Cell Motility and the Cytoskeleton 33:38-51 (1996) -TubuIin Redistribut ion in TaxoI-Treated Mitotic Cells Probed by Monoclonal Antibodies Martina Novakova, Eduarda Draberova, Wolfgang Schurmann, Gerhard Czihak, Vladimir Viklicky, and Pavel Draber Institute of Molecular Genetics, Academy of Sciences of the Czech Republic, Czech Republic (M.N., E.D., V.V., P.0 , ) ; and Institute of Developmental Biology and Genetics, University of Salzburg, Salzburg, Austria (W.S., G.C.) Monoclonal antibodies were prepared against conserved synthetic peptide from the C-terminus of the y-tubulin and their specificity was confirmed by immunoblotting, competitive enzyme-linked immunosorbent assay (ELISA) and immunofluorescence. The antibodies decorated interphase centrosomes as well as halfspindles and midbodies in mitotic cells of various origin. The prepared antibodies were used to study the y-tubulin distribution in nocodazole and taxol-treated cells. In the cells recovering from the nocodazole treatment, y-tubulin was found in centers of all microtubule asters. Examination of relative location of y-tubulin and microtubule asters in taxol-treated mitotic cells 3T3, HeLa and PtK, revealed that the number of taxol-induced microtubule asters exceeded the number of y-tubulinpositive spots. The y-tubulin was often found in the periphery of microtubule asters. Centrosomal phosphoprotein epitope detected by MPM-2 antibody colocalized with y-tubulin in taxol-treated mitotic cells. The presented data suggest that taxol-induced microtubule asters are in vivo nucleated independently of y-tubulin, and other minus-end nucleator(s)are necessary for formation of such asters. Alternatively, y-tubulin is present in subthreshold amounts undetectable by immunofhorescence. 0 1996 Wiley-Liss, Inc. Key words: mitotic cells, microtubule organizing centers, monoclonal antibodies, taxol, y-tubulin INTRODUCTION Organization of microtubular network is, in animal cells, controlled by microtubule organizing centers (MTOCs) [Pickett-Heaps, 19691. Despite their remarkable structural variations in different species and cell types, MTOCs have similar functions. They nucleate microtubule assembly, establish polarity of microtubules with their fast growing (plus) ends distal to MTOCs and determine the number of protofilaments in microtubules that assemble from them. Centrosomes and ciliary or flagellar basal bodies are the two most frequently studied MTOCs [Brinkley , 1985; Vorobjev and Nadezhdina 19871. Centrosomes are generally composed of two centrioles surrounded by electron-dense pericentriolar material which nucleates the cytoplasmic network of microtubules. Microtubules can be nucleated also in centrosomes that lack centrioles, as for example, in mam0 1996 Wiley-Liss, Inc. malian oocytes undergoing meiosis [Szollosi et al., 19721. One of the almost ubiquitously present components of MTOCs is y-tubulin, a member of the tubulin superfamily [Oakley and Oakley, 19891. It is a minor protein compared with a- and P-tubulin, making up less than 1% of total tubulin [Stearns et al., 19911, and it is phylogenetically highly conserved. The y-tubulins of humans and Xenopus laevis share 98% identity [Oakley, 19941. The amino acid identity with a- and P-tubulins is in the range of 28-35% [Oakley, 19941. Immunolocalization Received August 24, 1995; accepted September 5, 1995. Address reprint requests to Pavel DrBber, Ph.D., Institute of Molecular Genetics, Academy of Sciences of the Czech Republic, VidefiskB 1083, 142 20 Praha 4, Czech Republic. y-Tubulin in Taxol-Treated Mitotic Cells of y-tubulin has shown that it was associated with centrosomes in interphase and with spindle poles in mitosis [Joshi et al., 1992; Stearns et al., 1991; Zheng et al., 19911. In mitotic cells it was also found in half-spindles [Lajoie-Mazenc et al., 19941 and in midbodies during cytokinesis [Julian et al., 19931. The y-tubulin was also reported in acentriolar MTOCs during mouse development [Gueth-Hallonet et al., 1993; Palacios et al., 19931. Moreover, y-tubulin was found in the vicinity of basal bodies in retina photoreceptors and ciliated epithelia [Muresan et al., 19931 and in the centriolar basal body of the mouse sperm [Palacios et al., 19931. Microinjection of mammalian cells with anti-y-tubulin antibody has been shown to prevent the formation of both cytoplasmic and mitotic microtubules [Joshi et al., 19921. In the light of collected data it appears that y-tubulin could be a universal component of MTOCs in vivo and may be responsible for their nucleating activity independently of their association with centrioles. More recent in vitro experiments on reconstitution of centrosome assembly in cell free extracts indicate that y-tubulin might be required not only for centrosome function but also for centrosome formation [Fklix et al., 1994; Stearns and Kirschner, 19941. The y-tubulin can form multiprotein complexes with centrosomal proteins [Raff et al., 19931 and it was suggested that some of associated proteins either on their own or in association with y-tubulin might be directly involved in the nucleation of microtubules [Sunkel et al., 19951. Major tools for studying the distribution of y-tubulins are affinity-purified polyclonal antibodies. Monoclonal antibodies against y-tubulin have not yet been described. As y-tubulin is a minor protein and because it was not obtained from cells in purified form, fusion proteins or synthetic peptides were used for preparation of antibodies. The antibodies were raised against various bacterially produced fusion proteins [Oakley et al., 1990; Stearns et al., 19911 and highly conserved aminoterminal [Joshi et al., 1992; Julian et al., 19931 and carboxyterminal sequences [Julian et al., 1993; Lajoie-Mazenc et al., 19941 of y-tubulin. Here we report on preparation and characterization of monoclonal antibodies directed against a conserved carboxyterminal sequence of y-tubulins. Using these antibodies we showed that y-tubulin is associated