Identification of Tau and MAP2 as novel substrates of Rho-kinase and myosin phosphatase

Authors


Address correspondence and reprint requests to Kozo Kaibuchi, Department of Cell Pharmacology, Graduate School of Medicine, 65 Tsurumai, Showa-ku, Nagoya, Aichi 466–8550, Japan. E-mail: kaibuchi@med.nagoya-u.ac.jp

Abstract

Rho-kinase and myosin phosphatase are implicated in the phosphorylation-state of myosin light chain downstream of Rho, which is thought to induce smooth muscle contraction and stress fibre formation in non-muscle cells. Here, we found that microtubule-associated proteins, Tau and MAP2, interacted with the myosin-binding subunit (MBS) of myosin phosphatase, and were the possible substrates of both Rho-kinase and myosin phosphatase. We determined the phosphorylation sites of Tau (Thr245, Thr377, Ser409) and MAP2 (Ser1796) by Rho-kinase. We also found that Rho-kinase phosphorylated Tau at Ser262 to some extent. Phosphorylation by Rho-kinase decreased the activity of Tau to promote microtubule assembly in vitro. Substitutions of Ala for Ser/Thr at the phosphorylation sites of Tau (Tau-AAA) did not affect the activity to promote microtubule assembly, while substitutions of Asp for Ser/Thr (Tau-DDD), which are expected to mimic the phosphorylation-state of Tau, slightly reduced the activity. When Tau, or mutated forms of Tau, were expressed in PC12 cells, followed by treatment with cytochalasin D, they promoted extension of the cell process in a cytochalasin-dependent manner. However, Tau-DDD showed the weaker activity in this capacity than wild-type Tau or Tau-AAA. These results suggest that the phosphorylation-state of these residues of Tau affects its activity both in vitro and in vivo. Thus, it is likely that the Rho-kinase/MBS pathway regulates not only the actin-myosin system but also microtubule dynamics.

Abbreviations used
ANK

ankyrin repeat

AP sequence

assembly promoting sequence

BSA

bovine serum albumin

CAT

catalytic domain

DMEM

Dulbecco's modified Eagle's medium

ERM

Ezrin/Radixin/Moesin

FBS

fetal bovine serum

GST

glutathione-S-transferase

MAPs

microtubule associated proteins

MBP

maltose binding protein

MBS

myosin binding subunit of myosin phosphatase

MLC

myosin light chain

PHF

paired helical filament

SDS–PAGE

sodium dodecyl sulfate-polyacrylamide gel electrophoresis

Rho-associated kinase (Rho-kinase/ROCK/ROK) and myosin binding subunit (MBS) of myosin phosphatase are effectors of Rho small GTPase and thought to transmit signals from Rho to cytoskeletal proteins. Rho-kinase and MBS have been shown to be implicated in many cellular processes; stress fibre and focal adhesion formation, smooth muscle contraction, neurite retraction and cell migration (for reviews, see Kaibuchi et al. 1999; Amano et al. 2000). Rho-kinase regulates the phosphorylation state of myosin light chain (MLC) by the direct phosphorylation of MLC (Amano et al. 1996), and by the inactivation of myosin phosphatase, which is composed of MBS, a 37-kDa type 1 phosphatase catalytic subunit, and a 20-kDa subunit, through the phosphorylation of MBS (Kimura et al. 1996). In addition to MLC, the Ezrin/Radixin/Moesin (ERM) family proteins and adducin are found to be the substrates of both Rho-kinase and myosin phosphatase; Rho-kinase phosphorylates ERM and adducin, and myosin phosphatase that interacts with ERM and adducin through MBS dephosphorylates ERM and adducin phosphorylated by Rho-kinase (Fukata et al. 1998; Kimura et al. 1998; Retzer and Essler 2000). Thus, Rho-kinase and MBS are thought to co-operatively control the phosphorylation level of the subset of substrates and to regulate the cytoskeletal organization in vivo.

To further understand the functions of Rho-kinase and MBS, we attempted to identify MBS-interacting molecules other than MLC, ERM, and adducin. We purified MBS-interacting proteins with molecular masses of about 280 and 290 kDa, and identified them as MAP2. We also found that Tau interacted with MBS. Tau and MAP2 are the members of microtubule associated proteins (MAPs) which modulate microtubule structure and dynamics (for reviews, see Hirokawa 1994; Mandelkow and Mandelkow 1995). Tau and MAP2 are the best-known MAPs in neurons. They share a homologous assembly promoting sequence (AP sequence) composed of a repeated sequence motif in the COOH-terminus, and possess the respective projection domains in the NH2-terminus. Several lines of evidence indicate that the binding of MAPs to microtubules, or their capacity to modulate microtubule dynamics, is regulated by phosphorylation. However, it is difficult to elucidate which kinases are responsible for the phosphorylation of MAPs and when they phosphorylate and regulate MAPs in vivo.

In the present study, we found that MBS interacts with Tau and MAP2, that Rho-kinase phosphorylates them, and that myosin phosphatase dephosphorylates Tau phosphorylated by Rho-kinase. We determined the major sites of phosphorylation of Tau and MAP2 by Rho-kinase and evaluated the effects of the phosphorylation of Tau on its activity in vitro and in vivo.

