Casein kinase 1 delta phosphorylates tau and disrupts its binding to microtubules.

Tau hyperphosphorylation precedes neuritic lesion formation in Alzheimer's disease, suggesting it participates in the tau fibrillization reaction pathway. Candidate tau protein kinases include members of the casein kinase 1 (CK1) family of phosphotransferases, which are highly overexpressed in Alzheimer's disease brain and colocalize with neuritic and granulovacuolar lesions. Here we characterized the contribution of one CK1 isoform, Ckidelta, to the phosphorylation of tau at residues Ser202/Thr205 and Ser396/Ser404 in human embryonic kidney 293 cells using immunodetection and fluorescence microscopy. Treatment of cells with membrane permeable CK1 inhibitor 3-[(2,3,6-trimethoxyphenyl)methylidenyl]-indolin-2-one (IC261) lowered occupancy of Ser396/Ser404 phosphorylation sites by >70% at saturation, suggesting that endogenous CK1 was the major source of basal phosphorylation activity at these sites. Overexpression of Ckidelta increased CK1 enzyme activity and further raised tau phosphorylation at residues Ser202/Thr205 and Ser396/Ser404 in situ. Inhibitor IC261 reversed tau hyperphosphorylation induced by Ckidelta overexpression. Co-immunoprecipitation assays showed direct association of tau and Ckidelta in situ, consistent with tau being a Ckidelta substrate. Ckidelta overexpression also produced a decrease in the fraction of bulk tau bound to detergent-insoluble microtubules. These results suggest that Ckidelta phosphorylates tau at sites that modulate tau/microtubule binding, and that the expression pattern of Ckidelta in Alzheimer's disease is consistent with it playing an important role in tau aggregation.

AD 1 is a progressive neurodegenerative disease characterized in part by the appearance of neurofibrillary tangles, neuritic plaques, and neurophil threads (1). Each manifestation of neuritic pathology is comprised of tau protein aggregated into filaments. Because the spatial distribution of tau fibrillization appears stereotypically, correlates with neuronal cell loss, and parallels cognitive decline, it is a useful marker of and may contribute to degeneration in AD and other dementias.
CK1 levels appear to be regulated within the cell. For example, Cki␦ levels are elevated in both AD brain (22,23) and the spinal cords of a murine model of amyotropic lateral sclerosis (24). Moreover, isoforms Cki␣ and Cki␦ associate with the neurofibrillary lesions of AD (25,26), the pathological hallmarks in AD and other tauopathic neurofibrillary degenerations diseases. These observations led to the hypothesis that CK1 enzymes regulate tau phosphorylation in vivo, and that their up-regulation in level or activity contributes to tau hyperphosphorylation in disease (22).
Here we test this hypothesis by examining the ability of one CK1 isoform, Cki␦, to phosphorylate tau under basal and overexpression conditions in situ. The data suggests that CK1 activity makes a major contribution to basal levels of tau phosphorylation, and that Cki␦ can modulate microtubule stability by direct interaction with tau protein.

EXPERIMENTAL PROCEDURES
Materials-Polyclonal antibody against the C terminus of p35 (C- 19) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA), whereas monoclonal antibodies AT8, Tau5, PHF1, and 128A were ob- Recombinant Proteins-Recombinant htau40 containing a His 6 tag was expressed and purified as described (28). Poly-His free recombinant htau40 was prepared the same way, except that Escherichia coli expression was from vector pET-14b (EMD Biosciences, San Diego, CA), and protein isolated from immobilized metal affinity chromatography * This work was supported by National Institutes of Health Grant AG14452 (to J. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Human brain Cki␦ cDNA served as template for PCR amplification of C-terminal-truncated Cki␦ (Cki␦-⌬317). The sense primer (5Ј-ACG-GATCCCATATGGAGCTGAGAGTC) incorporated an NdeI site at the start codon, whereas the downstream primer (5Ј-CTCGAGTTAGT-GTCTCCAGCCGCTC) contained a TTA stop codon downstream of the codon for His 317 followed by an XhoI site. The amplified PCR product was subcloned directly into TOPO2.1 (Invitrogen, Carlsbad, CA) and sequenced in both directions for errors. The Cki␦-⌬317-TOPO2.1 cDNA was digested with NdeI and XhoI, the approximate 0.9-kb Cki␦-⌬317 fragment was gel purified and subcloned into E. coli expression vector PT7C (28). Recombinant Cki␦-⌬317 was purified by immobilized metal affinity chromatography and gel filtration chromatography as described previously for tau protein (28).
