Antibodies to synthetic peptides from the tubulin regulatory domain ...

1 downloads 0 Views 1MB Size Report
(assembly inhibition/anti-peptide antibodies/tubulin carboxyl-terminal domain/microtubule ... Exopeptidases that remove the last eight amino acid residues from ...
Proc. Nati. Acad. Sci. USA Vol. 85, pp. 6763-6767, September 1988

Cell Biology

Antibodies to synthetic peptides from the tubulin regulatory domain interact with tubulin and microtubules (assembly inhibition/anti-peptide antibodies/tubulin carboxyl-terminal domain/microtubule surface/immunofluorescence)

JUAN C. VERA, CORALIA 1. RiVAS, AND RICARDO B. MACCIONI* Department of Biochemistry, Biophysics, and Genetics, B-121, University of Colorado Medical School, Denver, CO 80262

Communicated by Keith R. Porter, May 9, 1988

ABSTRACT The carboxyl-terminal region of tubulin a and ( subunits plays a major role in regulating its assembly into microtubules and constitutes an essential domain for the selective interaction of microtubule-associated proteins (MAPs). With the goal of understanding the structural basis of the regulatory function of the carboxyl-terminal domains of tubulin subunits, we have produced rabbit antisera against two MAP-interacting peptides Lys-Asp-Tyr-Glu-Glu-Val-Gly-ValAsp-Ser-Val-Glu of a-tubulin and Tyr-Gln-Gln-Tyr-Gln-AspAla-Thr-Ala-Asp-Glu-Gln-Gly of fl subunit. The affinitypurified a and (3 anti-peptide antibodies interacted specifically with tubulin and with the respective peptide antigens but did not interact with MAPs. Substoichiometric amounts of both antibodies showed the capacity to inhibit in vitro MAP-induced tubulin assembly and to promote a fast depolymerization of preassembled microtubules. Taxol-promoted assembly of pure tubulin was not inhibited by the antibodies. In the presence of MAP-2 and taxol, the antibodies decreased the MAP-2 content of taxol-promoted microtubules. The interaction with microtubules was corroborated by immunofluorescence experiments in HeLa and NE-18 lung carcinoma cells. The epitopes recognized by the a and .8 anti-peptide antibodies appear to be located in the outer surface of the microtubular structure. A general overview of the essential features of tubulin structure and functions has emerged over the past two decades (1, 2). One question of current interest is the elucidation of the regulatory pathways for the assembly of tubulin into microtubules. Understanding of the control of microtubule assembly should provide clues about the functional versatility of microtubules. We have demonstrated that the C-terminal moieties of tubulin subunits constitute essential sites for the interaction of microtubule-associated proteins (MAPs) (3). Limited proteolysis of tubulin with subtilisin produces the cleaved tubulin dimer and two 4-kDa fragments containing the C-terminal domain of a and ,8 subunits. Removal of the 4-kDa fragment facilitates tubulin self-association and makes the assembly independent of the MAPs (4, 5). This 4-kDa tubulin domain appears to hinder the interactions responsible for its self-assembly into microtubules (4-6) and plays a regulatory role in microtubule formation (7). Exopeptidases that remove the last eight amino acid residues from the C-terminal moiety of both tubulin subunits do not relieve the modulatory effect of the 4-kDa C-terminal region and the MAP dependence of the assembly (8), demonstrating that these residues are not directly implicated in the interaction of tubulin with MAPs. This is further supported by observations indicating that antibodies that recognize the C-terminal tyrosine of a tubulin (9) do not affect

microtubule assembly (10).

Peptides corresponding to short discrete sequences of the C-terminal domain, a(430-441) and 1(422-434) bind to MAP-2 and tau, but a preferential interaction of the 1 peptide with these MAPs components was observed (8). These peptides correspond to the lowest homology region between the a- and 3-tubulin. Thus, the various MAP components may bind to a- and/or f-tubulin subunits with different affinities (8). To define the substructure of the C-terminal tubulin domain, we produced rabbit antisera against these MAPinteracting peptides. These anti-peptide antibodies exhibited a remarkable capacity to inhibit in vitro microtubule assembly. Addition of tubulin to the assembly mixture overcomes the inhibitory effect, but addition of purified MAP-2 or total MAPs did not have any effect on the assembly inhibition. Interestingly, substoichiometric amounts of either the a or ,1 anti-peptide antibodies induced a fast depolymerization of preassembled microtubules. The striking properties of these antibodies, along with their usefulness for immunolabeling microtubules in cultured cells, provide a powerful probe for further studies of the in vivo functions of cytoskeletal and mitotic microtubules.

