A discrete repeated sequence defines a tubulin binding domain on microtubule-associated protein tau

ARCHIVES

OF BIOCHEMISTRY

AND BIOPHYSICS

Vol. 275, No. 2, December, pp. 56%5’79,1989

A Discrete Repeated Sequence Defines a Tubulin Binding Domain on Microtubule-Associated Protein Tau RICARDO B. MACCIONI,*~~~‘JUAN C. VERA,*,? JORGE DOMINGUEZ,* AND JESUS AVILAt$ *University

of Colorado Health Sciences Center, B-121, Denver, Colorado 80262; tlnternatimal Biology (ICC), Casilla 70111, Santiago 7, Chile; and +Centro de Biologia Molecular, Universidad Au&ma, 28049 Madrid, Spain

Center

for Cancer and Developmental

Received May 10,1989, and in revised form August ‘7,1989

The protein domain responsible for the interaction of tau with tubulin has been identified. Biophysical studies indicated that the synthetic peptide Va1187-Gly204(VRSKIGSTENLKHQPGGG) from the repetitive sequence on tau binds to two sites on the tubulin heterodimer and to one site on each of the microtubule-associated protein-interacting C-terminal tubulin peptides a(430-441) and p(422-434). The binding data showed a relatively stronger interaction of VaF- Glyzo4with ,8(422-434) as compared to that with ~~(430-441).The interaction of this tau peptide with either o( or p tubulin peptides appears to be associated with conformational changes in both the tau and the tubulin peptides. The 0 tubulin peptide also appears to induce a structural change of tau fragment Va121s-Gly235. Interestingly, tau peptides Va1187-Gly204 and Va1218-Gly235 induced tubulin self-assembly in a cold-reversible fashion, and incorporated into the assembled polymers. The specificity of the interaction of the tau peptide was supported by the competition of tau protein for the interaction with the tubulin polymer. In addition, the tau peptide appears to contain the principal antigenic determinant(s) recognized by antiidiotypic antibodies that react with the tubulin binding domains on microtubule-associated proteins. The present findings together with the demonstration of the presence of multiple sites for the binding of the a(430-441) and p(422-434) tubulin fragments to tau, and the existence of repetitive sequences on tau, strongly support the hypothesis that the region of tau defined by the repetitive sequences is involved in its interaction with tubulin. o 1989 Academic PRSS. IIK. ples of heterologous interacting systems of relevance in the control of microtubule assembly. MAPS consist of high-molecularweight components, namely MAP-l, MAP2, and MAP-4 (1,2), and a set of intermediate-molecular-weight proteins including tau (3, 4) and the immature component “small MAP-2” (5). These proteins induce tubulin assembly and provide stabilization to the microtubular structure (6-8). The diversity of cellular functions of microtubules requires that their assembly and organization be finely regulated by specific molecular signals. The selective interaction of tubulin with MAPS appears to be

Different lines of evidence point to the importance of protein-protein interactions in controlling a wide variety of cellular processes. The interaction of tubulin with microtubule-associated proteins (MAPs)~ is one of the most striking exam’ To whom correspondence should be addressed at: International Center for Cancer and Developmental Biology (ICC), Casilla 70111, Santiago 7, Chile. 2 Abbreviations used: MAPS, microtubule-associated proteins; Mes, 4-morpholineethanesulfonic acid; EGTA, ethylene glycol bis(P-aminoethyl ether) N,N’tetraacetic acid; SDS, sodium dodecyl sulfate; TFA, trifluoroacetic acid; ELISA, enzyme-linked immunosorbent assay. 0003-9861/89 $3.00 Copyright All rights

0 1989 by Academic Press, Inc. of reproduction in any form reserved.

568

TUBULIN

BINDING

DOMAIN

modulated by the differential affinity of the various MAP components with either (Yor fl tubulin subunits (9, 10). The nature of the enzymatic phosphorylation of both tubulin and MAPS (7, 11, 12) may actively modulate tubulin-MAP interactions. After the initial findings that a 4-kDa domain from the C-terminal moiety of tubulin subunits plays a major role in regulating microtubule assembly (13,14) and is involved in the binding of MAPS (Xi), significant efforts have been directed at examining the structural, genetic, and molecular biological implications of the regulatory features of the C-terminal tubulin domains. The current information indicates that the extreme C-terminal moieties of tubulin subunits define tubulin isotypes found in several vertebrate species (16, 17). However, these tubulin moieties do not appear to be involved in the interaction of MAPS (9,18). On the basis of binding experiments of synthetic tubulin peptides and immunological studies, we have recently shown that the tubulin subdomains directly involved in the selective interaction of MAPS are defined by the lowhomology region between 01and p subunits, ~(430-441) and @(422-434) (10, 19). Our studies proceeding from the information on tubulin’s C-terminal sequence (20) to antibody production using synthetic peptides as antigens allowed us to obtain antisera containing antibodies to the tubulin peptides (21) and also anti-idiotypic antibodies that react with MAP-l, MAP-2, and tau (19). The uniqueness of these antibodies strongly indicates that the tubulin subdomains of the synthetic peptide sequences are directly implicated in the interaction of MAPS. Despite the increasing amount of knowledge on the substructure of the MAP-binding domain on tubulin (9, 10, 13, 15, 18, 19, 21-23) little information is available on tubulin binding sites on MAPS and their structural features. Rather large peptides containing the tubulin binding domain on the high-molecular-weight MAPS (27-35 kDa) (24-27) or tau (14 kDa) (28) have been identified after limited proteolysis. However, no information is available on the detailed structure of tubulin binding sites on

