HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 279, Issue 25, 26105-26114, June 18, 2004

This Article Abstract Full Text (PDF) All Versions of this Article: 279/25/26105    most recent M308298200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Yuan, Z. Articles by Paudel, H. K. Search for Related Content PubMed PubMed Citation Articles by Yuan, Z. Articles by Paudel, H. K. Social Bookmarking                      What's this?

14-3-3 Binds to and Mediates Phosphorylation of Microtubule-associated Tau Protein by Ser⁹-phosphorylated Glycogen Synthase Kinase 3 in the Brain^*

Zongfei Yuan, Alka Agarwal-Mawal, and Hemant K. Paudel¶

From the Bloomfield Center for Research in Aging, Lady Davis Institute for Medical Research, Sir Mortimer B. Davis-Jewish General Hospital and the Department of Neurology and Neurosurgery, McGill University, Montreal, Quebec H3T 1E2, Canada

Received for publication, July 30, 2003 , and in revised form, February 16, 2004.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
In mammalian brain, tau, glycogen synthase kinase 3 (GSK3), and 14-3-3, a phosphoserine-binding protein, are parts of a multiprotein tau phosphorylation complex. Within the complex, 14-3-3 simultaneously binds to tau and GSK3 (Agarwal-Mawal, A., Qureshi, H. Y., Cafferty, P. W., Yuan, Z., Han, D., Lin, R., and Paudel, H. K. (2003) J. Biol. Chem. 278, 12722–12728). The molecular mechanism by which 14-3-3 connects GSK3 to tau within the complex is not clear. In this study, we find that GSK3 within the tau phosphorylation complex is phosphorylated on Ser⁹. From extracts of rat brain and rat primary cultured neurons, Ser⁹-phosphorylated GSK3 precipitates with glutathione-agarose beads coated with glutathione S-transferase-14-3-3. Similarly, from rat brain extract, Ser⁹-phosphorylated GSK3 co-immunoprecipitates with tau. In vitro, 14-3-3 binds to GSK3 only when the kinase is phosphorylated on Ser⁹. In transfected HEK-293 cells, 14-3-3 binds to Ser⁹-phosphorylated GSK3 and does not bind to GSK3 (S9A). Tau, on the other hand, binds to both GSK3 (WT) and GSK3 (S9A). Moreover, 14-3-3 enhances the binding of tau with Ser⁹-phosphorylated GSK3 by 3-fold but not with GSK3 (S9A). Similarly, 14-3-3 stimulates phosphorylation of tau by Ser⁹-phosphorylated GSK3 but not by GSK3 (S9A). In transfected HEK-293 cells, Ser⁹ phosphorylation suppresses GSK3-catalyzed tau phosphorylation in the absence of 14-3-3. In the presence of 14-3-3, however, Ser⁹-phosphorylated GSK3 remains active and phosphorylates tau. Our data indicate that within the tau phosphorylation complex, 14-3-3 connects Ser⁹-phosphorylated GSK3 to tau and Ser⁹-phosphorylated GSK3 phosphorylates tau.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Tau is a neuronal microtubule-associated protein. It is involved in the regulation of microtubule dynamics, axonal transport, and neuronal morphology. In the brain, tau binds to microtubules, promotes microtubule assembly, and stabilizes microtubule structure (for reviews, see Refs. 1–3). Tau is phosphorylated in vivo, and phosphorylation reduces the affinity of tau for microtubules (1, 3). Recently, tau has become a center of investigation due to its role in the various neurological disorders collectively called tauopathies, which include Alzheimer's disease (AD)¹ and FTDP-17 (2, 4). Tau mutations have been discovered in familial FTDP-17. These mutations cause tau dysfunction and neurodegeneration in cell and animal models (2, 4). In AD, abnormally hyperphosphorylated tau (phosphorylated on more sites than normal tau) aggregates and forms paired helical filaments (PHFs) (5, 6). PHFs are the main structural component of neurofibrillary tangles, which are one of the characteristic neuropathological lesions found in the brains of patients suffering from AD (2, 3). Since abnormally hyperphosphorylated tau, which does not bind to microtubules, regains its biological activity upon dephosphorylation, abnormal hyperphosphorylation has been suggested as being one of the central events in the development of neurofibrillary pathology (2, 3). Abnormal tau phosphorylation causes loss of tau function, microtubule instability, and neurodegeneration in AD brain. Preventing or reducing excessive tau phosphorylation has been suggested as a possible therapeutic strategy in AD research (2). A loss in the regulatory mechanism that controls tau phosphorylation in normal brain was suggested to cause abnormal tau phosphorylation in AD brain (2, 3). The mechanism by which tau is phosphorylated in the brain is not clear.

Tau is phosphorylated by many kinases in vitro. These kinases include microtubule-associated protein kinase, cAMP-dependent protein kinase, calmodulin-dependent protein kinase 2, cyclin-dependent protein kinase 5, and glycogen synthase kinase 3 (GSK3) (see Ref. 7 for a list). Biochemical studies have shown that GSK3 is physically complexed with tau in the brain and phosphorylates tau in vitro (8, 9) and in vivo (10–13). GSK3 is a major tau kinase in the brain (8).