Materials and methods

Materials and chemicals

Glutathione-S-transferase (GST)-Rho-kinase-catalytic domain (CAT) (6–553 amino acids) and rat full-length MBS were produced in Sf9 cells with a baculovirus system (Matsuura et al. 1987) and purified on a glutathione-Sepharose 4B column (Matsui et al. 1996) and a Mono S column (Fukata et al. 1998), respectively. GST-MBS-ankyrin repeat (ANK) and GST-MBS-C were prepared as described (Kimura et al. 1998). Pig MAP2 and Tau were prepared as described (Sloboda and Rosenbaum 1982). Myosin phosphatase was prepared from chick gizzard as described (Shimizu et al. 1994). Anti-MAP2, anti-Tau, and anti-β-tubulin monoclonal antibodies, anti-Tau-pSer262 polyclonal antibody, and cytochalasin D were purchased from Sigma (St Louis, MO, USA). Anti-HA polyclonal antibody was purchased from Medical and Biological Laboratories Co. Ltd. (Nagoya, Japan). [γ-32P]ATP was purchased from Amersham Pharmacia Biotech (Piscataway, NJ, USA). All materials used in the nucleic acid study were purchased from Takara Shuzo Co. (Kyoto, Japan). Other materials and chemicals were obtained from commercial sources.

Plasmid constructs

The cDNAs encoding the NH2-terminal region (MAP2-N1; 1–803 aa), the central region (MAP2-N2; 804–1668 aa) and the COOH-terminal AP region (MAP2-AP; 1669–1828 aa) of mouse MAP2 were subcloned from pβ-act-MAP2-myc into the BamHI site of pGEX-2T. The COOH-terminal AP region (MAP2-AP; 1669–1828 aa) of mouse MAP2 was subcloned from pβ-act-MAP2-myc into the BamHI site of pMal-c2. The cDNAs encoding the NH2-terminal region (Tau-N; 1–241 aa), and the COOH-terminal AP region (Tau-AP; 241–432 aa) of rat Tau were subcloned from pβ-act-tau-myc into the BamHI site of pGEX-2T. The cDNA encoding the full length of rat Tau was subcloned from pβ-act-Tau-myc into the BamHI site of pGEX-2T or pEF-BOS-HA. The cDNAs of TauT245A,T377A,S409A and TauT245D,T377D,S409D, in which Ala or Asp is substituted for Thr245, Thr377, and Ser409, were generated with a site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA), and subcloned into pGEX-2T or pEF-BOS-HA vector. The amino acid numbers of Tau constructs follow that of Kanai et al. (Kanai et al. 1989) (Tau-N and Tau-AP correspond to 1–250 and 250–441 aa of human tau40 longest form, respectively), while those of sites of phosphorylation and mutation follow that of human tau40 longest form for generalization.

Affinity column chromatography and in vitro binding assay

Affinity column chromatography of MBS was performed as described (Kimura et al. 1998).

GST-Tau full length expressed in Escherichia coli and recombinant MBS (1–976 aa) expressed in Sf9 cell were used for in vitro binding assay. GST-Tau or GST was immobilized onto 40 μL of glutathione-Sepharose 4B beads. The immobilized beads were incubated with the MBS (0.5 μm) in buffer A (20 mm Tris/HCl at pH 7.5, 1 mm EDTA, 1 mm DTT, 5 mm MgCl2) containing 1 mg/mL bovine serum albumin (BSA) for 1 h at 4°C. The beads were washed five times with 500 μL of buffer A, and the bound proteins were eluted three times with GST-Tau by the addition of 115 μL (3.3 volumes) of Buffer A containing 10 mm glutathione. The first eluate was subjected to sodium dodecyl sulfate – polyacrylamide gel electrophoresis (SDS–PAGE) and the proteins were detected by silver staining.

Identification of p280 and p290

The affinity-purified p280 and p290 were dialyzed three times against distilled water and concentrated by freeze-drying. The concentrated samples were separated by SDS–PAGE and transferred onto a polyvinylidene difluoride membrane (Problot, Applied Biosystems, CA, USA). The immobilized proteins were reduced, S-carboxymethylated, and digested in situ with Achromobacter protease I (a Lys-C) (Iwamatsu 1992). Molecular mass analyses of Lys-C fragments were performed by Matrix-assisted Laser Desorption/Ionization time-of-flight (MALDI-TOF) mass spectrometry using a PerSeptive Biosystem Voyager-DE/RP (Jensen et al. 1996). Proteins were identified by comparing the molecular weights determined by MALDI-TOF/MS and theoretical peptide masses from the proteins registered in NCBInr (10 April 1999).

Kinase assay

The kinase reaction for GST-Rho-kinase-CAT was carried out in 50 μL of kinase buffer (50 mm Tris/HCl at pH 7.5, 1 mm EGTA, 1 mm EDTA, 5 mm MgCl2) containing 100 μm[γ-32P]ATP (1–20 GBq/mmol), recombinant kinase, and substrates. After incubation for 10 min to 1 h at 30°C, the reaction mixtures were boiled in SDS sample buffer and subjected to SDS–PAGE. The radiolabelled bands were visualized by an image analyzer (Fuji, Japan).