To generate a cell line stably expressing human htau40 (stable tau cells), 5 g of tau-pcDNA3.1 plasmid was linearized with BglII and transfected into HEK-293 cells. Forty h after transfection, cells were cultured in selective medium containing 500 g/ml G418.
To assay CK1 inhibitor IC261, tau stable cells treated with different IC261 concentrations were harvested after 30, 60, 120, 240, and 480 min of incubation. IC261 was stored as a 50 mM stock solution in Me 2 SO and diluted to a working stock of 1 mM in H 2 O before treatment.
To assess toxicity of various treatments, cells were stained with the DNA-specific fluorochrome Hoechst 33258 (Molecular Probes, Eugene, OR) and examined under fluorescence for chromatin condensation (29). Cells with uniformly distributed chromatin were scored as "viable." Immunoprecipitation-Cells were harvested and disrupted in Lysis buffer (20 mM Tris-HCl, pH 7.4, 140 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 1 mM Na 3 NO 4 , 10 mM NaF, 0.5% (v/v) Nonidet-40, 1 mM EDTA, and 1ϫ complete protease inhibitors), and cell debris was removed by centrifugation (10 min at 10,000 ϫ g) at 4°C. Protein concentrations were determined using the Coomassie Blue binding method with bovine serum albumin as standard (30). For immunoprecipitation experiments, extracts (500 g) were incubated with 2 g of each monoclonal antibody examined (anti-Cki␦, 128A; anti-tau, Tau5; anti-p25, C-19) for 2 h at 4°C. Protein G-agarose beads (60 l of 25% (w/v) slurry) were then added to the samples and incubation was continued for 1 h at 4°C. The resultant immunocomplexes were collected by centrifugation, washed three times with Lysis buffer, and subjected to immunoblot or in vitro protein kinase assays as described below.
In Vitro Protein Kinase Assay-Purified protein kinase (Cki␦-⌬317) or immunocomplexes derived from immunoprecipitation experiments were incubated with 2 g of htau40 in 20 l of Kinase buffer (50 mM Tris, 200 M ATP, and 1 mM dithiothreitol) at 37°C for up to 4 h. When necessary, reactions were supplemented with 1.25 Ci of [␥-32 P]ATP. All reactions were stopped by addition of 10 l of 4ϫ SDS-PAGE Loading buffer. Following SDS-PAGE, gels were subjected to immunoblot analysis as described above or stained and dried for autoradiography using a Molecular Imager FX Pro Plus multi-imager system (Bio-Rad).
Microtubule Binding Assay-To assay the ability of tau to bind microtubules in situ, HEK-293 cells transfected with pcDNA3.1/tau or Cki␦/tau for 24 h were harvested in 500 l of Microtubule-stabilizing buffer (80 mM PIPES, pH 6.8, 1 mM GTP, 1 mM MgCl 2 , 1 mM EGTA, 0.5% Triton X-100, 30% glycerol, and 1ϫ complete protease inhibitors) in the presence of 10 M Taxol at 37°C. After cell debris was removed by centrifugation (10,000 ϫ g for 10 min) at 37°C, supernatant fractions were collected and centrifuged again at 100,000 ϫ g for 1 h at 35°C (5). The resultant pellets (P1) were washed twice with and resuspended (500 l) in Microtubule-stabilizing buffer, sonicated for 10 s, and then subjected along with 100,000 ϫ g supernatants (S1) to SDS-PAGE and immunoblot analysis as described above.
To assay the ability of tau to bind microtubules in vitro, tubulin (1 mg/ml) was first assembled in Microtubule assembly buffer (80 mM PIPES, pH 6.8, 1 mM MgCl 2 , 1 mM EGTA, 1 mM GTP) containing 10 M Taxol for 30 min at 30°C. Resultant microtubules were then incubated (37°C for 30 min) with supernatants (S1; 500 g) from the in situ microtubule binding assays described above, and then harvested by centrifugation (100,000 ϫ g for 1 h at 25°C). Pellet fractions (P2) were resuspended in Microtubule-assembly buffer so that their final volumes equaled those of supernatant fractions (S2). Equal amounts of supernatant (S2) and pellet (P2) fractions were then analyzed by SDS-PAGE and the distribution of tau protein was examined by immunoblot analysis using monoclonal antibody Tau5.