GSK3 phosphorylates various proteins and regulates diverse cellular processes such as glycogen metabolism, cell fate specification, microtubule dynamics, oncogenesis, and apoptosis (14, 15). GSK3 contains a C-terminal catalytic domain and an N-terminal regulatory loop (16, 17). Phosphorylation on Ser⁹, located within the N-terminal loop, inhibits GSK3 activity in vitro and in various cellular settings (15, 17–23). Ser⁹ phosphorylation is one of the major mechanisms regulating GSK3 activity (17–19). During glycogen metabolism, insulin activates AKT (protein kinase B) (20). AKT then phosphorylates GSK3 on Ser⁹ (15, 20). This phosphorylation down-regulates GSK3 activity and allows glycogen synthase to escape GSK3-mediated inhibition (15–20). In neurons, however, although insulin-activated AKT phosphorylates GSK3 on Ser⁹ (21, 23), this phosphorylation stimulates GSK3-catalyzed tau phosphorylation and dissociation of tau from microtubules (24–26). These observations suggest that in neurons, Ser⁹-phosphorylated GSK3 may phosphorylate tau. As discussed above, Ser⁹ phosphorylation is implicated in down-regulating GSK3 activity. These observations raise the question of how Ser⁹-phosphorylated GSK3 could phosphorylate tau in the brain.

Recently, we found that GSK3 and tau are parts of a 500-kDa multiprotein tau phosphorylation complex (8). More recently, we reported 14-3-3 as the third component of the complex (9). Within the complex, 14-3-3 binds to GSK3, connects it to tau, and stimulates GSK3-catalyzed tau phosphorylation (9). The biochemical mechanism by which 14-3-3 interacts with GSK3 is not known. A characteristic feature of 14-3-3 is that it binds to a phosphorylated serine within its target protein (27–31). Studies based on synthetic peptides derived from Raf suggest that the 14-3-3-binding region consists of a short stretch of amino acid residues containing a phosphoserine (27, 30). Within this sequence, an Arg residue at position –3 relative to the phosphoserine was suggested to be critical for 14-3-3 binding (30). GSK3 has an Arg at position –3 relative to Ser⁹ (16). Many 14-3-3-binding proteins contain a Ser or a Thr at position –1 and/or –2 with respect to phosphoserine involved in 14-3-3 binding (see Ref. 27 for a list). GSK3 contains a Thr at both positions –1 and –2 with respect to Ser⁹ (16). Like BAD, tyrosine phosphatase PTH1, and Wee1, which contain a hydrophobic residue at position +1 with respect to their 14-3-3-binding phosphoserine (32–34), GSK3 contains a Phe residue at position +1 with respect to Ser⁹ (16). These observations suggest that GSK3 contains a 14-3-3-binding sequence around its Ser⁹. Since GSK3 is phosphorylated on Ser⁹ in vivo (19–21), we have investigated the interaction of 14-3-3 and GSK3 within the tau phosphorylation complex. Herein, we report that 14-3-3 binds to phosphorylated Ser⁹ of GSK3 in the brain. We also show that 14-3-3 connects Ser⁹-phosphorylated GSK3 to tau and facilitates tau phosphorylation by Ser⁹-phosphorylated GSK3. Our results suggest a novel mechanism that explains how GSK3, upon phosphorylation of Ser⁹, in response to various growth factors, may phosphorylate tau in the brain (24–26).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
cDNA Cloning and Plasmids—Human 14-3-3 in pGEX6p vector was subcloned into EcoRI/XbaI sites of a pcDNA3.1/Myc vector as described previously (9) using forward primer 5'-CCC CGG AAT TCG GAT AAA AAT GAG CTG GTT-3' and reverse primer 5'-CCC GCT CTA GAG TTA ATT TTC CCC TCC TTC-3'. pcDNA3.1 vector containing hemagglutinin (HA)-GSK3 (S9A) and pcEF vector containing HA-AKT were gifts from Dr. J. R. Woodgett (University of Toronto) and Dr. Silvio Gutkind (National Institutes of Health, Bethesda, MD). Other vectors, pcDNA3.1-FLAG-tau, pcDNA3.1-HA-GSK3, and 14-3-3-pGEX6p, are described previously (9).

Cell Culture and Transfection—To culture primary neurons, 15 fresh brains of 1-day postnatal rat pups were cut into small pieces and then transferred into a 50-ml Falcon tube containing 30 ml of PBS (25 mM Na₂HPO₄ (pH 7.4), 137 mM NaCl). After adding 3 ml of trypsin (2.5%), the tube was incubated for 15 min at 37 °C with occasional mixing by tube inversion. To the incubated mixture, 3 ml of fetal bovine serum (HyClone, Logan, UT) followed by 4 ml of DNase I (1 mg/ml) were added. The tube was subjected to five or six inversions followed by gentle trituration using a 25-ml glass pipette until the mixture became homogeneous. The mixture was filtered twice through a nylon membrane. The neurons in the filtrate were pelleted by centrifugation for 10 min at 10 °C. The pellet was washed twice with PBS and then dispersed in 45 ml of culture medium (minimum essential medium (Invitrogen) supplemented with 30 mM glucose, 2 mM glutamine, 1 mM pyruvate, and 10% fetal bovine serum). Neurons in the dispersed solution were plated at 3 x 10⁶ cells/mm in 35-mm poly-D-lysine-coated dishes and placed in aCO₂ incubator maintained at 37 °C. The culture medium was replaced with fresh culture medium supplemented with 1 mM fluorodeoxyuridine (Sigma) on the second and the third days. On the fourth day, the culture medium was again replaced with fresh culture medium without fluorodeoxyuridine. On the seventh day, neurons were scraped and suspended in 0.5 ml of lysis buffer (50 mM Tris-HCl (pH 7.4), 0.15 M NaCl, 25 mM -glycerophosphate, 1 mM EDTA, 1 mM EGTA, 10 mM NaF, 10 mM MgCl₂, 0.1% Nonidet P-40, 25 nM okadaic acid, 4 nM cypermethrin, 1 mM phenylmethylsulfonyl fluoride, and 2 µg/ml each of leupeptin, pepstatin, and aprotinin). The suspension was incubated on ice with occasional shaking for 30 min and centrifuged, and the supernatant was used for a GST pull-down assay.