Phosphatase assay

GST-Tau (1.08 μg of protein) was phosphorylated with GST-Rho-kinase-CAT (0.45 μg of protein) in 30 μL of kinase buffer (50 mm Tris/HCl at pH 7.5, 1 mm EGTA, 1 mm EDTA, 1 mm DTT, 5 mm MgCl2) containing 100 μm[γ-32P]ATP for 1 h at 30°C, and the reaction was stopped by the addition of 1 μm staurosporine. Myosin phosphatase from chick gizzard was pre-incubated in a reaction mixture containing 50 mm Tris/HCl at pH 7.5, 5 mm MgCl2, 30 mm KCl, 1 mg/mL BSA, 0.3 mm CoCl2, and 10 μm adenosine 5′-O–(3-thiotriphosphate)(ATPγS) in the presence or absence of GST-Rho-kinase-CAT for 10 min at 30°C, and the reaction was stopped by the addition of 1 μm staurosporine. The phosphatase assay was started by the addition of 10 μL of myosin phosphatase solution containing a given amount of myosin phosphatase to 30 μL of phosphorylated GST-Tau. The reaction mixture was then boiled in sample buffer for SDS–PAGE and resolved by SDS–PAGE. The 32P-labelled band corresponding to GST-Tau was visualized and estimated with an image analyzer (Fuji, Japan).

Microtubule assembly assay

Tubulin was purified from twice-cycled microtubule protein with a P11 phosphocellulose column (Whatman) as described (Shelanski et al. 1973). Microtubule assembly assay was carried out by sedimentation (Gaskin 1982). Tubulin (0.5 mg/mL) was mixed with phosphorylated or non-phosphorylated GST-Tau or GST-Tau mutants in the presence of 100 mm MES, 1.25 mm MgCl2, 1 mm EGTA and 1 mm GTP. This mixture was incubated at 37°C for 20 min and centrifuged at 100 000 g for 10 min at 30°C. The pellet and supernatant were subjected to SDS–PAGE.

Production of site- and phosphorylation state-specific antibodies for Tau and MAP2

Rabbit polyclonal Abs against MAP2 phosphorylated at Ser1796 (anti-MAP2-pS1796 antibody) and Tau phosphorylated at Thr377 (anti-Tau-pT377 antibody) were raised as described (Inagaki et al. 1997). The phosphopeptides Cys-Ser1791-Pro-Arg-Arg-Leu-phosphoSer1796-Asn-Val-Ser-Ser-Ser1801 for MAP2 and Cys-Glu372-Thr-His-Lys-Leu-phosphoThr377-Phe-Arg-Glu-Asn381 for Tau were chemically synthesized as antigens and bound to the carrier protein, keyhole limpet hemocyanin, at the NH2-terminal cysteine residue, by Peptide Institute Inc. (Osaka, Japan). Because we confirmed that anti-MAP2-pS1796 antibody recognized Tau-pS409 as well as MAP2-pS1796, we hereafter refer to this antibody as anti-Tau-pS409 antibody. Rabbit polyclonal antibody against Tau phosphorylated at Thr245 (anti-Tau-pT245 antibody) was produced against phosphopeptide Gly-Cys-Lys240-Ser-Arg-Leu-Gln-phosphoThr245-Ala-Pro-Val-Pro-Met250 by Biologica Co. (Aichi, Japan). The antisera obtained were then affinity-purified against the respective phosphopeptides.

Transfection into COS7 cells

COS7 cells were seeded on a 60-mm dish at a density of 5 × 105 cells/dish in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal bovine serum (FBS) and cultured overnight. The medium was renewed 2 h prior to transfection. Transfection of plasmids into COS7 cells was carried out using LIPOFECTAMINETM. reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. A LIPOFECTAMINETM./DNA mixture was prepared and added to the dish with gentle agitation. After a 6-h incubation, the cells were grown in DMEM with 10% FBS for 1 day and then in DMEM for 1 day. In some experiments, cells were treated with calyculin A (0.1 μm) for 10 min. The cells were treated with 10% (w/v) trichloroacetic acid. The resulting precipitates were subjected to immunoblot analysis.

Transfection into PC12 cells and immunofluorescent staining

PC12 cells were seeded at a density of 1.7 × 104 cells onto 13-mm glass coverslips coated with polylysine (Sigma) in DMEM with 10% FBS and 5% horse serum, and cultured overnight. The medium was renewed 2 h prior to transfection. Transfection of plasmids into PC12 cells was carried out using LIPOFECTAMINE2000TM. reagent (Invitrogen) according to the manufacturer's protocol. A LIPOFECTAMINE2000TM./DNA mixture was prepared and added to the medium with gentle agitation. The cells were grown in DMEM with 10% FBS and 5% horse serum for 2 days and then treated with cytochalasin D (2 μm) or vehicle for 1 h. The cells were fixed with 3.7% formaldehyde in PBS for 10 min, washed with PBS, and permeabilized with PBS containing 0.2% Triton X-100 and 2 mg/mL BSA for 10 min. After permeabilization, the cells were double stained with anti-β-tubulin monoclonal antibody and FITC-conjugated anti-mouse antibody and with anti-HA polyclonal antibody and Texas red-conjugated anti-rabbit antibody. After being washed with PBS three times, the cells were examined using a Zeiss axiophoto microscope.

