FRAT-2 preferentially increases glycogen synthase kinase 3 beta-mediated phosphorylation of primed sites, which results in enhanced tau phosphorylation.

Tau is a microtubule-associated protein found primarily in neurons, and its function is regulated by site-specific phosphorylation. Although it is well established that tau is phosphorylated at both primed and unprimed epitopes by glycogen synthase kinase 3 beta (GSK3 beta), how specific proteins that interact with GSK3 beta regulate tau phosphorylation has not been thoroughly examined. Members of the FRAT (frequently rearranged in advanced T-cell lymphoma) protein family have been shown to interact with GSK3 beta, and FRAT-1 has been shown to modulate the activity of GSK3 beta toward tau and other substrates. However, the effects of FRAT-2 on GSK3 beta activity and tau phosphorylation have not been examined. Therefore in this study the effects of FRAT-2 on GSK3 beta activity and tau phosphorylation were examined. In situ, FRAT-2 significantly increased GSK3 beta-mediated phosphorylation of tau at a primed epitope while not significantly affecting the phosphorylation of unprimed sites. Co-immunoprecipitation studies revealed that association of FRAT-2 with GSK3 beta resulted in a significant increase in phosphorylation of a primed substrate but did not alter phosphorylation of an unprimed substrate. Further, in vitro assays using recombinant proteins directly demonstrated that FRAT-2 enhances GSK3 beta-mediated phosphorylation of a primed substrate to a greater extent than an unprimed substrate. In addition, FRAT-2 is phosphorylated by GSK3 beta. This is the first demonstration of a protein differentially regulating the activity of GSK3 beta toward primed and unprimed epitopes.

Glycogen synthase kinase 3␤ (GSK3␤) 1 is a ubiquitous protein kinase that is expressed at high levels in the brain and specifically within neurons (for a review see Ref. 1). GSK3␤ is a unique Ser/Thr protein kinase that phosphorylates both primed (the target Ser/Thr is 4 amino acids N-terminal to a prephosphorylated Ser/Thr) and unprimed (the target Ser/Thr is flanked by a Pro) substrates (for a review see Ref. 2). GSK3␤ activity is regulated multiple ways, including phosphorylation and protein-protein interactions. Phosphorylation at Tyr 216 in-creases GSK3␤ activity perhaps by increasing its stability (GSK3␤ is a constitutively active kinase) (3,4), whereas phosphorylation at Ser 9 inhibits its activity (5,6). Recent work has suggested that phosphorylation of Tyr 216 is predominantly due to autophosphorylation (3), although other protein kinases may also be able to phosphorylate this site (7). Phosphorylation of Ser 9 on GSK3␤ is catalyzed by several kinases, with Akt being the most notable (reviewed in Ref. 2). Phosphorylation of Ser 9 causes the N terminus of GSK3␤ to act as a pseudosubstrate with the Ser(P) 9 occupying the primed site-binding pocket. This blocks substrate access to the kinase domain, and prevents GSK3␤ from efficiently phosphorylating its substrates (5,8,9). GSK3␤ is also regulated by protein-protein interactions. Perhaps the best characterized signaling pathway that regulates GSK3␤ via protein-protein interactions is the canonical wnt pathway. In the absence of a wnt signal, GSK3␤ binds to the scaffolding protein axin, which also binds ␤-catenin, along with other proteins, and facilitates the GSK3␤ phosphorylation of ␤-catenin (reviewed in Ref. 10). Phosphorylated ␤-catenin is then ubiquitinated and degraded by the proteasome (10). Upon wnt stimulation, the intracellular protein Dvl (dishevelled) is activated and works in concert with the GSK3␤-binding protein FRAT (frequently rearranged in advanced T-cell lymphoma) to displace GSK3␤ from the axin-␤-catenin complex. This displacement results in a decrease in ␤-catenin phosphorylation and dissociation from the complex. ␤-Catenin then translocates to the nucleus, where it initiates transcription of ␤-cateninresponsive genes, such as Myc and cyclin D1 (reviewed in Ref. 10).
A pivotal intracellular protein involved in wnt signaling is FRAT. Currently three isoforms of FRAT have been identified in mouse (FRAT-1, -2, and -3) (11,12), and two FRAT genes have been identified in humans (FRAT-1 and FRAT-2) (13-15), with each isoform containing a conserved GSK3␤-interacting domain (16). FRAT-1 competes with axin for binding of GSK3␤, because the binding sites on GSK3␤ for these proteins overlap (17,18). Most of the studies examining FRAT function to date have been carried out with FRAT-1 (16,19,20). Although both FRAT-1 and FRAT-2 are expressed ubiquitously, tissue expression analysis shows that FRAT-2 is expressed at higher levels in brain than FRAT-1 (11,15) and therefore may play a more prominent role in modulating neuronal GSK3␤ substrates. Recently, van Amerongen et al. (11) demonstrated that FRAT-2 is significantly less efficient than FRAT-1 at propagating the wnt signal. Given these and other findings, the authors suggest that FRAT-2 may play more of a role in regulating GSK3␤ function outside of the wnt signaling pathway (11).