HEK-293 cells maintained in Dulbecco's modified high glucose Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum were plated on 10-cm dishes. Once 80% confluent, cells were transfected with 0.2–10 µg of cDNA using LipofectAMINE Plus reagent (Invitrogen) following the manufacturer's instructions. The amount of DNA used for each transfection in various experiments was as follows: 1 µg of tau, 1 µg of HA-GSK3 (WT), 3 µg of HA-GSK3 (S9A), 6 µg of HA-AKT, and 0.2 µg of Myc-14-3-3. For each transfection, the amount of total DNA never exceeded 10 µg. Cells were harvested 72 h after transfection and used for GST pull-down assays, immunoblottings, and immunoprecipitations.

Proteins and Antibodies—Recombinant GST-GSK3 and GST-14-3-3 were purified from bacterial lysates overexpressing respective proteins by glutathione-agarose chromatography (8, 9). Following purification, GST tag was removed as described (35). Protein phosphatase 1 was purified from bacterial lysate overexpressing human protein phosphatase 1 (36, 37). Tau phosphorylation complex was purified from fresh bovine brain extract as described previously (8, 9). Effluent fractions from phosphocellulose chromatography containing tau phosphorylation complex were combined and concentrated by dialysis using Aquacide III (Calbiochem) and used to generate Fig. 7A. To prepare Ser⁹-phosphorylated GSK3, GSK3 was phosphorylated by AKT at 30 °C for 1 h in a mixture containing 25 mM Hepes (pH 7.2), 0.1 mM EDTA, 0.1 mM DTT, 0.2 mM [-³²P]ATP, 1 mM MgCl₂, 0.2 mg/ml GSK3, and 1 µg/ml AKT (Upstate Biotechnology, Inc., Lake Placid, NY). Phosphorylated GSK3 was dialyzed against 25 mM Hepes (pH 7.2), 1 mM EDTA, and 1 mM DTT for 16 h at 4 °C. To prepare dephosphorylated GSK3, the above phosphorylated GSK3 was incubated at 30 °C for 1 h in a mixture containing 25 mM Hepes (pH 7.2), 1 mM EDTA, 1 mM DTT, 0.5 mM MnCl₂, 0.1 µg/ml protein phosphatase 1, and 0.15 mg/ml GSK3. Samples were analyzed by SDS-PAGE followed by autoradiography to monitor phosphorylation and dephosphorylation of GSK3. AKT incorporated 0.3 mol of phosphate/mol of GSK3, whereas protein phosphatase 1 almost completely dephosphorylated AKT-phosphorylated GSK3.

View larger version (29K): [in this window] [in a new window]   FIG. 7. Tau-associated GSK3 in the brain is Ser9-phosphorylated. To examine whether tau bound GSK3 in the brain is Ser9-phosphorylated, we either analyzed a partially purified tau phosphorylation complex from bovine brain extract by FPLC gel filtration or performed a co-immunoprecipitation. A, FPLC gel filtration. Gel filtration was carried out at 4 °C using an Amersham Biosciences FPLC system and a FPLC Superose 12 gel filtration column (50 x 1.6 cm) equilibrated in 25 mM MOPS (pH 7.2), 0.1 mM EDTA, 0.1 mM DTT, 200 mM NaCl, 50 mM -glycerophosphate, and 10 mM NaF. Fractions (1 ml each) were collected, and aliquots (15 µl each) from the indicated fractions were used for GSK3 activity assay (8) or immunoblot analysis using the indicated antibodies. B–D, co-immunoprecipitation. A fresh rat brain extract was subjected to immunoprecipitation using the indicated antibodies. Each resulting immune complex was then immunoblotted with the indicated antibody. Similar observations were made in at least four independent experiments.

Immunoprecipitation—HEK-293 cells transfected with various genes were washed with cold phosphate-buffered saline, scraped, and suspended in 0.8 ml of cold lysis buffer. The suspension was incubated on ice for 30 min with occasional gentle shaking and then centrifuged at full speed using a bench top centrifuge at 4 °C. The supernatant was used for immunoprecipitation, GST pull-down assay, and immunoblotting. The procedure for immunoprecipitation is essentially as described (9).

To generate Figs. 1 and 7, B–D, 4.5 g of fresh rat brain was cut into small pieces and then homogenized in 4 ml of lysis buffer using a glass homogenizer. The resulting homogenate was centrifuged at 27,000 x g for 40 min at 4 °C. The supernatant was used for immunoprecipitation using the indicated antibodies as described (8).

