Interaction of Neuronal Cdc2-like Protein Kinase with Microtubule-associated Protein Tau*

Neuronal Cdc2-like protein kinase (NCLK), a ; 58-kDa heterodimer, was isolated from neuronal microtubules J. Biol. Chem. 267, The biochemical nature of NCLK-microtubule association is not known. In this study we found that NCLK is released from microtubules upon microtubule disassembly as a 450-kDa species. The 450-kDa species is an NCLK z tau complex, and NCLK-bound tau is in a nonphosphorylated state. Tau phosphorylation causes NCLK z tau complex dissociation, and phosphorylated tau does not bind to NCLK. In vitro , the Cdk5 subunit of NCLK binds to the microtu-bule-binding region of tau and NCLK associates with microtubules only in the presence of tau. Our data indicate that in brain extract NCLK is complexed with tau in a tau phosphorylation-dependent manner and that tau anchors NCLK to microtubules.

The neuronal Cdc2-like protein kinase (NCLK) 1 (also known as tau kinase II and brain proline-directed protein kinase) is a heterodimer of cyclin-dependent kinase 5 (Cdk5) and a neuronal-specific activator p25 subunit (reviewed in Ref. 1). The p25 subunit is a proteolytic fragment of a 35-kDa protein, p35 (1,2). Targeted disruption of the cdk5 gene in mice results in unique lesions in the central nervous system, lack of cortical laminar structure, as well as cerebellar foliate and prenatal death (3). In differentiating neurons, neurite outgrowth is inhibited by the introduction of a dominant negative Cdk5 mutant and is enhanced by Cdk5 and p35 overexpression (4). Similarly, mice lacking p35 do not show NCLK activity and display cortical lamination defects, seizures, and adult lethality (5). These observations indicate that NCLK plays an important role in brain development, neuronal differentiation, and neurite outgrowth. In neurons NCLK phosphorylates inhibitor 1 (6) and DARPP-32 (7), the two inhibitory proteins that upon phosphorylation suppress phosphoprotein phosphatase 1 activity. Munc18 (8) and synapsin (9), which respectively regulate synaptic vesicle exocytosis and neurotransmitter release, are phosphorylated by NCLK. NCLK also phosphorylates the microtubule-associated protein tau (10,11) and neurofilaments (12,13). These data indicate that NCLK regulates neural signaling, vesicular exocytosis, neurotransmitter release, and microtubule dynamics.
Neurofibrillary tangles are one of the two characteristic pathological lesions found in the brains of patients suffering from Alzheimer's disease (AD) (14). Paired helical filaments (PHFs) are the major fibrous component of neurofibrillary tangles and are composed of abnormally phosphorylated tau (i.e. containing more phosphate than normal tau) (15). It is thought that in AD abnormal tau phosphorylation causes a loss of tau function, cytoskeletal instability, PHF formation, and neurodegeneration (14).
NCLK activity has been reported to be higher in AD brain compared with normal controls (11,16). In cultured neurons, NCLK activation leads to tau hyperphosphorylation, microtubule disruption, and induction of apoptosis (11). Mice overexpressing the p25 subunit to activate NCLK display hyperphosphorylated tau and disrupted neuronal cytoskeleton (17). In vitro, NCLK phosphorylates tau on sites that are abnormally phosphorylated in PHF-tau (10), and NCLK phosphorylation promotes tau dimerization (18), a biochemical step involved in the conversion of tau to PHFs (19). These observations suggest that NCLK is also involved in AD pathogenesis (11).