Other procedures

Determination of phosphorylation sites of Tau (Kawano et al. 1999) and SDS–PAGE (Laemmli 1970) was performed as described previously.

Results

Identification of MBS-interacting molecules

To search for novel MBS-interacting molecules, bovine brain membrane extract was loaded onto a glutathione-Sepharose affinity column on which GST, GST-MBS-ANK, or GST-MBS-C (Fig. 1a) was immobilized. The proteins bound to the affinity columns were then eluted by the addition of 200 mm NaCl. Proteins with molecular masses of about 280 kDa (p280) and 290 kDa (p290) were well detected in the eluate from the GST-MBS-ANK affinity column along with adducins (Fig. 1b), less so in that from the GST-MBS-C affinity column (data not shown), and neither in that from the GST nor GST-Notch-ANK affinity column (Fig. 1b).

Figure 1.

(a)  Structure of rat MBS. MBS contains an ankyrin repeat in the NH 2 -terminus and leucine zipper-like motif in the COOH-terminus. The ankyrin repeat region (39–295 aa) and COOH-terminal region (699–976 aa) of MBS were fused with GST to produce GST-MBS-ANK and GST-MBS-C. (b)  Purification of MBS-ANK-interacting molecules. The bovine brain membrane extract was loaded onto a glutathione-Sepharose column containing the indicated GST fusion proteins. The bound proteins were eluted by the addition of 200 m m NaCl. Aliquots of the eluates were resolved by SDS–PAGE, followed by silver staining. Closed arrowheads denote the position of MAP2. Open arrowheads denote α-, β-, and γ-adducins. (c)  Immunoblot analysis of MAP2 and Tau. Aliquots of the eluates from the affinity columns were resolved by SDS–PAGE, followed by immunoblot analysis with anti-MAP2 Ab (upper) or with anti-Tau Ab (bottom). Arrowheads denote the position of MAP2 or Tau. The results are representative of three independent experiments. (d)  Interaction of Tau with MBS in vitro was examined. MBS expressed in and purified from Sf9 cells was incubated with GST or GST-Tau immobilized glutathione-Sepharose. After a wash, the bound proteins were eluted with GST or GST-Tau by the addition of glutathione. The eluates were subjected to SDS–PAGE and detected by silver staining. Arrow denotes the position of MBS. The results are representative of three independent experiments.

To identify p280 and p290, molecular mass analyses were conducted as described in Materials and methods. The molecular weights of peptides derived from p280 and p290 were determined, and found to be identical to those from human MAP2. p280 and p290 were recognized by anti-MAP2 antibody (Fig. 1c). Bands recognized by anti-MAP2 antibody were detected in eluate from GST-MBS-ANK column, but neither in eluate from GST, GST-MBS-C nor GST-Nothch-ANK. Although the calculated MW of MAP2 is 198.8 kDa, the apparent value based on SDS–PAGE has been reported at 280 kDa. We thus concluded that p280 and p290 were bovine counterparts of MAP2 and hereafter refer to them as MAP2. MAP2 is a member of the MAP family that includes MAP2, MAP1, and Tau, and is implicated in the regulation of microtubule dynamics. We further examined whether Tau interacts with MBS. The six bands recognized by anti-Tau antibody (45–70 kDa) were detected only in eluate from the GST-MBS-ANK column (Fig. 1c). It is well known that Tau from brain makes multiple bands on SDS–PAGE, which is thought to be the results of both alternative splicing and phosphorylation. These results indicate that MAP2 and Tau specifically interact with MBS through its ankyrin repeat domain.

To examine the interaction between MBS and MAPs in a cell-free system, recombinant rat Tau was prepared as GST-fusion protein. Recombinant MBS purified from Sf9 cells was incubated with GST or GST-Tau immobilized beads. After the beads had been washed, the GST fusion proteins were eluted by the addition of glutathione. MBS was co-eluted with GST-Tau, but not with GST (Fig. 1d). This result indicates that recombinant Tau binds directly to MBS. Because the COOH-terminal AP region is well conserved between Tau and MAP2 and there is little homology in the NH2-terminal projection region, it is likely that Tau and MAP2 interact with the ankyrin repeat region of MBS through their AP regions.

In vitro phosphorylation of MAPs by Rho-kinase

We next examined whether Rho-kinase phosphorylates MAPs in a cell-free system. The constitutively active form of recombinant Rho-kinase (GST-Rho-kinase-CAT) purified from Sf9 cells phosphorylated both Tau and MAP2 purified from pig brain (data not shown). We examined whether GST-Rho-kinase-CAT phosphorylates various fragments of Tau and MAP2 (Fig. 2a) and found it to phosphorylate recombinant GST-Tau-N, GST-Tau-AP, or GST-MAP2-AP, but barely GST-MAP2-N1 or GST-MAP2-N2 (Fig. 2b). The maximum amounts of phosphates incorporated into GST-Tau-N, GST-Tau-AP, and GST-MAP2-AP were approximately 0.9, 1.8, and 0.6 mol per 1 mol of protein, respectively.

Figure 2.