ON TAU

569

PROTEIN

these MAP fragments. Here we present evidence that the synthetic tau peptides VaF - Glyzo4 (VRSKIGSTENLKHQPGGG) and Va1218-Gly235(VTSKCGSLGNIHHKPGGG) from the repetitive tau sequence (29, 30) define tubulin binding domains on tau, a finding supported by the recently published analysis of MAP sequences (31). Studies of tubulin-tau interaction have been developed by analyzing the binding of the synthetic tau peptides to tubulin and to the MAP-interacting cu(430441) and p(422-434) tubulin peptides. The tau peptide promotes tubulin assembly in a cold reversible fashion and incorporates into the assembled tubulin polymers. In addition, we have identified the reactive epitopes on tau, using anti-idiotypic antibodies which react specifically with the domains on MAPS involved in the interaction with tubulin and with the microtubule surface (19,21). MATERIAL

AND

METHODS

Protein putijkatimz. Microtubular protein was prepared from cow brains by two temperature-dependent cycles of assembly and disassembly. Immediately before use, the pellets were resuspended in assembly buffer (0.1 M Mes, pH 6.8,l mM M$t, and 2 mM EGTA) and a third cycle was performed (23). Tubulin depleted of MAPS (96% pure) was obtained by phosphocellulose chromatography (3). The microtubuleassociated proteins MAP-2 and tau were purified according to Vera et ul. (32). Purity was confirmed in all cases by SDS-acrylamide gel electrophoresis. Protein determinations. Protein concentration of tubulin and tau were determined spectrophotometritally (33). In some experiments the amount of protein was quantitated by turbidimetry after 15% trichloroacetic acid precipitation (18). The concentrations of the synthetic tau and tubulin peptides were determined by the fluorescamine method. Peptide synthesis. The tau peptides Val’87-Gly204 and Va[%Gly’“” as well as the 01and p tubulin peptides were synthesized by the solid-phase system of Merrifield (34). The peptides were purified using a reverse-phase (C-18) column and using gradients from 0.1% trifluoroacetic acid (TFA) to 60% acetonitrile in 0.1% TFA (10). The resulting peptides were analyzed with respect to amino acid composition after hydrolysis in 6 M HCl for 24 h at 110°C in evacuated sealed tubes. In all cases the composition found was equivalent to the theory. Purity of peptides was 99%. Biophysical studies. Binding of the tau peptides to tubulin under nonassembly conditions was quanti-

570

MACCIONI

tated by ultracentrifugation in a Beckman Airfuge as described by Howlett et al (35) but using sucrose instead of dextran to provide density stabilization (23). Centrifugation was carried out for 3 h at room temperature at 130,OOOgon solutions containing 3.0 mg/ ml tubulin and the desired concentration of tau peptide admixed with tracer amounts of “H-acetylated peptide (10). Under these conditions more than 99% of tubulin sedimented out of the lo-~1 aliquot removed from the top of the sedimentation column, while blanks containing no tubulin permitted correction for the small amount (2-4%) of unbound peptide sedimented during the ultracentrifugation. Fluorescence spectra were obtained in a SLM Aminco (SLM Instruments, Inc.) spectrofluorometer using double density software disks and a IBM personal computer as information recording unit. The temperature of the bath was maintained at 21°C by a thermostated chamber. Binding of Val’H7-Gly2”4 peptide to either 01or fl C-terminal tubulin peptides was monitored by the perturbation of fluorescence intensities from the spectra of the intrinsic fluorophores of tubulin peptides, taking advantage of the fact that tau peptide does not have any aromatic amino acid residues. The excitation wavelength was 278 nm, using emission and excitation bandwidths of 4 nm. The Raman spectrum was substracted from the emission spectra obtained after addition of tau peptide to the synthetic tubulin peptides. Increasing amounts of tau peptide (5 ~1) were added to a fixed concentration of either 01or fi tubulin peptides (50 pM, 1 ml) and the emission spectra recorded in the range 280-360 nm. The association constant, K,,, can be obtained from the data as indicated by Lehrer and Fasman (36). CD spectra were recorded in a Cary Model 60 spectropolarimeter with a Model 6001 CD unit atachment, the sample compartment being maintained at 27°C. Slits were porogramed to yield 15A bandwidth at each wavelength. Path lengths were 5-0.5 mm depending upon wavelength range. Mean residue ellipticities : [H]mrw (deg cm’/dmol) were calculated in the usual fashion (23). Ultraviolet spectra and differential spectroscopic analysis were recorded in a Beckman spectrophotometer with l-cm cells. The differential spectra were obtained with o and /3 peptide concentrations of 55 pM and tau peptide concentrations ranging from 10 to 120 pM. Repeated scans were usually recorded without differences between scans. Assembly assays The turbidimetric method was employed for all the kinetics of assembly of tau and tau peptide-induced tubulin polymerization. The assays with appropriate blanks were carried out in a Beckman spectrophotometer. Tubulin in the assembly buffer was admixed with either tau or the tau peptide and incubated in a thermostated cuvette at 37°C for 2 min; GTP was added to a final concentration of 1 mM to initiate assembly. The turbidity increase was recorded at 350 nm starting immediately after the ad-