Although an increasing amount of data clearly demonstrate that NCLK is an important enzyme, little is known about how it functions in the brain. Initially NCLK was isolated from the microtubule fraction purified from brain extract (20) indicating that a considerable amount of the kinase is bound to brain microtubules. Later, NCLK immunoreactivity was shown to be present in microtubules purified by repeated cycles of assembly/disassembly (10). These observations indicate that NCLK is an integral part of neuronal microtubules. The nature of NCLK association with microtubules and the significance of this association are unknown. In this study we analyzed microtubule-associated NCLK. Herein we show that microtubule-associated NCLK is a large molecular complex bound to microtubuleassociated protein tau. Our data suggest that NCLK associates with microtubules through tau. Tau Deletion Mutant Plasmids-Five synthetic primers, P1 (5Ј-CGC CAT ATG GCT GAG CCC CGC-3Ј), P2 (5Ј-CGC CAT  ATG ACA GCC CCC GTG-3Ј), P3 (5Ј-CGG GAT CCT CAC TGC AGG  CGG CTC T-3Ј), P4 (5Ј-CGG GAT CCT CAT TTA TTT CCT CCG C-3Ј), and P5 (5Ј-CGG GAT CCT CAC AAA CCC TGC TTG G-3Ј) were used in the following combinations to generate three tau deletion mutants: R-tau-(1-244) (forward P1 and reverse P3), R-tau-(245-369) (forward P2 and reverse P4), and R-tau(245-441) (forward P2 and reverse P5). By using the longest isoform of human tau cDNA (21) as the template, Pfu DNA polymerase-catalyzed polymerase chain reactions (26 cycles each) were performed at 95°C for 45 s, 56°C for 45 s, and 72°C for 1 min followed by 10 min with Taq polymerase at 72°C. Each polymerase chain reaction product was ligated into a pGEM-T Easy vector (Promega) and amplified. The pGEM-T vector inserts were excised with NdeI and BamHI and ligated into pET-3a expression vectors. Recombinant pET-3a plasmids were transformed into BL21 (DE3) cells for protein overexpression. All deletion mutants were sequenced using a Thermo Sequenase fluorescent labeled primer cycle sequencing kit (Amersham Pharmacia Biotech).

Construction of
Proteins and Peptides-Tubulin, MAP2, and tau were purified from bovine brain microtubules through phosphocellulose chromatography as described (10). NCLK was purified from fresh bovine brain extract (22). Recombinant tau (R-tau) and tau deletion mutants were purified from the respective bacterial lysates (23). Preparation of antibodies against bovine brain tau and a synthetic peptide derived from the N-terminal region of Cdk5 has been described previously (6,10). Anti-NCLK C-terminal antibody was from PharMingen. Monoclonal antibody against Cdk5 was from Upstate Biotechnology, Inc. (New York). Monoclonal antibody against tubulin, MAP2, and MAP1 were from Sigma. Monoclonal antibody Tau-1 was purchased from Roche Molecular Biochemicals. Monoclonal antibody PHF-1 was a gift from Dr. Peter Davis (Albert Einstein College of Medicine, New York).
Protein and Peptide Concentrations-Concentrations of tau and Rtau were determined spectrophotometrically (24). Concentrations of tau deletion mutants and phosphorylated R-tau were determined by Bio-Rad protein assay using R-tau as the standard. Concentrations of all other proteins were determined by Bio-Rad protein assay using BSA as a standard. Concentration of NCLK was based on its activity (22).
Preparations of GST Fusion Proteins-Recombinant glutathione Stransferase (GST), GST-p25, and GST-Cdk5 were purified from their respective bacterial lysates as described (25) except the lysis buffer also contained 1.6 M urea. Each bacterial lysate was centrifuged at 27, 000 ϫ g at 4°C for 15 min. The supernatant was dialyzed and mixed with glutathione-agarose beads (Sigma). The beads were recovered by centrifugation, washed, and used to generate Figs. 6, 7, and 9C.
GST-Pull Down Assay-Glutathione-agarose beads (50 l each) coated with GST, GST-Cdk5, or GST-p25 were mixed with 400 l of tau solution (50 mM Tris-HCl (pH 7.4), 50 mM NaCl, 5 mM EDTA, 0.3 M sucrose, 0.05% Tween 20, 0.3% BSA, and 20 g/ml indicated R-tau species), and the mixture was incubated for 5 h at 4°C with end-overend shaking. After incubation, beads were recovered by centrifugation and washed twice with washing buffer (50 mM Tris-HCl (pH 7.4), 50 mM NaCl, 5 mM EDTA, 0.3 M sucrose, 0.05% Tween 20, and 0.3% BSA). The beads were washed two more times with washing buffer without BSA and supplemented with 2% Nonidet P-40. The washed beads were mixed with 50 l of SDS-PAGE sample buffer, boiled, and centrifuged, and 10 l of supernatant was analyzed by immunoblot analysis using an anti-tau antibody.