The microtubule-associated protein tau is both an in vitro and in vivo substrate of GSK3␤ (21)(22)(23)(24). GSK3␤ phosphorylates numerous sites on tau, which are present in hyperphosphorylated tau found in Alzheimer's disease brain. Members of the wnt pathway have been shown to influence GSK3␤-mediated tau phosphorylation. Previously, our group demonstrated that axin potently inhibits tau phosphorylation by GSK3␤, likely by sequestering GSK3␤ away from tau (25). Dvl-1, -2, and -3, which are present in brain, have also been shown to decrease tau phosphorylation by GSK3␤ (26), despite the fact that Dvl and GSK3␤ do not interact (20). These data indicate that Dvl may be regulating other kinases that phosphorylate tau in addition to GSK3␤, because tau phosphorylation at the 12E8 (Ser(P) 262 ) site was decreased by expression of Dvl (26), despite the fact that Ser 262 is not phosphorylated by GSK3␤ (27). Interestingly, a short FRAT-1 peptide (FRATtide) was shown to inhibit GSK3␤-mediated phosphorylation of unprimed sites on tau while not significantly altering primed site phosphorylation (16). Similarly, Culbert et al. (19), using adenovirally expressed FRAT-1 in PC-12 cells, demonstrated that full-length FRAT-1 inhibits GSK3␤ phosphorylation of an unprimed site on tau while not affecting glycogen synthase (a primed substrate) activity. These results suggest that FRAT-1 selectively inhibits GSK3␤-mediated phosphorylation of unprimed substrates. Despite the homology between FRAT-1 and FRAT-2, there is little data on the function of FRAT-2, and its effects on tau phosphorylation by GSK3␤ have not been examined.
Although FRAT-1 and FRAT-2 are homologous, they are not identical and therefore may differentially regulate GSK3␤ activity. Given the fact that FRAT-1 regulation of GSK3␤-dependent tau phosphorylation has been examined previously, the focus of this study was on determining the effects of FRAT-2 on GSK3␤-mediated tau phosphorylation. In contrast to the results obtained with FRAT-1 (16,19), FRAT-2 facilitated GSK3␤-mediated tau phosphorylation at a primed site (Thr 231 ) while not significantly affecting phosphorylation at an unprimed site (Ser 396/404 ). Co-immunoprecipitation assays demonstrated that FRAT-2 significantly increased GSK3␤-mediated phosphorylation of a primed peptide but not of an unprimed substrate (recombinant tau). Recombinant FRAT-2 also increased activity of recombinant GSK3␤ toward a primed substrate in an in vitro assay. Furthermore, FRAT-2 was found to be a substrate of GSK3␤. These results suggest that, like p53 (28), FRAT-2 may enhance GSK3␤ activity and that this enhancement is selectively directed toward primed substrates. In addition, these results are consistent with an alternate role for FRAT-2 other than acting in the wnt signaling pathway (11).
Constructs and Recombinant Protein Preparation-The plasmids encoding FRAT-2-green fluorescent protein (GFP) (15), tau (encoding the longest human tau isoform (441 amino acids)) (29), and GSK3␤ (HAtagged, with the S9A mutation and thus constitutively active) (30) were described previously. FRAT-2 was subcloned into pGEX-6P glutathione S-transferase (GST) fusion vector (Amersham Biosciences). Using FRAT-2-GFP as the template, PCR was performed (forward primer, 5Ј-CGC GGA TCC ATG CCG TGC CGG AGG GAG-3Ј; reverse primer, 5Ј-CGT GAA TTC TCA GAG CAA GGA GCC TGA GGG CTG CAG GGC AAT-3Ј), and the PCR product was digested with BamHI and EcoRI and ligated into the corresponding cloning sites in pGEX-6P vector. Plasmid integrity was verified using DNA sequencing analysis. GST-FRAT-2 was purified according to the manufacturer's protocol, with the exception that cleared bacterial lysates were incubated with preblocked glutathione-Sepharose beads (Amersham Biosciences) for 1 h with rotation at 4°C. For controls, GST was prepared in the same manner. To determine the concentration of the purified GST-FRAT-2, the samples were diluted in 2ϫ SDS stop buffer and electrophoresed on a 10% SDS-polyacrylamide gel; in addition aliquots containing known amounts of bovine serum albumin (Fisher) were also run on the gel. The gel was Coomassie-stained and destained and dried. The gel was scanned and quantitated using UNSCANIT software (Silk Scientific, Inc.), and the amount of GST-FRAT-2 in the sample was determined as a function of the bovine serum albumin standards.