Cell Extraction and Immunofluorescence Staining-Wild-type HEK-293 cells were cultured on coverslips for 24 h and transfected with Cki␦/tau or PcDNA3.1/tau for 24 h as described above. To extract soluble proteins, cells were washed with phosphate-buffered saline and incubated with Microtubule-stabilizing buffer containing 10 M Taxol for 2 min at room temperature. Cells were washed once with Microtubule-stabilizing buffer without detergent. Both the extracted cells and non-extracted cells were fixed with methanol. Following the phosphatebuffered saline washes, cells were sequentially incubated with phosphate-buffered saline containing 3% bovine serum albumin for 1 h, Tau5 antibody (1:1000) overnight at 4°C, and fluorescein-labeled goat anti-mouse IgG secondary antibody (1:1000) for 1 h. Immunostaining was visualized by fluorescence microscopy (Eclipse 800, Nikon, Japan), captured with a SPOT II digital camera (Diagnostic Instruments, Inc., Sterling Heights, MI), and processed with Metamorph imaging software (Universal Imaging Corp., Downingtown, PA).
Analytical Methods-Hyperbolic inhibition curves were fit to the rectangular hyperbola: y ϭ ax/(b ϩ x), where y is immunoreactivity determined at inhibitor concentration x, and constant b corresponds to x at 50% y max (i.e. IC 50 ).
Probability values were determined by Student's t test for single comparison and one-way analysis of variance with Tukey's post-hoc test for multiple comparisons. All analyses were performed using the InStat Software program.

Cki␦ Phosphorylates Tau in Vitro-To determine whether
Cki␦ could directly phosphorylate tau, a truncated form of the human enzyme (Cki␦-⌬317) was expressed and purified as a His 6 fusion protein and used to phosphorylate tau protein in vitro. Truncated enzyme was used because it was easily prepared after heterologous expression in E. coli while retaining the protein kinase activity of full-length protein (33). Incubation of htau40 with purified Cki␦-⌬317 in the presence of radioactive nucleotide substrate led to detectable phosphate incorporation within 30 min (Fig. 1A). Detection of total tau protein on Western blots using monoclonal antibody Tau5, the reactivity of which is not phosphorylation dependent (28), revealed a substantial electrophoretic mobility shift by 4 h incubation (Fig. 1B). These slowly migrating tau species were strongly labeled by monoclonal antibodies AT8 and PHF1, which selectively bind phospho-tau residues Ser 202 /Thr 205 (34) and Ser 396 /Ser 404 (35), respectively. These data show that the CK1 isoform Cki␦ could directly phosphorylate tau protein, and that monoclonal antibodies AT8 and PHF1 could be used as probes for Cki␦-mediated phosphorylation reactions.
Cki␦ Phosphorylates Tau in Situ-To determine whether Cki␦ could modulate tau phosphorylation in situ, HEK-293 cells were transiently transfected with plasmids encoding htau40 and Cki␦, p25 (an activator of Cdk5), or empty pcDNA vector. Transfection times of 24 h were employed because cells remained viable during this time frame in all studies reported herein (data not shown). Cdk5 activator p25 was included because of its ability to activate endogenous Cdk5 in situ (36) and therefore serve as an alternate source of tau hyperphosphorylation activity. Total tau protein was detected with monoclonal antibody Tau5, whereas phospho-tau was detected with antibodies AT8 and PHF1. All cells transfected with htau40 constructs contained detectable levels of tau protein ( Fig. 2A). When these cells were co-transfected with empty vector (i.e. in the absence of exogenous protein kinase), a portion of total tau was PHF1 and AT8 positive, reflecting basal levels of tau phosphorylation. Basal levels of Cki␦ also were substantial in these cells as demonstrated by Western analysis using monoclonal antibody 128A. However, cotransfection with DNA constructs encoding Cki␦ significantly increased the amounts of both Cki␦ and phospho-tau detectable by both the AT8 and PHF1 antibodies (Fig. 1A). These data recapitulate the pattern seen in vitro (Fig. 1A) and suggest that Cki␦ overexpression significantly increased tau phosphorylation in situ (Fig. 2B). Expression of p25 also induced a large increase in tau phosphorylation, primarily because of the absence of endogenous p25 in non-transfected cells (Fig. 2B).