Microtubule Purification-Microtubules were purified essentially as described previously (10). A fresh bovine brain (600 g) was homogenized with a blender for 1 min in 900 ml of cold PEM buffer (0.1 M PIPES (pH 6.6), 1 mM EGTA, 1 mM MgSO 4 , and 1 mM ␤-mercaptoethanol) at 4°C. The homogenate was centrifuged at 16,000 ϫ g for 45 min, and the supernatant (H in Fig. 1 and Table I) was adjusted 1 mM in GTP (Sigma) and incubated in a water bath at 37°C for 30 min with occasional gentle shaking. The incubated sample was transferred to four 250-ml centrifuge bottles. Each sample in the bottle was under layered with 20 ml of prewarmed 10% sucrose in PEM buffer containing 1 mM GTP. Samples were then centrifuged for 45 min at 16,000 ϫ g at 37°C. Supernatant (S1) was discarded, and the microtubule pellet was dispersed in 85 ml of ice-cold PEM buffer containing 1 mM GTP using a glass homogenizer and was designated as P1. The homogenized sample was incubated at 0°C for 30 min to solubilize microtubules and centrifuged at 27,000 ϫ g for 45 min to remove insoluble materials. The supernatant containing soluble microtubules was incubated at 37°C to induce microtubule assembly for 30 min, and the assembled microtubules were recovered by centrifugation at 37°C for 30 min. The supernatant (S2) was discarded, and the pellet was dissolved in 40 ml of ice-cold PEM/GTP buffer as above and designated as P2. The P2 fraction was subjected to a third cycle of assembly as above, and the supernatant (S3) was discarded and the pellet dissolved in 20 ml of ice-cold PEM/GTP buffer and designated as P3. P3 was immediately processed for NCLK purification without freezing.
Partial Purification of NCLK from Microtubules-All procedures were carried out at 4°C. The above microtubule fraction P3 (20 ml containing 2.5 mg/ml protein) was centrifuged at 10 5 ϫ g for 30 min, and the supernatant was loaded onto a phosphocellulose (Whatman) column (20 ϫ 5 cm) previously equilibrated in PEM buffer containing 0.1 mM GTP. Tubulin was recovered in the flow-through fraction. The column was washed with 4 column volumes of PEM buffer and then eluted with 300 ml of NaCl gradient (0 -1 M) in PEM buffer. Effluent fractions containing NCLK activity were combined (20 ml), concentrated by aquacide III (Calbiochem) to 5 ml, and loaded onto a FPLC Superose 12 gel filtration column (50 ϫ 1.6 cm), equilibrated, and eluted with buffer A (25 mM MOPS (pH 7.4), 50 mM ␤-glycerophosphate, 10 mM NaF, 0.2 M NaCl, 10 mM MgCl 2 , 1 mM EDTA, and 1 mM dithiothreitol). Active effluent fractions were combined, concentrated, adjusted to 2 M in NaCl, and incubated on ice for 2 h. Incubated sample was chromatographed through an FPLC Superose 12 gel filtration column equilibrated and eluted with buffer A containing 1 M NaCl. Effluent fractions containing NCLK activity were combined and used to generate Fig. 5.
Microtubule Sedimentation Assay-A sedimentation assay was carried out to monitor the microtubule assembly essentially as described previously (10) using a Sorval RC M120 EX micro-ultracentrifuge. Microtubules were assembled at 37°C for 30 min in a mixture containing 1 mg/ml tubulin, 1 mM GTP, and 30 M taxol (Molecular Probes) in PEM buffer. In the meantime, 10 l of NCLK (10 units) was mixed with 10 l of either tau (0.5 M), MAP2 (0.5 M), or buffer and incubated for 1 h in ice followed by 10 min at 37°C. To each of these incubated samples, 50 l of the above assembled microtubules at 37°C were added, and incubation at 37°C was continued for another 30 min. The samples were then centrifuged at 10 5 ϫ g at 37°C for 30 min. Supernatants were withdrawn and stored in ice, whereas pellets were dispersed in 50 l of ice-cold buffer A and incubated in ice for 1 h. Aliquots (10 l each) from both supernatants and dispersed pellets were subjected to SDS-PAGE, immunoblot, and NCLK activity analyses. The immunoblots were scanned, and the immunoreactive bands were quantitated by a personal Densitometer SI (Molecular Dynamics).
Other Methods-NCLK activity was assayed using a synthetic peptide substrate (22). SDS-PAGE, immunoblot, and slot-blot analyses were performed essentially as described (23). To perform immunoprecipitation, samples (ϳ 0.5 ml each) were precleared with 25 l of protein A-agarose beads (Sigma), pre-equilibrated in buffer A, and divided into equal halves. To each half, 25 l of either preimmune serum or the indicated primary antiserum was added, and both halves were shaken end over end at 4°C. After 4 h shaking, 25 l of protein A-agarose beads pre-equilibrated in buffer A were added to each half, and shaking was continued for another hour. Protein A-agarose beads were collected by centrifugation, washed 5 times with ice-cold buffer A, and immunoblotted using the indicated second antibody.

