Brain Proline-directed Protein Kinase Phosphorylates Tau That Are ...

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 268, No. 31, Issue of November 5, p p . 23512-23516,1993 Printed in U. S. A.

Brain Proline-directed Protein Kinase Phosphorylates Tau on Sites That Are Abnormally Phosphorylated in Tau Associated with Alzheimer’s Paired Helical Filaments*

(Received for publication, April 16, 1993, and in revised form, June 7, 1993)

Hemant K. Paudel, John LewS, Zenobia AliS, and Jerry H. WangQ From the Medical Research Council Group in Signal Transduction, Department of Medical Biochemistry, Faculty of Medicine, University of Calgary, Calgary, Alberta T2N 4 N i , Canada

Brain proline-directed protein kinase (BPDK), which contains a catalytic subunit homologous to and displaying site-specific phosphorylation similar to ~ 3 4 ’ ~ ’ ~ kinase (Lew, J., Winkfein, R. J., Paudel, H. K., and Wang, J. H. (1992) J. Biol. Chem. 267, 25922- 26926), has been examined for possible involvement in tau phosphorylation. Immunoblot analyses using peptide antibodies specific for BPDK have revealed the presence of the kinase in bovine brain microtubules purified extensively by repeated polymerization and depolymerization cycles. When the microtubule proteins are separated into the tubulin and microtubule- associated protein fractions, BPDK is found exclu- sively in the latter fraction. BPDK phosphorylates both tau and MAP2, the former protein being phosphorylated to a stoichiometry of 3.8 mol of phosphate/mol of tau. Analysis of the phosphopeptides isolated from the tryptic digest of the phosphorylated bovine tau has revealed seven phosphorylation sites. Based on the sequence alignment between bovine and human tau proteins, these sites correspond to Ser-195, Ser-202, Thr-206, Thr-231, Ser-235, Ser-396, and Ser-404 of human tau. Mass spectrometric analysis of the tau protein isolated from Alzheimer’s paired helical filaments (PHFs) has determined three abnormal phosphorylation sites and two phosphopeptides containing a total of five abnormal phosphates (Hasegawa, M., Morishima-Kawashima, M., Takio, K., Suzuki, M., Ti- tani, K., and Ihara, Y. (1992) J. Biol. Chem. 267, 17047-17054). Two of the sites in tau phosphorylated by BPDK, Thr-231 and Ser-236, are among the abnormal phosphorylation sites, and the other sites phosphorylated by BPDK are within phosphopeptides from PHF-tau. These results suggest that BPDK may be one of the kinases responsible for the abnormal phosphorylation-associated PHF-tau.

Neurofibrillary tangles and amyloid plaques are the two prominent neuropathological hallmarks associated with Alz-

* This work was supported by grants from the Medical Research Council of Canada, National Cancer Institute of Canada, and Alz- heimer Society of Canada. 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.

$ Alberta Heritage Foundation Medical Studentship Awardee. Q Alberta Heritage Foundation Medical Research Scientist; to

whom correspondence should be addressed Dept. of Medical Bio- chemistry, University of Calgary, 3330 Hospital Drive, N. W., Cal- gary, Alberta T2N 4N1, Canada. Tel.: 403-220-3041; Fax: 403-283- 4740.

heimer’s disease (for reviews, see Refs. 1 and 2). The main structural components of neurofibrillary tangles are the paired helical filaments (PHFs)’ that have been shown to contain predominantly the microtubule-associated protein tau (3-9). Tau associated with PHF exists as a unique form, distinct from normal tau in being insoluble (8-lo), displaying retarded mobility on SDS-PAGE (3, 4, l l ) , and is abnormally phosphorylated (3,7, 11-15). Studies have indicated that abnormal phosphorylation of tau is a major factor in the transformation of normal tau to PHF-tau (1-3,7,11-15). Attempts have been made by various investigators to identify the protein kinases responsible for the tau phosphorylation in Alzheimer’s disease. In vitro, CAMP-dependent protein kinase (16-18), protein kinase C (16, 19), casein kinase I1 (16), calmodulin- dependent protein kinase I1 (16, 20), and glycogen synthase kinase-3 (21) phosphorylate tau. Calmodulin-dependent protein kinase I1 (16, 20), CAMP-dependent protein kinase (17, 18), and glycogen synthase kinase-3 (21) have been shown to cause retardation of tau on SDS-PAGE. Recently, Ishiguro et al. (22) have isolated two kinases, tau kinase I and tau kinase 11, from bovine brain microtubules. Although both these kinases phosphorylate tau, only tau kinase I causes a significant change in the mobility of tau on SDS-PAGE.

Various biochemical and immunological approaches have been taken to identify abnormal phosphorylation sites associated with PHF-tau (3, 23-27). Most notable is the recent mass spectrometric study that has revealed several abnormally phosphorylated sites within PHF-tau (23). These studies have indicated that most of the abnormal phosphorylation sites within PHF-tau have consensus sequence for proline- directed kinases (3, 23-27). Among proline-directed kinases, MAP-kinase has been shown to phosphorylate tau and con- vert it to a species with retarded mobility on SDS-PAGE (21, 28). The cell cycle regulator p34cdc28 from yeast (29) and ~ 3 4 ” ~ “ ’ kinase from starfish oocytes (28) and mouse FM3A cells (30) have also been reported to phosphorylate tau.

Recently, we have purified a novel brain proline-directed protein kinase (BPDK) from bovine brain (31). BPDK has two subunits with molecular masses of 33 and 25 kDa (31). The 33-kDa subunit is catalytic and is highly homologous to ~ 3 4 ~ ~ “ ’ kinase (32). The homologues of the 33-kDa subunit have been cloned in species such as rat (33) and human (34, 35) and are designated as neuronal cdc2-like kinase (331, PSSALRE kinase (34), and cdk5 (35). In this study we have

The abbreviations used are: PHF, paired helical filaments; PAGE, polyacrylamide gel electrophoresis; MOPS, 4-morpholinepropanesul- fonic acid PIPES, 1,4-piperazinediethanesulfonic acid HPLC, high performance liquid chromatography; PTH, phenylthiohydantoin; BPDK, brain proline-directed protein kinase; MAP, microtubule- associated protein; TFA, trifluoroacetic acid.