Transient Transfections-Tau, GSK3␤, and FRAT-2-GFP were transiently transfected into HEK cells using FuGENE 6 transfection reagent (Roche Applied Sciences), according to the manufacturer's protocol. Forty-eight hours after transfection, the cells were washed once with ice-cold phosphate-buffered saline and collected, as described below, for the different assays.
Immunoprecipitation and Immunoblotting-The cells were collected in lysis buffer, containing 0.5% Nonidet P-40, 150 mM NaCl, 10 mM Tris-Cl, pH 7.4, 1 mM EGTA, 1 mM EDTA, 0.1 mM phenylmethylsulfonyl fluoride, 1 M okadaic acid, 10 g/ml each of aprotinin, leupeptin, and pepstatin. The samples were sonicated on ice for 10 s and centrifuged at 16,000 ϫ g for 10 min. The protein concentrations of supernatants were then determined by the bicinchoninic acid assay (Pierce).
For immunoprecipitation, the samples were diluted to a final concentration of 0.5 mg/ml, in a final volume of 200 l using lysis buffer containing protease inhibitors. Monoclonal anti-GFP antibody (1.5 g) (Roche Applied Science) was added to each sample, and the samples were incubated at 4°C on a rotational shaker overnight. The samples were then incubated with precleared protein G-Sepharose beads (Amersham Biosciences) for 3 h at 4°C with rotation, and the beads were rinsed three times with lysis buffer and boiled in 2ϫ SDS buffer (0.25 M Tris-Cl, pH 7.5, 2% SDS, 25 mM dithiothreitol (DTT), 5 mM EGTA, 5 mM EDTA, 10% glycerol, and 0.01% bromphenol blue). The supernatants were collected and electrophoresed on a 10% SDS-polyacrylamide gel and transferred to nitrocellulose membranes. The nitrocellulose membranes were then blocked for 1 h with TBST (20 mM Tris-Cl, pH 7.6, 137 mM NaCl, 0.05% Tween 20) and 5% milk and incubated with the indicated antibodies overnight. The blots were then rinsed with TBST and incubated with appropriate horseradish peroxidase-conjugated secondary antibody for 1 h at room temperature and rinsed with TBST and developed with enhanced chemiluminescence (Amersham Biosciences). For immunoblotting the samples were collected directly into 2ϫ SDS buffer without dye or DTT, and the samples were processed as above, except they were boiled immediately following sonication. The samples were then diluted in 2ϫ SDS buffer with dye and DTT and electrophoresed on 10% SDS-polyacrylamide gels and transferred to nitrocellulose membranes and blotted as above. To evaluate the phosphorylation state of tau the following antibodies were used: AT180 (Pierce/Endogen), which recognizes tau phosphorylated on Thr 231 (31, 32) (numbering based on longest human tau isoform (33)) and PHF-1 (a gift from Dr. P. Davies), which recognizes tau phosphorylated on Ser 396/404 (34,35). Total tau levels were determined using a combination of the phosphoindependent monoclonal tau antibodies Tau 5 (a gift from Dr. L. Binder) (36) and 5A6 (37). Levels of transfected GSK3␤ and FRAT-2-GFP were detected using monoclonal anti-GSK3␤ (Transduction Laboratories) and monoclonal anti-GFP (Roche Applied Science), respectively. Quantitation of tau phosphorylation at AT180 and PHF-1 epitopes was carried out by scanning immunoblots and determining the relative optical density with UNSCANIT (Silk Scientific Inc) software, normalizing the values to total tau levels, and expressing the data as percentages of control (Tau only).