To confirm that the increased levels of Cki␦ observed upon overexpression were accompanied by increases in phosphotransferase activity, Cki␦ was immunoprecipitated and subjected to in vitro kinase assays using htau40 and [␥-32 P]ATP as substrates 24 h after transient transfection with Cki␦/tau or pcDNA3.1/tau constructs. Radioactive assay reaction products were separated by SDS-PAGE, stained with Coomassie Blue, and visualized on a phosphorimager. Although CK1 activity was found in both the Cki␦/tau overexpressing and pcDNA3.1/tau control cells (Fig. 3A), quantitation of the data showed that Cki␦ overexpression led to ϳ2-fold increases in recoverable phosphotransferase activity (Fig. 3B). These data confirmed that Cki␦ overexpression lead to increased Cki␦ activity within HEK-293 cells, but also showed that significant levels of Cki␦ activity were present under basal conditions.
Cki␦ and Tau Directly Interact in Situ-The Cki␦-mediated tau phosphorylation observed in HEK-293 cells could be direct or arise from an indirect protein phosphorylation cascade. To distinguish these alternatives, tau and Cki␦ were separately immunoprecipitated from stable tau cells and subjected to Western analysis using monoclonal antibodies 128A or Tau5. Results showed that tau and Cki␦ coimmunoprecipitated in both experimental paradigms (Fig. 4, C and D). These data suggest that tau associates with Cki␦ in situ, and therefore potentially serves as a direct substrate for Cki␦-mediated phosphorylation reactions.
To determine the region of Cki␦ involved in binding tau protein, htau40 (non-His 6 tagged) was incubated with or without the truncation mutant Cki␦-⌬317 in vitro and immunoprecipitated with antibody 128A. Subsequent Western analysis with antibodies 128A and Tau5 confirmed direct association A, cells transfected with empty vector showed measurable basal levels of phospho-tau and Cki␦ but not of p25. Transfection of cells with Cki␦ or cdk5 activator p25 significantly increased phospho-tau over basal levels. B, total tau (Tau5 epitope) and phospho-tau (PHF1 epitope) levels were quantified densitometrically from three individual experiments. When normalized for total tau content, phospho-tau was found to increase significantly with both p25 and Cki␦ transfection. *, p Ͻ 0.01; **, p Ͻ 0.001 when compared with empty vector control. between truncated Cki␦ and htau40 (Fig. 4B). These data suggest that at least a portion of the amino acid sequences mediating direct binding of Cki␦ to tau protein are located in the protein kinase catalytic domain.
CK1 Inhibitor IC261 Reverses Tau Hyperphosphorylation Induced by Cki␦ Overexpression-IC261 is a small molecule, membrane-permeable, ATP-competitive inhibitor of CK1 isoforms including Cki␦ (14,37). To determine whether inhibitor treatment could selectively reverse Cki␦-mediated tau hyperphosphorylation, HEK-293 cells were co-transfected with Cki␦/ tau, P25/tau, or pcDNA3.1/tau for 24 h, treated with and without IC261 (10 M for 30 min), and then probed for phospho and total tau levels with antibodies PHF1, AT8, and Tau5. The resultant Western blots showed that IC261 inhibited tau phosphorylation induced by Cki␦ overexpression without affecting expression levels of Cki␦ (Fig. 5A). Inhibition was selective for CK1, as shown by the failure of IC261 to significantly modulate tau phosphorylation induced by p25 overexpression (Fig. 5B). These findings are consistent with the inhibitory selectivity of IC261 determined both in vitro and in situ (37,38), and confirm the utility of IC261 for inhibition of CK1 activity in culture. Moreover, they suggest that CK1-mediated tau phosphorylation is opposed by endogenous phosphoprotein phosphatase activity that can decrease occupancy of phosphorylation sites in response to IC261 over time courses as short as 30 min.