(a)  Structures of Tau and MAP2. Tau and MAP2 are composed of the NH 2 -terminal projection domain and the COOH-terminal microtubule (MT)-binding domain/AP sequence. Full lengths of Tau, Tau-N (1–250 aa), Tau-AP (250–441 aa), MAP2-N1 (1–803 aa), MAP2-N2 (804–1668 aa), and MAP2-AP (1669–1828 aa) were constructed as GST fusion proteins and obtained from E. coli . R1–4 indicate the repetitive sequences. The amino acid numbers of Tau constructs follow that of the human tau40 longest form, and Tau-N and Tau-AP correspond to 1–241 aa and 241–432 aa of Tau of Kanai et al. (1989 ), respectively. (b)  Phosphorylation of GST-MAPs by Rho-kinase. GST-Tau-N, -Tau-AP, -MAP2-N1, -MAP2-N2, or -MAP2-AP was phosphorylated by GST-Rho-kinase-CAT. The phosphorylated proteins were resolved by SDS–PAGE and visualized by an image analyzer. Arrowheads denote the phosphorylated Tau or MAP2. The arrow denotes the autophosphorylation of GST-Rho-kinase-CAT. The results are representative of three independent experiments.

We used GST-Tau-N, GST-Tau-AP, and maltose binding protein (MBP)-MAP2-AP to determine the sites of phosphorylation of MAPs by Rho-kinase. The major phosphorylation sites were identified as Thr245, Thr377, and Ser409 for Tau (the amino acid numbers follow that of human tau40 longest form), and Ser1706) for MAP2 (Fig. 3). We also found that Ser262 of Tau was phosphorylated to some extent, and Thr175 and Ser198 were weakly phosphorylated by Rho-kinase by using phosphorylation-state specific antibodies. Tau and MAP2 share a sequence similarity in the AP region (56% identical). The Ser409 of Tau is thought to correspond to Ser1706) of MAP2. However, there are no Ser or Thr in MAP2 corresponding to Thr245 and Thr377 of Tau. It is noted that we were not able to determine all the phosphorylation sites of Tau and MAP2, and that there may be several sites other than those above.

Figure 3.

The sites of phosphorylation of Tau (Thr245, Thr377, Ser409 and Ser262) and MAP2 (ser1796) by Rho-kinase are indicated (shaded). Phosphorylation state-specific antibodies were prepared against Tau-pT245 (anti-Tau-pT245 antibody), Tau-pT377 (anti-Tau-pT377 antibody) or MAP2-pS1796 (anti-MAP2-pS1796/Tau-pS409 antibody). Underlines indicate the sequences for antigen.

In vitro dephosphorylation of Tau by myosin phosphatase

Because Rho-kinase and myosin phosphatase seem to share their substrates and co-operatively regulate the phosphorylation state of a subset of substrates, we then examined whether myosin phosphatase dephosphorylates Tau phosphorylated by Rho-kinase. Myosin phosphatase showed the phosphatase activity toward GST-Tau phosphorylated by GST-Rho-kinase-CAT (Fig. 4). We also found that pre-incubation of myosin phosphatase with GST-Rho-kinase-CAT and ATPγS, which led to the thiophosphorylation of MBS, resulted in a decrease of the phosphatase activity of myosin phosphatase toward GST-Tau (Fig. 4). These results indicate that myosin phosphatase dephosphorylates Tau phosphorylated by Rho-kinase, and that its activity is modulated through the phosphorylation of MBS by Rho-kinase.

Figure 4.

Dephosphorylation of Tau by myosin phosphatase. GST-Tau was phosphorylated by GST-Rho-kinase-CAT, and then incubated with indicated doses of myosin phosphatase pre-incubated with ATPγS (closed circle) or with ATPγS and GST-Rho-kinase-CAT (open circle) for 10 min at 30°C. The proteins were resolved by SDS–PAGE and radiolabelled GST-Tau was quantified by an image analyzer. All data are means ± SD of triplicate determinations.

Effect of Tau phosphorylation on the microtubule assembly activity of Tau

To evaluate the effect of Tau phosphorylation by Rho-kinase on the activity of Tau, microtubule assembly assay was performed using GST-Tau and GST-Rho-kinase-CAT. As shown in Fig. 5, GST-Tau promoted the assembly of microtubule in a dose-dependent manner as described (Weingarten et al. 1975). Phosphorylation of GST-Tau by GST-Rho-kinase-CAT (2.7 mol of phosphate per 1 mol of GST-Tau under the conditions) decreased its activity to promote microtubule assembly (Fig. 5).

Figure 5.

Effect of phosphorylation of Tau by Rho-kinase on the ability of microtubule-assembly. Assembly reactions were performed at tubulin concentration of 0.5 mg/mL, and were started by the addition of GST-Tau pre-incubated with GST-Rho-kinase-CAT (open square) or with ATP and GST-Rho-kinase-CAT (closed square), GST-Tau-AAA pre-incubated with GST-Rho-kinase-CAT (open circle) or with ATP and GST-Rho-kinase-CAT (closed circle), or GST-Tau-DDD pre-incubated with GST-Rho-kinase-CAT (open triangle) or with ATP and GST-Rho-kinase-CAT (closed triangle). All data are means ± SD of triplicate determinations.