GSK3␤ Activity Assays-HEK cells were transfected as above, and the lysates were immunoprecipitated using 4.9 g of the monoclonal anti-HA antibody (Sigma) (to immunoprecipitate GSK3␤) or 2.5 g of the monoclonal anti-GFP antibody (to immunoprecipitate FRAT-2 and the associated GSK3␤) precoupled to protein G-Sepharose beads (Amersham Biosciences). The cell lysates were incubated for 3 h at 4°C with the antibody-protein G complexes. After washing the protein G-Sepharose beads containing the immuno-mobilized GSK3␤ twice with lysis buffer, the beads were washed two additional times with kinase buffer (20 mM Tris-Cl, pH 7.5, 5 mM MgCl 2 , and 1 mM DTT) and incubated in 30 l of kinase buffer with 0.1 g/l recombinant tau (Panvera), 0.182 Ci/l [␥-32 P]ATP (Amersham Biosciences), and 125 M ATP (Sigma) for 30 min at 30°C. The samples were centrifuged at 16,000 ϫ g for 1 min, the supernatant was collected, and 25 l of 2ϫ SDS buffer containing dye and DTT was added prior to incubation in a boiling water bath and electrophoresed on an 8% SDS-polyacrylamide gels. The gels were Coomassie-stained and destained overnight, dried, and exposed to phosphorus imaging screens prior to quantitation of 32 P incorporation using phosphorus imaging (Molecular Devices Storm Scanner model 840 and Image Quant 5.0 software). To determine the amount of GSK3␤ in the immunoprecipitates, 2ϫ SDS buffer was added to the protein G beads containing the immunoprecipitated GSK3␤, and the samples were boiled then centrifuged at 16,000 g for 1 min. The supernatants were electrophoresed on 8% SDS-polyacrylamide gels and transferred to nitrocellulose membranes. The blots were then immunoblotted for GSK3␤ as described above. GSK3␤ activity was then normalized to GSK3␤ levels in the sample, and the data were expressed as percentages of control (GSK3␤ alone). Assays using the phospho-glycogen synthase peptide were carried out as above, with the exception that 100 M phospho-glycogen synthase peptide 2 (Tocris), 250 M ATP, and 0.091 Ci/l [␥-32 P]ATP were used. After incubation, the samples were centrifuged at 16,000 ϫ g for 1 min, the supernatant was collected, and 9 l of the sample was pipetted in triplicate onto separate P81 filter papers (1 cm in diameter) (Whatman). Filter papers were allowed to dry for ϳ1 min and washed in 0.5% o-phosphoric acid four times for 15 min, each wash, on a benchtop rotational shaker. Filter papers were then washed in 95% ethanol for 15 min and allowed to dry. Dried filter papers were then subjected to scintillation counting. The counts/min for each sample were normalized to GSK3␤ levels as above, and each value was expressed as a percentage of control (GSK3␤).
In Vitro GSK3␤ Activity Assays-In vitro activity assays were carried out using 60 units of recombinant GSK3␤ (New England Biolabs)/ reaction for all conditions. Prior to the initiation of the reaction, recombinant GSK3␤ without GST-FRAT-2 or with the indicated amount of GST-FRAT-2 was preincubated for 15 min at 30°C in kinase buffer (8 l of total volume) without substrate or ATP. Following preincubation, 22 l of kinase buffer with other reagents was added to yield final concentrations of 100 M phospho-glycogen synthase peptide 2 (Tocris), 250 M ATP, and 0.091 Ci/l [␥-32 P]ATP in a final volume of 30 l. The samples were then incubated at 30°C for 30 min, and 9 l of sample was spotted onto P81 filter paper (1-cm diameter) in triplicate and allowed to dry for ϳ1 min. Filter papers were then processed as described above and counted. The assays were also carried out using recombinant tau as the substrate (0.1 g/l), and the reactions were stopped by adding 20 l of 2ϫ SDS stop buffer and boiled for 5 min. The samples were then processed and quantitated as described above. For all in vitro experiments with GST-FRAT-2, separate control reactions were carried out with equimolar amounts of GST and quantitated in an identical manner. The activity of GSK3␤ obtained with GST-FRAT-2 was then normalized to the values obtained with an equimolar concentration of GST.
Data Analysis-The data were analyzed using an unpaired, twotailed t test between individual groups. The conditions were considered significantly different when p Ͻ 0.05.

FRAT-2 Increases GSK3␤-mediated Phosphorylation of Tau at a Primed Epitope While Not Affecting Phosphorylation of an
Unprimed Site-Previous studies have provided evidence that FRAT-1 inhibits GS3␤ phosphorylation of tau at unprimed sites (16,19) while not affecting primed site phosphorylation; however, the effects of FRAT-2 on GSK3␤-mediated tau phosphorylation have not been examined. Given that FRAT-2 likely functions differently than FRAT-1 (11), we examined the effects of FRAT-2 on GSK3␤-mediated tau phosphorylation. HEK 293 cells were transiently transfected with tau in the absence or presence of GSK3␤ and in the presence or absence of FRAT-2-GFP, and tau phosphorylation was examined by immunoblotting with phospho-specific antibodies (Fig. 1A). The effects of FRAT-2 on GSK3␤ phosphorylation of tau at the primed site AT180 (Thr(P) 231 ) and the unprimed site of PHF-1 (Ser(P) 396/404 ) were quantitated, normalized to the levels of total tau (Tau 5/5A6), and expressed as percentages of the tau only condition (Fig. 1, C and D). Expression of FRAT-2 consistently resulted in a decrease in the levels of tau expression (even when the amount of tau transfected in was increased). When FRAT-2 was expressed alone with tau, there was no significant difference in the phosphorylation state of tau at either the primed AT180 (Thr(P) 231 ) epitope or the unprimed PHF-1 (Ser(P) 396/404 ) epitopes (Fig. 1, C and D). As expected, when GSK3␤ was expressed with tau, phosphorylation at both epitopes increased significantly (Fig. 1, C and D). Surprisingly, when FRAT-2 was expressed with GSK3␤, phosphorylation of tau at the unprimed PHF-1 epitope did not change significantly compared with what was observed with GSK3␤ alone (Fig. 1C).