with the microtubule-nucleating material in both the normal and nocodazole-treated cells but not in taxol-treated mitotic cells. MATERIALS AND METHODS Cells Mouse embryonic fibroblasts 3T3, mouse neuroblastoma Neuro-2a, mouse embryonal carcinoma P19X1, 39 human epidermoid carcinoma A-43 1, human amnion cells AMA, human cervical adenocarcinoma HeLa S3, rat kangaroo kidney PtK, and turkey embryonic fibroblast TC-3 1 cells were grown on coverslips at 37°C in air with 5% CO, in RPMI-1640 medium supplemented with 3 mM L-glutamine, 1 mM sodiumpyruvate, penicillin (100 i.u./ ml), streptomycin (100 pg/ml), gentamycin (20 pg/ml) and 10% (v/v) heat-inactivated foetal calf serum. Neuro-2a cells were also cultured in serum-free medium for 3 days to allow formation of long projections. To some of the cultures, taxol (National Cancer Institute, Bethesda, MD) and nocodazole (Sigma Chemie, Deisenhofen, Germany) at 10 pM were added for 1, 6, 18 and 24 hours. To wash off nocodazole, the cells were transferred to a new medium and incubated for 3 or 6 min at 37°C before fixation. Peptide Synthesis To prepare anti-y-tubulin antibodies, a 16-aminoacid peptide EYHAATRPDYISWGTQ corresponding to the human y-tubulin sequence 434-449 [Zheng et al., 19911 was synthesized. A cystein (C) had been added to the amino terminus of the peptide in order to allow oriented coupling to the carrier proteins. In competitive assays a 16-amino-acid peptide EEFATEGTDRKDVFFY corresponding to the human y-tubulin sequence 38-53 [Zheng et al., 19911 and an 8-amino-acid peptide GEEEGEEY corresponding to the a-tubulin sequence 444-451 [Ponstingl et al., 19811 were also used. The peptides were prepared by Dr. Ivan Blaha (Institute of Organic Chemistry and Biochemistry, Academy of Sciences of the Czech Republic, Prague) by the solid-phase method [Merrifield, 19631 on the classic Merrifield’s chlormethylated carrier. Details of peptide synthesis and HPLC purification were described previously [Viklickf et al., 19881. Antibodies Peptide (C)434-449 was covalently coupled to maleimide activated keyhole limpet hemocyanin (KLH) or bovine serum albumin (BSA; Imject Activated Immunogen Conjugation Kit, Pierce, Rockford, IL) at a ratio of 2 mg peptide/2 mg activated carrier protein according to manufacturer’s directions and used for immunization of rabbits and mice. Balbk 6-8-weeks-old mice were immunized intraperitoneally (i.p.) with 100 pg of KLH with coupled in Freund’s complete adjuvant. Booster injections of 100 pg of antigen in Freund’s incomplete adjuvant were given i.p. at two-week intervals. Sera were monitored for antibody activity by enzyme-linked immunsorbent assay (ELISA) on BSA with coupled peptide and by indirect immunofluorescence on A-43 1 cells fixed by cold methanol/acetone. Four days prior to fusion, the mice were injected intravenously with 100 pg 40 Novakova et al. of antigen in phosphate-buffered saline (PBS), pH 7.2. The fusion with mouse myeloma cells Sp2/0, screening by ELISA, cloning and production of ascitic fluids in Balb/c mice have been described previously [Viklicky et al., 1982; Draber et al., 19881. The subclasses of monoclonal antibodies were identified by isotyping kit (ISO- 1; Sigma Chemie, Deisenhofen, Germany). Preparation of rabbit antibodies against (C)434-449 peptide coupled to KLH, as well as following affinity purification on Sepharose 4B column with immobilized peptide coupled to BSA, was performed essentially as described previously [Draber et al., 19911. Microtubular structures were detected with affinity-purified polyclonal antibody against ap-tubulin heterodimer [Draber et al., 19911, mouse monoclonal antibody TU-01 (IgGl) against a-tubulin [Viklickg et al., 19821 and mouse monoclonal antibody TU-04 (IgM) against a-tubulin [Draber et al., 19891. The last one was conjugated with lissamine rhodamine B sulphonylchloride [Brandtzaeg, 19731. A monoclonal antibody MPM-2 against phosphorylated epitope in mitotic cells [Davis et al., 19831 was kindly provided by Drs. P. Rao and J. Kuang (University of Texas, Texas Medical Center, Houston). A monoclonal antibody HTF-14 (IgG1) against human transferrin [Viklickg et al., 19831 was used as a negative control. Fluorescein isothiocyanate (F1TC)-conjugated anti-rabbit Ig antibody as well as FITC-conjugated and horseradish peroxidase-conjugated anti-mouse Ig antibodies were from Sevac (Prague, Czech Republic) , tetramethylrhodamine isothioc yanate (TR1TC)-conjugated anti-rabbit Ig antibody was from SANBIO (Uden, Holland) and anti-mouse Ig antibody conjugated with alkaline phosphatase was purchased from Promega Biotec (Madison, WI). measurements were made for each test point. Standard errors of the means of triplicate absorbance measurements were less than 3%. Electrophoresis and lmmunoblotting SDS-polyacrylamide gel electrophoresis (SDSPAGE) was carried out according to Laemmli . Protein concentration in SDS-sample buffer was determined as described [Driber, 19911. Separated proteins were transferred onto nitrocellulose sheets by electroblotting [Towbin et al., 19791. Details of the immunostaining procedure using secondary antibody labeled with alkaline phosphatase are described elsewhere [Draber et al., 19881. The monoclonal antibody TU-01 was used as an ascitic fluid diluted 1 :1,000; the monoclonal anti y-tubulin antibodies were used as undiluted supernatants. Isoelectric points of the monoclonal antibodies were determined by isoelectric focusing in a horizontal thin layer with the help of PI markers (Pharmacia LKB Biotechnology AB, Bromma, Sweden). lmmunofluorescence Extraction and fixation steps were carried out in a microtubule stabilizing buffer (MSB) consisting of 0.1 M KMes, 2 mM EGTA, 2 mM MgCl,, 4% polyethylene glycol 6000, pH 6.9. Fixed and unfixed cytoskeletons were prepared as described [Draber et al., 19891. A brief description of the fixation protocols, with abbreviations in parentheses, follows. Cells grown on coverslips were fixed for 10 min in methanol at -20°C