Endogenous CK1 Contributes to Basal Levels of Tau Phosphorylation-The above data showed that basal levels of both Cki␦ and phospho-tau were substantial in HEK-293 cells expressing tau protein, and were consistent with previous studies showing that phosphorylation sites Ser 396 and Ser 202 were partially occupied in both stable and transiently transfected 3T3, Chinese hamster ovary, and SH-SY5Y cells (39 -41). To assess the contribution of endogenous Cki␦ activity to basal levels of tau phosphorylation, an HEK-293 cell line stably expressing htau40 was generated and employed in two experimental approaches. First, the dependence of tau phosphorylation on activity contributed by all CK1 isoforms was estimated by treating stable tau-HEK-293 cells with varying concentrations of IC261. In the absence of inhibitor, significant amounts of tau were detectable in this cell line, with a portion being PHF1 reactive (Fig. 6A). Although addition of IC261 for 30 min did not change total tau levels (i.e. detectable with antibody Tau5), it did result in large decreases in levels of phospho-tau detected by antibody PHF1 (Fig. 6A). Inhibition of tau phosphorylation was dose dependent, with an IC 50 of 1.5 Ϯ 0.5 M and up to 71.9 Ϯ 7.5% of PHF1 reactivity at saturation (Fig.  6B). In vitro, IC261 has been shown to inhibit purified recombinant Cki␦ with an IC 50 of 1.1 Ϯ 0.1 M and 98.1 Ϯ 3.2% of activity at saturation (Fig. 6B; data replotted from Fig. 1 of Ref. 37). These data suggest that CK1 activity makes a major contribution to basal levels of tau phosphorylation in HEK-293 cells.
As a second approach, the specific contribution of CK1 isoforms Cki␦/⑀ to basal tau phosphorylation was assessed by siRNA-mediated down-regulation conducted 48 h after transfection. RNA interference promotes hydrolysis of targeted mRNA in a sequence-specific reaction (42,43). The levels of total tau, phosphorylated tau, and Cki␦ were then detected by immunoblot analysis as a negative control, siRNA oligonucleotides selective for luciferase (GL2) were employed. Compared with GL2 treatment, Cki␦/⑀ siRNA lowered Cki␦ levels ϳ30% 48 h after transfection and decreased phosphorylated tau about 25% as well (Fig. 7B). RNA interference did not change the expression levels of total tau (Fig. 7A). These data indicate that either down-regulating Cki␦ activity or its protein level decreases tau phosphorylation, and suggests that Cki␦/⑀ contrib- FIG. 4. Cki␦ associates with tau in situ. Lysates (Lys) prepared from stable tau or wild-type HEK-293 cells were immunoprecipitated with anti-tau (Tau5) or anti-Cki␦ (128A) monoclonal antibodies as described under "Experimental Procedures" and subjected to immunoblot analysis using: A, 128A; or B, Tau5 as probes. A, both wild-type and stable tau cells contained Cki␦ that could be immunoprecipitated with 128A. B, in contrast, only stable tau cells contained detectable tau protein after immunoprecipitation with Tau5. Cki␦ and tau coimmunoprecipitated with either 128A or Tau5 antibodies in stable tau cells, suggesting direct association in situ. C, htau40 (1 g/ml) without His 6 tag incubated (4°C overnight) with or without truncation mutant Cki␦-⌬317 (1 g/ml) in Lysis buffer was immunoprecipitated with antibodies Tau5 or 128A and subjected to immunoblot analysis with anti-tau (Tau5) and anti-Cki␦ (128A) antibodies. Tau and Cki␦-⌬317 coimmunoprecipitated under these conditions, suggesting direct interaction between tau and the catalytic domain of Cki␦ in vitro. utes at least a portion of the basal CK1 activity detectable in HEK-293 cells.
Cki␦-mediated Tau Phosphorylation Disrupts Microtubule Binding-The microtubule binding activity of tau protein is modulated by phosphorylation (4,5). To assess the effect of Cki␦-induced phosphorylation on tau/tubulin interactions, HEK-293 cells transiently co-transfected with Cki␦/tau or pcDNA/tau constructs were examined by fluorescence microscopy after extraction with detergent in the presence of Taxol (a microtubule stabilizing agent). In the absence of extraction, total tau immunofluorescence in both Cki␦/tau or pcDNA/tau transfections were similar, again demonstrating that Cki␦ overexpression did not markedly change tau levels over the time course of the experiment (Fig. 8A). But in detergentextracted cells, which retained a detergent-insoluble cytoskeleton, levels of tau immunofluorescence were significantly lower in the presence of Cki␦ (Cki␦/tau transfection) than in its absence (pcDNA/tau transfection). These data suggested that the fraction of detergent-stable (i.e. microtubule-associated) tau decreased in response to phosphorylation induced by Cki␦ overexpression.