To further examine the effect of the phosphorylation sites of Tau identified above, we produced Tau mutant with Ala substituted for Thr245/Thr377/Ser409 (Tau-AAA) which is expected to be a non-phosphorylated form, and that with Asp substituted for Thr245/Thr377/Ser409 (Tau-DDD) which is expected to mimic a phosphorylated form. GST-Tau-AAA showed almost the same ability to promote microtubule assembly as GST-Tau wild type, whereas GST-Tau-DDD showed slightly weaker ability than GST-Tau wild type (Fig. 5). Incorporation of phosphate into GST-Tau-AAA or GST-Tau-DDD was 1.4 or 1.2 mol per 1 mol of protein, respectively, when they were incubated with GST-Rho-kinase-CAT and ATP under the same conditions as GST-Tau wild type (2.7 mol of phosphate per 1 mol of protein). Phosphorylation of GST-Tau-DDD by Rho-kinase reduced the activity to the level of phosphorylated GST-Tau wild type, while phosphorylation of GST-Tau-AAA partially reduced the activity (Fig. 5). These results indicate that Rho-kinase decreases the activity of Tau to promote microtubule assembly through the phosphorylation of it, partly at the sites of Thr245/Thr377/Ser409, and that other sites of phosphorylation by Rho-kinase also contribute to the modulation of the activity of Tau.

Phosphorylation of Tau in COS7 cells

To examine the phosphorylation-state of Tau and MAP2 phosphorylated at the Rho-kinase sites in vivo, we prepared rabbit polyclonal antibodies against Tau-pT245, -pT377, or MAP2-pS1796/Tau-pS409. These antibodies specifically recognized GST-Tau phosphorylated by Rho-kinase, but neither GST-Tau without phosphorylation, GST-Tau-AAA nor GST-Tau-DDD phosphorylated by Rho-kinase (Fig. 6a). Anti-Tau-pS409 antibody recognized MBP-MAP2-AP phosphorylated by Rho-kinase (data not shown). We also examined dephosphorylation of Tau by myosin phosphatase for each site, and found that myosin phosphatase dephosphorylated Tau at these three sites (Thr245, Thr377, Ser409), as well as Ser262 (Fig. 6b). Myosin phosphatase preferentially dephosphorylated Tau at Thr245 and Thr377, and moderately at Ser409 and Ser262.

Figure 6.

(a)  The specificity of anti-Tau-pT245, -pT377 and -pS409 antibodies. Protein (750 fmol) of GST-Tau wild type without phosphorylation (lane 1), GST-Tau wild type (lane 2), GST-Tau-AAA (lane 3) or GST-Tau-DDD (lane 4) phosphorylated by GST-Rho-kinase-CAT was resolved by SDS–PAGE and subjected to immunoblot analysis using anti-Tau-pT245, -pT377, or -pS409 antibody. Arrows denote the position of phosphorylated GST-Tau. The results are representative of three independent experiments. (b)  Dephosphorylation of Tau by myosin phosphatase at Thr245, Thr377, Ser409 and Ser262. Protein (750 fmol) of GST-Tau wild type phosphorylated by Rho-kinase was incubated with indicated amounts of myosin phosphatase, and then subjected to immunoblot analysis using anti-Tau-pT245, -pT377, -pS409 or -pS262 antibody. Arrows denote the position of phosphorylated GST-Tau. Stoichiometries of whole sites measured by [ 32 P] are indicated below. (c)  Phosphorylation of Tau in COS7 cells expressing Rho and/or Rho-kinase. The level of Tau phosphorylation was examined in COS7 cells expressing HA-Tau and HA-Rho V14 and/or myc-Rho-kinase by immunoblot analysis. COS7 cells were co-transfected with pEF-BOS-HA-Tau and pEF-BOS-myc (lane 1), pEF-BOS-myc treated with calyculin A (0.1 μ m ) (lane 2), pEF-BOS-HA-Rho V14 (lane 3), pEF-BOS-myc-Rho-kinase (lane 4), pEF-BOS-HA-Rho V14  + pEF-BOS-myc-Rho-kinase (lane 5), or pEF-BOS-myc-Rho-kinase-CAT (lane 6). Precipitates treated with 10% trichloroacetic acid from the cells were subjected to SDS–PAGE, and relative amounts of phosphorylated Tau were determined by immunoblot analysis with anti-Tau-pT245, -pT377 or -pS409 antibody. The relative amounts of HA-Tau were confirmed by immunoblot analysis with anti-HA antibody (data not shown). The extra bands below Tau in the blot are thought to be degradation forms of Tau. The results are representative of three independent experiments.

To determine whether Rho-kinase regulates the level of Tau phosphorylation in vivo, anti-Tau-pT245, -pT377 and -pS409 antibodies were used. Because Tau is not expressed in COS7 cells, pEF-BOS-HA-Tau was co-transfected with plasmids carrying the RhoV14 (the dominant active form of Rho) and/or Rho-kinase cDNAs into COS7 cells. The level of Tau phosphorylation was low in serum-starved COS7 cells expressing HA-Tau alone. When the cells were treated with calyculin A, which is a phosphatase inhibitor, the level of Tau phosphorylation was increased at Thr245, Thr377 and Ser409 (Fig. 6c), indicating that the activities of phosphatases are involved in the regulation of phosphorylation level of Tau in the resting COS7 cells. The expression of RhoV14 or full-length Rho-kinase slightly increased the phosphorylation level of Tau at Thr245 and Ser409. Co-expression of RhoV14 and full-length Rho-kinase further enhanced, and that of CAT (constitutively active form of Rho-kinase) markedly increased, the phosphorylation of Tau at these sites (Fig. 6c). However, the expression of CAT increased the Tau phosphorylation, but that of RhoV14 and/or full-length Rho-kinase had little effect on the Tau phosphorylation at Thr377 (Fig. 6c). Taken together, these results provide evidence that the Rho/Rho-kinase pathway regulates the level of Tau phosphorylation in intact cells, especially at Thr245 and Ser409. As to Ser262, we could not observe an elevation of the signal in the cells expressing either RhoV14, Rho-kinase or CAT (data not shown), possibly because anti-Tau-pS262 antibody weakly recognized the non-phosphorylated form of Tau in vitro.