23512

Phosphorylation of T a u by Brain Proline-directed Kinase 23513

examined if BPDK phosphorylates tau. Herein, we report that BPDK is a microtubule-associated protein that phosphorylates both tau and MAPZ. Moreover, all the sites phosphorylated by PBDK in tau are among the phosphorylation sites of PHF-tau.

MATERIALS AND METHODS

Proteins and Peptides-Brain proline-directed kinase (BPDK) was purified from bovine brain as described (31). Tau protein was purified essentially as described (36). Microtubules were purified from bovine brain homogenate by three cycles of temperature-induced polymerization and depolymerization in the presence of 1 mM GTP. The final depolymerized fraction was placed in a boiling water bath for 4 min, and the heat-stable tau and MAP2 were purified from the supernatant of the boiled sample by phosphocellulose chromatography. Tau and MAP2 co-eluted, and the mixture of the two proteins was used to generate phosphorylation as shown in Fig. 3, A and B. Tau was separated from MAP2 by gel filtration of the above protein mixture on a Sepharose 4B column. Purified tau showed five major bands with apparent molecular masses ranging from 55 to 65 kDa on SDS- PAGE and was dialyzed against 50 mM MOPS (pH 7.2), 0.1 mM EDTA, and stored frozen at -70 "C until use. The ~34'~'' kinase was purified from mitotic human HeLa cells using ~ 1 3 ~ " " affinity chromatography as described (31). The MAP-kinase ( ~ 4 4 " ~ ~ ) , purified from sea star oocytes, and polyclonal antibody against synthetic peptide derived from residues 333-367 of the rat MAP-kinase (~43"") that recognizes MAP-kinases, p43"", p42""*, and ~ 4 4 " ~ ~ were gener- ous gifts from Dr. S. Pelech, University of British Columbia. Tryp- sin and bovine serum albumin were from Sigma. Synthetic peptides derived from amino (EKIGEGTYGVVYK) and carboxyl (NDLDNQIKKM) termini of HeLa cell ~ 3 4 ~ ' kinase, carboxyl terminus (FSDFCPP) of BPDK, and ~ 3 4 ' ~ ' kinase phosphorylation site of histone H1 (KTPKKAKKPKTPKKAKKL) were synthesized in an Applied Biosystem 431A peptide synthesizer using Fmoc (N-(9- fluoreny1)methoxycarbonyl) chemistry in the Medical Research Council, Signal Transduction Peptide Synthesis Core Facility, The University of Calgary. All the peptides derived from ~ 3 4 ' ~ ' kinase and BPDK contained an additional cysteine residue at the carboxyl termini. Peptides were purified by HPLC, and their purities were verified by their amino acid analyses. Peptides were coupled to keyhole limpet hemocyanin through the carboxyl-terminal cysteine residues, and the antibodies were raised against each peptide in New Zealand White rabbits and purified as described previously (31). All peptide antibodies were prepared in the Medical Research Council, Signal Transduction Antibody Facility, The University of Calgary.

Protein and Peptide Concentrations-Amount of BPDK is ex- pressed in activity units, and 1 unit of BPDK corresponds to the amount of kinase that transfers 1 pmol of phosphate to the synthetic peptide derived from histones in 1 min. The specific activity of the purified enzyme is -4000 units/pg (31). Concentration of tau protein is based on its absorption at 280 nm (37). All other protein concentrations were determined by Bradford protein assay (38) using bovine serum albumin as standard. Concentrations of all synthetic peptides are based on their amino acid analyses.

Phosphorylation of Tau and MAP2 by BPDK-Samples containing the mixture of tau and MAP2, or tau alone, were incubated at 30 "C in reaction mixtures that contained all the components of the phosphorylation except the kinase that was added to initiate the reaction. The final concentrations of the various components in the phosphorylation were 50 mM MOPS (pH 7.2), 0.1 mM EDTA, 10 mM Mg(CH3C00)2,0.2 mM [Y-~'P]ATP, 0.5 mg of tau/ml, and 4000 units BPDK/ml. When the mixture of tau and MAP2 was phosphorylated, the combined concentration of the two proteins in the assay was 0.5 mg/ml. After indicated time points, aliquots were taken out, mixed with an equal volume of SDS-PAGE sample buffer (100 mM Tris- HC1 (pH 8.0), 2% SDS, 0.1% bromphenol blue, 0.1 mM EDTA, 5% P-mercaptoethanol, and 25% glycerol), heated at 100 "C for 2 min, and electrophoresed. Gels were stained and destained, and the radioactive proteins were visualized by autoradiography of the gels.

Purification of Phosphpeptides-Tau (0.5 mg) was phosphorylated by BPDK for 120 min as described above except the concentration of tau was 1 mg/ml. Phosphorylated tau was desalted on a Sephadex G-25 column previously equilibrated in 25 mM NH4HCOs (pH 8.0), lyophilized, redissolved in 200 pl of 50 mM Tris-HC1 (pH 8.2) containing 50 pg of trypsin/ml, and incubated at 37 "C. After 15 h of incubation, the sample was acidified by adding 200 p1 of 20% acetic

acid and fractionated in a Sephadex G-25 column equilibrated in 10% acetic acid. An aliquot from each fraction was counted in a liquid scintillation counter to determine the amount of radioactivity. Radio- active fractions were pooled and lyophilized. The lyophilized sample was dissolved in 0.2 ml of 0.1% TFA and injected into a HPLC reverse phase column equilibrated in 0.1% TFA. The column was eluted with a linear gradient of 0-70% acetonitrile in 0.1% TFA. An aliquot from each fraction was counted in a liquid scintillation counter, and the radioactive fractions were pooled.

Peptide Sequencing-Peptides were covalently attached to Seque- lon-AA (Millipore) following manufacturer's instructions and se- quenced by Edman degradation in an Applied Biosystem 470 A Gas- phase Sequencer. PTH-derivatives released after each cycle were identified by a Beckman HPLC system using reverse phase Cla column (Alltech Econosphere cartridge, 5 pm, 46 X 250 mm). An aliquot from each cycle was counted in a liquid scintillation counter to determine the amount of radioactivity. Peptide sequencing was performed by Protein Sequencing Facility, The University of Calgary.