View larger version (43K): [in this window] [in a new window]   FIG. 1. Ser9-phosphorylated GSK3 precipitates with GST-14-3-3 from brain extract. Glutathione-agarose beads coated with GST or GST-14-3-3 were incubated with fresh rat brain extract. After incubation, beads were washed and subjected to SDS-PAGE or immunoblotted with the indicated antibodies. A, Coomassie Brilliant Blue-stained SDS gel; B and C, immunoblots (IB). Anti-GSK3 cross-reacts with GSK3, whereas anti-GSK3-pS9 is specific for GSK3 and GSK3 phosphorylated on Ser9 and Ser21, respectively. Similar observations were made in three different experiments.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Ser⁹-phosphorylated GSK3 Precipitates with GST-14-3-3 from Brain Extract—14-3-3 binds to GSK3 in vivo (9). Because 14-3-3 is a phosphoserine-binding protein (28–31) and GSK3 is phosphorylated on Ser⁹ in vivo (18, 20), we asked whether 14-3-3 binds to Ser⁹-phosphorylated GSK3. We incubated glutathione-agarose beads coated with GST-14-3-3 or GST with the lysates from a fresh rat brain extract. Beads were washed and analyzed by SDS-PAGE and immunoblotted with anti-GSK3 antibody. A protein-stained gel showed a major 55-kDa GST-14-3-3 band and several protein bands of various sizes in the lane containing GST-14-3-3 (Fig. 1A, lane 3). This observation is as expected, since 14-3-3 binds to various proteins of different sizes (27, 31). Importantly, an immunoblot analysis indicated that GSK3 also came down with GST-14-3-3 from brain extract (Fig. 1B, lane 3). Furthermore, GSK3, which bound to GST-14-3-3 was immunoreactive to a monoclonal antibody, anti-GSK3-pS9, that specifically recognizes GSK3 phosphorylated on Ser⁹ and GSK3 phosphorylated on Ser²¹ (Fig. 1C, lane 3).

To determine whether the above observation is due to 14-3-3-bound GSK3 being phosphorylated on Ser⁹ or due to a nonspecific interaction of the antibody used to generate Fig. 1C, we repeated the above experiment using a polyclonal antibody specific for GSK3 phosphorylated on Ser⁹ and GSK3 phosphorylated on Ser²¹. Again, GSK3 that came down with glutathione-agarose beads coated with GST-14-3-3 displayed immunoreactivity (data not shown). To substantiate the above observations, we mixed glutathione-agarose beads coated with GST or GST-14-3-3 with fresh lysates from cultured primary rat neurons. Beads were washed and then immunoblotted with anti-GSK3 or anti-GSK3-pS9 antibody. As expected, GSK3 specifically bound to GST-14-3-3 and was phosphorylated on Ser⁹ (data not shown). These observations suggested that 14-3-3 binds to Ser⁹-phosphorylated GSK3 in the brain.

14-3-3 Does Not Bind to GSK3 (S9A)—If 14-3-3 binds to Ser⁹-phosphorylated GSK3, mutation of Ser⁹ to Ala should block the interaction of 14-3-3 with GSK3. To test this, we transfected HEK-293 cells with HA-GSK3 (WT) or HA-GSK3 (S9A). Cells were lysed, and glutathione-agarose beads coated with GST-14-3-3 were incubated with each lysate. Beads were then washed and immunoblotted with anti-HA antibody to detect the bead-bound GSK3 species. As shown in Fig. 2A, HA-GSK3 (WT) bound to GST-14-3-3 (lane 3). HA-GSK3 (S9A), on the other hand, was not detected in the glutathione-agarose beads coated with both GST (lane 4) and GST-14-3-3 (lane 5). These observations indicate that 14-3-3 does not bind to GSK3 (S9A).

View larger version (36K): [in this window] [in a new window]   FIG. 2. Ser9 to Ala mutant of GSK3 does not bind to 14-3-3. HEK-293 cells transfected with the indicated constructs were lysed, and each lysate was subjected to GST pull-down assay or co-immunoprecipitation. A, GST pull-down assay. Glutathione-agarose beads coated with GST or GST-14-3-3 were incubated with lysates of cells transfected with either HA-GSK3 (WT) or HA-GSK3 (S9A). Beads were washed and immunoblotted (IB) with anti-HA antibody. Both HA-GSK3 (WT) and HA-GSK3 (S9A) migrate with similar sizes on a SDS-gel, are HA-tagged, and are recognized by anti-HA antibody. B and C, co-immunoprecipitation. HA-GSK3 (WT) and HA-GSK3 (S9A) were immunoprecipitated (IP) from respective cell lysates using anti-HA antibody. Each immune complex was then immunoblotted with the indicated antibody. D and E, immunoblots. Cell lysates used to generate B and C were immunoblotted with the indicated antibodies to monitor expression of the indicated genes. This experiment was repeated at least four times with similar results.

GSK3 Is Phosphorylated on Ser⁹ in HEK-293 Cells—We argued that 14-3-3 binds to phosphorylated Ser⁹ of GSK3. Therefore, within the HEK-293 cells used to generate Fig. 2, HA-GSK3 (WT) must have been phosphorylated on Ser⁹ by an endogenous kinase(s), and that may have allowed Myc-14-3-3 to bind to Ser⁹-phosphorylated HA-GSK3. HA-GSK3 (S9A), on the other hand, could not be phosphorylated on Ser⁹ and hence could not bind to Myc-14-3-3. To confirm this view, we co-transfected HA-GSK3 (WT) or HA-GSK3 (S9A) with HA-AKT into HEK-293 cells. Transfected cells were lysed, and each lysate was analyzed for HA-GSK3 phosphorylation by immunoblot analysis using anti-GSK3-pS9 antibody. Indeed, HA-GSK3 became phosphorylated on Ser⁹ when transfected alone (Fig. 3C, lane 3). When co-transfected with HA-AKT, HA-GSK3 was 3-fold more phosphorylated than when transfected alone (compare lanes 3 and 4 in Fig. 3C). HA-GSK3 (S9A), on the other hand, remained unphosphorylated when transfected alone or with HA-AKT (data not shown).

