copy number is not changed in sporadic AD brains (61,62), and tau mRNA has 5Ј toplike structure that is preferentially regulated by the mTorC1-S6K pathway (36,42). Inhibition of mTor with rapamycin induces a decrease in the level of total tau both in vitro and in vivo (36,43,51). In the present study, we found that constitutive overexpression of mTor increased total tau, whereas constitutive deletion of mTor or constitutive overex- pressions of the inactive form of mTor and S6K decreased total tau (Figs. 3 and 4). Taking into consideration the previous data reported in our laboratory that mTor signaling is up-regulated in AD brains (20,21), it is suggested that mTorC1-mediated tau translation contributes to the significant amount of normal tau remaining in the homogenates of AD brains (59,60).
We reported that a ϳ10-fold increase in AD brains is selectively found for p-tau epitopes Thr-217, Ser-202, Thr-231, and Thr-231/Ser-235, sites that are located at the flanking region (46). Pseudophosphorylating tau at the flanking and repeat regions (Thr(P)-212, p-Thr(P)-231, and Ser(P)-262) not only results in the loss of its normal function but also the gain of a toxic activity that causes disruption of the microtubule network and cell death (10). Tau in SH-SY5Y cells is hyperphosphorylated at some of the same sites as AD soluble p-tau purified from AD brains such as Ser-262, Thr-231/Ser-235, Ser-396/404, and Ser-214 (55,63), and it does not bind to Taxol-stabilized microtubules but fully inhibits tau-promoted microtubule assembly in vitro, although about 3 times more SH-SY5Y tau was required to achieve the same inhibition as AD p-tau. Interestingly, results from the in vitro phosphorylation assay and cellular models in the present study showed that mTor phosphorylates tau at Ser-214, Thr-231, and Ser-356 (Figs. 2-4), a similar group of epitopes shared by other prosurvival signals such as Akt and S6K (36,64,65) that exclusively mediate the phosphorylation sites of tau at flanking and repeat regions, resulting in inhibition of microtubule binding (66,67). These signals are critical for converting tau into a toxic molecule.
Hyperphosphorylation contributes to the aberrant formation of insoluble tau aggregates (10,68), but the mechanisms that bridge the hyperphosphorylation and aggregation of tau are poorly understood. Epitopes of tau such as Ser-214 at ϳ36 kDa or Ser-356 at ϳ50 kDa induced by mTor were principally extracted in the Triton X-100-insoluble pellet, suggesting that mTor mediates the process of converting tau from a soluble hyperphosphorylated form to an aggregated form.
Efforts aiming to elucidate the mechanisms of abnormal tau hyperphosphorylation, a major step of tau deposition in neurons of AD brains, have led to the identification of several tau protein kinases, PKA, Akt, GSK-3␤, S6K, and Cdk5, and a tau protein phosphatase, PP2A (21,27,28,31,35,55). Results from the present study confirmed that the level of phosphorylated GSK-3␤ is regulated by mTorC2. It is known that GSK-3␤ is constitutively phosphorylated at Tyr-216 site when it is expressed (29,32), and GSK-3␤ Ser-9 is sensitive to phosphorylation regulation by Akt via phosphoinositidedependent kinase 1 and phosphoinositide-dependent kinase 2 (mTorC2) in response to fluctuations in upstream signals (22,23).
Data from the present study indicated that genetic up-regulation of mTor increases the levels of phosphorylated Akt and GSK-3␤, the levels of both total and phosphorylated PP2A, and the protein levels of both PKA␣ and PKA␤, but the level of the PI3K-mTorC1 downstream target Cdk5 was decreased (Fig. 8). It has also been shown that mTor phosphorylates and suppresses PP2A activity, and inhibition of both mTorC1 and mTorC2 with LY294002 and selective inhibition of mTorC1 with rapamycin decrease the phosphorylation of PP2A catalytic subunit (41). The impact of mTor on the immunoreactivities of these enzymes suggests that the end phosphorylation status of tau in neurons with up-regulated mTor activity observed in AD brains is a synergistic action of multiple upand or downstream signaling targets of mTor including itself with the exception of GSK-3␤ and Cdk5 because they were suppressed by up-regulated mTor. There are discrepancies regarding the effects of mTor on the activities of PKA, Akt, and PP2A and the various aspects of tau homeostasis among m-WT, m-S, and m-SD cells. It is speculated that mutation of mTor at rapamycin-resistant site changes its mechanism of interaction with tau and the upstream signaling that regulates tau phosphorylation.
In summary, we have established a simple but highly reproducible cellular system of tau hyperphosphorylation and aggregation that recapitulates some key features of AD. We have used this system to gain important insights into the role of mTor in tau hyperphosphorylation at the flanking and microtubule binding regions. Taking into consideration that mTor is a major intracellular hub integrating divergent intracellular and extracellular signals and that it is activated by these signals in neurons in the aging process, our results suggest a central role for mTor in the onset and progression of tau pathology in sporadic AD. Our cell-based models not only provide a valuable tool for studying the pathogenesis of tau abnormal translation, hyperphosphorylation, and aggregation but also offer a potential system for identifying therapeutic strategies against neurodegenerative tauopathies such as small molecules inhibiting translation, hyperphosphorylation, and fibrillation of tau or promoting microtubule stability.