To confirm these findings, HEK-293 cells transiently cotransfected with Cki␦/tau or pcDNA/tau for 24 h were detergent extracted, fractionated into soluble (S1) and particulate (P1) fractions (the latter containing microtubules), and subjected to Western analysis using antibodies Tau5 and PHF1.
Phospho-tau (detected by PHF1) was found exclusively in the soluble fraction, with greater amounts recovered from cells transfected with Cki␦/tau than with pcDNA/tau (Fig. 8B). In contrast, significant amounts of total tau (detected by Tau5) were recovered in both particulate and soluble fractions (Fig.  8B). However, the subcellular distribution of total tau shifted toward the soluble pool in cells transfected with Cki␦/tau compared with pcDNA/tau cells (Fig. 8C).
Together these data suggest that the microtubule binding activity of tau decreases in response to Cki␦-mediated phosphorylation. To test this hypothesis, soluble tau prepared from each of the cell populations described above (fraction S1) was subjected to in vitro microtubule binding assays. Results showed that a portion of soluble tau from pcDNA/tau cells retained an ability to associate with synthetic microtubules (Fig. 8D), consistent with tau overexpression leading to saturation of endogenous microtubules. In contrast, soluble tau prepared from Cki␦/tau cells was almost devoid of microtubule binding activity (Fig. 8D), despite accumulating to higher levels than in control extracts (Fig. 8, C and D). These data suggest that the accumulation of soluble tau observed in situ upon Cki␦ overpression results from phosphorylation of tau protein at sites that directly modulate microtubule binding. DISCUSSION CK1 Isozymes and Tau Phosphorylation-Because of its natively unfolded structure and high complement of hydroxyamino acids, tau is an efficient substrate for many protein kinases in vitro (44), including tissue-derived CK1 activity (13,16).
Here it was found that CK1 serves as a tau protein kinase in situ as well. First, studies with the selective inhibitor IC261 showed that over 70% of basal tau phosphorylation at Ser 396 / Ser 404 sites in HEK-293 cells stems from CK1 activity. Supporting this conclusion, the IC 50 for IC261 in situ was only slightly higher than the value determined in vitro with purified show that IC261 inhibited basal and Cki␦-induced tau phosphorylation without modulating the levels of total tau or Cki␦. In contrast, neither p25 levels nor p25-induced tau phosphorylation was significantly affected by IC261. B, phospho-tau levels were quantified by densitometry from three independent experiments, normalized to levels of total tau, and plotted as percent phospho-tau in untreated, empty vector/tautransfected cells. **, p Ͻ 0.01 compared with untreated empty vector/ tau-transfected cells. ##, p Ͻ 0.01 compared with untreated Cki␦/tautransfected cells.
FIG. 6. CK1 inhibitor IC261 decreases tau phosphorylation. A, stable tau cells were grown to 80% confluence and then incubated with 0, 1, 3, or 10 M inhibitor IC261 for 30 min. All reactions were controlled for Me 2 SO, the vehicle for IC261. Treated cells were harvested, lysed, and subjected to SDS-PAGE. A, representative immunoblots probed with monoclonal antibodies PHF1, Tau5, and 128A. IC261-mediated inhibition of phospho-tau levels (i.e. PHF1 immunoreactivity) was concentration dependent. B, phospho-tau levels were quantified by densitometry, normalized to levels of total tau, and plotted as percent inhibition, where 0% inhibition corresponds to PHF1 immunoreactivity in the absence of IC261. Each point represents mean Ϯ S.E. of four independent experiments, whereas the line represents best fit of the data to a rectangular hyperbola (f). IC261 inhibited tau phosphorylation (PHF1 epitope) with an IC 50  Cki␦. Moreover, the IC261 inhibition isotherm was consistent with a single class of binding site. Because IC261 inhibits most CK1 isoforms with similar potency (37), basal levels of tau phosphorylation could potentially result from the activity of multiple CK1 isoforms. But the decrease in phospho-tau accompanying down-regulation of Cki␦/⑀ using RNA interference suggests that these specific isoforms compose at least a portion of IC261-sensitive tau phosphorylation activity under basal conditions. Indeed, Cki␦ and tau could be co-immunoprecipitated from stable tau cells containing only basal levels of Cki␦. These data demonstrate that tau and Cki␦ directly associate in situ, and are consistent with their colocalization to microtubules (16).