followed by 6 min in acetone at -20°C (MIA). Cells were also directly fixed in methanol for 10 min at -20°C (M) or in acetone for 10 rnin at -20°C (A). Alternatively cells were extracted for 1 min with 0.2% Triton X-100 at 37°C and fixed for 20 min in 3% formaldehyde at the same temELISA perature (Tx/F). Cells were also fixed for 30 rnin in 3% Quantitative ELISA with monoclonal antibodies formaldehyde before extraction for 4 rnin with 0.5% Triwas performed essentially according to Landsdorp et al. ton X-100 at 37°C (F/Tx); or cells were fixed at 37°C for 119801 with 10 pgiml of BSA with coupled (C)434-449 3 rnin in 0.25% glutaraldehyde/0.5% Triton X-100 mixpeptide adsorbed on 96-well plastic plates (NUNC, ture, followed by 20 rnin incubation at the same temperRoskilde, Denmark). Competitive ELISA was per- ature in 1% glutaraldehyde and then three times (5 min formed as described [Grimm et al., 19871 using HPLC- each) in 0.5 mg/ml NaBH, (G/Tx). Cells fixed by Tx/F pure peptides, tubulin or BSA and prediluted superna- procedure were also postfixed by cold methanol (Tx/F/ tants. Values for the maximal optical density were M). To prepare unfixed cytoskeletons, cells were exbetween 0.8 and 0.9. Antibodies were mixed with vari- tracted for 1 rnin in 0.2% Triton X-100 at 37°C in the able concentrations of peptides, phosphocellulose-puri- presence of 10 pM taxol (unfixed). fied porcine brain tubulin or BSA and incubated for 60 All antibody dilutions were made with 2% BSA in rnin at room temperature. The remaining antibody bind- PBS; all subsequent incubations were at room temperaing activity was detected in ELISA using peptide immo- ture. Fixed cytoskeletons were incubated for 45 rnin with bilized on carrier. Bound antibodies were detected with mouse antibodies (undiluted hybridoma supernatants, asanti-mouse antibody conjugated with horseradish perox- citic fluids diluted 1:500), washed (three times, 5 min idase and o-phenylene-diamine as chromogen. Optical each) in PBS, incubated for 45 min with FITC-conjudensity at 490 nm was measured with a Microelisa Mini gated anti-mouse Ig antibody (dilution 1 :30) and washed Reader (Dynatech Laboratories, Alexandria, VA). Three again. In experiments on unfixed cytoskeletons, cover- y-Tubulin in Taxol-Treated Mitotic Cells slips were incubated for 60 min at 37°C with primary antibodies and fixed in formaldehyde [DrAber et al., 19891. For double-label fluorescence with two mouse monoclonal antibodies, the remaining binding sites on the FITC-conjugated antibody were blocked by 30-min incubation with normal mouse serum diluted l:lO, prior to 30-min incubation with rhodamine-conjugated antitubulin TU-04 antibody diluted 1:50 [Drhberova and DrAber, 19931. For double-label immunofluorescence with affinity-purified rabbit antibody against y-tubulin and mouse monoclonal antibodies, the coverslips were first incubated with polyclonal antibody (dilution 1: 10) and after washing with monoclonal antibody. Slides were thereafter incubated simultaneously with TRITCconjugated anti-rabbit Ig antibody and FITC-conjugated anti-mouse Ig antibody. The slides were incubated 10 min with HOECHST 33258 dye (Serva Feinbiochemica, Heidelberg, Germany) at a concentration of 1 pg/ml in PBS to detect DNA, mounted in MOWlOL 4-88 (Calbiochem AG, Lucerne, Switzerland) and examined with a Leitz Orthoplan microscope equipped with 50/1 .OO Fluorescence water-immersion objective, epi-illumination using the filter combination cubes N2, I2 and D, and an Orthomat 35 mm camera. Photographs were taken on Kodak TriXpan 400 film and printed on hard paper. Neither the control antibody HTF-14 nor the conjugates alone gave any specific staining. Tracing of prints prepared from overexposed negatives helped to place the arrows and arrowheads into identical positions in each pair of prints. RESULTS Characterization of Antibodies Against y-Tubulin In order to have specific and standard reagents for detection of y-tubulin, mouse monoclonal antibodies against phylogenetically conserved peptide from the C-terminal region of the molecule were prepared. The antibodies TU-30 (IgG2b, PI 7.8-8.4), TU-31 (IgG2b, PI 7.6-8.2) and TU-32 (IgG1, pI 6.6-7.0) reacted specifically in ELISA with corresponding peptide bound to the carrier and decorated centrosomes in methanol/acetone-fixed A-431 cells that were used for screening. In whole-cell lysates of various cultured animal cell lines, the monoclonal antibodies reacted only with a protein of relative mobility corresponding to 48 kDa polypeptide and did not react with other proteins present in the whole-cell lysate. Example of immunoblotting on 3T3 cell lysate is shown in Figure 1. The antibodies reacted in immunoblotting neither with high-speed extracts of 3T3 cells nor with porcine brain microtubule protein prepared by repeated cycles of polymerization-depolymerization (not shown). The specificity of monoclonal antibodies was further confirmed by competitive ELISA. The anti- 41 Fig. 1. Immunoblot of total extract of 3T3 cells with monoclonal antibodies against y-tubulin. Lane 1: Coomassie Blue staining. Lanes 2-5: Immunoreactivity with antibodies TU-01 against a-tubulin, TU30, TU-31 and TU-32. Proteins were separated on 5% to 15% linear polyacrylamide gel. Lanes 1-2 contained 35 p g of protein; lanes 3-5 contained 90 p g of protein. Bars on left margin indicate position, from top to bottom, of specific molecular mass markers (205 kDa, 116 kDa, 97.4 kDa, 66 kDa, 45 kDa, 29 ma). bodies were strongly inhibited by C-terminal peptide EYHAATRPDYISWGTQ of y-tubulin that was used for immunization; however, N-terminal peptide EEFATEGTDRKDVFFY of y-tubulin as well as C-terminal peptide GEEEGEEY of tyrosinated a-tubulin did not exert any inhibitory effect at the highest concentration tested (10 pM). Similarly, purified brain tubulin and control bovine serum albumin were without effect at this concentration. Figure 2 shows a typical example of competitive ELISA with antibody TU-30. The