MATERIALS AND METHODS Immunization Protocol. The immunization protocol consisted essentially of (i) subcutaneous injections of 0.2 mg each of a or ,8 peptide, coupled to Affigel (Bio-Rad), and emulsified in Freund's adjuvant into rabbits. The peptides were coupled directly to Affigel according to the manufacturer's instructions. (ii) Three subcutaneous booster injections in incomplete Freund's adjuvant were given every 10 days. Dot Immunobinding Analysis. The antigens were prepared in 0.1 M 2-(N-morpholino)ethanesulfonic acid (Mes) buffer (pH 6.8) and dotted onto nitrocellulose followed by blocking with 0.3% Tween 20 in 15 mM phosphate/0.5 M NaCl (blocking buffer) (11). The dilution used for the a and 1 peptide antisera was 1:1000 in blocking buffer, while dilution for the anti-peptide antibodies purified by affinity chromatography was 1:300. After incubation of the nitrocellulosebound antigens with the respective antibodies for 16 hr at room temperature, the blots were exposed to the second antibody, goat anti-rabbit IgG for 4 hr. Finally, the blots were developed with the peroxidase-antiperoxidase complex using 3,3'-diaminobenzidine. For the competition experiments, the a or 1 peptides were mixed with affinity-purified anti-a- or anti-,8-peptide antibody (aAbl or 83Abl) and allowed to compete with each peptide immobilized on nitrocellulose. The antibodies were affinity-purified from the IgG fraction of either the a or ,B antisera with nitrocellulose paper previously Abbreviations: aAbl, affinity-purified anti-a-peptide antibody;

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

,BAbl, affinity-purified anti-p-peptide antibody; MAPs, microtubuleassociated proteins. *To whom reprint requests should be addressed.

6763

6764

Cell Biology: Vera et al.

saturated with tubulin (16). For affinity purification in peptide-Sepharose gels, the antisera were passed through the column (0.9 x 8 cm), and the bound aAbl and BAbl were selectively eluted with 0.2 M acetic acid, brought to neutral pH, and dialyzed three times against 1 liter of Mes (pH 6.8). Enzyme-Linked Immunoassay. Polyvinyl microtiter wells were incubated overnight with 100 ,ul of either the a or P peptide at a concentration of 10 jug/ml in carbonate buffer (pH 9.4). The remaining sites in the plates were blocked with 0.05% Tween 20 plus 2 mg of bovine serum albumin per ml in phosphate-buffered saline solution. Serial dilutions of either aAbl (50 1.l) or (3Abl (50 1.d) were added to the wells and bound antibody was determined by the indirect peroxidase technique and o-phenylenediamine. For competition experiments, serial dilutions of either the a or A peptides, tubulin, or MAPs were admixed with the respective Abl antibody and allowed to compete with the immobilized peptide in the microtiter wells in a total vol of 50 Al. Immunoblotting Experiments. MAP components and tubulin were separated by polyacrylamide electrophoresis in the presence of 0.2% NaDodSO4 with 5-15% acrylamide gradients. The separated proteins were electrotransferred to nitrocellulose membranes in 25 mM Tris.HCl/0.375 M glycine, pH 8.3, containing 0.2% NaDodSO4 and 20%o methanol at 0.1 A during 14 hr. After electrotransfer, the membranes were processed as indicated above for dot immunobinding. Protein Purification. Tubulin from cow brains was prepared by in vitro polymerization-depolymerization cycles followed by phosphocellulose chromatography to remove MAPs (3). MAPs were obtained according to Vera et al. (12). Microtubular protein and MAP concentrations were determined turbidimetrically after 5% trichloroacetic acid precipitation using appropriate standard curves. Tubulin, MAP-2, and tau concentrations were determined spectrophotometrically (13). The concentrations of the synthetic peptides were determined by the fluorescamine method. Assembly Assays. The maximum extent oftubulin assembly was assayed by sedimentation of polymers with a Beckman air-driven ultracentrifuge for 10 min at 140,000 x g as described (5). The assembly kinetics of tubulin were analyzed by the turbidimetric procedure (3). Cold reversibility was tested by cooling the assembly mixture to 0°C for 10 min and remeasuring the turbidity. Samples were obtained both before and after bringing the assembly mixture to cold, fixed with glutaraldehyde, and analyzed in the electron microscope. Peptide Synthesis. The a- and 83-tubulin peptides were synthesized by the solid-phase system of Merrifield (14) and the synthetic products were purified by HPLC with a high pore (330 A) Bio-Rad RP-304 reversed-phase column (C4). The running buffer for the separation of the peptide was a mixture of acetonitrile with 0.1% aqueous solution of trifluoroacetic acid with ratios of 15:85 or 8:92 (vol/vol) for the a and (3 peptides, respectively. Indirect Immunofluorescence. The NE-18 lung squamous carcinoma and HeLa (strain CL-2) cells were grown on coverslips in Petri dishes in RPMI 1640 medium supplemented with 10% fetal calf serum. For immunofluorescence, the medium was decanted and the attached cells were washed twice with warm microtubule stabilizing buffer (MSB: 0.05 M Mes, pH 6.8/10 mM EGTA/2 mM MgCl2/0.2 mM GTP/5% dimethyl sulfoxide), extracted with the same buffer containing 0.3% Triton X-100, and fixed by immersion in MSB containing 3% paraformaldehyde. Fixed cells were incubated with the peptide-specific antibody followed by detection with fluorescein isothiocyanate-conjugated goat anti-mouse IgG (15). In controls, the cells were incubated in MSB or in buffer (0.05 M Mes, pH 6.8/2 mM CaCl2) at 4OC to depolymerize the microtubules prior to the fixation step. In other control experiments, fixed cells were incubated with the peptide-