View larger version (31K): [in this window] [in a new window]   FIG. 3. GSK3 is phosphorylated on Ser9 in HEK-293 cells. Lysates of HEK-293 cells transfected with the indicated constructs were analyzed by immunoblot (IB) analysis using the indicated antibodies to monitor HA-GSK3 phosphorylation or expression of various constructs. A and B, expression of transfected genes. C, HA-GSK3 phosphorylation. Note that both HA-GSK3 and HA-AKT are immunoreactive against anti-HA antibody but migrate with different sizes on an SDS gel. Similar observations were made in at least three experiments.

View larger version (31K): [in this window] [in a new window]   FIG. 4. Effect of phosphorylation on the interaction of 14-3-3 with GSK3. Lysates of HEK-293 cells transfected with the indicated constructs and used to generate Fig. 3 were incubated with glutathione-agarose beads coated with GST or GST-14-3-3. Beads were washed and immunoblotted (IB) with anti-HA antibody to detect HA-GSK3. This experiment was repeated at least three times with similar results.

View larger version (28K): [in this window] [in a new window]   FIG. 5. Phosphorylation enhances the interaction of GSK3 with 14-3-3. Lysates of HEK-293 cells transfected with the indicated constructs were either immunoblotted (IB) to analyze GSK3 phosphorylation and expression of various genes or used to assay binding of Myc-14-3-3 to HA-GSK3 by co-immunoprecipitation. Blots corresponding to HA-GSK3 and Myc-14-3-3 were scanned, and band intensity values of various bands were obtained and used to determine the fraction of GSK3 phosphorylated or fraction of 14-3-3 bound to HA-GSK3. A, HA-GSK3 phosphorylation on Ser9. B–D, expression of various genes. E, co-immunoprecipitation (IP). GSK3 was immunoprecipitated from indicated lysates, and the resulting immune complex was immunoblotted with anti-Myc antibody. Note that both HA-GSK3 and HA-AKT are immunoreactive to anti-HA antibody. Therefore, we have used anti-GSK3 polyclonal antibody to immunoprecipitate HA-GSK3 in this experiment. F, a comparison between the relative fraction of HA-GSK3 phosphorylated and the relative fraction of 14-3-3 bound to HA-GSK3. To determine the relative fraction of GSK3 phosphorylated on Ser9, band intensity values of lanes 3 and 4 from A representing GSK3 phosphorylation were divided by respective band intensity values from B representing total GSK3. Each resulting value was then divided by the resulting value for lane 3. Similarly, to determine the relative fraction of Myc-14-3-3 bound to HA-GSK3, band intensity values of lanes 3 and 4 from E representing 14-3-3 co-immunoprecipitating with HA-GSK3 were divided by respective band intensity values from D, representing total Myc-14-3-3. Each resulting value for lanes 3 and 4 was then divided by the resulting value for lane 3. The values with S.E. represent three independent experiments.

Ser⁹-phosphorylated GSK3 Binds to 14-3-3—Finally, to show that 14-3-3 binds to Ser⁹-phosphorylated GSK3, we mixed 14-3-3 with GSK3, Ser⁹-phosphorylated GSK3, or Ser⁹-phosphorylated GSK3 that was dephosphorylated by protein phosphatase 1. We immunoprecipitated 14-3-3 from each mixture and then immunoblotted each resulting immune complex with anti-GSK3 antibody that recognizes both phosphorylated and nonphosphorylated GSK3. GSK3 co-immunoprecipitated with 14-3-3 from mixture containing 14-3-3 and Ser⁹-phosphorylated GSK3 (Fig. 6, lane 3) but not from those containing 14-3-3 and GSK3 (Fig. 6, lane 2) or 14-3-3 and dephosphorylated GSK3 (Fig. 6, lane 4). Thus, 14-3-3 did not bind to nonphosphorylated GSK3 but bound to Ser⁹-phosphorylated GSK3, and this binding was abolished when GSK3 was dephosphorylated. Based on these observations and the data from Figs. 1, 2, 3, 4, 5, we concluded that 14-3-3 binds to Ser⁹-phosphorylated GSK3 in the brain.

View larger version (28K): [in this window] [in a new window]   FIG. 6. 14-3-3 binds to Ser9-phosphorylated GSK3 in vitro. 14-3-3 was mixed with GSK3, Ser9-phosphorylated GSK3, or dephosphorylated GSK3 in a mixture containing 25 mM Tris-HCl (pH 7.5), 0.2 M NaCl, 0.1 mM EDTA, 0.1 mM DTT, 1 mg/ml bovine serum albumin, 10 µg/ml 14-3-3, and 10 µg/ml GSK3, Ser9-phosphorylated GSK3, or dephosphorylated GSK3. 14-3-3 was then immunoprecipitated (IP) from each mixture with anti-14-3-3 antibody. Each resulting immune complex was then immunoblotted (IB) with anti-GSK3 antibody. Similar observations were made in three different experiments.

To substantiate the above observation, we immunoprecipitated tau from a fresh rat brain extract. The anti-tau immune complex was then analyzed by immunoblot analysis. As observed previously (8, 9), both GSK3 (Fig. 7B, lane 3) and 14-3-3 (Fig. 7D, lane 3) co-immunoprecipitated with tau. Importantly, GSK3 in the anti-tau immune complex was Ser⁹-phosphorylated (Fig. 7C, lane 3). Thus, tau-bound GSK3 in the brain is Ser⁹-phosphorylated.