Second, overexpression of Cki␦ increased tau phosphorylation at sites it phosphorylated in vitro. Again, IC261 totally reversed the tau hyperphosphorylation induced by Cki␦ overexpression. However, the recognition sequence mediating the substrate selectivity of Cki␦ is not understood in the context of tau phosphorylation. On the basis of studies with short peptides, CK1 isoforms are phosphotropic kinases that recognize the motif S/T(P)XXS/T (45,46). Nonetheless, priming of substrates with phosphate is not a strict requirement, and motifs consisting of acidic residues N-or C-terminal to the phosphorylatable residue also are phosphorylated by CK1 (20,47,48 (17), the relationship between these sites and the motifs summarized above is weakly apparent only for Ser 396 /Ser 404 . It appears that recognition of full-length tau as substrate differs from that of short peptides. In any event, phosphorylation of tau at Ser 202 / Ser 205 and Ser 396 /Ser 404 by Cki␦ does not require priming by other protein kinases.
Consequences of CK1-mediated Tau Phosphorylation-Tau protein binds and stabilizes microtubules (49). However, this microtubule stabilizing function is regulated by its phosphorylation state. For example, hyperphosphorylated tau binds to microtubules weakly, but strong binding can be restored after dephosphorylation (4,50,51). Thus, the stability of the axonal in the absence of detergent extraction, no change in tau immunofluorescence was observed with Cki␦ overexpression, showing that transfection of cells with Cki␦ did not change total tau levels. c and d, in detergent-extracted cells, however, expression of Cki␦ led to decreased tau association with detergent-insoluble cytoskeletons. These data are consistent with a migration of tau from detergent-insoluble to detergent-soluble fractions of the cell in response to Cki␦ overexpression. B, HEK-293 cells transfected (24 h) with pcDNA/tau or Cki␦/tau were harvested, lysed, and separated into soluble (S1) and pellet (P1) fractions by centrifugation as described under "Experimental Procedures." Total (Tau5 epitope), phospho-tau (PHF1 epitope), and ␣-tubulin levels were then detected by immunoblot analysis. In pcDNA/tau cells, tau was found distributed between soluble and particulate fractions. In Cki␦/tau, however, tau distribution shifted toward the soluble pool, consistent with a decrease in microtubule association. C, total tau levels in particulate and soluble fractions were quantified by densitometry. Each bar represents the mean Ϯ S.E. of four independent experiments (**, p Ͻ 0.01 compared with empty vector/tau control). D, microtubule binding assays were performed using Taxol-stabilized microtubules prepared in vitro and equal volumes of the soluble, cell-derived fractions (S1) described above. After centrifugation, the resultant pellet (P2; containing microtubules) and supernatant (S2) fractions were analyzed for tubulin by SDS-PAGE (Coomassie Blue) and total tau (Tau5) by immunoblotting. Whereas a portion of soluble tau prepared from cells transfected with pcDNA/tau was capable of binding microtubules, soluble tau prepared from Cki␦/tau cells was almost devoid of microtubule binding activity. cytoskeleton may be influenced the equilibrium between tau phosphorylation and dephosphorylation. Phosphorylation sites that mediate this activity include Ser 199 , Ser 202 , Ser 231 , Thr 205 , Ser 396 , and Ser 404 (5,52). Cki␦ phosphorylates most of these sites in situ, and as a result tau/microtubule equilibrium shifts markedly toward free tau in response to Cki␦ overexpression. Similar shifts have been observed following Cdk5/p25, GSK3␤, or protein kinase A-induced phosphorylation of tau (5,36,(53)(54)(55)(56). Moreover, Cdk5/p39-induced tau phosphorylation reduces its affinity for microtubules during development in a transgenic mouse model (57). These data suggest that control of microtubule stability may involve common phosphorylation sites that are differentially regulated through the action of multiple protein kinases.