Proc. Nati. Acad. Sci. USA 85 (1988)

specific antibody in the presence of the peptide (2 mg/ml) or the antibodies were previously absorbed with tubulin immobilized in nitrocellulose. RESULTS Production of Anti-peptide Antibodies. Two synthetic peptides from the 4-kDa C-terminal tubulin region exhibiting the lowest homology between a and (8 subunit sequences were used: a(430-442), Lys-Asp-Tyr-Glu-Glu-Val-Gly-Val-AspSer-Val-Glu; (3(422-434), Tyr-Gln-Gln-Tyr-Gln-Asp-AlaThr-Ala-Asp-Glu-Gln-Gly. Antibodies to the synthetic peptides were produced by injecting the tubulin peptides into different groups of rabbits. The immunization protocol allowed us to obtain anti-peptide antibodies with high titer in less than 1 month, as determined by the enzyme-linked immunoassay. During the screening, we found antibodies reacting with tubulin and with either the a or the ( peptides, but we also found antibodies reacting with MAPs. Two different populations of antibodies were separated by affinity chromatography and were characterized by enzyme-linked immunoassay. The first subset of antibodies purified by peptide-Sepharose affinity chromatography correspond to the a and (8 anti-peptide antibodies (aAbl and (BAbl). The second fraction, not bound to tubulin peptide-affinity matrix, proved to be autoantibodies to the a or (3 anti-peptide antibodies, characterized as anti-idiotypic antibodies. The latter antibodies were purified from both the a and ( antisera by MAP affinity chromatography (aAb2 and (BAb2) and exhibited a strong interaction with MAP-1, MAP-2, and tau (16). In this study, we have focused mainly on the first antibody fractions from both the a and (3 antisera. These fractions interacted with the a and 8 peptides, respectively. No reaction of the affinity-purified a- and (Abl with MAPs was observed. The ability of tubulin, synthetic peptides, and MAPs to inhibit the binding of a- and (Abl to either peptide was compared by enzyme-linked immunoassay. The a peptide-aAbl complex formation was inhibited by free a peptide and tubulin, but not by MAPs and the (8 peptide, while the (3 peptide-(3Abl complex was completely inhibited by free (3 peptide and tubulin but was not inhibited by the a peptide or by MAPs (Table 1). The immunoblotting of a-tubulin by aAbl and the blotting of( subunit with (Abl provide further support for the specificity of the interaction of the antibodies with the respective peptide and the lack of cross-reactivity. The specificity of the antibodies was also confirmed in immunofluorescence experiments (see below). Inhibition of MAP-Induced Tubulin Polymerization by the Anti-peptide Antibodies. Nonimmune serum did not affect MAP-induced assembly of tubulin. However, an almost complete inhibition was attained in the presence of small amounts of either aAbl (1:30, mol/mol with respect to tubulin dimer) or,(Abl (1:150 with respect to tubulin). The extent of inhibition increased with the amount of antibody in the range 1-10 jig (Fig. 1 A and B). The simultaneous addition of both antibodies (1 ,ug of aAbl and 0.7 Ag of (BAbl) had a summative effect on the inhibition of MAP-induced assembly (60%6). Doubling the aAbl to 2 ug while keeping (BAbl constant (0.7 ,ug) increased inhibition to 78%, but an increase of aAbl by a factor of 2.3 and of (3Abl by a factor of 2 gave 98% inhibition of the assembly (Fig. 1C). The concentrationdependent inhibition of tubulin-MAP assembly by aAbl or Abl was confirmed by using the sedimentation assay to quantitate the extent of assembly (Fig. 1D). As described above, (3Abl was more effective in blocking microtubule assembly as compared with aAbl. A readdition of free tubulin subunits to the assembly assay partially prevented the inhibition by both a and (3 antibodies, while a 3-fold increase of MAP concentration in the assay did not overcome the