NCLK in Microtubule
Fraction-When microtubules in a fresh brain homogenate were subjected to three cycles of temperature-induced assembly and disassembly, ϳ40, ϳ22, and ϳ20% of total NCLK in the brain homogenate was recovered within first (P1), second (P2), and third (P3) microtubule pellets, respectively (Table I). NCLK-specific activity in P1 was ϳ10-fold higher than in brain homogenate and increased and became stable in P2 and P3 (Table I).
SDS-PAGE of various fractions showed that P1 that contained ϳ4% of total protein in the homogenate (Table I) was highly enriched with microtubules (Fig. 1A). The few minor contaminants in P1 decreased in P2. P3 microtubules appeared quite pure and contained essentially tubulin and other microtubule-associated proteins (26). Tubulin, the major component of microtubules, and the two most prominent microtubuleassociated proteins, MAP2 and tau, were concentrated during each step of purification (Fig. 1, B-D). The amounts of tubulin in P1, P2, and P3 were ϳ36, ϳ27, and 22% of the total, respectively (Fig. 1F). The amounts of tau and MAP2 in P1, P2, and P3 were ϳ30 and ϳ25%, ϳ26 and 23%, and 15 and ϳ20% of the total, respectively. These values seem to be reasonable for a standard microtubule purification procedure (26). More importantly, the amounts of NCLK in P1, P2, and P3 were ϳ40, 22, and 19% of total, respectively. Thus, the fraction of NCLK in various microtubule pellets resembles that of tubulin, MAP2, and tau. These observations are consistent with previous reports (10,20) and indicate that NCLK is stably bound to microtubules in brain extract.
Microtubule-associated NCLK Is a Relatively Large Molecular Size Species-To learn more about microtubule-associated NCLK, microtubules purified after three cycles of assembly/ disassembly were chromatographed through a phosphocellulose column. Tubulin was recovered within the flow-through fraction, whereas column-bound NCLK was eluted with a NaCl gradient. SDS-PAGE showed that MAP1, MAP2, and tau were also present in effluent fractions containing NCLK activity (data not shown). Fractions containing NCLK activity were combined, concentrated, and analyzed by FPLC Superose 12 gel filtration chromatography ( Fig. 2A). NCLK eluted from the column within fractions 38 -50 with peak fraction 42. The size and the Stokes radius of NCLK in Fig. 2A was estimated to be ϳ450-kDa and ϳ61 Å, respectively. Since NCLK is a 58-kDa heterodimer with a Stokes radius of ϳ35 Å (22), these observations indicated that microtubule-associated NCLK is a relatively large molecular species probably bound to some other biological molecule(s).
SDS-PAGE (Fig. 2B) and immunoblot analyses using antibodies against tau, MAP2, and MAP1 (data not shown) of various column fractions showed that MAP1 and MAP2 eluted near void volume within column fractions 38 -46 with peak fraction 40 (Fig. 2B). Similarly tau was present within fractions 38 -50 with peak fraction 42. A protein band of slightly higher molecular size than tau was also present within fractions 38 -46. This band and some bands that migrated in between MAP2 and tau on the SDS gel displayed cross-reactivity with anti-MAP2 antibody (data not shown) and therefore may be proteolytic products of MAP2. There were a few other faint high molecular size bands that migrated in between the tau and the MAP2 bands. These bands may represent MAP3, MAP4, and proteolytic fragments of MAP1.
To determine whether NCLK was associated with any of the MAPs, column fractions containing NCLK activity in Fig. 2A were combined and immunoprecipitated by using an anti-NCLK antibody. Immunoblot analyses using monoclonal antibodies against MAP1 and MAP2 did not detect any MAP1 or MAP2 in the above anti-NCLK immune complex (data not shown). However, when probed by using an anti-tau antibody, an intense immunoreactivity was observed (Fig. 3A). To confirm these data, we immunoprecipitated tau using an anti-tau antibody from the same column fractions and analyzed the anti-tau immune complex with anti-NCLK monoclonal antibody. As expected, NCLK co-immunoprecipitated with tau (Fig.  3B). These observations indicated that 450-kDa species in Fig.  2A contains NCLK complexed with tau.
Dissociation of NCLK⅐Tau Complex-To examine further the 450-kDa complex, fractions 38 -46 containing NCLK activity in Fig. 2A were pooled, concentrated, treated with 2 M NaCl, and analyzed by the same gel filtration column as in Fig. 2A but equilibrated and eluted with buffer containing 1 M NaCl. NCLK eluted from the column as a symmetrical peak within fractions 50 -58 with peak fraction 54 (Fig. 4A). The size of NCLK in Fig.  4A (fraction 54) is ϳ60-kDa. These observations indicated that NCLK has dissociated from the 450-kDa complex in the presence of high salt concentration.