concentrations of C-terminal peptide giving 50% inhibition of binding for antibodies TU-30, TU-31 and TU-32 were 5 , 15 and 4 nM, respectively. Monoclonal antibodies decorated during the cell cycle all typical y-tubulin-containing structures. Figure 3 shows triple-label staining with anti-a-tubulin antibody, anti-y-tubulin antibody and fluorescent DNA-binding dye in 3T3 cells. The anti-y-tubulin antibodies stained centrosome within interphase and duplicated centrosomes migrating to the opposite sites of nucleus within prophase (Fig. 3a-c), spindle poles and two half-spindles within metaphase (Fig. 3d-f) and microtubule bundles forming midbodies within telophase (Fig. 3g-i). The staining was prevented by preabsorption of antibodies with C-terminal peptide of y-tubulin, but not with N-terminal peptide of y-tubulin (not shown). The antibodies did not decorate interphase microtubules or short aster microtubules at the spindle poles. In interphase cells two 42 Novakova et al. 1.o lo-’ 100 10’ 102 103 104 Peptide concentration (nM) Fig. 2 . Comparison of the reactivity of TU-30 antibody with peptides by indirect competitive ELISA. C-terminal peptide of y-tubulin EYHAATRPDYISWGTQ (a),N-terminal peptide of y-tubulin EEFATEGTDRKDVFFY (m). adjacent spots could be detected reflecting probably centrosomes after replication. The monoclonal antibodies reacted with centrosomes in all tested cell lines of various species and different tissue origin. The following cell lines were immunostained: human A-431, AMA and HeLa cells, mouse 3T3, Neuro-2a and P19/X1 cells, rat kangaroo PtK, cells and turkey TC-3 1. The cell lines differed in the size of y-positive dots in interphase cells. For example, in PtK2 cells the antibodies decorated large dots (Fig. 4a-C) and in Neuro2a smaller dots (Fig. 4d-f) in comparison with interphase 3T3 cells. In Neuro 2a, cultivated under conditions when long processes were observed, y-tubulin was detected only in cell bodies and not in the projections (Fig. 4d,e). Testing of the accessibility of antigenic determinants recognized by monoclonal antibodies in animal cells prepared under various fixation conditions and in unfixed cells revealed that cold methanol/acetone provided the best visualization of y-tubulin. This fixation was therefore routinely used in immunofluorescence experiments. When cells were extracted by Triton X-100 and fixed by formaldehyde or glutaraldehyde, the antibodies did not decorate centrosomes in indirect immunofluorescence with anti-mouse antibody conjugated with FZTC. The same results were obtained in unfixed, detergent-extracted cells. When cytoskeletons prepared by aldehyde fixations were postfixed by cold methanol, the y-tubulin was available for binding of antibodies. Affinity-purified polyclonal antibody against crp-tubulin heterodimer decorated microtubular structures in all fixed cytoskeleton preparation as well as in unfixed cells. The effect of fixation on binding the monoclonal antibodies is summarized in Table I. Distribution of y-Tubulin in Nocodazole-TreatedCells When 3T3 cells were incubated with 10 p M nocodazole for 1 , 4 , 18, or 24 hours and thereafter released from the nocodazole treatment for 3 min or 6 min before fixation, a regrowth of microtubules from centrosomes was observed. Triple-label staining showed that y-tubulin was limited to the centers of microtubule asters as shown in Figure 5a-c. In a small number of cells (less than 1 % in cell population) as many as 8 microtubule nucleation sites were found and y-tubulin was again located in centers of all microtubule asters. Figure 5d-f shows the beginning of microtubule regrowth in such cells 3 min after nocodazole release. No substantial differences in the intensity of y-tubulin immunofluorescence was observed among individual nucleating sites. Distribution of y-Tubulin in Taxol-Treated Mitotic Cells Taxol, a microtubule stabilizing drug, causes in interphase cells a rearrangement of microtubules into bundles and in mitotic cells induces spindle disruption y-Tubulin in Taxol-Treated Mitotic Cells Fig. 3. Immunofluorescence triple-label staining of 3T3 cells in various stages of the cell cycle with antibody TU-30 against y-tubulin. Cells were stained with anti-a-tubulin antibody TU-04 (a, d, g), an- 43 tibody TU-30 (b, e, h) and DNA-binding dye (c, f, i). The cells in(a-c) are in interphase and prophase; the cell in (d-f) is in metaphase, and the cell in (g-i) is in telophase. Bar = 10 pm. and assembly of a new microtubule asters [De Brabander was thus limited to only a few asters in one mitotic cell. et al., 19861. 3T3 cells treated with 10 pM taxol for 18 In the remaining asters no y-tubulin was detected using hours were characterized by multiple nuclei and by typ- immunofluorescence microscopy. The y-tubulin was ofical microtubule arrangements (Fig. 6). Triple-label ten found in the periphery of the microtubule asters as staining with anti-a-tubulin antibody, anti-y-tubulin an- demonstrated on distribution of y-tubulin in the cell with tibody and DNA-binding dye showed that in interphase multiple microtubule asters (arrows in Fig. 6g, h). The cells y-tubulin was distributed randomly and usually out- y-tubulin was also found outside the center of microtuside of microtubule bundles (Fig. 6a-c). In cells that bule asters in cells treated with taxol for a shorter ( 1 and entered mitosis, y-tubulin was located in the area where 4 hours) or longer (24 hours) time. This finding indicates induced aster-like aggregates of short microtubules were that in taxol-treated cells y-tubulin does not nucleate the found (arrows in Fig. 6d, e). Some 4-20 microtubule formation of microtubule asters. asters were found in a single cell. Antibodies against Differential localization of microtubule asters and y-tubulin, however, stained only 1-4 spots in such cells. y-tubulin in taxol-treated mitotic cells was not limited The localization of y-tubulin into the microtubule asters just to 3T3 cells. Similar distribution of y-tubulin was 44 Novakova et al. Fig. 4. Immunofluorescence triple-label staining of PtK, cells and Neuro-2a cells with antibodies TU-3 1 and TU-32 against y-tubulin. Rat kangaroo kidney PtK, cells (a-c) and mouse neuroblastoma cells Neuro-2a (d-f) were stained with anti-a-tubulin antibody TU-04 (a, d), antibody TU-31 (b), TU-32 (e) and DNA-binding dye (c, f ) . Arrows indicate the same positions. Bar = 10 pm. TABLE I. Effect of Fixation Procedure on Reactivity of Antibodies With Centrosomes of 3T3 Cells Fixation Drotocol" Antibody TU-30 TU-31 TU-32 PoIyclonaI MIA' ++d ++ ++ + M A + + + + + + + + TxiF FlTx GlTx - - - - - - - - ++ ++ ~ ++ TxlFIM + + + ++ Unfixed - ++ aDescription of fixation protocols is in Materials and Methods. bStaining of microtubular structures with affinity-purified polyclonal antibody against ap-tubulin heterodimer. 