Cell Biology: Vera et al.

Proc. Natl. Acad. Sci. USA 85 (1988)

6765

Table 1. Specificity of anti-peptide antibodies

Antigen a peptide-(430-441)

Antibody Anti-a-(430-441)

Competitor a(430-441)

Competition assay Concentration, x 10-5 M 7.7 7.7 38.0 2.0 2.0

Inhibition, % 100 0 .8(422-434) 0 0(422-434) Tubulin 100 MAP 0 8 peptide-(422-434) Anti-1-(422-434) 7.7 a(430-441) 0 a(430-441) 38.0 0 7.7 100 ,8(422-434) Tubulin 2.0 100 MAP 2.0 0 The indicated concentrations of either peptide, tubulin, or MAPs were mixed with aAbl or PAbl and allowed to compete with the respective peptide immobilized on polystyrene plates or nitrocellulose. Molar concentration of tubulin was determined by using a molecular weight of 100,000. Molar concentration of MAPs was estimated from the sum of the molar contributions of tau and MAP-2 (molar ratio, 2:1) using molecular weights of 66,000 and 270,000 for tau and MAP-2, respectively.

inhibitory effect. Particularly, MAPs have been used in this experiment instead of MAP-2 or tau to ascertain the effect of the a- and BAbl on the microtubule system containing several MAP components. MAP-2 or tau-induced microtubule assembly was also affected by the antibodies. The inhibition of MAP-induced assembly of tubulin by both a and (3 antibodies was also correlated to observations in the electron microscope of antibody-promoted blockage of microtubule formation. No microtubules were found after assembly in the presence of 6-8 ,g of either a- or Abl (Fig. 2). Depolymerization of Preformed Microtubules by the a- and ,BAbl. Not only did aAbl and O3Abl inhibit MAP-induced A

B

0

0.2

0.2

a b

b 0

0

-C~~~~~~ 0.1

0.1

1M_*

~~d

0.0

d

0.0 O

5 Timemin

10

0

5

10

Time~min

microtubule assembly they also depolymerized preassembled microtubules. A fast disassembly of microtubules formed from tubulin and total MAPs occurred after addition of either 20 ,ug of aAbl or 6 ,ug of 3Abl (Fig. 3 A and B). Similar results were obtained with microtubules assembled in the presence of either pure MAP-2 or tau. Addition of 20 ug of nonimmune IgG produced a decrease in turbidity of only 20%, which accounted for a dilution effect, since simply adding 0.1 ml of assembly buffer to 0.3 ml of assembly mixture produced an identical decrease in turbidity (Fig. 3C). Assembly analysis by the sedimentation procedure showed that the disassembling effect of a or A antibodies was dependent on the total concentration of the respective antibody in the assembly assay (Fig. 3D). Taxol-Induced Tubulin Self-Assembly Is Not Affected by aand/or ,BAbl. Inhibition of tubulin assembly and disassembly of preformed polymers may be a consequence of the antibody competition with MAPs for binding to tubulin or that antibody could hinder the polymerization of tubulin into microtubules by a steric effect. To address this possibility, phosphocellulose-purified tubulin was induced to polymerize in the presence of taxol. Under these experimental conditions, the polymerization of tubulin was not affected by a- or is

A