ET AL. dition of GTP. Cold reversibility was measured by cooling the assembly mixture at 0°C and remeasuring the turbidity. Samples were obtained both before and after cooling, fixed with glutaraldehyde, stained with uranyl acetate, and analyzed under the electron microscope. For the experiments of incorporation of tau peptide, tubulin samples were assembled in the presence of 0.08 mg/ml “H-acetylated tau peptide (6900 of cpm/pg), 5 yM taxol, and increasing concentrations tau as competitor. The polymer from each sample was sedimented in the Airfuge and the resulting pellets were washed in warm assembly buffer and disolved in 0.2 ml of 0.1 M Mes, pH 6.8, containing 0.1% SDS. Aliquots were used to determine total protein by turbidity (18). Radioactivity of incorporated tau peptide was measured in different aliquots by scintillation counting. Immunological procedures. The anti-idiotypic antibodies were prepared and characterized as described (19,21). For the ELISA procedure, polyvinyl microtiter wells were incubated overnight with 100 ~1 purified tau, tau peptide, or the 01or /3 tubulin peptides at 10 pg/ml in carbonate buffer, pH 9.4. After a wash with phosphate-saline buffer containing 0.05% Tween 20 and 2 mg/ml bovine serum albumin, serial dilutions of the antibody (50 ~1) were added to the wells and the bound antibody was determined using the indirect peroxidase technique (19). In the competition assays, the inhibitor was added to the antibodies in test tubes and the tubes were incubated at 37°C for 1 h before the antibodies were added to the plates. RESULTS

Binding of the Tau Peptide VaE1”7-Gl~k to Tub&in Aiming at the identification of tau domains for the interaction with tubulin, we analyzed the binding of the synthetic peptides Val’s7-Gly204 and Va1218-Gly235 from the repetitive sequence on tau to the tubulin dimer. Binding data were obtained in an Airfuge, as described under Material and Methods, under nonassembly conditions over a ZO-fold range of constituent concentrations of tau peptide Va11s7-Gly204 (1.33 to 26.7 PM) using a concentration of tubulin dimer of 2.9 PM (Fig. 1). Over this range the number of nanomoles of the tau peptide bound per nanomole of tubulin increased from 0.41 to 1.67 without evidence, within experimental error, for cooperative binding. Under the assumption of statistical binding, the best linear fit to experimental data in a double-reciprocal plot

TUBULIN

BINDING

DOMAIN

ON TAU

PROTEIN

571

peptide (Fig. ZA). Although to a lesser extent, quenching was also observed upon interaction of tau fragment with p(422-434) tubulin peptide (Fig. 2B). Thus, binding of tau peptide produced perturbation of fluorescence intensity of the tubulin peptides. For both the Q and p tubulin peptides the amount of quenching was proportional to the concentration of the tau peptide added. The fluorescence intensity decreased linearly with the increase of tau peptide, 0.75 0 0.25 0.50 until the ratio of moles of tau Va1187-Gly204, peptide to moles of a tubulin peptide was one. Subsequent additions of tau peptide did not decrease significantly the (Ytubulin FIG. 1. Binding of the tau peptide Va1’87-Gly204 to peptide fluorescence (Fig. 2C). Thus the tubulin determined by ultracentrifugation in an Airquenching data for cytubulin peptide plotfuge. Double-reciprocal plot of moles of tau peptide ted in a titration curve as a function of the bound per mole of tubulin heterodimer (u) against the ratio Va1187-Gly204/Cr(430-441)are indicaequilibrium concentration of tau peptide. Each value tive of a binding stoichiometry of 1 mol tau corresponds to the average of three determinations. peptide/mol tubulin fragment. These data allowed us to calculate a binding constant of 4 X lo4 M-~ (Fig. ZC, inset). Due to the gave a value of 1.85 +-0.23 binding sites and low level of fluorescence quenching associan intrinsic binding constant, K, = 2.5 ated with the interaction of tau peptide X lo5 M-‘. An analysis of data according to with the ,0tubulin peptide, no quantitative a Hill plot allowed us to determine a nn analysis of these results was feasible. = 1.06 suggesting the noncooperative naThe effects of p tubulin peptide interacture of the binding. Under the same experi- tion on the conformation of tau peptide mental conditions, the peptide Va1218- Va1187-Gly204 was further examined using G1y235(1.33-26.7 pM) also bound to the tu- far uv circular dichroism. The results are bulin dimer (2.9 PM). Data plotting in a shown in Fig. 3A. The CD spectra of /3 tudouble-reciprocal plot allowed us to calcu- bulin peptide and tau fragment Va11g7late a value of 1.74 f 0.21 binding sites and Gly204were indicative of random confora binding constant, K, = 3.7 X lo5 M-’ for mations. The CD spectrum of tau peptide the interaction (data not shown). in the presence of the p tubulin fragment exhibited spectral differences, mainly in the range of 220-242 nm. Since the specInteraction of Tau Peptides with, a(@Otrum of p tubulin peptide was negligible at 441) and p(422-434) Tub&in Peptides the concentrations of /3 peptide utilized Having established that the tau peptide (0.02 mg/ml), these spectral differences Va1187-Gly204binds to the tubulin hetero- suggest that binding of /3 tubulin peptide dimer, we then investigated whether tau induces a structural change on tau peptide. peptide interacts with the tubulin peptides This is very interesting considering that in from the low-homology C-terminal region this experiment the ratio of @tubulin pepthat defines MAPS binding domains on tu- tide to tau peptide was 0.1 mol/mol. To furbulin (9, 10, 19, 21). Tubulin-tau interacther explore the interaction with repetition was studied by binding analysis of tive tau sequences, the effect of p tubulin synthetic peptides from binding domains peptide on the conformation of Va1218in both proteins using fluorescence spec- G1y235 was analyzed. In the absence of the p troscopy. Emission spectra showed a fluo- tubulin peptide, the Va1218-Gly235 fragment rescence quenching effect associated with has a random conformation. However, in tau peptide interaction with a(430-441) the presence of the C-terminal p tubulin