Tau Binds to Both GSK3 (WT) and GSK3 (S9A)—To elucidate the biochemical significance of the presence of Ser⁹-phosphorylated GSK3 within the tau phosphorylation complex (Fig. 7), we asked whether the binding of tau to GSK3 requires phosphorylation of GSK3 on Ser⁹. We co-transfected tau in HEK-293 cells with HA-GSK3 (WT) or HA-GSK3 (S9A). Transfected cells were lysed and subjected to immunoprecipitation using anti-HA antibody. Immunoblot analysis of each immune complex showed that tau co-immunoprecipitated with both HA-GSK3 (WT) (Fig. 8C, lane 2) and HA-GSK3 (S9A) (Fig. 8C, lane 3). Blot band intensity quantification indicated that only 1.3-fold more tau bound to HA-GSK3 (WT) than to HA-GSK3 (S9A) (Fig. 8E). These data indicate that tau binds to HA-GSK3 (WT) only slightly better than to GSK3 (S9A). This means that tau binds to both GSK3 (WT) and GSK3 (S9A), and GSK3 does not need to be phosphorylated on Ser⁹ in order to bind to tau.

View larger version (35K): [in this window] [in a new window]   FIG. 8. Effect of mutation of GSK3 Ser9 to Ala on the interaction of tau with GSK3. Lysates of HEK-293 cells transfected with the indicated constructs were either immunoblotted (IB) to monitor expression of various genes or subjected to co-immunoprecipitation (IP) to evaluate binding. The resulting blots were used to determine the fraction of tau bound to each GSK3 species. A and B, immunoblots. Aliquots (20-µl each) from the indicated fractions were immunoblotted with the indicated antibody to monitor expression of indicated constructs. C and D, co-immunoprecipitation. Indicated cell lysates were subjected to immunoprecipitation using anti-HA antibody. Each resulting anti-HA immune complex was then immunoblotted with the indicated antibody. Note that both HA-GSK3 (WT) and HA-GSK3 (S9A) efficiently immunoprecipitate with anti-HA antibody (D). E, fraction of tau bound. The fraction of tau bound to the indicated GSK3 species was determined by dividing the band intensity value of tau in lanes 2 and 3 of C, representing the amount of tau co-immunoprecipitating with each GSK3 species, by the corresponding tau band intensity in B, representing total tau. The values are averages of three independent determinations.

To test the above idea, we transfected HEK-293 cells with tau, Myc-14-3-3, and HA-GSK3 or HA-GSK3 (S9A). Transfected cells were lysed and subjected to immunoprecipitation using anti-HA antibody. This antibody effectively immunoprecipitated both HA-GSK3 (WT) and HA-GSK3 (S9A) from respective cell lysates in comparatively equal amounts (Fig. 9D, lanes 2 and 3). Myc-14-3-3 co-immunoprecipitated with HA-GSK3 (WT) (Fig. 9E, lane 3) but not with HA-GSK3 (S9A) (Fig. 9E, lane 2), indicating that 14-3-3 does not bind to GSK3 (S9A) significantly even in the presence of tau. Tau on the other hand, co-immunoprecipitated with both HA-GSK3 (WT) (Fig. 9F, lane 3) and HA-GSK3 (S9A) (Fig. 9F, lane 2). Intensities of various bands determined that 6-fold more tau co-immunoprecipitated with GSK3 (WT) than with HA-GSK3 (S9A) (Fig. 9G). These observations indicate that in the presence of 14-3-3, tau binds to HA-GSK3 (WT) much more efficiently than to HA-GSK3 (S9A). We then compared the fraction of tau bound to HA-GSK3 (WT) and HA-GSK3 (S9A) in the absence of 14-3-3 (Fig. 8E) with that in the presence of 14-3-3 (Fig. 9G). As shown in Fig. 9H, the fraction of tau bound to HA-GSK3 (WT) in the absence and the presence of Myc-14-3-3 is 0.6 and 2.2, respectively. The fraction of tau bound to HA-GSK3 (S9A), on the other hand, is 0.46 and 0.33 in the absence and the presence of Myc-14-3-3, respectively. This means 14-3-3 enhances the interaction of GSK3 (WT) with tau by more than 3-fold, whereas the interaction of tau with GSK3 (S9A) is insensitive to 14-3-3.

View larger version (32K): [in this window] [in a new window]   FIG. 9. Effect of 14-3-3 on the interaction of tau with GSK3 (WT) and GSK3 (S9A). Lysates from HEK-293 cells transfected with the indicated constructs were subjected to immunoblotting (IB) and immunoprecipitation (IP). Using the band intensity values of the resulting blots, fractions of tau bound to indicated GSK3 species were determined. A–C, immunoblots. Indicated cell lysates were immunoblotted with the indicated antibodies to monitor expression of indicated genes. D–F, co-immunoprecipitations. HA-GSK3 (WT) and HA-GSK3 (S9A) were immunoprecipitated from respective cell lysates using anti-HA antibody. Resulting immune complexes were immunoblotted using antibodies, as indicated. G, fraction of tau bound. To determine the fractions of tau bound to the indicated GSK3 species, band intensity values of tau bands on lanes 2 and 3 in F were divided by the tau band intensity of the corresponding lane in A. The values represent three independent determinations. H, a comparison of fractions of tau bound to GSK3 (WT) or GSK3 (S9A) in the presence of 14-3-3 or alone (in the absence of 14-3-3). Note that fraction values of tau bound to GSK3 (WT) and GSK3 (S9A) in the presence of 14-3-3 are from G. Fraction values of tau bound to GSK3 (WT) and GSK3 (S9A) alone are from Fig. 8E.