SDS-PAGE, Immunoblots, and Phosphoamino Acid Analysis- SDS-PAGE was carried out on 10% gels using the discontinuous system of Laemmli (39). Gels were stained with 0.2% Coomassie Brilliant Blue and destained with 20% methanol and 7% acetic acid. Autoradiography was performed using Kodak AR-5 films. Proteins were transferred to polyvinylidene difluoride membranes (Millipore) and immunoblotted with indicated antibody as described previously (31). Phosphoamino acid analyses were performed essentially as described (40).

RESULTS

Protein Kinases Associated with Microtubules-To demon- strate the association of protein kinases with microtubules, the purification procedure for tau as described under "Mate- rials and Methods" was modified. The microtubules in the brain homogenate were subjected to temperature-induced polymerization in the presence of l mM GTP and recovered as a pellet by ultracentrifugation. The pellet was resuspended in the buffer at 4 "C to depolymerize microtubules, and the insoluble materials were removed by ultracentrifugation. The supernatant-containing microtubules were subjected to two more cycles of polymerization and depolymerization. Micro- tubules purified by this procedure practically contain only tubulin and other microtubule-associated proteins (36, 41). An aliquot from the purified microtubule fraction was incubated with the phosphorylation mixture containing M?' and [y-32P]ATP as described under "Materials and Methods" except no exogenous kinase was added and the protein concentration of the mixture was 1 mg/ml. The result, which was analyzed by SDS-PAGE followed by autoradiography, indicated a high level of endogenous phosphorylation (data not shown). When microtubules were fractionated into tubulin and microtubule-associated proteins by phosphocellulose chromatography (see below), the endogenous protein phosphorylation activity was associated with the microtubule- associated protein fraction and was not detected in tubulin fraction (data not shown). These results indicated that protein kinase activity is tightly associated with microtubules.

To examine if BPDK is associated with microtubules, an aliquot of the purified microtubule fraction was electrophoresed and immunoblotted with the human ~34'~" ' kinase amino-terminal peptide antibody that recognizes BPDK (31). As shown in Fig. lA, an -33-kDa immunoreactive protein band, which comigrates with the catalytic subunit of BPDK, is present in both initial homogenate (lane 2 ) and purified microtubules ( l a n e 3 ) . There was no significant loss in the amount of this protein in each cycle of polymerization, depolymerization, and washing of microtubules during purification (data not shown). These results suggest that a 33-kDa protein similar to the catalytic subunit of BPDK is associated with microtubules. When a similar experiment was performed using MAP-kinase peptide antibody, two immunoreactive

23514 Phosphorylation of Tau by Brain Proline-directed Kinase 1 2 3 4 5

A BPDK-

1 2 3 B '"- - -

YAP-kinase-

FIG. 1. Immunoblots of various bovine brain fractions with peptide antibodies directed to the amino-terminal region of p34'*'* kinase and MAP-kinase. Unless otherwise stated, all procedures were performed at 4 "C; and all ultracentrifugations were carried out in a Beckman L8-80 ultracentrifuge using Ti-45 rotor a t lo6 X g for 30 min. Fresh bovine brain (200 g) was homogenized in 200 ml of homogenizing buffer (0.1 M PIPES (pH 6.4), 2 mM EGTA, 1 mM MgSOJ for 1 min. The homogenate was centrifuged at 16,000 X g in a Beckman J2-21M centrifuge using JA-10 rotor for 30 min. The supernatant was subjected to ultracentrifugation. The pellet was discarded, and the supernatant was diluted with an equal volume of 8 M glycerol in homogenizing buffer and called "initial homogenate." Initial homogenate was preincubated at 37 "C for 2 min, and then an aliquot from a stock GTP solution was added to bring the final GTP concentration to 1 mM. The microtubules in the initial homogenate were allowed to polymerize by incubating the entire solution at 37 "C for 30 min and recovered by ultracentrifugation a t 37 "C. The pellet was dispersed in cold homogenizing buffer (1/10 of the volume of the initial homogenate) and cooled on ice for 30 min to depolymerize and solubilize microtubules. The insoluble materials in the sample were removed by ultracentrifugation. The supernatant containing microtubules was subjected to two more cycles of temperature-induced polymerization and depolymerization as described above and resulted in pure microtubules. The purified, depolymerized microtubules were fractionated on a phosphocellulose column essentially as described (36) and resolved in two protein peaks, tubulin fraction and microtubule-associated protein fraction. Aliquots from various fractions were electrophoresed on 10% SDS-gels and immunoblotted with peptide antibodies. A, immunoblot with ~ 3 4 ' ~ ' ~ kinase amino-terminal peptide antibody. Lane 1, purified BPDK (400 units); lane 2, initial homogenate (30 pg); lane 3, microtubules (15 pg); lane 4, tubulin fraction (15 pg); and lane 5, microtubule-associated protein fraction (10 pg). B, immunoblot with MAP-kinase peptide antibody. Lane 1 , BPDK (4000 units); lane 2, MAP-kinase (~44"~') (50 ng); and lane 3, microtubules (15 pg).

bands were detected (Fig. lB, lane 3 ) . Because the size of the MAP-kinase ( ~ 4 4 ~ ~ ~ ) control (lane 2 ) is 44 kDa, the apparent molecular masses of the two immunoreactive bands are esti- mated to be =42 and 43 kDa. Previously, MAP-kinases, ~ 4 3 ~ ' ~ ' and have been demonstrated to be relatively abundant in rat brain and spinal cord (42-44). Therefore, these two immunoreactive bands are likely to be ~ 4 3 ~ ' ~ ' and p4Zerk2. Thus, consistent with a previous report (21), MAP-kinases also appear to be associated with microtubules.