572

MACCIONI

ET AL.

-C 1.5

I;r p

.

1.0 iO.5 0 0 I

300

320

340

300

320

0

340

Wavelength

20

40

q&.3Y2&J~ , 1

12

3

“467- ~Y2o,/cQ‘,O -u)

FIG. 2. Interaction of tau peptide with a(430-441) and @(422-434) tubulin peptides analyzed by fluorescence spectroscopy. (A) Fluorescence spectra of (Ytubulin peptide (50 pM) in the absence (curve 1) or the presence of 20 pM (curve 2), and 50 fiM (curve 3) of tau peptide. (B) Fluorescence spectra of fl tubulin peptide (50 pM) in the absence (curve 1) or the presence of 20 FM (curve 2) and 50 pM (curve 3) of tau peptide. (C) Fluorescence titration of o( tubulin peptide (50 FM) with tau peptide (lo-120 PM). The emission wavelength was 305 nm. The inset shows the data analyzed according to Lehrer and Fasman (36) using a I:1 (mole:mole) stoichiometry of tau peptide/a tubulin peptide. Other conditions were as described under Material and Methods.

. . . . . ..* .......* ,/---I 200

220

240

Wavelength, nm

260

200

220

240

260

Wavelength,nm

FIG. 3. Interaction of tau peptides with @tubulin peptide analyzed by circular dichroism. (A) CD spectra of Val ‘s”-Gly204 peptide (0.3 mg/ml) in the absence (solid line) or the presence of 0.02 mg/ml fl tubulin peptide (dotted line). The spectrum of fl tubulin peptide alone (0.05 mg/ml) is also shown (dashed line). This is a comparison of normalized spectra (deg cm’/dmol protein vs wavelength). (B) CD spectrum of Va12’8-Gly2”5 peptide (0.3 mg/ml) in the absence (solid line) or the presence of 0.02 mg/ml p tubulin peptide (dotted line). The spectrum of /3 tubulin peptide at a concentration of 0.02 mg/ml gave a nearly basal signal. In all cases the buffer was 10 mM Mes, pH 6.8, containing 1 mM Mg”f.