'M, methanol; A, acetone; Tx, Triton X-100; F, formaldehyde; G , glutaraldehyde. d+ , strong; + , moderately strong; -, negative. + also observed in taxol-treated mitotic HeLa and PtK, cells. Figure 7 demonstrates the results of triple-label staining of HeLa cells with anti-a-tubulin antibody, antiy-tubulin antibody and DNA-binding dye. In all taxoltreated mitotic cells the number of microtubule asters exceeded that of y-tubulin-positive spots (arrows in Fig. 7a, b). At higher magnification, y-tubulin did not seem to be located in the center of the asters (arrows in Fig. 7d, e). Triple-label staining of taxol-treated mitotic 3T3 cells was performed with anti-y-tubulin antibody, the MPM-2 antibody recognizing centrosomal phosphopro- y-Tubulin in Taxol-Treated Mitotic Cells 45 Fig. 5 . The distribution of y-tubulin in 3T3 cells recovering from nocodazole arrest. Cells were treated with 10 pM nocodazole for 18 hours, fixed after a 6-min ( a x ) or 3-min (d-f) incubation in medium without nocodazole and triple-label-stained with anti-a-tubulin antibody TU-04 (a), anti-y-tubulin antibody TU-30 (b) and DNA-binding dye (c). Bar = 10 pm. tein epitope that is known to increase substantially in mitosis [Davis et al., 19831, and DNA-binding dye in order to compare the locations of corresponding antigens in microtubule asters. Association of MPM-2 epitope with spindle poles in untreated mitotic cells is shown in Figure 8a-c. In taxol-treated mitotic cells MPM-2 staining was located in the centers of some microtubule asters only (arrows in Fig. 8d, e) and its distribution was similar to that of y-tubulin. To see whether these centrosoma1 antigens codistribute in taxol-treated mitotic cells, triple-label staining was performed with polyclonal affinity-purified anti-y-tubulin antibody, MPM-2 antibody and DNA-binding dye. In all tested samples the MPM-2 epitope colocalized with y-tubulin staining in mitotic cells (arrows in Fig. 8g, h). However, in taxol-treated interphase cells only y-tubulin was detected (arrowheads in Fig. 8g, h). The colocalization of MPM-2 antigen and y-tubulin in taxol-treated mitotic cells was observed also in PtK, cells (not shown). The results shown in Figure 8 indicate that the two centrosomal antigens are not associated with microtubule aster formation induced by taxol in vivo. DISCUSSION y-Tubulin has been shown to be a ubiquitous component of MTOC and it has been proposed that it plays a key role in the nucleation of microtubule arrays in vivo [Joshi, 19931. As y-tubulin has not been purified in functional form, in vitro evidence on direct association of y-tubulin with ap-heterodimer in the course of microtubule nucleation is still missing. Immunofluorescence studies with polyclonal antibodies revealed that y-tubulin is present not only in MTOC but also in mitotic spindle [Lajoie-Mazenc et al., 19941 and in midbody of mammalian cells during cytokinesis [Julian et al., 19931. The localization and timing of y-tubulin during the cell cycle is probably highly regulated and its physiological role could be more diverse than initially assumed [LajoieMazenc et al., 19941. Recent data demonstrated that y-tubulin was required not only for nucleation of cytoplasmic microtubules but also for formation of mature centrosomes in cell-free systems [Felix et al., 1994; Stearns and Kirschner, 19941 and for the formation of normal MTOC in Drosophila [Sunkel et al., 19951. Here 46 Novakova et al. Fig. 6. The distribution of y-tubulin in taxol-treated mitotic 3T3 cells. Cells were incubated for 18 hours with 10 (*M taxol, fixed and triplelabel-stained with anti-a-tubulin antibody TU-04 (a, d, g), anti- y-tubulin antibody TU-30 (b, e, h) and DNA-binding dye (c, f, i). The interphase cell is in a<, mitotic cells are in d-i. Arrows in each pair of prints are placed in identical position. Bar = 10 kni. we report on redistribution of y-tubulin detected by monoclonal antibodies in taxol-treated mitotic cells. Monoclonal antibodies were prepared against synthetic peptide from the carboxyterminal region of human y-tubulin and the specificity of antibodies was confirmed by double immunofluorescence with affinity-purified polyclonal antibody against y-tubulin, by immunoblotting and competitive ELISA. In 3T3 cells the antibodies stained centrosomes within interphase and spindle poles, half-spindles and midbodies within mitosis. No staining of interphase microtubules was detected. During mitosis the amount of y-tubulin observed in spindle poles was much higher than the amount detected in the interphase centrosomes. Antibodies decorated only half-spindles which participate in the structure of spindles and not astral microtubules emanating from spindle poles. Preincubation of the antibodies with peptide used for immunization precluded the labelling of all above mentioned structures. However, preincubation of antibodies with peptide corresponding to the aminoterminal region of y-tubulin or with purified porcine brain ap-tubulin heterodimer did not affect the staining. The antibodies therefore do not cross-react with tubulin heterodimer; this finding is in agreement with the results of