Expression of Tau mutants in PC12 cells

J. G. Leger and colleagues reported that PC12 cells expressing full-length Tau extend microtubule-dependent processes only in the presence of cytochalasin B, an actin-depolymerizing drug (Leger et al. 1994). Here, we used this assay for Tau mutants substituted with Ala or Asp for Thr245/Thr377/Ser409. PC12 cells expressing wild-type Tau extended microtubule-dense processes after the 1-h treatment with cytochalasin D, as described previously (Fig. 7). We confirmed that the expression of GST did not result in a significant promotion of process formation. Both Tau mutants altered at phosphorylation sites also promoted process formation in a cytochalasin D-dependent manner. However, Tau-DDD showed a significantly weaker activity in this capacity than wild-type Tau or Tau-AAA (Fig. 7), suggesting that substitution of Asp for Thr245/Thr377/Ser409 reduced the activity of Tau to promote the formation of processes. These results are in good agreement with the model that phosphorylation of Tau by Rho-kinase leads to a decrease in its activity to promote microtubule assembly in vivo.

Figure 7.

Effect of mutations of phosphorylation sites of Tau on the promotion of process formation in PC12 cells. (a) pEF-BOS-HA-Tau wild type (1, 2), pEF-BOS-HA-Tau-AAA (3), or pEF-BOS-Tau-DDD (4) was transfected into PC12 cells by lipofection. After a 48-h incubation, the cells were treated with vehicle (1) or 2 μ m cytochalasin D (2–4) for 1 h, and then fixed and doubly stained with anti-β-tubulin monoclonal antibody (1–4) and anti-HA polyclonal antibody (data not shown). Arrows show the transfected cells. Scale bar, 20 μm (b) The percentages of cells with processes > 30 μm in HA-positive cells with (hatched column) or without (open column) cytochalasin D are indicated. Data are means ± SD of at least triplicate determinations. **Significantly different ( p  < 0.01) by t -test.

Discussion

MAPs are isolated from tissues as phosphoproteins and good substrates for many protein kinases including MARK/PAR-1 (Drewes et al. 1995), PKA (Theurkauf and Vallee 1983), PKC (Mori et al. 1991), GSK-3 (Hanger et al. 1992; Mandelkow et al. 1992), and PKN (Taniguchi et al. 2001) in vitro. MARK/PAR-1 phosphorylates Ser262, Ser293, Ser324, and Ser356 of Tau (the amino acid numbers follow that of the human tau40 longest form), and Ser1680, Ser1711, and Ser1743 of MAP2 (KXGS motif in the repeated sequence). There are four KXGS motives in Tau, and three in MAP2. It has been reported that phosphorylation of the first KXGS (Ser262) of Tau eliminates its stabilizing activity, and that the others have only a modulatory influence. PKC phosphorylates Ser324 of Tau, and Ser1703, Ser1711, and Ser1728 of MAP2 (Correas et al. 1992). PKA phosphorylates Ser212, Ser324, Ser356, Ser409, and Ser416 of Tau (Scott et al. 1993; Drewes et al. 1995). PKN, the other effector molecule of Rho, phosphorylates Ser258, Ser320, and Ser352 (Taniguchi et al. 2001). The phosphorylation and splicing patterns of Tau and MAP2 are regulated developmentally (Riederer and Matus 1985; Riederer et al. 1995; Imahori and Uchida 1997; Spillantini and Goedert 1998). The hyperphosphorylation of Tau in Alzheimer's disease, in which Tau is phosphorylated by proline-directed protein kinases such as GSK-3β and Cdk5 and forms paired helical filaments (PHFs), is the best-studied case (for review, see Goedert 1993; Billingsley and Kincaid 1997). The phosphorylation of Tau also appears to be involved in the asymmetric localization of Tau in neurons during axonogenesis. It has been reported that Tau is dephosphorylated predominantly in the nascent axon and phosphorylated in the somatodendrite at the TAU-1 site (Ser202, Thr205) (Binder et al. 1985; Papasozomenos and Binder 1987; Mandell and Banker 1996). However, the underlying mechanisms or physiological significance of these phenomena remain largely unknown. Rho-kinase preferentially phosphorylates the COOH-terminal AP region of Tau and MAP2, which interacts with microtubules. Among the sites phosphorylated by Rho-kinase, Ser409 of Tau is reported to be the site of phosphorylation by PKA (and probably the same for Ser1796 of MAP2). As far as we know, the sites of Thr245 and Thr377 in Tau have not yet been referred to. Thus, the patterns of phosphorylation by Rho-kinase are unique compared with those for other kinases. Because phosphorylation at Ser262 of Tau is reported to be crucial for the promotion of microtubule polymerization, we also examined whether Rho-kinase phosphorylates and myosin phosphatase dephosphorylates at this site. We found that Rho-kinase phosphorylated and myosin phosphatase dephosphorylated Tau at Ser262 to some extent, but exact stoichiometries were not determined. Phosphorylation of Tau at these sites resulted in a moderate decrease in the activity of Tau to promote microtubule assembly.