The microtubule-associated proteins dissociate from tubulin upon depolymerization of microtubules, and the tubulin dimer can be separated from most of the associated proteins by phosphocellulose chromatography (36). To characterize further the 33-kDa microtubule-associated protein, the purified depolymerized microtubule fraction was chromato- graphed on a phosphocellulose column. Two protein peaks eluted from the column were designated as peak 1 and peak 2 in order of their elution from the column (data not shown). Peak 1 and peak 2 contained essentially tubulin and a mixture of MAP2, tau and other microtubule-associated proteins, respectively. An aliquot from each of the peaks was immunoblotted with the ~ 3 4 ' ~ ~ ' kinase amino-terminal peptide antibody. As shown in Fig. lA, the 33-kDa protein band is undetectable in peak 1 (lane 4 ) and elutes with microtubule-

associated proteins including MAP2 and tau in peak 2 (lane 5).

To test further the suggestion that the 33-kDa microtubule- associated protein is BPDK, aliquots from purified microtubule fraction were electrophoresed and immunoblotted with three different antibodies that recognize distinct segments of BPDK and human HeLa cell ~ 3 4 ' ~ ~ ~ kinase. Previous study has shown that the peptide antibody directed against the amino-terminal region of human ~ 3 4 ' ~ ~ ' kinase displays cross- reactivity toward BPDK, whereas one directed toward the carboxyl-terminal region of ~ 3 4 ~ ~ ' ' kinase does not react (31). An additional antibody, which we have recently raised against the peptide derived from the carboxyl-terminal region of the 33-kDa subunit of BPDK, reacts with BPDK on immunoblot, but not with the human ~ 3 4 ' ~ ' ~ kinase (data not shown). As shown in Fig. 2, the 33-kDa protein cross-reacts with peptide antibody directed against the amino terminus of ~34'~'' kinase (panel A, lane 3 ) and the carboxyl termini of BPDK (panel C, lane 3 ) and is not recognized by the peptide antibody directed to the carboxyl terminus of ~ 3 4 ~ ~ ' ~ kinase (panel B, lane 3 ) . Thus, the 33-kDa protein displays immunoreactivities toward the three peptide antibodies, characteristic of BPDK. Based on these results, we concluded that the 33-kDa microtubule-associated protein is BPDK.

Phosphorylation of Tau and MAP2 by BPDK-Tau and MAP2 are the two major microtuble-associated proteins (41). We, therefore, examined if these two proteins are phosphorylated by BPDK. An aliquot from the mixture of tau and MAP2, purified by heat denaturation of microtubules followed by phosphocellulose chromatography as described under "Mate- rials and Methods," was incubated with purified BPDK in the presence of Mg+, [r3*-P]ATP, and other components of the phosphorylation reaction. After indicated time points, the phosphorylation was analyzed by SDS-PAGE, followed by autoradiography of the phosphorylation mixture. As shown in Fig. 3B, lane 3, both tau and MAP2 are phosphorylated by BPDK. To examine further the phosphorylation of tau in detail, we performed a similar phosphorylation experiment except the mixture of tau and MAP2 was replaced by purified tau. As shown in Fig. 3C, with increasing periods of time a progressive increase in the amount of radioactivity incorporated into tau is observed (lanes 3-7). Both control lanes, 1 and 2, containing BPDK and tau, respectively, do not show any radioactivity even after incubation for 90 min. When the phosphorylation of tau by BPDK was carried out in the presence of 0.2 mM peptide derived from histone, an excellent

A B C

1 2 3 1 2 3 1 2 3

pmmq

FIG. 2. Immunoblots of microtubules with peptide antibodies directed against amino and carboxyl termini of ~ 3 4 ~ ' ~ kinase and BPDK. Purified microtubules were immunoblotted with various peptide antibodies as described under "Materials and Meth- ods." A, immunoblot with p34'd'Z amino-terminal peptide antibody that recognizes both 34'd'Z kinase and BPDK. B, immunoblot with ~ 3 4 ~ ' ' carboyxl-terminal peptide antibody that recognizes ~34'~'' kinase and does not cross-react with BPDK. C, immunoblot with BPDK carboxyl-terminal peptide antibody that recognizes BPDK and does not cross-react with ~ 3 4 ~ ~ ~ ' kinase. Lane 1 , human HeLa cell ~34'~'' kinase (50 ng); lane 2, BPDK (400 units); and lane 3, microtubules (10 pg).

Phosphorylation of Tau by Brain Proline-directed Kinase 23515

A B 1 2 3 1 2 3

MAPP+ “

Tau+ [ Bg

C

1 2 3 4 5 6 7

Tau+ [ I

FIG. 3. Phosphorylation of MAP2 and tau by BPDK. Phos- phorylations of tau alone and the mixture of MAP2 and tau were carried out as described under “Materials and Methods” except Mg(CH3C00)2 was replaced by MgC& in the tau phosphorylation mixture. After the indicated time, reactions were terminated, and the phosphorylation was analyzed by SDS-PAGE followed by autoradiography. A, SDS-gel of the phosphorylation of the mixture of MAP2 and tau. Lune 1, BPDK control (40 units) incubated in the phosphorylation mixture for 15 min; lane 2, mixture of MAP2 and tau control (5 pg) incubated in phosphorylation mixture for 15 min; and lane 3, mixture of MAP2 and tau (5 pg) + BPDK (40 units) incubated in phosphorylation mixture for 15 min. B, autoradiography of A. C, autoradiography of the SDS-gel of tau phosphorylation. Lane 1, BPDK control (40 units) incubated in phosphorylation mixture for 90 min; lane 2, tau control (5 pg) incubated in phosphorylation mixture for 90 min; and lanes 3-7, tau (5 pg each) removed from phosphorylation mixture containing tau + BPDK after 5, 15, 30, 60, and 90 min, respectively.

substrate of BPDK (31), the phosphorylation was completely suppressed (data not shown). These results indicated that tau is phosphorylated by BPDK. However, MAP-kinases, which have very similar substrate specificity as ~ 3 4 ‘ ~ ‘ ~ kinase (re- viewed in Ref. 42) and are quite abundant in mammalian brain (43, 441, have been reported to phosphorylate tau (21, 28). Therefore, it became important to examine whether our kinase preparation was contaminated with any MAP-kinases. ==4000 units of BPDK were electrophoresed and immunoblotted with MAP-kinase peptide antibody. As shown in Fig. lB, the antibodies detected 50 ng of MAP-kinase as an intense band in the control lane (lane 2) and completely failed to show any immunoreactivity in the lane containing BPDK (lane 1 ), indicating that our BPDK preparation is free from ~ 4 3 ” ~ ’ and p42erk2, the two most abundant MAP-kinases in mammalian brain (43,44). From these observations, we concluded that the kinase in our preparation that phosphorylates tau and MAP2 is indeed BPDK.