immuno- y-Tubulin in Taxol-Treated Mitotic Cells 47 Fig. 7. The distribution of y-tubulin in taxol-treated mitotic HeLa cells. Cells were incubated for 18 hours with 10 p M taxol, fixed and triple-label-stained with anti-a-tubulin antibody TU-04 (a, d), anti-y-tubulin antibody TU-30 (b, e) and DNA-binding dye (c, 0 . Arrows in each pair of prints are placed in identical position. Bar = 10 pm. blotting on whole-cell lysates and competitive ELISA. The observed immunofluorescence staining pattern is in accordance with the described staining of interphase centrosomes [Stearns et al., 1991; Zheng et al., 1991; Joshi et al., 19921, half-spindles [Lajoie-Mazenc et al., 19941 and midbodies [Julian et al., 19931 with affinity-purified polyclonal antibodies against different y-tubulin sequences. Failure to detect y-tubulin in half-spindles and midbodies in previous immunological studies with polyclonal antibodies [Stearns et al., 1991; Zheng et al., 1991; Joshi et al., 19921 could reflect lower titers of used antibodies. The monoclonal antibodies were raised against phylogenetically highly conserved peptide and stained centrosomes, mitotic spindles and midbodies in all tested cell lines from evolutionally distant species including avian cells. Antibodies also immunostained cultured plant cells from Nicotiana tabacum (A. Smertenko and P. DrBber, unpublished data). In mouse neuroblastoma cells Neuro2a-cultured for several days in serum-free medium, long projections were observed. These were stained with antibody against a-tubulin, but not by monoclonal anti-y-tubulin antibodies. As in non- neuronal cells, y-tubulin was located only in pericentriolar region. We thus confirm on a different model the previous finding of Baas and Joshi [ 19921, obtained on rat sympathetic neurons by immunoelectron microscopy with polyclonal anti-y-tubulin antibody. To determine whether the exposure of antigenic determinants, recognized by corresponding antibodies, is dependent on fixation procedure, different fixation protocols were applied to interphase 3T3 cells. We observed that centrosome staining with monoclonal antibodies as well as with affinity-purified polyclonal antibody against y-tubulin (not shown) was fixation-dependent (Table I). In preparation fixed only in formaldehyde (30 min) or glutaraldehyde (20 min) centrosomes were not detected. Neither was staining observed in unfixed detergent-extracted cells. Centrosomes were, however, visualized after cold methanol or acetone treatment or in cell fixed with formaldehyde and postfixed with methanol or acetone. On the other hand anti-a-tubulin monoclonal antibodies of IgG as well as IgM class or affinity-purified antibody against a@-tubulin heterodimer stained cytoplasmic microtubules in all tested fixation protocols. 48 Novakova et al. Fig. 8. The distribution of MPM-2 antigen in 3T3 cells. Untreated cells (a<) or cells incubated 18 hours with 10 FM taxol (d-i) were fixed and triple-label-stained with anti-a-tubulin antibody TU-04 (a, d), MPM-2 antibody (b, e) and DNA-binding dye (c, f ) . Localization of MPM-2 antigen and y-tubulin in taxol-treated cells was visualized by triple-label staining with polyclonal affinity-purified anti-y-tubulin antibody (g), MPM-2 antibody (h) and Hoechst 33258 (i). Arrows and arrowheads indicate the same position in mitotic or interphase cells, respectively. Bar = 10 Fm. Similarly, microtubules were stained with polyclonal antibody to brain tubulin in detergent-extracted cells. These data indicate that changes in y-tubulin conformation or changes in distribution of proteins associated with y-tubulin are necessary for exposure of the carboxyterminal region of the molecule for binding of antibodies. Differential exposure of tubulin epitopes on cytoplasmic microtubules was previously described in fixed or in detergent-extracted cells [DrBber et al., 19891. In nocodazole-treated 3T3 cells, y-tubulin was located at microtubule nucleating sites in cells recovering from the nocodazole treatment. In a small number of cells, multiple nucleation sites were observed, all of them y-tubulin-positive. Multiple nucleation sites could result from centrosomal duplication or centrosomal fragmentation after a prolonged nocodazole treatment. A diminished y-tubulin staining of some nucleation sites, reflecting centrosome fragmentation, was described in the nocodazole-treated Xenopus XTC cells [Stearns et al., 19911. However, we observed an identical staining of all y-tubulin-positive sites in the cells treated with nocodazole for various time intervals. In contrast to cells recovering from the nocodazole treatment, y-tubulin in the taxol-treated mitotic cells was usually located outside microtubule asters. Such a localization of y-tubulin was found in all tested cultured animal cell lines. In tdxol-treated cells the localization of y-tubulin was identical with that of the centrosomal y-Tubulin in Taxol-Treated Mitotic Cells phosphoprotein epitope recognized by the MPM-2 antibody [Davis et al., 19831 in mitotic cells. In all tested cases the y-tubulin staining appeared to be identical with that of MPM-2. These results demonstrate that in tested cell lines y-tubulin is not involved in nucleation of microtubule asters free of centrosomes or is only present in quantities undetectable by immunofluorescence. Previously it has been shown that also another marker of pericentriolar matrix, 5051 antigen detected by human autoimmune serum [Calarco-Gillan et al., 19831, was not located in centers of taxol-induced microtubule asters [Kallajoki et al., 19921. It has been proposed that in taxol-induced microtubule asters the minus ends of microtubules were situated at the aster centers [De Brabander et al., 1986; Maekawa et al., 19911. In this respect taxol asters resemble the centrosomes. Several proteins are known to locate to the centers of taxol-induced microtubule asters and they could thus be directly or indirectly involved in microtubule nucleation. These include the 0013 