TUBULIN

BINDING

DOMAIN ON TAU PROTEIN

573

p(422-434) tubulin peptide/Va1’87-Gly204 (Fig. 4B). In both cases the absorbance change was dependent on the increase of tau peptide concentration until the ratio of moles tau peptide to moles of either 01or /3 tubulin peptide was one. Subsequent additions of tau peptide did not change the absorbance significantly, pointing to a 1:l stoichiometry for the binding of tau peptide to (Yand p tubulin peptides. Quantitative analysis of the data (37) gave approximate association constants of 4 X 104-1 X lo5 M-’ and 9 X 104-4 X lo5 M-l for the binding of tau peptide to the cxand p tuI I I I I -0.011 bulin peptides respectively. A more accu280 300 260 rate determination of the association conWavelength stants was not possible due to the low values of the differential spectra. The FIG. 4. Interaction of tau peptide with CYand /3 tuassociation constant for the (Ytubulin pepbulin peptides analyzed by differential spectroscopy. Difference spectra of (Yand fl tubulin peptides in the tide is in good agreement with that obpresence of increasing amounts of tau peptide. The tained from the fluorescence experiments. solvent was 0.05 M Mes, pH 6.8, containing 1 mM M$+. As stated above, the tau peptide does not (A) a tubulin peptide (55 ELM) in the absence (curve 1) contain aromatic amino acid residues, and or the presence of 10 pM (curve 2), 20 flM (curve 3), 30 accordingly does not show any absorbance pM (curve 4), and 40 pM (curve 5) of tau peptide. (B) /3 in the range 255-320 nm. However, solutubulin peptide (55 PM) in the absence (curve 1) or the tions of this peptide showed uv light absorpresence of IO yM (CUrVe 2), 20 pM (CUrVe 3), 30 pM bance properties in the range 225-245 nm. (curve 4), and 40 pM (curve 5) of tau peptide. The absorbance in this region increased linearly with the increase of tau peptide concentration (data not shown). In the fragment, the tau peptide Va1218-Gly235 ex- presence of a fixed concentration of either hibits spectral characteristics indicative of (Yor 0 tubulin peptides, an increase of tau a new conformation suggesting that 0 tu- peptide concentration resulted in an increase of absorbance which was lower than bulin peptide may induce a conformation that of controls using tau peptide alone. change of tau peptide (Fig. 3B). This negative effect was more pronounced The structural aspects of the interaction between peptides from interacting do- in the presence of the p peptide as commains on tubulin and tau were also exam- pared with the (Y peptide. The spectroined using differential spectroscopy. Tak- scopic data suggest that formation of tau ing advantage of the fact that the tau pep- peptide-tubulin peptide complexes may intide does not contain aromatic amino acid duce some kind of structural changes in residues, the differential spectra of LYand the tau and/or tubulin peptides. None of fl tubulin peptides in the presence and the the changes described above were observed absence of tau peptides should provide in- when the experiments were carried out in formation on solvent exposure of the chro- buffers containing 0.3 M NaCl, an experimophores of the tubulin fragments. A pos- mental condition which is used to block tuitive differential spectrum was obtained in bulin-MAP interactions (24). the range 255-300 nm for the pair (~(430441)/Va1’87-Gly204,with an increase of ab- Incorporation of Tau Peptides into Tubulin sorbance as concentration of tau peptide Polymers: Competition with Tau increased (Fig. 4A). On the other hand, a negative differential spectrum was obThe analysis of the predicted tau repetitained in the range 255-295 for the pair tive sequences together with the binding

MACCIONI ET AL.

574

B

0.10

$3 0.10 2

E

E 5: $ 0.05

0

ifi a 0.05

4

0.0 ~ 0.0

0.0 10

20 Time, min

30

0.0

0.1 Tau Peptide,

0.2 mM

FIG. 5. Samples of tubulin (1.2 mg/ml) were allowed to polymerize and the assembly was assayed by the turbidimetric method in the absence (curve f) or the presence of 0.01 mM tau protein (curve a), or in the presence of tau peptide at final concentrations of 0.15 mM (curve b), 0.10 mM (curve c), 0.08 mM (curve d), and 0.02 mM (curve e). (B) Plot of the extent of assembly (A&.& as a function of concentration of tau peptide.

experiments suggest that the sequence represented in the tau peptides may define, or be part of, the tubulin binding domain on tau. It is therefore important to assess whether these fragments are capable of inducing tubulin assembly into microtubules or related structures. As shown in Fig. 5A, the peptide Va11s7-Gly204induced tubulin assembly in a peptide concentration-dependent fashion (Fig. 5B). After addition of the peptide to the assembly system, the turbidity increase was less steep than that in the presence of tau with a slow approach to the plateau. The assembly product was highly sensitive to cold depolymerization (Fig. 5A) and the free tubulin was able to reassemble in the presence of the peptide. Electron micrographs of tau peptide-induced polymers showed the presence of microtubule-related structures exhibiting irregularities along the polymer surface and frayed ends (Fig. 6). The presence of filamentous structures accompanying the assembled polymers was also observed, suggesting a diminished stability of these polymers as compared with microtubules formed in the presence of native tau protein. Our previous results on the binding of the ,8 tubulin peptide to tau (9, 10) along

with the findings on tau sequence strongly suggest that the other two repetitive sequences on tau, Va1218-Gly235(VTSKCGSLGNIHHKPGGG) and Va1250- G1y2’j7 (VQSKIGSLDNITHVPGGG) also may induce tubulin assembly. We found that Va1218-Gly235 promoted assembly to around 90% of the assembly level produced by Va1’87-Gly204.However, no assembly activity leading to microtubules was detected in the presence of the third peptide Va1250G1y267, suggesting that this tau moiety may play another role, possibly stabilization of the microtubule structure. In addition to the effect of tau in promoting tubulin assembly it is important to determine if tau peptide Val’s7-Gly204 binds to microtubules during assembly and disassembly. Thus, we examined the incorporation of tau peptide (Va1187-Gly204)into microtubules and the competition of tau with the tau peptide-polymer interaction. The stoichiometry of the binding of 3Hacetylated tau peptide to tubulin polymers was 0.16-0.19 mol/mol tubulin dimer. When the assembly was performed in the presence of increasing concentrations of tau the amount of tau peptide incorporated into the polymer decayed with a hyperbolic curve (Fig. 7).