View larger version (31K): [in this window] [in a new window]   FIG. 10. Effect of 14-3-3 on tau phosphorylation by GSK3 (WT) and GSK3 (S9A). Lysates of HEK-293 cells transfected with tau, Myc-14-3-3, HA-GSK3 (WT), and HA-GSK3 (S9A) in the indicated combinations were analyzed by immunoblot (IB) analysis using the indicated antibodies to monitor expression of various constructs and tau phosphorylation. Monoclonal antibody PHF1 is specific for phosphorylated tau. Blots representing tau were scanned, and resulting band intensity values were used to determine the relative amount of tau phosphorylated. A–D, immunoblots. E, relative amount of tau phosphorylated. To determine the relative amount of tau phosphorylated, the band intensity value of tau in each of lanes 3–8 of A representing phosphorylated tau were divided by the band intensity values of tau in the corresponding lanes of B representing total tau. The resulting value for each lane was then divided by the resulting value for lane 3. The values are averages of three independent determinations.

Phosphorylation of Tau by Ser⁹-phosphorylated GSK3— Ser⁹ phosphorylation has been suggested to down-regulate GSK3 activity (17, 18). However, our data suggest that 14-3-3 facilitates tau phosphorylation by Ser⁹-phosphorylated GSK3. Therefore, to examine how phosphorylation of GSK3 on Ser⁹ affects tau phosphorylation by this kinase, we transfected HEK-293 cells with HA-GSK3, tau, HA-AKT, and Myc-14-3-3 in various combinations. Transfected cells were lysed, and cell lysates were evaluated for both phosphorylation of GSK3 on Ser⁹ and tau phosphorylation. Based on tau band intensities, the relative amounts of tau phosphorylated in various transfected cells were determined.

GSK3 was phosphorylated on Ser⁹ by an endogenous kinase in cells transfected with HA-GSK3 and tau (Fig. 11A, lanes 5) and cells transfected with HA-GSK3, Myc-14-3-3, and tau (Fig. 11A, lane 9). When co-transfected with HA-AKT, HA-GSK3 was highly phosphorylated on Ser⁹ (Fig. 11A, lanes 6–8).

View larger version (37K): [in this window] [in a new window]   FIG. 11. Effect of 14-3-3 on tau phosphorylation by Ser9-phosphorylated GSK3. HEK-293 cells transfected with the indicated constructs were lysed, and each lysate was analyzed by immunoblot (IB) analysis to evaluate HA-GSK3 phosphorylation on Ser9, expressions of various genes, tau phosphorylation, and relative amount of tau phosphorylated. A–F, immunoblots. G, relative amount of tau phosphorylated. Blots E and F were scanned, and intensity values of each tau band in lanes 2–9 from F were divided by the intensity value of the corresponding lane from E. Each resulting value was then divided by the resulting value for lane 2. The values are averages of three independent determinations.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
GSK3 has a kinase domain and a short N-terminal regulatory loop containing Ser⁹ (16, 17). In the nonphosphorylated state, the N-terminal loop does not interact with the kinase domain, and the kinase is fully active. Upon phosphorylation on Ser⁹, the N-terminal domain binds to the kinase domain, and phosphorylated Ser⁹ occupies the catalytic cleft, thus blocking substrate entry (17, 18). Therefore, in vitro phosphorylation of GSK3 on Ser⁹ inactivates its kinase activity almost completely (17, 19, 22). Ser⁹ phosphorylation has been considered as one of the mechanisms suppressing GSK3 activity in vitro and in vivo (14, 15, 17, 18).

In various cell types, AKT phosphorylates GSK3 on Ser⁹ but causes only partial inhibition of GSK3 activity. For example, in L8 myotubes, insulin stimulates AKT by more than 10-fold and causes phosphorylation of 60–100% of total GSK3 on Ser⁹, and only 40–50% of total GSK3 activity is inhibited (20). In NIH 3T3 cells, GSK3 is highly phosphorylated on Ser⁹ when cells are exposed to epidermal growth factor with only a 50% drop in GSK3 activity (38). In A431 cells, epidermal growth factor, which is known to stimulate AKT and phosphorylation of GSK3 on Ser⁹ (38), activates GSK3 (39). Several other studies have reported similar results in various cell lines (38–40). Thus, although in vitro Ser⁹-phosphorylated GSK3 is inactive (17, 19, 22), a pool of Ser⁹-phosphorylated GSK3 in vivo remains active (20, 38–41). These observations suggest that Ser⁹ phosphorylation of GSK3 has a role other than inhibiting kinase activity and that there is a mechanism to maintain GSK3 active upon its phosphorylation on Ser⁹.

GSK3 is a component of cell signaling pathways mediated by various growth factors including insulin, insulin-like growth factor (IGF), and fibroblast growth factor (FGF). These growth factors activate AKT, which then phosphorylates GSK3 on Ser⁹ (15, 20, 21). Mammalian brain is rich in insulin, IGF, and FGF receptors, and studies indicate that tau phosphorylation is regulated by these growth factors in mature and developing brains as well as in differentiating neurons in culture (21, 24–26, 42). When human neuroblastoma cells or rat cortical neurons are treated with insulin or IGF, GSK3 phosphorylates tau (24, 25). Similarly, GSK3 phosphorylates tau in neuronal progenitor cells in response to FGF exposure (26). As discussed above, AKT phosphorylates GSK3 on Ser⁹ when neurons are exposed to insulin, IGF, or FGF (14, 15, 21, 43). These observations suggest that in neurons, GSK3 phosphorylates tau upon its phosphorylation on Ser⁹.