antigen [Gosti-Testu et al., 19861, CTR2611 antigen [Buendia et al., 19901, calmodulin [De Brabander et al., 19861, and nuclear protein that associates with the mitotic apparatus and that was described under different names such as NuMA, SPN, SP-H, IH1 antigen, W1, centrophilin [for review see Compton and Cleveland, 19941. It was proposed that NuMA might act as a microtubular minus-end organizer in mitotic cells in vivo [Kallajoki et al., 1992; Maekawa et al., 19911. Microtubules are nucleated in the absence of centrioles in mammalian oocytes and early embryos, in contrast to adult somatic cells. It has been shown that in mouse oocytes y-tubulin was associated both with spindle poles and cytoplasmic MTOC [Gueth-Hallonet et al., 1993; Palacios et al., 19931. In taxol-treated oocytes y-tubulin was found in the center of many cytoplasmic asters induced by taxol [Gueth-Hallonet et al., 19931. The 505 1 antigen, another component of pericentriolar matrix, was also located to the center of taxol-induced microtubule asters in mouse oocytes [Maro et al., 19851. Early mouse embryos contain a large pool of soluble y-tubulin and it is possible that some of the material found in the asters was recruited from a soluble stock of molecules [Gueth-Hallonet et al., 19931. However, in PtK, cells 5051 antigen [Kallajoki et al., 19921 and y-tubulin (this work) were usually found outside taxol-induced microtubule asters. This suggests an involvement of different sets of proteins in the formation of taxolinduced microtubule asters in oocytes and in differentiated cells. It has been shown that taxol induced the formation of microtubule asters in cell-free extracts of Xenopus eggs [Verde et al., 19911. Aster assembly requires phosphorylation and it is dependent on the mitotic state. The 49 asters do not grow from preformed nucleation centers but microtubules aggregate and induce the reorganization of pericentriolar material through the action of the minus end-directed motor cytoplasmic dynein. A similar effect in mitotic Xenopus egg extract has dimethyl sulfoxide (DMSO) [Stearns and Kirschner, 19941. The y-tubulin as well as MPM-2 antigen(s) were localized to the centers of DMSO or taxol-induced asters [Stearns and Kirschner, 19941. This implied that y-tubulin was able to associate with the minus end of microtubules made in solution from mitotic extracts. Our finding that y-tubulin is not associated with taxol-induced asters in mitotic cells of various cultured cell lines could reflect the differences between the used models, i.e., animal somatic cells, versus Xenopus eggs. Alternative explanations would imply that the formation of taxol-induced asters in vivo is a more complicated process. In conclusion, our data indicate that in taxol-treated mitotic cells microtubule asters are nucleated independently of y-tubulin or that y-tubulin is present in amounts undetectable by immunofluorescence microscopy. Other minus-end nucleator(s) are probably necessary for generation of taxol-induced microtubule asters in vivo in cultured somatic cell lines. Prepared monoclonal antibodies could be a useful tool for further elucidation of y-tubulin function and regulation in different cell models. ACKNOWLEDGMENTS Taxol was a generous gift of the Drug Synthesis and Chemistry Branch, Division of Cancer Treatment, National Cancer Institute (Bethesda, MD). We thank Drs. P. Rao and J. Kuang for a gift of MPM-2 antibody. This work was supported in parts by grants 552407 and 752101 from the Czech Academy Grant Agency and by a grant from the Austrian Ministerium fur Wissenschaft und Forschung. REFERENCES Baas, P.W., and Joshi, H.C. (1992): y-Tubulin distribution in the neuron: Implication for the origins of neuritic microtubules. J. Cell Biol. 119:171-178. Brandtzaeg, P. (1973): Conjugates of immunoglobulin G with different fluorochromes. I. Characterization by anionic exchange chromatography. Scand. J. Immun., Suppl. 2:273-290. Brinkley, B.R. (1985): Microtubule organizing centers. Annu. Rev. Cell Biol. 1:145-172. Buendia, B., Antony, C., Verde, F., Bornens, M., and Karsenti, E. (1990): A centrosomal antigen localized on intermediate filaments and mitotic spindle poles. J. Cell Sci. 97:259-271. Calarco-Gillan, P.D., Siebert, M.C., Hubble, R., Mitchison, T., and Kirschner, M. (1983): Centrosome development in early mouse embryos as defined by an autoantibody against pericentriolar material. Cell 35:621-629. Compton, D.A., and Cleveland, D.W. (1994): NuMA, a nuclear pro- 50 Novakova et al. kin involved in mitosis and nuclear reformation. Cum. Opinion Cell Biol. 6:343-346. Davis, F.M., Tsao, T.Y., Fowler, S.K., and Rao, P.N. (1983): Monoclonal antibodies to mitotic cells. Proc. Natl. Acad. Sci. U. S ,A. 8012926-2930, De Brabander, M., Geuens, G., Nuydens, R., Willebrords, R., Aerts, F., and De Mey, J. (1986): Microtubule dynamics during the cell cycle: The effects of taxol and nocodazole on the microtubule system of PtK2 cells at different stages of the mitotic cycle. Int. Rev. Cytol. 101:215-274. Draber, P. (1991): Quantitation of proteins in sample buffer for sodium dodecyl sulfate-polyacrylamide gel electrophoresis using colloidal silver. Electrophoresis 12:453-456. Drabber, P., Lagunowich, L.A., Driberova, E., Viklickf , V., and Damjanov, I. (1988): Heterogeneity of tubulin epitopes in mouse fetal tissues. Histochemistry 89:485-492. Drfiber, P., Draberovi, E., Linhartovfi, I., and Viklickf, V. (1989): Differences in the exposure of C- and N-terminal tubulin domains in cytoplasmic microtubules detected with domain-specific monoclonal antibodies. J. Cell Sci. 92519-528. Draber, P., Driberova, E., and Viklickf, V. (1991): Immunostaining of human spermatozoa with tubulin domain-specific monoclonal antibodies. Histochemistry 195519-524. Driberovl, E., and Draber, P. (1993): A microtubule-interacting protein involved in coalignment of vimentin intermediate filaments with microtubules. J. Cell Sci. 106:1263-1273. Fklix, M.A., Antony, C., Wright, M., and Maro, B. (1994): Centrosome assembly in vitro: Role of y-tubulin recruitment in Xenopus sperm aster formation. J. Cell Biol. 124:19-31. Gosti-Testu, F., Marty, M.