FIG. 6. Electron microscopy of the assembled products of tubulin (1.3 mg/ml) in the presence of 0.6 mg/ml tau protein (A) or in the presence of 0.20 mg/ml tau peptide (B and C). The control in the absence of tau and the peptide is shown in D. Bars represent 1000 A.

Interaction of MAP-Reacting Antiidiotypic Antibodies with the Peptide Va1187-Gl~04

We have recently shown the presence of MAP-reacting anti-idiotypic antibodies in

the sera of rabbits immunized with either the a(430-441) or ,6(422-434) tubulin peptides (19, 21). These unique antibodies interact specifically with the tubulin binding domain on MAPS. The antibodies present in the serum of rabbits immunized with

576

MACCIONI ET AL.

complex, but a peptide concentration of 11 was required for a 50% inhibition of tau protein-antibody interaction (Fig. 8B). This is an interesting observation that could be related to differences in the binding affinity of the antibody and/or the presence of multiple binding sites on tau. PM

DISCUSSION

The present data show that the tau peptides Va1’87-Gly204and Va1218-Gly235 interact with tubulin and with two synthetic peptides, ~~(430-441)and /3(422-434) from CompetingProtein.pg the C-terminal regulatory domain of tuFIG. 7. Effect of tau protein on the association of bulin subunits. These tau peptides promote and incorporate the tau peptide Val 1s7-Gly204 with microtubules. Sam- tubulin polymerization ples of tubulin (1.2 mg/ml) were induced to assemble specifically into the assembled polymers. in the presence of 3H-acetylated-tau peptide (6900 These findings provide strong evidence cpm/rg), 5 pM taxol, and increasing concentrations of that the predicted amino acid sequences of tau protein (0) or bovine serum albumin (0). The the tau peptides define, or are part of, the polymer was sedimented in an Airfuge and the radiotubulin binding domain on tau. Several activity associated with the assembled protein tolines of evidence support this conclusion, gether with the amount of polymerized protein were i.e., ultracentrifugation analysis, fluoresdetermined as indicated under Material and Methods cence spectroscopy, uv differential specto calculate the stoichiometry of the peptide incorporation into the tubulin polymer. Results are ex- troscopy, circular dichroism and the data on tubulin assembly in the presence of the pressed as percentages of labeled peptide incorporated as a function of the competitor concentration, tau fragment. tau protein. The binding analysis using Airfuge ultracentrifugation indicated a stoichiometry of 2 mol of tau fragment per mole of the p tubulin peptide were purified using tubulin with a binding constant of 2.5-3.7 tau-affinity columns. When tested in solid- X lo5 M-‘. This stoichiometry, along with phase ELISA using tau immobilized on previous results showing the interaction of tau with both a and p tubulin subunits and polystyrene plates, a highly specific interaction of tau and the tau peptide was ob- the data on MAP-reacting anti-idiotypic served (Fig. 8A). In order to further dem- antibodies (10, 19) strongly suggest the presence of one tau binding site per tubulin onstrate the specific nature of the interaction, competition experiments using the subunit. The experiments on the binding of tau synthetic peptide were carried out. The tau fragments to (Y and p C-terminal tutau peptide, at 200 PM, inhibited by 95% bulin peptides corroborated this conclusion. Analysis of the fluorescence experithe interaction between the anti-idiotypic antibody and the immobilized tau peptide ments indicated a molar stoichiometry of (Figure 8B). Interestingly, tau peptide at 1:l for the binding of tau fragment (Va11g7concentrations of 500 PM inhibited by 97% Gly204)to CYtubulin peptide. Further support for the stoichiometric interaction of the tau protein-anti-idiotypic antibody complex formation. This indicates that the tau peptide with CYand p tubulin peptides amino acid sequence of the peptide Va11s7- is provided by results of the differential spectroscopy. The binding constant for the Glym4 contains the antigenic determinant(s) recognized by the anti-idiotypic association of the tau fragment to (Y tuantibodies on tau. It is important to note bulin peptide was slightly smaller than that the peptide, at 3 PM, inhibited by 50% that of tau peptide interaction with the tubulin dimer as determined by fluorescence the tau peptide-anti-idiotypic antibody

TUBULIN

0.01 0.1

577

BINDING DOMAIN ON TAU PROTEIN

1

Antibody,

10

100

pg/ml

0.1

1

10

W87-%0L,

100

1000

I.IM

FIG. 8. (A) Binding of tau affinity-purified anti-idiotypic antibody to tau protein (O), tau peptide Va1’87-Gly”“4 (m), and @(422-434)tubulin peptide (0) immobilized on polystyrene plates. Each point represents the average of three determinations. Inset, SDS gradient gel (7-15% acrylamide) of the purified tau used in the experiments. (B) Inhibition of the binding of tau affinity-purified antiidiotypic antibodies (8 pg/ml) to tau (0) and to tau peptide (m) by increasing amounts of tau peptide, using solid-phase ELISA.