SDS-PAGE (Fig. 4B) and immunoblotting using anti-tau antibody (data not shown) of various column fractions showed that tau had eluted within column fractions 40 -52 with a peak fraction 48. The size of tau in Fig. 4 (fraction 48) was determined to be ϳ165 kDa. Since tau is an asymmetric molecule and elutes as a relatively large molecule from a gel filtration column (18, 24), we could not ascertain the oligomeric structure of tau from Fig. 4A gel filtration data. However, tau that eluted as a 450-kDa species in Fig. 2A (fraction 42) became 165-kDa species in Fig. 4A. These observations indicated that tau also has dissociated from the 450-kDa complex in the presence of high salt.
Reconstitution of NCLK⅐Tau Complex-To reconstitute a NCLK⅐tau complex, fractions 52-58 from Fig. 4A were combined, concentrated, and mixed with purified bacterially expressed recombinant tau (R-tau). The mixture was dialyzed against buffer containing 0.2 M NaCl and analyzed by an FPLC Superose 12 gel filtration column. Control R-tau eluted from the column within fractions 38 -46 with peak fraction 42 (Fig.  5A, top panel). Similarly, control NCLK mixed with BSA was recovered within fractions 50 -60 with peak fraction 56 (Fig.  5A, middle panel). NCLK mixed with R-tau, however, eluted from the column within fractions 38 -44 with peak fraction 40 (Fig. 5A, lower panel). This NCLK must be R-tau-bound because R-tau co-eluted with NCLK within fractions 38 -44 with peak fraction 40.
To demonstrate that tau and NCLK bind to each other, BSA, tubulin, and R-tau were immobilized on a membrane in a slot-blot apparatus. Purified NCLK was layered over each immobilized protein and analyzed by slot-blot analysis using an anti-NCLK antibody. As shown in Fig. 5B, NCLK bound to R-tau but not to BSA or tubulin.
The Catalytic Cdk5 Subunit of NCLK Binds to Tau-NCLK is composed of a catalytic Cdk5 and regulatory p25 subunit (1, 2, 12). To determine which subunit binds to tau, we performed a GST-pull down assay. As shown in Fig. 6, R-tau bound to GST-Cdk5 and did not bind to GST or GST-p25. These observations indicated that tau binds to Cdk5 subunit of NCLK. Cdk5 Binds to the Microtubule-binding Region of Tau-The tau molecule contains an N-terminal projection domain, a microtubule-binding region, and a C-terminal tail (14). The microtubule-binding region of the longest tau isoform contains four microtubule-binding repeats (14,21). To identify the Cdk5-binding region within tau, we evaluated the binding to GST-Cdk5 of three R-tau deletion mutants: R-tau-(1-244) con-taining the N-terminal projection domain, R-tau-(245-441) containing the microtubule-binding region and C-terminal tail, and R-tau-(245-369) containing only the microtubule-binding region (Fig. 7A). As shown in Fig. 7, GST-Cdk5 bound to wild type R-tau (B) but failed to bind to R-tau-(1-244) (C) indicating that the Cdk5-binding region is not located within the Nterminal projection domain of tau. In a similar binding experiment, however, both R-tau-(245-369) (D) and R-tau-(245-441) (E) bound to GST-Cdk5. Since R-tau-(245-441) contains the microtubule-binding region and C-terminal tail and R-tau-(245-369) only contains the microtubule-binding region, these observations indicated that the R-tau microtubule-binding region is sufficient to bind to Cdk5.
Phosphorylation Causes NCLK⅐Tau Dissociation-Because tau is a phosphoprotein, we examined if tau phosphorylation may interfere with NCLK⅐tau interaction. FPLC Superose 12 gel filtration fractions containing the 450-kDa complex ( Fig.  2A) were combined and divided into equal halves. One-half was incubated with [␥-32 P]ATP/Mg 2ϩ and another with a nonhydrolyzable ATP analog AMP-PNP/Mg 2ϩ . Each half was analyzed by FPLC gel filtration in low salt conditions as in Fig. 2A. As shown in Fig. 8A (open circles) and B, tau and NCLK co-eluted in sample incubated with AMP-PNP/Mg 2ϩ within fractions 38 -48 with peak fraction 42 indicating that 450-kDa complex is intact under these conditions. The sample incubated with [␥-32 P]ATP/Mg 2ϩ , however, contained NCLK activity peak  Table I for the nomenclature of various fractions.