In contrast, FRAT-2 significantly increased tau phosphorylation by GSK3␤ at the primed AT180 epitope (Fig. 1D), indicating a selective facilitation of GSK3␤-mediated tau phosphorylation at the primed AT180 epitope by FRAT-2. Fig. 1B shows immunoblots of both FRAT-2 and GSK3␤ levels, indicating similar expression levels of expression between conditions. Immunoblot analysis of cell lysates showed that GSK3␤ phosphorylation at either Ser 9 or Tyr 216 was not altered in the presence of FRAT-2 (data not shown). Together, these data clearly demonstrate that FRAT-2 facilitates GSK3␤ phosphorylation of a primed epitope in tau while not affecting the phosphorylation of an unprimed epitope.
FRAT-2 Increases GSK3␤ Phosphorylation of a Primed Peptide in Vitro-Because FRAT-2 selectively facilitated GSK3␤ phosphorylation of a primed site in tau in transfected cells, further studies were carried out to determine whether this was due to a direct effect of FRAT-2 on GSK3␤. GSK3␤ was immunoprecipitated from transfected cell lysates with a monoclonal HA antibody and used to 32 P phosphorylate a primed glycogen synthase peptide in vitro. When GSK3␤ was immunoprecipitated from HEK cell lysates and used in the assay, the presence of FRAT-2 did not significantly affect GSK3␤ activity toward the substrate ( Fig. 2A). However, GSK3␤ is robustly expressed, and in the immunoprecipitates it is likely that only a small percentage of the GSK3␤ was bound by FRAT-2. Indeed, when the GSK3␤ immunoprecipitates were probed for FRAT-2 the levels appeared to be quite low (data not shown). Therefore, to more accurately determine the effects of FRAT-2 on GSK3␤ activity, FRAT-2 was immunoprecipitated with a monoclonal GFP antibody and used to measure GSK3␤ activity using the primed peptide. When the values were normalized to the levels of GSK3␤ in the precipitates, FRAT-2 significantly increased GSK3␤ activity (Fig. 2A). The total levels of transfected proteins are shown in Fig. 2B. These data clearly demonstrate that FRAT-2 facilitates GSK3␤ phosphorylation of a primed substrate.
FRAT-2 Does Not Significantly Affect GSK3␤ Phosphorylation of Recombinant Tau in Vitro-To further characterize the effects of FRAT-2 on GSK3␤-mediated tau phosphorylation, assays were carried out using recombinant tau as a substrate. Recombinant tau is a good substrate for GSK3␤ in vitro (21,38,39) and is unprimed. When GSK3␤ was immunoprecipitated from cell lysates that were transfected with GSK3␤ and FRAT-2 with a monoclonal HA antibody, it phosphorylated the recombinant tau to the same extent as the GSK3␤ immunoprecipitated from cells expressing just GSK3␤ (Fig. 3), similar to what was observed with the primed peptide. However, in contrast to the findings with the primed peptide, immunoprecipitation of FRAT-2 with the anti-GFP antibody followed by measurement of GSK3␤-mediated phosphorylation of recombinant tau did not result in a significant increase in GSK3␤ activity, although there was a trend toward increased phosphorylation of recombinant tau (Fig. 3). These data suggest that FRAT-2 significantly increases GSK3␤ activity toward primed substrates but is not as effective at increasing GSK3␤ activity toward unprimed substrates.