Transgenic mice overexpressing GSK3 (WT) display up-regulated GSK3 activity and neurodegeneration (11). Surprisingly, transgenic mice overexpressing GSK3 (S9A) show high GSK3 activity in their brain extract but develop normally without any neurodegeneration (12) (for commentary, see Ref. 44). More importantly, tau is hyperphosphorylated in the brains of mice overexpressing GSK3 (WT) (11). In the brains of transgenic mice overexpressing GSK3 (S9A), tau is not phosphorylated significantly even until the age of 5–7 months (12). The reason why GSK3 (WT) and GSK3 (S9A) have such opposite effects in the brain is not known. However, the sole difference between GSK3 (WT) and GSK3 (S9A) is that only the former can be phosphorylated on Ser⁹. Thus, despite being constitutively active, GSK3 (S9A) is not capable of phosphorylating tau significantly in a manner similar to GSK3 (WT) in vivo. Consistent with the view that Ser⁹ phosphorylation is important for GSK3 to phosphorylate tau in the brain, starvation induces Ser⁹ phosphorylation of GSK3 with a concomitant increase in GSK3-catalyzed tau phosphorylation in the brains of adult mice (45).

Hyperphosphorylation of tau is associated with AD pathogenesis. Studies indicate that GSK3 phosphorylates tau in normal and AD brain (2, 8, 9, 11, 12). Therefore, the mechanism by which GSK3 phosphorylates tau in vivo is of considerable interest. Previous studies have shown that tau, GSK3, and 14-3-3 are components of a tau phosphorylation complex (8, 9). Within the complex, the 14-3-3 dimer simultaneously binds to and enhances the interaction of GSK3 with tau (9). In this study, we find that GSK3 within the tau phosphorylation complex is phosphorylated on Ser⁹. In HEK-293 cells transfected with tau, 14-3-3, GSK3 (WT), and GSK3 (S9A) in various combinations, tau interacts with both GSK3 (WT) and GSK3 (S9A) (Fig. 8E). However, 14-3-3 interacts with Ser⁹-phosphorylated GSK3 but not with GSK3 (S9A) (Fig. 2). Furthermore, in the absence of 14-3-3, Ser⁹-phosphorylated GSK3 has very low activity and does not phosphorylate tau significantly (Fig. 11G). In the presence of 14-3-3, Ser⁹-phosphorylated GSK3 displays high activity and phosphorylates tau (Fig. 11G). Taken together, our results indicate that within the tau phosphorylation complex, 14-3-3 connects Ser⁹-phosphorylated GSK3 to tau, and Ser⁹-phosphorylated GSK3 phosphorylates tau.

This study emphasizes the mechanism by which 14-3-3 promotes binding of Ser⁹-phosphorylated GSK3 with tau. In this mechanism, 14-3-3 simultaneously binds to tau and Ser⁹-phosphorylated GSK3. Thus, GSK3 and tau do not interact directly but thorough 14-3-3. Furthermore, GSK3 must be phosphorylated on Ser⁹ to be able to bind to 14-3-3. However, in a previous study, using recombinant proteins, we showed that in vitro, 14-3-3 promotes binding of tau with nonphosphorylated GSK3 as well (9). This observation indicates that 14-3-3 also promotes binding of tau with GSK3 by a mechanism that does not require phosphorylation of GSK3 on Ser⁹.

At this moment, we do not know the mechanism by which 14-3-3 enhances binding of GSK3 with tau in a Ser⁹ phosphorylation-independent manner. However, 14-3-3 binds to GSK3 only when GSK3 is phosphorylated on Ser⁹ (Fig. 6), and therefore 14-3-3 cannot bind and connect GSK3 to tau when the kinase is not phosphorylated on Ser⁹. Instead, previous studies have shown that GSK3, in addition to interacting with tau via 14-3-3, also binds to tau directly with low affinity (8, 9). More importantly, this binding does not require GSK3 to be phosphorylated on Ser⁹ (8). 14-3-3, on the other hand, binds to tau and changes tau conformation (7). Moreover, GSK3 and 14-3-3 binding sites within tau do not overlap (7, 8). It is therefore possible that 14-3-3 binds to tau and induces a conformational change thereby enhancing the affinity of tau for GSK3. As a result, more tau binds to GSK3 in the presence of 14-3-3 than in the absence (9). This means that 14-3-3 not only connects Ser⁹-phosphorylated GSK3 to tau but also enhances the affinity of tau for GSK3. Although this may explain why in vitro more tau binds to GSK3 in the presence of 14-3-3 than in the absence (9), we find that the binding of tau with GSK3 (S9A) is insensitive to 14-3-3 in intact cells (Fig. 9). These observations suggest that in vivo, 14-3-3 regulates the interaction of tau with GSK3 only when the kinase is phosphorylated on Ser⁹ and are consistent with the idea that 14-3-3 binds to Ser⁹-phosphorylated GSK3 first and then subsequently targets the kinase to tau.

Within the tau phosphorylation complex, 14-3-3 is the central molecule and holds GSK3 and tau together (9). As discussed above, 14-3-3 is a phosphoserine-binding protein (27–29), and GSK3 must be phosphorylated on Ser⁹ before 14-3-3 can bind to and target this kinase to the tau phosphorylation complex in the brain. GSK3 is known to be phosphorylated on Ser⁹ in response to various extracellular signals (15, 17, 18, 21, 43). It is possible that the tau phosphorylation complex is in dynamic equ