FIG. 2. FPLC gel filtration of microtubule-associated NCLK.
Microtubules obtained after three cycles of assembly/disassembly were fractionated over a phosphocellulose column. NCLK recovered from phosphocellulose chromatography was analyzed by FPLC Superose 12 gel filtration column (50 ϫ 1.6 cm) chromatography. FPLC was carried out at 4°C using Amersham Pharmacia Biotech FPLC system at a 1 ml/min flow. Fractions (1 ml each) were collected, and aliquots (10 l each) from indicated fractions were withdrawn and subjected to NCLK activity assay, SDS-PAGE, or immunoblotting using anti-NCLK Nterminal polyclonal antibody. A, gel filtration column profile; B, SDS-PAGE; C, immunoblot.
within fractions 54 -62 with peak fraction 56 (Fig. 8A, solid  circles). Similarly, tau eluted from the column with a peak fraction of 44 (Fig. 8C). The tau blot autoradiography showed that tau was phosphorylated in these fractions (Fig. 8D). These observations indicated that in sample incubated with [␥-32 P]ATP/Mg 2ϩ , NCLK and tau have dissociated from each other with concomitant tau phosphorylation.
If tau phosphorylation causes NCLK⅐tau complex dissociation, NCLK bound tau will be expected to be in a nonphosphorylated state. Similarly, tau in the 450-kDa complex of Fig. 2A should be in a nonphosphorylated state. To test these possibilities, we immunoprecipitated NCLK from brain extract using an anti-NCLK antibody. The immune complex was probed with phosphorylation sensitive anti-tau monoclonal antibodies. Tau-1 recognizes nonphosphorylated tau, whereas PHF-1 cross-reacts with only phosphorylated tau (11). As shown in Fig. 9A, tau co-immunoprecipitated with NCLK cross-reacted with tau-1 but not with PHF-1 (Fig. 9B).
To demonstrate that tau in 450-kDa complex is in a nonphosphorylated state, column fractions in Fig. 2A were immunoblotted using Tau-1 and PHF-1 antibodies. Tau-1 antibody displayed intense cross-reactivity with tau within fractions 38 -48, whereas the PHF-1 antibody showed no reaction (data not shown).
Finally, to substantiate the notion that phosphorylated tau does not bind to NCLK, we analyzed the binding of GST-Cdk5 with NCLK phosphorylated R-tau by GST-pull down assay. As shown in Fig. 9C, GST-Cdk5 bound to R-tau (lane 4) and completely failed to bind to phosphorylated R-tau (lane 6).
NCLK Binds to Microtubules through Tau-To determine the biochemical basis of the NCLK-microtubule association, NCLK, tau, and MAP2 were added to taxol-stabilized microtubules in various combinations, and microtubules were recovered as a pellet by centrifugation. Both pellet and supernatant were analyzed by immunoblot analysis using an anti-NCLK antibody. As shown in Fig. 10B, ϳ75% NCLK stayed in the supernatant (lane 2) and ϳ25% was present in microtubule pellet (lane 6) when NCLK alone was incubated with microtubules. When a similar experiment was performed in the presence of MAP2, ϳ30% NCLK associated with microtubules (lane 8) and ϳ70% stayed in the supernatant (lane 4). When tau was included in the incubation mixture, however, ϳ85% NCLK was recovered in the microtubule pellet (lane 7), and only ϳ15% was left in the supernatant (lane 3). As shown in Fig. 10A, This tau-induced enhancement of NCLK association with microtubules is not due to a difference in microtubule assembly as under all combinations the amount of assembled microtubules was almost identical (lanes 5-8). To substantiate these data, we performed a similar experiment but assayed NCLK activity in the microtubule pellet. As shown in Fig. 10C, very little NCLK activity was recovered in microtubule pellet in which NCLK alone was incubated with microtubules. However, a robust NCLK activity was detected in the microtubule pellet when NCLK was incubated with microtubules in the presence of tau. These observations indicated that NCLK associates with microtubule only in the presence of tau. Since NCLK directly binds to tau and does not bind tubulin (Fig. 5B) and tau is known to bind to microtubules (14), our data strongly argue that NCLK binds to microtubules through tau. DISCUSSION Previously, NCLK was purified from bovine brain extract by two different approaches. In the first approach NCLK was purified from microtubules and was found to be ϳ50 kDa (20). In the second approach, NCLK was purified from brain extract and was reported to be a relatively large size complex (22). To  Fig. 2A (fractions 38 -46) were concentrated to ϳ5 ml, adjusted to 2 M in NaCl, and incubated in ice for 2 h. The incubated sample was analyzed by a FPLC Superose 12 gel filtration column. All chromatographic conditions were identical to those of Fig. 2A except the NaCl concentration used in the column buffer was 1 M. Indicated column fractions (10 l each) were subjected to NCLK activity assay, SDS-PAGE, and immunoblotting using an anti-NCLK N-terminal antibody. A, column profile; B, SDS-PAGE; C, immunoblot.