or uv differential spectroscopy. This is not unexpected if we consider that in one case we are analyzing the interaction of tau fragment with the tubulin heterodimer, and in other case we are measuring the binding of a small fragment from a discrete sequence of one of the tubulin subunits. Interestingly, the uv differential spectroscopy appears to indicate a preferential interaction of the tau fragment Val’*?Glyzo4with the fi tubulin peptide, which is in agreement with our previous studies showing a preferential interaction of tau and MAP-2 with the p tubulin peptide (9, 10). Both fluorescence and differential spectroscopic experiments are indicative of perturbation effects on the chromophore groups of (Yand p tubulin fragments (tyrosine residues) induced by the tau peptide. This suggests a minor change in the microenvironment around those chromophores, since the tau peptide does not contain any aromatic amino acid residues. On the other hand, the CD experiments along with the spectroscopic data showing a significant absorbance change around 227 nm strongly suggest a structural change associated with the interaction of tau peptide with the tubulin fragment. Particularly, the CD experiment suggests that p tubulin

peptide promotes a conformational change in tau. The localization of the differences in ellipticities in the region 220-242 nm, when the CD spectrum of tau peptide plus p tubulin fragment was compared with the spectrum of tau alone, may be interpreted by the effect of p tubulin peptide in inducing a structure in tau fragments Valls7-G~Y’~~which could be a /3 Gly204and Va1218 turn type of conformation (38) or another similar structure. The tau peptides do not have secondary structure as revealed by CD analysis, which is consistent with the observations by Cleveland et al (39) that tau does not have a helical conformation and predictions of secondary structure according to the sequence data (29). In previous studies we have shown that P(422-434) peptide from the regulatory moiety of tubulin binds cooperatively to three to four tubulin binding sites on tau. Recently, Lee et al (29) found an H&amino acid sequence which is repeated three times in the mouse tau sequence. These repetitive sequences are highly homologous and contain only some minor conservative replacements. These sequences are also present in the human brain tau (30). Our studies using cow brain tau confirm that tubulin binding sites on tau are highly conservative among species. Now we have

578

MACCIONI ET AL.

demonstrated that peptides representing two of these sequences, Val’s’-Gly204 and Va1218-Gly235,bind to tubulin and specifically to two synthetic tubulin peptides, ~~(430-441) (KDYEEVGVDSVE) and /3(422-434) (YQQYQDATADEQG) that define the MAPS binding domain on tubulin. Useful and complementary information was also obtained from the analysis of the assembly of tubulin in the presence of tau peptide. The specificity of the interaction of tau peptide with tubulin in the assembled polymer was confirmed by the effect of tau in competing with tau peptide for the interaction with the tubulin polymer. It is noteworthy that both tau peptides Val”?Glyzo4and Va1218-Gly235 induce tubulin assembly, independently of the presence of each other, suggesting that multiple tau peptides defined by the repetitive sequences are not required for promoting assembly. These findings raise the possibility of an allosteric action of tau which could result in the stabilization of a tubulin conformation needed for the assembly. All these findings indicate that the tau peptides Val’s7-Gl$04 and Va1218-Gly235 appear to define a region on tau involved in its ability to promote microtubule assembly. Recent observations by Aizawa et al. (28) that a 14-kDa fragment containing the first two repetitive sequences on tau induces tubulin polymerization and those on sequences of MAPS (31) provide complementary support to our findings. The 14kDa fragment of tau is common to all tau components. A cyanogen bromide fragment containing the only cysteine residue on tau also interacts with microtubules (I. Correa, E. Montejo, and J. Avila, unpublished). In addition to the structural studies, the anti-idiotypic antibodies that recognize the tubulin binding domain on MAPS interacted with tau and the tau peptide Val’87-Gly204, thus confirming that this peptide represents a tubulin binding domain on tau. Interestingly, these antibodies also reacted with 18- and 20-kDa fragments containing the tubulin binding region of MAP-2 (unpublished), suggesting the presence of similar antigenic determi-

nants at the level of tubulin binding domains of both tau and MAP-2. ACKNOWLEDGMENTS This research has been supported by grants from The Council for Tobacco Research, USA, Inc., the American Heart Association of Colorado, and the Alzheimer’s Disease and Related Disorders Association (to R.B.M) and a grant from Comision Asesora, CAICYT (to J.A.). We thank Mrs. I. A. Maccioni for help in the preparation of the manuscript. REFERENCES 1. MURPHY, D. B., VALLEE, R., AND BORISY, G. G. (1977) Biochemistl-y 16,2598-2605. 2. PARYSEK, L. M., ASNES, C. F., AND OLMSTED, J. B. (1984) J. Cell Biol. 99,2287-2296. 3. WEINGARTEN, M., LOCKWOOD, A., Hwo, S., AND KIRSCHNER, M. W. (1975) Proc. NatL Acad. Sci USA 72,1858-1862. 4. MARECK, A., FELLOUS, A., FELLON, J., AND NUNEZ, J. (1980) Nature (London) 284,353-355. 5. RIEDERER, B., AND MATUS, A. (1985) Proc. Nat1 Acad Sci. USA 82,6006-6009. 6. WICHE, G. (1985) Trends Biochem. Sci. 10,67-70. 7. OLMSTED, J. B. (1986) Annu. Rev. Cell Biol. 2,421457. 8. MACCIONI, R. B. (1986) Molecular Cytology of Mi-