FRAT-2 and GSK3␤ Co-immunoprecipitate-Lysates from cells transfected with GSK3␤, FRAT-2, and tau were immunoprecipitated with the anti-GFP antibody and probed for GFP (FRAT-2), GSK3␤, or Tau5/5A6 (Fig. 4). Both endogenous (lower band) and HA-tagged GSK3␤ (upper band) co-immunoprecipitated with FRAT-2, indicating that FRAT-2 interacts with GSK3␤. Reprobing of the same blots for tau showed that tau did not interact with FRAT-2 at a detectable level (Fig. 4). These data clearly demonstrate that FRAT-2 and GSK3␤ interact in cells and that FRAT-2 does not facilitate tau phosphorylation by acting as a scaffold for tau and GSK3␤.

FRAT-2 Increases GSK3␤ Phosphorylation of Primed Sites Recombinant FRAT-2 Directly Activates GSK3␤ in a Primed
Substrate Assay-To further confirm that FRAT-2 increases GSK3␤-mediated phosphorylation of primed substrates, phosphorylation assays were carried out using purified GST-FRAT-2 and recombinant GSK3␤. To account for any effects of GST on the assay, GST was used at equimolar concentrations as GST-FRAT-2, and the data obtained with the GST-FRAT-2 were normalized relative to the GSK3␤ activity in the presence of GST only and expressed as percentages of GSK3␤ alone (control). The addition of GST-FRAT-2 resulted in a significant dose-dependent increase in GSK3␤ activity with the maximum activation being ϳ8-fold (Fig. 5). These data demonstrate that FRAT-2 alone is sufficient to robustly increase GSK3␤ activity toward primed substrates.
Recombinant FRAT-2 Activates GSK3␤ toward Recombinant Tau Substrate but Competes with Tau for GSK3␤ Activity at Higher Concentrations-The effect of GST-FRAT-2 on the GSK3␤ phosphorylation of recombinant tau, which is an unprimed substrate, was investigated next. At low concentrations, GST-FRAT-2 increased GSK3␤ phosphorylation of re- FIG. 1. FRAT-2 differentially affects phosphorylation of primed and unprimed sites on tau by GSK3␤. A, representative immunoblots showing total tau levels (Tau 5/5A6), and phospho-tau levels at the primed AT180 site and unprimed PHF-1 site. HEK cells were transfected with tau and the other constructs as indicated under the panels. Tau protein levels were normalized for the immunoblots. Expression of exogenous GSK3␤ resulted in a decrease in the electrophoretic mobility of tau, indicative of increased phosphorylation, and phosphorylation at both the PHF-1 and AT180 epitopes increased as expected. B, representative immunoblots showing levels of transfected FRAT-2-GFP and GSK3␤. C and D, quantitation of immunoblot data for PHF-1 (C) and AT180 (D) epitopes. Phospho-tau immunoreactivity was normalized to total tau levels in each sample and expressed as percentages of control (Tau alone). Expression of GSK3␤ or FRAT-2 and GSK3␤ significantly increased phosphorylation at the PHF-1 epitope to the same extent (C). The presence of FRAT-2 significantly increased the GSK3␤-mediated phosphorylation of tau at the AT180 epitope (D). The data are presented as the means Ϯ S.E. from three to five experiments. *, p Ͻ 0.05 compared with GSK3␤ only .   FIG. 2. FRAT-2 potentiates GSK3␤ phosphorylation of a primed peptide. A, cells were transfected with the constructs indicated and immunoprecipitated (IP) with either the HA antibody to directly precipitate exogenous GSK3␤ or with the GFP antibody to immunoprecipitate FRAT-2. The immunoprecipitates were used to phosphorylate a primed glycogen synthase peptide. When GFP-FRAT2 was immunoprecipitated, and therefore all the GSK3␤ in the assay was bound to FRAT-2, GSK3␤ phosphorylated the primed peptide to a significantly greater extent compared with when only GSK3␤ was present. The data were normalized to immunoprecipitated GSK3␤ levels and expressed as percentages of the values obtained with GSK3␤ alone (immunoprecipitated with HA antibody). The data are presented as the means Ϯ S.E. of three to five experiments. *, p Ͻ 0.05 compared with GSK3␤ alone. B, representative immunoblots showing levels of transfected proteins in cell lysates used for the GSK3␤ assays. combinant tau ϳ2.5-fold (Fig. 6A). However, this increase was attenuated when 1 g of GST-FRAT-2 was added to the reaction, and the phosphorylation of recombinant tau by GSK3␤ was significantly decreased by the addition of 5 g of GST-FRAT-2 (Fig. 6A). When developing the autoradiographs, it became apparent that GST-FRAT-2 was also being phosphorylated by GSK3␤ (data not shown), indicating that the reason that GST-FRAT-2 decreased tau phosphorylation by GSK3␤ when present at higher concentrations was because it was competing with tau as a substrate. To unequivocally demonstrate that GST-FRAT-2 was being phosphorylated by GSK3␤, assays were carried out using GST, GST-FRAT-2, or tau as substrates. These data demonstrate GST-FRAT-2 is phosphorylated by GSK3␤, whereas GST is not phosphorylated (Fig.  6B). This is the first direct demonstration that FRAT-2 is a substrate of GSK3␤. These results clearly indicate that in the assays where recombinant tau is used as a substrate, FRAT-2 at the higher concentrations competes with tau as a substrate of GSK3␤, which leads to the decrease in tau phosphorylation. In the assays where the primed peptide was used as the substrate for GSK3␤, no inhibition by FRAT-2 was observed. This is likely due to the fact that primed substrates are phosphorylated more efficiently by GSK3␤ than unprimed substrates (reviewed in Ref. 2). Another contributing factor is the amount of substrate used in each assay. In the primed peptide assay, the substrate is present at 100 M, whereas the recombinant tau substrate is present at 2.1 M, and thus the concentration is ϳ50-fold less. Therefore, the primed substrate would be in excess of FRAT-2, even at 5 g (3.3 M), whereas the recombinant tau would be present at 2.1 M, and GST-FRAT-2 would be present in excess of tau at 3.3 M. Nonetheless, at the lower concentrations of FRAT-2, the magnitude of activation of GSK3␤ by FRAT-2 toward the unprimed substrate (recombinant tau) was significantly less than that of the primed substrate.