investigate the cause of the disparity in the size of NCLK in the above two reports, we modified the purification procedure used in the first approach (20). We excluded the (NH 4 ) 2 SO 4 fractionation step and separated NCLK and other MAPs from tubulin by temperature-induced depolymerization of microtubules followed by phosphocellulose chromatography. When NCLK recovered from phosphocellulose column was chromatographed through an FPLC gel filtration column, the kinase eluted as a 450-kDa species (Fig. 2A). Thus, microtubule-associated NCLK is also a large macromolecular complex. However, as observed previously (22) and in this study (Fig. 4A), NCLK dissociates from the complex in the presence of high salt concentration. It is very likely that (NH 4 ) 2 SO 4 fractionation used in the first approach (20) may have caused NCLK dissociation from 450-kDa complex.
Until now the biochemical mechanism of the microtubule association of NCLK was unknown. Data presented here demonstrate that microtubule-associated NCLK is complexed with tau. In vitro, NCLK binds to tau and does not bind to tubulin (Fig. 5B), and tau is required for microtubule association of NCLK (Fig. 10). These data indicate that tau anchors NCLK to microtubules. Previously, protein kinase A, Cdc2 kinase, and MAP kinase were also reported to associate with microtubules. Protein kinase A and MAP kinase bind to microtubules through MAP2 (27,28), whereas Cdc2 kinase associates microtubules through MAP4 (29). Thus it appears that association of kinases to microtubules through MAPs is a general phenomenon.
Tau has three to four microtubule-binding repeats (14,30). These microtubule-binding repeats act independently, and only one repeat is sufficient to anchor tau to microtubules and promote microtubule assembly (30,31). A number of proteins such as actin (32), calmodulin (33,34), ␣-synuclein (35), phosphatase 2A (36), and NCLK (this study) bind to the microtubule-binding repeats of tau. These observations indicate that tau uses its microtubule-binding repeats for interacting with not only microtubules but other proteins also. Furthermore, tau is a naturally denatured protein without any compact folding (37,38) and hence has conformational flexibility to act as a docking protein and bind microtubules and other protein(s) at the same time. Thus, tau may bind to a protein through one of the microtubule-binding repeats and use other microtubulebinding repeats for microtubule binding. Consistent with this idea we found that tau simultaneously binds to microtubules and NCLK (Fig. 10).
NCLK binds to the microtubule-binding repeats of tau through its Cdk5 subunit (Fig. 6). Microtubule-binding repeats Other chromatographic conditions were the same as in Fig. 2A. Fractions (0.25 ml each) were collected, and indicated fractions were analyzed by NCLK activity assay. Similar results were obtained when the experiment was repeated one more time using different NCLK and R-tau preparations. B, slot-blot binding assay. Indicated proteins were spotted on a polyvinylidene difluoride membrane, layered with NCLK, and subjected to slot-blot analysis using anti-NCLK N-terminal antibody. The amounts of various proteins spotted on the membrane were as follows: BSA 5 g, tubulin 5 g, NCLK (positive control) 200 units, and indicated amounts of R-tau. are rich in basic residues and recognize the acidic region of target proteins containing several acidic residues situated in a row (39,40). We do not yet know the regions within Cdk5 involved in this binding. However, Cdk5 contains a stretch of acidic residues, 38 DDDDE 42 (12). This stretch is located within the kinase subdomain II and is highly variable among members of different kinase families (41) but identical in Cdk5 from human (42), mouse (43), rat (44), bovine (12), and Xenopus (45) and has only one conserved substitution, Asp 39 3 Glu in Drosophila (46). Experiments are underway in our laboratory to determine if this stretch is involved in tau binding.