Tau was phosphorylated by BPDK under standard conditions for 2 h, except ATP concentration was 1.5 mM. The radioactivity in tau was quantitated by filter paper assay (45). The results indicated that 3.8 mol of phosphate/mol of tau were incorporated (data not shown).

Determination of Phosphorylation Sites-Tau (0.5 mg) was phosphorylated by BPDK for 2 h as described under “Mate- rials and Methods. Phosphorylated tau was desalted and digested with trypsin. The tryptic digest was fractionated by a Sephadex G-25 column. As shown in Fig. 4, inset, only one radioactive peak was eluted from the column. The radioactive fractions were pooled, lyophilized to remove acetic acid, redissolved in 200 pl of 0.1% TFA, and fractionated over a reverse phase Cl8 HPLC column previously equilibrated in

E n 4 0

n ” 0 10 20 3 0 4 0 5 0 8 0

Effluent (ml)

FIG. 4. HPLC of the tryptic digest of tau phosphorylated by BPDK. Tau was phosphorylated by BPDK and digested with trypsin as described under “Materials and Methods.” The digest was fractionated in a Sephadex G-25 column (1 X 36.5 cm), equlibrated, and eluted with 10% acetic acid at a flow rate of 0.25 ml/min. Fractions (0.5 ml) were collected, and 10 pl from each fraction were counted in a liquid scintillation counter to determine the amount of radioactivity. Radioactive fractions were pooled, redissolved in 200 pl of 0.1% TFA, and injected into a Waters Vydac CI8 reverse phase column (catalog no. 218 TP 54) equilibrated in 0.1% TFA. The column was eluted with a linear gradient of 0-70% acetonitrile in 0.1% TFA at a flow rate of 1 ml/min. Fractions (1 ml) were collected, and 10 p1 from each fraction were counted in a liquid scintillation counter to determine the amount of radioactivity. Inset, gel filtration profile of the tryptic digest of tau phosphoryated by BPDK.

Effluent (ml)

FIG. 5. Purification of phosphopeptides 1.2, and 3. Pools, I and I1 from Fig. 4, were lyophilized, redissolved in 0.1 ml of 0.1% TFA, and rechromatographed by HPLC. Chromatographic conditions were the same as described in the legend to Fig. 4, except the acetonitrile gradient was 0-60% in 60 min. A and C, HPLC profile of pool I and pool 11, respectively. B and D, distribution of radioactivity in fractions from A and C, respectively.

0.1% TFA. The column was eluted with a linear gradient of acetonitrile (0-70%) in 0.1% TFA in 60 min. As shown in Fig. 4, two major and one minor radioactive peaks were resolved. The radioactive fractions in each peak were pooled. Of the total radioactivity injected into the HPLC column, 53,32, and 3% of activities were recovered in pools I, 11, and 111, respectively. Each pool was lyophilized, redissolved in 100 pl of 0.1% TFA, and rechromatographed by HPLC as above, except the acetonitrile gradient was changed from 0 to 60% in 60 min. Pool I resolved into three peaks (Fig. 5A), and only one contained radioactivity (Fig. 5B). The radioactive peak was designated phosphopeptide 1. Similarly, rechromatography of pool I1 resulted in two radioactive peptides, phosphopeptides 2 and 3 that eluted first and later from the column, respectively (Fig. 5, C and D). When pool I11 was fractionated similarly, the radioactivity dispersed into many fractions, and no clear radioactive peptide was obtained (data not shown). Thus, the three phosphopeptides, 1, 2, and 3, recovered from the tryptic digest of phosphorylated tau, were selected for further investigation. All three phosphopeptides were sub-

23516 Phosphorylation of Tau by Brain Proline-directed Kinase TABLE I

Sequence determination of tryptic phosphopeptides derived from tau The amino acid sequence of each phosphopeptide was determined as described under "Materials and Methods." Yield, indicates pmol of

PTH-amino acid derivative released after each cycle; X , represents the residue whose PTH-derivative could not be identified. The amount of radioactivity released in each cycle was quantitated in a liquid scintillation counter.

Phosphopeptide 1 Phosphopeptide 2 Phosphopeptide 3

Residue Yield CPm Residue Yield CPm Residue Yield cpm

1 Thr 71 8170 X 5823 X 5315 2 Pro 57 2195 Pro 93 1070 Gly 13 795 3 Pro 68 1115 Val 121 825 TYr 9 390 4 LYS 55 1085 Val 151 700 Ser 6 415 5 X 58,005 Ser 54 905 Ser 5 404 6 Pro 46 21,465 GlY 62 610 Pro 7 415 7 Ser 27 5760 ASP 57 1785 Gly 8 345 8 Ala 50 2430 Thr 51 1390 X 1565 9 Ala 52 1030 X 8235 Pro 3 810

10 LYS 19 695 Pro 43 4835 Gly 4 455 11 -4% 82 2525 X 875

Cycle

12 Pro 3 525 13 GlY 4 410 14 Ser 2 270 15 Arg 7 220

jected to 15 cycles of Edman degradation in a gas-phase amino acid Sequencer. An aliquot from each cycle was counted in a liquid scintillation counter to determine the amount of radioactivity released in each cycle.