DISCUSSION
The purpose of this study was to examine the effects of FRAT-2 on GSK3␤-mediated tau phosphorylation and to determine whether FRAT-2 had a differential effect on primed site versus unprimed site phosphorylation by GSK3␤. In this study we demonstrate for the first time that FRAT-2 preferentially increases GSK3␤-mediated phosphorylation of primed sites. Further we show that the FRAT-2-induced increase in the phosphorylation of primed substrates is due to a direct interaction with GSK3␤ and that FRAT-2 is a substrate of GSK3␤. These are novel and important findings because they provide insight into the mechanisms that may regulate the phosphorylation of specific GSK3␤ substrates.
Although two FRAT genes have been identified in humans (FRAT-1 and FRAT-2) (13-15), the majority of previous studies have focused on FRAT-1 (16,19,20) or the Xenopus homolog, GSK3␤-binding protein (17,40), which negatively regulate GSK3␤ activity. In contrast, little is known about the functions of FRAT-2 and its role in regulating GSK3␤ activity. This is likely due to the fact that full-length human FRAT-2 was FIG. 3. FRAT-2 has no effect on GSK3␤-mediated phosphorylation of an unprimed substrate. The cells were transfected with the constructs indicated, and the cell lysates were immunoprecipitated (IP) with antibodies to either HA to precipitate exogenous GSK3␤, or GFP, to precipitate FRAT-2 and hence only GSK3␤ bound to FRAT-2. Precipitates were used to phosphorylate recombinant tau, which is an unprimed GSK3␤ substrate. The interaction of FRAT-2 with GSK3␤ did not significantly increase GSK3␤ phosphorylation of recombinant tau, although there was a trend toward increased activity. The data were normalized to immunoprecipitated GSK3␤ levels and expressed as percentages of values for GSK3␤ alone (immunoprecipitated with HA antibody). The data are presented as the means Ϯ S.E. of three to five experiments.

FRAT-2 Increases GSK3␤ Phosphorylation of Primed Sites
cloned in 2002 (15) and murine FRAT-2 in 2004 (11), and therefore only a few studies have been carried out. FRAT-2 mRNA is present in brain at higher levels than FRAT-1 mRNA (15), but cellular localization studies have not yet been carried out. Recently FRAT-2 was shown to be significantly less efficient in transducing a wnt signal than either FRAT-1 or FRAT-3 (11). Therefore, it is possible that FRAT-2 does not play a significant role in the wnt signaling pathway but rather regulates GSK3␤ activity outside of this signaling pathway (11). Results from our studies clearly showed that FRAT-2 selectively potentiates GSK3␤ phosphorylation of primed substrates. Although in vitro assays using recombinant proteins showed significantly increased GSK3␤ activity toward an unprimed substrate in the presence of low concentrations of FRAT-2, the magnitude of activation was far less than what was observed for a primed substrate. These findings clearly demonstrate that although FRAT-1 inhibits GSK3␤-mediated phosphorylation of substrates at unprimed sites, FRAT-2 preferentially increases GSK3␤ activity toward primed substrates.