Role of Tau Phosphorylation in Disease-The pathway through which tau fibrillizes in disease is not entirely clear but appears to begin with amorphous aggregation of tau protein into non-fibrillar deposits (58). Phosphorylation can play a major role in this process, by modulating the equilibrium between microtubule-bound and free populations as discussed above. In vitro, elevated concentrations of phospho-tau promotes amorphous aggregation (59). In vivo, amorphous deposits, frequently in conjunction with intracellular membranes, are where the earliest ␤-sheet enriched species are detectable using fluorescent dyes (60). Studies with synthetic inducers of tau fibrillization have revealed the existence of a partially folded, thioflavin S-positive intermediate in the reaction pathway, the formation of which appears to be essential for fibrillization (61). Therefore, the key roles of tau phosphorylation in the early stage disease appear to be modulation of the tau/ microtubule equilibrium to raise intracellular concentrations of free tau, and promotion of amorphous aggregation from which assembly competent intermediates can form. Consistent with this model, extensive hyperphosphorylation of tau in COS cells leads to thioflavin-S positive deposits that are not fibrillar in appearance and contain only modest amounts of tau filaments (40).
CK1 in Disease-The properties of CK1 isoforms are consistent with a role in disease. First, as shown here, CK1 composes the bulk of basal tau phosphorylation activity in HEK-293 cells, and phosphorylates tau on sites that modulate tau function. These observations are consistent with Cki␦ and tau being resident on microtubules (16). Second, CK1 isoforms correlate spatially with neurofibrillary lesions in both AD (22) and other tauopathies (26), and appear as major components in preparations of authentic, disease-derived filaments (25). One isoform, Cki␦, is a particularly robust marker of granulovacuolar degeneration bodies (22). Thus CK1 is associated with more than one lesion in AD. Third, the appearance of CK1-positive lesions correlates temporally with memory decline in longitudinal studies of AD, suggesting that CK1 is as powerful a marker of AD progression as is tau (62). Finally, levels of at least one CK1 isoform, Cki␦, are greatly increased in postmortem AD tissue (22). Unlike tau protein, overexpression of Cki␦ is observable at the mRNA level (23), suggesting that accumulation does not result solely from sequestration with insoluble proteinaceous inclusions. In situ, Cki␦ expression increases in response to drug-or ␥-irradiation-induced genotoxic stress suggesting DNA damage as a potential mechanism for the elevated levels seen in AD (19,63). Further study of Cki␦ expression in AD will be required to clarify this issue.
Despite these properties, CK1 is but one of several protein kinases capable of phosphorylating tau in situ. GSK3␤ and Cdk5 also physically associate with tau within cells (64) and can modulate tau phosphorylation in transgenic animal models (65,66). GSK3␤ most prominently catalyzes phosphorylation at Ser 202 , Ser 235 , Ser 396 , and Ser 404 (55), whereas Cdk5/p25 has been shown to phosphorylate tau at Thr 181 , Ser 202 , Thr 205 , Thr 212 , Thr 217 , Ser 396 , and Ser 404 (54,67,68). Thus GSK3␤ and Cdk5/p25 potentially join CK1 in modulating tau/microtubule equilibrium (69). Although the potential interplay among these enzymes is unknown, it is interesting to note that GSK3␤ functions primarily as a phosphate-directed protein kinase dependent on priming of substrates by other protein kinases (70). Because of its strong contribution to basal tau phosphorylation state and inducibility in response to stress, CK1 could potentially provide priming activity as it does with ␤-catenin phosphorylation (71). Yet CK1 also recognizes primed substrates as in the case of p53 (72), and in this substrate recognition mode could potentially further amplify disease-associated phosphorylation of sites found tightly clustered around the microtubule repeat region (2,3).
Finally, we note that CK1 phosphorylates proteins in addition to tau in vivo. In the cases of p53 (72), ␤-catenin (73), multidrug transporter Pdr5p (74), Hedgehog signaling effector Cubitus interruptus (75), and proteins involved in circadian rhythm control (76), CK1-mediated phosphorylation modulates protein turnover. This function may be part of the neuronal response to stress, accounting for both the induction of CK1 in AD and its presence in granulovacuolar degeneration bodies. Thus CK1-mediated tau phosphorylation may result from an attempt by the cell to modulate protein turnover in response to cellular distress.