As shown in Table I, the amino acid sequence of phosphopeptide I is TPPKXPSAAK. The fifth cycle released a very high amount of radioactivity, and the PTH-derivative in this cycle (indicated by X ) could not be identified. Thus, the 5th residue must be the phosphorylated amino acid whose PTH- derivative could not be identified by sequencer. Based on these observations and the published amino acid sequence of the longest isoform of bovine tau (46), Ser-242 is identified as the phosphorylation site within tau. The phosphoamino acid analysis of peptide 1 revealed that it contained both phosphoserine and phosphothreonine (data not shown). These results indicated that the 1st residue, the only threo- nine residue within this peptide, was also phosphorylated. However, the amount of radioactivity released in the first cycle is only ~ 1 4 % of the fifth cycle (Table I). Moreover, PTH-Thr was quantitatively identified as the 1st residue of this peptide. Based on these observations, we concluded that the 1st Thr residue in peptide 1 was partially phosphorylated. Thus, corresponding Thr-238 in bovine tau is also phosphorylated by BPDK.

The amino acid sequence of phosphoprotein 2, determined by the sequencer, is XPVVSGDTXPR (Table I). The first cycle contained several contaminant amino acids and released PTH-Gly, PTH-Ser, PTH-Pro, and PTH-Asp. Therefore, we could not assign any of these residues as the 1st residue of the peptide 2 with certainty. No PTH-derivative was detected in the ninth cycle, and both first and ninth cycles released a significant amount of radioactivity during Edman degradation (Table I). Based on these results we concluded that phosphopeptide 2 is phosphorylated on the 1st and 9th residues. Thus, phosphopeptide 2 extends from residues 403 to 413 of the tau sequence (46), and the phosphorylation sites are Ser-403 and Ser-411.

The amino acid sequence of phosphopeptide 3 determined is shown in Table I and contains 3 phosphorylated residues. This peptide extends from residues 202 to 216 in the bovine tau sequence (46), and the phosphorylation sites are Ser-202, Ser-209, and Thr-212. Thus, seven sites in total, Ser-202, Ser- 209, Thr-212, Thr-238, Ser-242, Ser-403, and Ser-411, within tau are phosphorylated by BPDK.

DISCUSSION

Although neurons are terminally differentiated and largely nonproliferating cells, the cell cycle regulator ~34'~'' kinase- like activity has been implicated in certain neuronal functions. Two neurofilament proteins, NF-H and NF-M, are heavily phosphorylated in axons at specific proline-directed sites resulting in SDS-PAGE mobility shift (47). Similar shifts can be produced by the in vitro phosphorylation of these proteins by p34'"' kinase (48). Similarly, tau is phosphorylated by yeast p34cdc28 kinase (29) and ~ 3 4 " ~ " ' kinase from starfish oocytes (28) and mouse FM3A cells (30). The sites within tau that are phosphorylated by ~34'~'' kinase from mouse FM3A cells have been identified to be Ser-202, Thr-205, Thr-231, and Ser-235 (30). These sites are among the abnormal phosphorylation sites found in PHF-tau (3, 23-27). In contrast, it has been shown that the expression of ~ 3 4 ' " ~ kinase is down- regulated during the differentiation of neuronal cells, and neither the mRNA nor the protein corresponding to ~34'~'' is present in detectable amounts in adult rat brain (49). These observations suggest that a cdc2-like kinase, rather than ~34'~" ' kinase itself, is responsible for the in vivo phosphorylation of neuronal-specific proteins at cdc2 kinase recognition sites.

The present as well as previous studies provide strong support for the notion that BPDK is, at least, partly responsible for the observed cdc2-like phosphorylation in neurons. The 33-kDa catalytic subunit of BPDK shows a high degree of amino acid sequence homology to ~34'~'' kinase displaying 58% amino acid identity (32). I n vitro, the purified BPDK phosphorylates the neurofilament proteins NF-M, NF-H (32), as well as tau protein (this study) at similar or identical sites as those phosphorylated by ~ 3 4 " ~ " kinase. Studies based on synthetic peptides have revealed a similar substrate specificity determinant for both kinases (52). Thus, these two kinases are structurally and functionally homologous. However, ~ 3 4 " ~ " ' kinase is enriched in proliferating cells and exists in negligible amounts in neurons (49), whereas BPDK appears to have a wider distribution being present in highest amounts in brain (33, 34). The co-existence and co-purification of BPDK and microtubule-associated proteins are further support for the functional role of BPDK in the phosphorylation of neuronal microtubule-associated proteins.

A number of protein kinases, including CAMP-dependent protein kinase (17, 18) and calmodulin-dependent protein

Phosphorylation of T a u by Brain Proline-directed Kinase 23517

kinase I1 (16, 20), have been shown to catalyze the in uitro phosphorylation of tau. Despite the observations that both CAMP-dependent protein kinase and calmodulin-dependent protein kinase I1 phosphorylate and shift the mobility of tau on SDS-PAGE, they do not appear to phosphorylate those sites that are characteristically phosphorylated in PHF-tau (23). A great majority of phosphorylation sites in PHF-tau, as determined by immunological methods (3,24-27) and mass spectrometric analysis (23), display the motif of the proline- directed phosphorylation site. Among proline-directed kinases, both MAP-kinases (21, 28) and ~ 3 4 " ~ " ' kinase (28- 30) have been shown to phosphorylate tau. On the basis of the preceding discussion and results of this study, BPDK is the most likely in uiuo kinase that catalyzes the cdc2-like phosphorylation of tau protein.

In this study, we have identified seven sites within bovine tau that are phosphorylated by BPDK in uitro. When the sequences of bovine (46) and human tau (50) are aligned, the BPDK phosphorylation sites correspond to Ser-195, Ser-202, Thr-205, Thr-231, Ser-235, Ser-396, and Ser-404 residues of the longest isoform of human tau (50). All of these sites appear to be phosphorylated in the tau protein isolated from PHFs (3, 23-27). The mass spectrometric analysis of PHF- tau has revealed three abnormal phosphorylation sites, Thr- 231, Ser-235, and Ser-262, as well as two phosphopeptides containing abnormal phosphates: peptides 191-225 and 386- 438 (23). Immunological studies have also indicated the presence of abnormal phosphorylations within the latter two segments of PHF-tau (3,24-27). The seven sites phosphorylated by BPDK include two of the identified sites, Thr-231 and Ser-235; three sites in peptide-(191-225), Ser-195, Ser- 202, and Thr-205; and two in peptide-(386-438), Ser-396 and Ser-404. These observations and the apparent association of BPDK with microtubules suggest that BPDK is one of the kinases that, in uiuo, target sites abnormally phosphorylated in PHF-tau.