crotubules, Basque Country University Press, Leioa, Spain. 9. MACCIONI, R. B., RIVAS, C., AND VERA, J. C. (1987)

in The Cytoskeleton in Cell Differentiation and Development (Maccioni, R. B., and Arechaga, J., Eds.), pp. 39-49, ICSU Press-IRL (Oxford Univ. Press), London/New York. 10. MACCIONI, R. B., RIVAS, C., AND VERA, J. C. (1988) EMBOJ 7.1957-1963. 11. GOLDENRING, J. R., VALLANO, M. L., AND DE LoRENZO, R. J. (1985) J. Neurochem. 45,900-905. 12. WANDOSELL, F., SERRANO, L., AND AVILA, J. (1987) J. Biol. Chem. 262,8268-8273. 13. SERRANO, L., DE LA TORRE, J., MACCIONI, R. B., AND AVILA, J. (1984) Proc. Nat1 Acad. Sci. USA 81,5989-5993. 14. MACCIONI, R. B., SERRANO, L., AVILA, J., AND CANN, J. R. (1985) Eur. J. Biochem 156, 375381. 15. SERRANO, L., AVILA, J., AND MACCIONI, R. B. (1984) Biochemistry 23,4675-4681. 16. LEWIS, S. A., LEE, M. G. S., AND COWAN, N. W. (1985) J. Cell BioL 101,852-861. 17. CLEVELAND, D. W. (1987)J. Cell Biol. 104,381-383. 18. VERA, J. C., RIVAS, C., AND MACCIONI, R. B. (1989) Biochemistry 28,333-339. 19. RXVAS, C. I., VERA, J. C., AND MACCIONI, R. B. (1988) Proc. NatL Acad Sci. USA 85,6092-6096.

TUBULIN

BINDING

DOMAIN

20. PONSTINGL, H., KRAUHS, E., LITTLE, M., KEMPF, M., HOFFER-WARBINECK, R., AND ADE, W. (1982) Cold Spring Harbor Symp. C&ant. BioL 46,191-198. 21. VERA, J. C., RIVAS, C. I., AND MACCIONI, R. B. (1988) Proc. NatL Acad. Sci. USA 85,6X3-6767. 22. SERRANO, L., MONTEJO, E., HERNANDEZ, M. A., AND AVILA, J. (1985) Eur. J. B&hem. 153,595600. 23. MACCIONI, R. B., CANN, J. R., AND STEWART, J. (1985) Eur. J. B&hem. 154,427435. 24. VALLEE, R. B. (1980) Proc. NatL Acad. Sci USA 77,3206-3210. 25. GOTTLIEB, R. A., AND MURPHY, D. B. (1985) J. Cell

BioL 101,1782-1’789. 26. FLYNN, G., JOLY, J. C., AND PURICH, D. L. Biochem. Biophys. Res. Commun. 148, 1459. 27. AIZAWA, H., MUROFUSHI, H., KOTANI, S., NAGA, S., HIROKAWA, N., AND SAKAI, H.

(1987) 1453-

PROTEIN

579

29. LEE, G., COWAN, N. W., AND KIRSCHNER, M. W. (1988) Science 239,285-288. 30. GOEDERT, M., WISCHIK, C. M., CROWTHER, R. A., WALKER, J. E., AND KLUG, A. (1988) Proc. NatL

Acad 55. USA 85,4051-4055. 31. LEWIS, S. A., WANG, D., AND COWAN, N. J. (1988) Science 242,936-939. 32. VERA, J. C., RIVAS, C., AND MACCIONI, R. B. (1988)

FEBS Z&t. 232,159-162. 33. HERNANDEZ, M. A., AVILA, J., AND ANDREU, J. M. (1986) Eur. J. B&hem 154.41-48. 34. STEWART, J. M., AND YOUNG, E. (1984) Solid Phase Peptide Synthesis, 2nd ed. Pierce Chemical Co., Rockford, IL. 35. HOWLETT, G. J., YEH, E., AND SCHACHMAN, H. K. (1978) Arch. B&hem Biophys. 190,809~819. 36. LEHRER, S. S., AND FASMAN, G. (1966) B&hem.

Biophys. Res. Commun. 23,133-138. 37. GORMAN, E. G., AND DARNALL, D. W. (1981) Bio-

HISA(1987)

J Biol Chem 262,3782-3767. 28. AIZAWA, H., KAWASAKI, H., MUROFUSHI, H., KoTANI, S., SUSUKI, K., AND SAKAI, H. (1988) J.

BioL Chem. 263,7703-7707.

ON TAU

chemistry 20,38-43. 38. GIERASCH, L. M., DEBER, C. M., MADISON, V., NIU, C. H., AND BLOUT, E. R. (1981) Biochemistry 20, 4730-4738. 39. CLEVELAND, D. W., Hwo, S. Y., AND KIRSCHNER, M. W. (1977) J. Mol. BioL 116,227-248.