-C., Berges, J., Maunoury, R., and Bornens, M. (1986): Identification of centrosomal proteins in a human lymphoblastic cell line. EMBO J. 5:2545-2550. Grimm, M., Breitling, F., and Little, M. (1987): Location of the epitope for the alpha-tubulin monoclonal antibody TU-01. Biochim. Biophys. Acta 914:83-88. Gueth-Hallonet, C., Antony, C., Aghion, J., Santa-Maria, A,, Lajoie-Mazenc, I., Wright, M., and Maro, B. (1993): y-Tubulin is present in acentriolar MTOCs during early mouse development. J. Cell Sci. 105:157-166. Joshi, H.C. (1993): y-Tubulin: The hub of cellular microtubule assemblies. BioEssays 15:637-643. Joshi, H .C., Palacios, M. J., McNamara, L . , and Cleveland, D. W. (1992): y-Tubulin is a centrosomal protein required for cell cycle-dependent microtubule nucleation. Nature 356:80-83. Julian, M., Tollon, Y . , Lajoie-Mazenc, I., Moisand, A., Mazarguil, H., Puget, A,, and Wright, M. (1993): y-Tubulin participates in the formation of midbody during cytokinesis. I . Cell Sci. 105:145-156. Kallajoki, M., Weber, K., and Osborn, M. (1992): Ability to organize microtubules in taxol-treated mitotic PtK2 cells goes with the SPN antigen and not with the centrosome. J. Cell Sci. 102:91102. Laemmli, U.K. (1970): Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680685. Lajoie-Mazenc, I., Tollon, I., Detraves, C., Julian, M., Moisand, A., Gueth-Hallonet, C., Debec, A., Salles-Passador, I., Puget, A,, Mazarguil, H., Raynaud-Messina, B., and Wright, M. (1994): Recruitment of antigenic gamma-tubulin during mitosis in animal cells: Presence of gamma-tubulin in the mitotic spindle. J. Cell Sci. 107:2825-2837. Landsdorp, P.M., Astaldi, G.C.B., Oosterhof, F., Janssen, M.C., and Zeijlemakers, W.P. (1980): Immunoperoxidase procedures to detect antibodies against cell surface antigens. Ouantitation v ” . of binding and staining of individual cells. J. Immunol. Methods 39:393-405. Maekawa, T., Leslie, R., and Kuriyama, R. (1991): Identification of a minus end-specific microtubule-associated protein located at the mitotic poles in cultured mammalian cells. Eur. J. Cell Biol. 54:255-267. Maro, B., Howlett, S . K . , and Webb, M. (1985): Non-spindle microtubule organizing centers in metaphase 11-arrested mouse oocytes. J. Cell Biol. 101:1665-1672. Memfield, R.B. (1963): Solid phase peptide synthesis. I. The synthesis of a tetrapeptide. J. Am. Chem. SOC.85:2149-2154. Muresan, V., Joshi, H.C., and Besharse, J.C. (1993): y-Tubulin in differentiated cell types: Localization in the vicinity of basal bodies in retinal phosphoreceptors and ciliated epithelia. J. Cell Sci. 104:1229-1237. Oakley, B.R. (1994): y-Tubulin. In Hyams, J.S., and Lloyd, C.W. (eds): “Microtubules.” New York: Wiley-Liss, Inc., pp. 3345. Oakley, C.E., and Oakley, B.R. (1989): Identification of y-tubulin, a new member of the tubulin superfamily encoded by mipA gene of Aspergillus nidulans. Nature 338:662-664. Oakley, B.R., Oakley, C.E., Yoon, Y., and Jung, M. (1990): y-Tubulin is a component of the spindle pole body that is essential for microtubule function in Aspergillus niduluns. Cell 61: 1289-1301. Palacios, M.J., Joshi, H.C., Simerly, C., and Schatten, G . (1993): y-Tubulin reorganization during mouse fertilization and early development. J. Cell Sci. 104:383-389. Pickett-Heaps, J.D. (1969): The evolution of the mitotic apparatus: An attempt at comparative ultrastructural cytology in dividing plant cells. Cytobios 3:257-280. Ponstingl, H., Krauhs, E., Little, M., and Kempf, T. (1981): Complete amino acid sequence of a-tubulin from porcine brain. Proc. Natl. Acad. Sci. U.S.A. 78:2757-2761. Raff, J.W., Kellog, D.R., and Alberts, B.M. (1993): Drosophilu gamma-tubulin is part of a complex containing two previously identified centrosomal MAPS. J. Cell Biol. 1212323-835. Steams, T., and Kirschner, M. (1994): In vitro reconstitution of centrosome assembly and function: The central role of y-tubulin. Cell 76:623-637. Steams, T., Evans, L., and Kirschner, M. (1991): y-Tubulin is highly conserved component of the centrosome. Cell 65:825836. Sunkel, C.E., Gomes, R., Sampaio, P., Perdigao, J., and Gonzfilez, C. (1995): y-Tubulin is required for the structure and function of the microtubule organizing centre in Drosophila neuroblasts. EMBO J. 14:28-36. Szollosi, D., Calarco, P., and Donahue, R.P. (1972): Absence of centrioles in the first and second meiotic spindles of mouse oocytes. J. Cell Sci. 11:521-541. Towbin, H., Stdehelin, T., and Gordon, J. (1979): Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: Procedure and some applications. Proc. Natl. Acad. Sci. U.S.A. 76:4350-4354. Verde, F., Berrez, J.-M., Antony, C., and Karsenti, E. (1991): Taxolinduced microtubule asters in mitotic extracts of Xenopus eggs: Requirement for phosphorylated factors and cytoplasmic dynein. J. Cell Biol. 112:1177-1187. Viklicky, V., Drabber, P., HaSek, J., and Bartek, J. (1982): Production and characterization of a monoclonal antitubulin antibody. Cell Biol. Int. Rep. 6:725-731. Viklickf, V., Bartek, J., Verlovi, H., Draber, P., and HoiejSi, V. (1983): Mapping of antigenic sites on human transferrin by monoclonal antibodies. In Boss, B.D., Langman, R., Trow- y-Tubulin in Taxol-Treated Mitotic Cells bridge, I., and Dulbecco, R. (eds): “Monoclonal Antibodies and Cancer.” New York: Academic Press, pp. 261-264. Viklickf, V . , Drhber, P., Bliha, I., and Linhartovh, I. (1988): Distribution of tyrosinated and detyrosinated a-tubulin revealed by mouse anti-peptide polyclonal antibodies. Fol. biol. (Praha) 34~380-389. 51 Vorobjev, I.A., and Nadezhdina, E.S. (1987): The centrosome and its role in the organization of microtubules. Int. Rev. Cytol. 106: 227-293. Zheng, Y., Jung, M.K., and Oakley, B.R. (1991): Gamma-tubulin is present in Drosophila melanogaster and Homo sapiens and is associated with centrosome. Cell 6 5 9 17-823.