One of the characteristics of PHF-tau is its observed retarded electrophoretic mobility compared with that of normal tau on SDS-PAGE (3,4, 11). This has been attributed to the abnormal phosphorylations associated with PHF-tau (1, 3, 7, 11-15). In this study, we have found that BPDK, which specifically phosphorylates almost all the sites that have been identified as being associated with PHF-tau (23), causes no apparent retardation in the electrophoretic mobility of tau (data not shown). Consistent with these observations, ~ 3 4 ~ ~ ~ ' kinase from starfish oocyte (28) and yeast ~ 3 4 ' ~ ' ' ~ kinase (29) have been reported not to change the SDS-PAGE mobility of tau. In contrast, phosphorylation of recombinant tau by ~34'~'' kinase from mouse FM3A cells causes mobility shift on SDS-PAGE (30). The cause of this discrepancy is not clear, but there are a number of explanations. For example, tau isolated from a natural source has been reported to contain a significant amount of endogenous phosphates that differ depending upon the purification procedure, source, and age of the animal (51). Therefore the tau, isolated from natural source, may not be very sensitive to change in the SDS-PAGE mobility by phosphorylation by BPDK. Alternatively, it is possible that mobility shift in tau may require simultaneous phosphorylation of several proline-directed sites in a tau molecule. Such a situation may not be attained in the in uitro phosphorylation of tau by BPDK. It should be noted that BPDK phosphorylates tau on seven sites to an average stoichiometry of 3.8 mol of phosphate/mol of tau, and, therefore, phosphorylated tau is expected to contain mixtures of incom- pletely phosphorylated tau proteins.

Two protein kinases capable of phosphorylating tau have

been purified recently from the microtubule protein fraction, designated as tau kinase I and tau kinase I1 (22). Several lines of evidence suggest that tau kinase I1 and BPDK either are identical or very similar proteins. Similar to BPDK that is composed of 33- and 25-kDa subunits, tau kinase I1 has been shown to contain 30- and 23-kDa subunits (22). Both kinases catalyze the phosphorylation of synthetic peptides at proline- directed sites and associate with microtubules. The identity of tau kinase I, on the other hand, is yet to be established. It appears to possess some attributes suggesting that it may belong to the MAP-kinase family. Its molecular mass on SDS- PAGE, 45 kDa, is similar to those reported for four MAP- kinases (42-44). Like MAP-kinase (28), it phosphorylates tau resulting in SDS-PAGE mobility shift (22). Moreover, MAP- kinases ~ 4 3 " ~ ' and ~ 4 2 ~ ' ~ ' are relatively abundant in brain (43, 44) and can be detected by Western immunoblot using MAP-kinase-specific antibody (Fig. lB, lane 3) (21). However, MAP-kinases phosphorylate synthetic peptides containing the phosphorylation site motif PXS/TP (42), but tau kinase I has been reported to have no activity toward peptides derived from tau containing such motif (22) suggesting that these two kinases may be different. The (Y and /3 isoforms of glycogen synthase kinase-3 with molecular masses of 45 and 51 kDa, respectively, have also been shown to co-purify with microtubules, phosphorylate tau, and cause SDS-PAGE mobility shift (21). Thus, tau kinase I may be an isoform of either MAP-kinase or glycogen synthase kinase-3 or an entirely different kinase. Further studies will be necessary to resolve this issue.

Acknowledgments-We thank Katherine Beaudette for her help in MAP-kinase immunoblots and Erwin Wirch for excellent technical assistance.

REFERENCES

2. Lee, V. M.-Y., andTrojanowski, J. Q. (1992) Curr. Opin. Neurobiol. 2,653- 1. Kosik, K. S. (1991) Trends Neurosci. 1 4 , 218-219

fi5fi 3. Lee:'G. M.-Y., Balin, B. J., Otvos, I., Jr., and Trojanowski, J. Q. (1991)

4. Greenherg, S. G., and Davies, P. (1990) Proc. Natl. Acad. Sci. U.S.A. 8 7 , Science 26 1,675-678

5. Wischik, C. M., Novak, M., Thagersen, H. C., Edwards, P. C., Runswick, 5827-5831

M. J., Jakes, R., Walker, J. E., Milstein, C., Roth, M., and Klug, A. (1988) Proc. Natl. Acad. Sci. U. S. A. 85,4506-4510

6. Ksiezak-Reding, H., and Yen, S:H. (1991) Neuron 6 , 717-728 7. Grundke-Iqbal, I., Iqhal, K., Tung, Y.-C., Quinlan, M., Wisniewski, H. M.,

8. Ksiezak-Reding, H., and Yen, S.-H. (1991) Neuron 6,717-728 9. Kondo, J., Honda, T., Mori, H., Hamada, Y., Miura, R., Ogawara, M., and

and Binder, L. I. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,4913-4917

Ihara, Y. (1988) Neuron 1,827-834 10. Selkoe, D. J., Ihara, Y., and Salazar, F. J. (1982) Science 2 1 6 , 1243-1245 11. Flament, S., Delacourte, A,, He'mon, B., and De'fossez, A. (1989) J. Neurol.

12. Ksiezak-Reding, H., Binder, L., and Yen, S.-H. (1990) J. Neurosci. Res.

13. Vincent, I., and Davies, P. (1990) Brain Res. 6 3 1 , 127-135 14. Greenberg, S. G., Davies, P., Schein, J. D., and Binder, L. I. (1992) J. Bid.

Chem. 267,564-569 15. Grundke-Iqhal, I., and Iqhal, K. (1986) Proc. Natl. Acad. Sci. U.S.A. 8 3 ,

4913-4917 16. Steiner, B., Mandelkow, E."., Biernat, J., Gustke, N., Meyer, H. E.,

Schmidt, B., Mieskes, G., Soiling, H. D., Drechsel, D., Kirschner, M. W.,

17. Litersky, J. M., and Johnson, G. V. W. (1992) J. Biol. Chem. 267 , 1563- Goedert, M., and Mandelkow, E. (1990) EMBO J. 9,3539-3544

Sci. 92,133-141

26,420-430

1 5 f i R 18.