We here purified MBS-interacting proteins by GST-MBS-ANK affinity column chromatography and identified them as Tau and MAP2. Our data demonstrate that Tau and MAP2 are putative substrates for both Rho-kinase and myosin phosphatase, suggesting that the Rho signalling pathway regulates the phosphorylation-state of Tau and MAP2 through Rho-kinase and myosin phosphatase. The pattern of phosphorylation of Tau by Rho-kinase is different from that of PHF-Tau, the Tau-1 site, or by any known kinase. Because the direct evidence for Rho-kinase to be involved in the regulation of neurite extension in PC12 cells was not available, it is necessary to examine whether Tau is phosphorylated at Thr245/Thr377/Ser409 in a Rho/Rho-kinase signalling-dependent manner in neuronal cells. Interestingly, myosin phosphatase dephosphorylated Tau with varying efficacies for each site: it preferentially dephosphorylated Tau at Thr245 and Thr377 in comparison with Ser409 and Ser262, indicating the balance of kinase and phosphatase intricately affect the state of phosphorylation of Tau. It should be noted that myosin phosphatase, whose activity is modulated by Rho-kinase through the phosphorylation of MBS, may dephosphorylate Tau and MAP2 phosphorylated not only by Rho-kinase but also by other kinases. It is possible that Rho/Rho-kinase/myosin phosphatase pathway co-ordinately controls phosphorylation of Tau and MAP2 with other kinases. Conceivably, MAPs also might function as anchor molecules for myosin phosphatase. Several phosphatases have been found to associate with microtubules to date. Type 1 phosphatase (PP1) was found to interact with Tau (Liao et al. 1998), and Type 2A phosphatase (PP2A) with both Tau and microtubules (Sontag et al. 1995, 1999). Thus, a lot of kinases and phosphatases seem to elaborately tune the phosphorylation-state of Tau and MAP2 in various aspects of neuronal development. This raises the fundamental question of when or where Rho-kinase and MBS would play a role in the regulation of the functions of Tau and MAP2, downstream of Rho.

In non-neuronal cells, Rho family GTPases have been known to be implicated in microtubule dynamics, as well as microfilament organization (see reviews, Gundersen 2002; Small and Kaverina 2003). In NIH3T3 cells, Rho induces long-term stabilization of a subset of microtubules in the lamella through mDia, the other effector of Rho (Palazzo et al. 2001). Conversely, disruption of microtubules resulted in the activation of Rho (Bershadsky et al. 1996; Enomoto 1996; Liu et al. 1998; Pletjushkina et al. 1998), and targeting interaction of microtubules to focal adhesion leads to the suppression of Rho/Rho-kinase and to adhesion turnover (Kaverina et al. 1999). These observations seem to be contradictory with our results in which Rho and Rho-kinase would decrease polymerization of microtubules via Tau and MAP2 phosphorylation. This might be explained by the difference of cell types. In fact, several lines of evidence indicate that the neurite retraction occurs when Rho and Rho-kinase were activated in neuroblastoma cells (Amano et al. 1998; Hirose et al. 1998; Katoh et al. 1998), which was accompanied by the suppression of assembly of microtubules and intermediate filaments (Hirose et al. 1998). We also found that the injection of active Rho protein into Vero fibroblast cell resulted in the rapid disruption of microtubules (Watanabe et al. unpublished observations). Because the regulation of microtubules and intermediate filaments is thought to be necessary to accomplish the shape change or movement for the cells in response to the signals, it is conceivable that Rho and Rho-kinase modulate microtubules and intermediate filaments as well as microfilaments in certain cases. Our present model, that the activation of Rho-kinase, which phosphorylates both MBS and MAPs, reduces the activity of MAPs agrees with the observation that the Rho/Rho-kinase pathway prevents the neurite extension in N1E-115 cells. However, further studies are necessary to understand the relationship between Rho-kinase/MBS and MAPs, and ultimately that between the Rho signalling pathway and cytoskeletons including microfilaments, microtubules and intermediate filaments.

Acknowledgements

We thank Drs Y. Kanai and N. Hirokawa (University of Tokyo) for providing us with the cDNA encoding Tau, Dr N. Cowan (New York University) for providing the cDNA encoding MAP2 and Dr M. Nakafuku (University of Tokyo) for providing the cDNA encoding Notch and PC12 cells. We also thank Dr E. Nishida (Kyoto University) for helpful discussion. We are grateful to M. Maeda, A. Takemura, and T. Ishii for secretarial assistance. This work was supported by grants-in-aid for scientific research from the Ministry of Education, Culture, Science and Technology of Japan, by Research for the Future programme from the Japan Society for the Promotion of Science, by Special co-ordination funds for promoting Science and Technology, and by the organization for Pharmaceutical Safety and Research.

Ancillary