We found that microtubule-associated NCLK is complexed with tau and that tau in the complex is in a nonphosphorylated state. The NCLK⅐tau complex dissociated upon tau phosphorylation (Fig. 8), and phosphorylated tau did not bind to NCLK (Fig. 9). Previously, phosphorylated tau has been shown to have a reduced affinity for microtubules (14,40,47,48). Together these observations suggest that in brain NCLK-tau association may be in a dynamic equilibrium regulated by tau phosphorylation/dephosphorylation. Nonphosphorylated tau may anchor NCLK to microtubules. Upon activation, NCLK may phosphorylate tau leading to the dissociation of tau from microtubules. The presence of tau and NCLK within the microtubule lattice may specify tau phosphorylation by NCLK in response to incoming intracellular signals that act through NCLK and regulate microtubule dynamics.
In AD tau becomes abnormally phosphorylated and aggregates into paired helical filaments (14,15). Because abnormal phosphorylation precedes tau aggregation in pretangle neurons (49,50), activation of tau-specific kinase(s) may be the earliest event in AD pathology. Elevated NCLK activity in postmortem FIG. 8. Effect of ATP/Mg 2؉ on NCLK⅐tau complex. NCLK⅐tau complex was isolated from microtubules by phosphocellulose followed by FPLC Superose 12 gel filtration chromatography as in Fig. 2A. Fractions containing complex were concentrated, and 0.5 ml of concentrated sample was incubated at 30°C for 30 min with either AMP-PNP/ Mg 2ϩ or [␥-32 P]ATP/Mg 2ϩ . The final concentrations of various components in the incubation mixture were 25 mM MOPS (pH 7.4), 0.1 mM EDTA, 1 mM dithiothreitol, 50 mM ␤-glycerophosphate, 10 mM NaF, 0.2 M NaCl, 10 mM MgCl 2 and 0.2 mM either [␥-32 P]ATP or AMP-PNP (Sigma catalog number A2647). Each incubated sample was analyzed by an FPLC Superose 12 gel filtration column, equilibrated, and eluted with buffer A. All chromatographic conditions were as in Fig. 2A. Indicated fractions (10 l each) were analyzed by NCLK activity assay and immunoblot analysis using anti-tau antibody. A, column profile; B and C, immunoblots of column fractions from samples incubated with AMP-PNP/Mg 2ϩ and [␥-32 P]ATP/Mg 2ϩ , respectively; D, autoradiography of C.
FIG. 9. Immunoblotting of anti-NCLK immune complex with phosphorylation-sensitive anti-tau monoclonal antibodies and GST-pull down assay. A and B, immunoblots. NCLK was immunoprecipitated from brain extract by using anti-NCLK C-terminal antibody, and the immune complex was immunoblotted with either monoclonal antibodies Tau-1 or PHF-1. Tau-1 recognizes nonphosphorylated tau, whereas PHF-1 cross-reacts with only phosphorylated tau. C, GSTpull down assay. Glutathione-agarose beads coated with GST or GST-Cdk5 were mixed with R-tau or phosphorylated R-tau, washed, and immunoblotted using anti-tau antibody that cross-reacts equally with both R-tau and phosphorylated R-tau. Lanes 1 and 2, R-tau and phosphorylated R-tau control markers; lanes 3 and 4, GST and GST-Cdk5 mixed with R-tau; lanes 5 and 6, GST and GST-Cdk5 mixed with phosphorylated R-tau.
FIG. 10. Microtubule sedimentation assay. Indicated proteins were incubated with taxol-stabilized microtubules, and the mixture was centrifuged. After centrifugation, pellet and supernatants were analyzed by SDS-PAGE, immunoblot analysis using anti-NCLK N-terminal antibody, and NCLK activity assay. A, SDS-PAGE; B, immunoblot; C, NCLK activity in microtubule pellets obtained in two different experiments. AD brain has been reported by two independent studies (11,16). Since activated NCLK in neurons causes hyperphosphorylation of tau, microtubule disruption, and apoptosis (11,17), an aberrant activation of NCLK was suggested to be involved in AD pathogenesis (11). However in addition to tau, NCLK also phosphorylates neurofilaments, Munc-18, synapsin, inhibitor 1, and DARPP-32 (6 -13). Surprisingly, in AD brain only tau phosphorylation is significantly up-regulated (14). These observations suggest that NCLK specifically targets tau in diseased brain. In this study, we demonstrated that tau and NCLK are complexed with each other and compartmented within microtubules. Within such a microcompartment, NCLK will only act upon tau and will have no access to other cellular proteins. Our study may provide further clues as to how in AD aberrantly activated NCLK leads to abnormal tau phosphorylation, cytoskeletal dysfunction, PHF formation, and perhaps neurodegeneration (11).