19.

20. 21.

22.

23.

24.

25.

Scott, C. W., Spreen, R. C., Herman, J. L., Chow, F. P., Davison, M. D.,

Hoshi, M., Nishida, E., Miyata, Y., Sakai, H., Miyoshi, T., Ogawara, H.,

Baudier, J., and Cole, R. D. (1987) J. Biol. Chem. 2 6 2 , 17577-17583 Mandelkow, E."., Drewes, G., Biernat, J., Gustke, N., Lint, J. V., Van-

denheede, J. R., and Mandelkow, E. (1992) FEES Lett. 314,315-321 Ishiguro, K., Takamatsu, M., Tomizawa, K., Omori, A,, Takahashi, M.,

Arioka, M., Uchida, T., and Imahori, K. (1992) J. Biol. Chem. 267 ,

Hasegawa, M., Morishima-Kawashima, M., Takio, K., Suzuki, M., Titani, 10897-10901

Biernat, J., Mandelkow, E."., Schroter, C., Lichtenberg-Kraag, B., Stei- K., and Ihara, Y. (1992) J. Biol. Chem. 267,17047-17054

ner, B., Berling, B., Meyer, H., Mercken, M., Vandermeeren, A,, Goedert, M., and Mandelkow, E. (1992) EMBO J. 11, 1593-1597

Kanemaru, K., Takio, K., Miura, R., Titani, K., and Ihara, Y. (1992) J. Neurochern. 5 8 , 1667-1675

"_- Young, J., and Caputo, C. B. (1992) J. Biol. Chem. 268,1166-1173

and Akiyama, T. (1987) FEES Lett. 2 1 7 , 237-241

23518 Phosphorylation of Tau by Brain Proline-directed Kinase 26. Kosik, K. S., Orrechio, L. D., Binder, L., Trojanowski, J. Q., Lee, V. M.-Y.,

and Lee, G. (1988) Neuron 1,817-825 27. Lichtenherg-Kraag, B., Mandelkow, E."., Biernat, J., Steiner, B., Schro-

ter, C., Gustke, N., Meyer, H. E., and Mandelkow, E. (1992) Proc. Natl. Acad. Sci. U.S.A. 89,5384-5388

28. Drewes, G., Lichtenberg-Kraag, B., Doring, F., Mandelkow, E."., Biernat, J.. Goris. J.. Dor'ee, M.. and Mandelkow, E. (1992) EMBO J. 11,2131- . , 2138

. .

29. Mawal-Dewan, M., Sen, P. C., Abdel-Ghany, M., Shalloway, D., andRacker,

30. Vulliet, R., Halloran, S. M., Braun, R. K., Smith, A. J., and Lee, G. (1992)

31. Lew, J., Beaudette, K., Litwin, C. M. E., and Wang, J. H. (1992) J. Biol.

E. (1992) J. Biol. Chem. 267, 19705-19709

J. Biol. Chem. 267,22570-22574

Chem. 267.13383-13390 32. Lew. J.. Winkfein.R. J.. Paudel, H. K.. and Wan% J. H. (1992) J. Biol.

Chm. 267,25922-25926

Acad. Sci. U.S.A. 89. 10867-10871

-.

33. Hellmich, M. R., Pant, H. C., Wada, E., and Battey, J. F. (1992) Proc. Natl.

34. Mayerson, M., Enders,.G. H., Wu, C.-L., Su, L. K., Gorka, C., Nelson, C.,

35. Xiong, Y., Zhang, H., and Beach, D. (1992) Cell 71, 505-514 36. Herzog, W., and Weber, K. (1978) Eur. J. Biochem. 92, 1-7 37. Cleveland, D. W., Hwo, S.-Y., and Kirschner, M. W. (1977) J. Mol. Biol.

Harlow, E. M., and Tsai, L.-H. (1992) EMBO J. 11,2909-2917

116,227-247

38.

40. 39.

41. 42. 43.

44. 45.

46.

47.

48.

49.

50.

51.

52.

Bradford, M. M. (1976) Anal. Biochem. 72,248-254 Laemmli, U. K. (1970) Nature 227,680-685 Hunter, T., and Sefton, B. (1980) Proc. Natl. Acad. Sci. U. S. A. 77,1311-

Matus, A. (1988) Annu. Reu. Neurosci. 11,29-44

Boulton, T. G., Yancopoulos, G. D., Gregory, J. S., Slaughter, C., Moomaw, Pelech, S. L., and Sanghera, J. S. (1992) Trends Biochem. Sci. 17,233-238

Boulton, T. G., and Cohbs, M. C. (1991) Cell Regul. 2,357-371 Reimann, E. M., Walsh, D. A,, and Krebs, E. G. (1971) J. Biol. Chem. 246,

Himmler, A., Drechsel, D., Kirschner, M. W., and Martin, D. W., Jr. (1989)

Nixon, R. A,, Brown, B. A., and Marotta, C. A. (1982) J. Cell Biol. 94,150-

Hisanaga, S., Kusuhata, M., Okumura, E., and Kishimoto, T. (1991) J.

Haves. T. E.. Valtz, N. L. M., and McKay, R. D. G. (1991) New Biol. 3,

1315

C., Hsu, J., and Cohb, M. H. (1990) Science 249,64-67

1986-1995

Mol. Cell. Biol. 9, 1381-1388

158

Biol. Chem. 266, 21798-21803

2591269

R. A. (1989) Neuron 3,519-526

S. (1990) J. Neurosci. 10, 3295-3304

20830

Goedert, M., Spillantini, M. G., Jakes, R., Rutherford, D., and Crowther,

Ueda, K., Masliah, E., Saitoh, T., Bakalis, S. L., Scoble, H., and Kosik, K.

Beaudette, K., Lew, J., and Wang, J. H. (1993) J. Biol. Chem. 268,20825-

Brain Proline-directed Protein Kinase Phosphorylates Tau That Are ...

Documents

Transcript of Brain Proline-directed Protein Kinase Phosphorylates Tau That Are ...