Even though FRAT-1 and FRAT-2 are members of the same family of proteins and share ϳ70% identity, there are key differences that could account for the opposing effects of the two proteins on GSK3␤ activity. The most striking differences are the regions present in FRAT-1, which are absent in FRAT-2. FRAT-2 (15) lacks the last 10 amino acids in the extreme C terminus, as well as two large regions just N-terminal of the GSK3␤ interacting domain of FRAT-2 compared with FRAT-1 (amino acids 141-152 and 155-166 in FRAT-1) (15). Although the function of these regions are not known, it is possible that deletion of these regions could allow FRAT-2 to activate GSK3␤ when it binds, instead of inhibit as is the case with FRAT-1. There are also key amino acids that are different within FRAT-2. For example, a Ser present at position 88 in FRAT-1 is a Pro at the corresponding amino acid in FRAT-2. This would result in a large structural shift within FRAT-2. Also present in FRAT-1 are two Pro (positions 108 and 190), which in FRAT-2 are either Ser or Ala. These amino acid differences could clearly result in structural differences between FRAT-1 and FRAT-2, because Pro residues add bends into the structure of a protein, so the presence or absence of Pro can result in large differences in protein structure between the two proteins. Future studies are clearly required to elucidate the structural differences between FRAT-1 and FRAT-2 that are responsible for their opposing effects on GSK3␤ activity.
Activation of GSK3␤ through binding of another protein is not unprecedented because the transcription factor p53 binds and increases GSK3␤ activity (28). In this study the in vitro GSK3␤-mediated phosphorylation of both recombinant tau and a primed substrate were increased in the presence of recombinant p53. This activation was shown to be independent of the phosphorylation state of GSK3␤ and therefore results from a direct activation of GSK3␤ by its interaction with p53 (28). The effect of p53 on GSK3␤ provides evidence that GSK3␤ activity can be up-regulated by its interaction with another protein. Nonetheless, the present finding that FRAT-2 preferentially increases GSK3␤-mediated phosphorylation of a primed substrate is the first demonstration that a protein binding to GSK3␤ can selectively facilitate the phosphorylation of primed sites by GSK3␤. Tau is a well characterized substrate of GSK3␤ both in situ and in vitro. The phosphorylation state of tau is directly correlated to its ability to bind and stabilize microtubules. Recently, it has been demonstrated that phosphorylation at specific epitopes is sufficient to regulate tau-microtubule interactions. One such site is the AT180 site (Thr 231 ). The AT180 site is a primed site, and when it is phosphorylated, tau-microtubule interactions are greatly decreased (29). In this study, we clearly demonstrate that GSK3␤ phosphorylation of primed tau epitopes is facilitated by FRAT-2. Given the finding that GSK3␤-mediated phosphorylation of unprimed Ser/Thr-Pro sites on tau does not significantly impact tau-microtubule interactions (30), it can be speculated that FRAT-2 plays a role in regulating the physiological functioning of tau. Indeed, there is evidence that increased phosphorylation of unprimed GSK3␤ sites in tau may be pathological. For example, pseudo-phosphorylation (changing the Ser to Glu or Asp) of the unprimed GSK3␤ sites Ser 396/404 makes tau more fibrillogenic (41) and thus may play a role in the development of tau pathology in Alzheimer's disease and other neurodegenerative conditions. Therefore, FRAT-2 may play a role in directing GSK3␤ away from potentially pathological phosphorylation sites on tau.
In summary, we demonstrated for the first time that FRAT-2 directly facilitates GSK3␤ activity, particularly at primed sites. Further we show that FRAT-2 itself is a substrate of GSK3␤. There is increasing evidence that the majority of physiological substrates of GSK3␤ are primed (42,43), and therefore it perhaps is not surprising that FRAT-2 preferentially increases GSK3␤-mediated phosphorylation of primed epitopes. The fact that FRAT-2 is less efficient than FRAT-1 at transducing the wnt signal is interesting because it points to a possible role for FRAT-2 outside of the wnt pathway as its primary function (11). Results from our study would support this idea, because FRAT-2 had of the opposite effect of FRAT-1 (19) on GSK3␤ activity. Therefore it can be proposed that outside of the wnt signaling pathway FRAT-2 binds GSK3␤ and directs it toward primed substrates rather than just preventing ␤-catenin phosphorylation by displacing GSK3␤ from axin (16,17,19,20). In this situation the interaction of FRAT-2 with GSK3␤ may regulate the phosphorylation of substrates at the appropriate epitopes. In the case of tau, the interaction of FRAT-2 with GSK3␤ may facilitate physiological phosphorylation events that regulate tau-microtubule interactions and attenuate phosphorylation at unprimed sites that have the potential to be more pathological. Overall, it is likely that FRAT-2 plays a unique and important role in the regulation of substrate phosphorylation by GSK3␤.