A centrosomal antigen localized on intermediate filaments and mitotic spindle poles BRIGITTE BUENDIA1, CLAUDE ANTONY2, FULVIA VERDE1, MICHEL BORNENS3 and ERIC KARSENTI1 1 European Molecular Biology Laboratory, D-6900 Heidelberg, Federal Republic of Germany lnstitut J. Monod, CNRS, Paris VII, Tour 43, 2 place Jussieu, 75005 Paris, France ^Centre de Ginitique MoUculaire du CNRS, 2 avenue de la terrasse, 91190 Gifsur Yvette, France 2 Summary A monoclonal antibody (CTR2611) raised against centrosomes isolated from human lymphocytes (KE37) stains the pericentriolar material and intermediate filaments in the same cells. In MDCK cells, where most of the microtubules do not originate from the pericentriolar region during interphase, the antigen is distributed along intermediate filaments. At the onset of mitosis, a large fraction of the CTR2611 antigen associates with the minus-end domain of the microtubules of the mitotic spindle but not with the pericentriolar region itself. Treatment of mitotic MDCK cells with taxol leads to the assembly of many microtubule asters in the cytoplasm at the expense of the mitotic spindle. The CTR2611 antigen is present in the center of each of these asters. Similar asters can also be produced in vitro by adding taxol to concentrated Xenopus egg mitotic cytoplasm. Again, the antigen is found close to the center of the asters. These results suggest that CTR2611 antigen is associated with a material involved in microtubule nucleation or microtubule minus-end stabilization. The monoclonal antibody recognizes a 74xlO3Mr polypeptide and other polypeptides at 120xl03Mr and 170xl03Mr. The 74xlO3Mr polypeptide is found in all species examined so far, suggesting that it contains a highly conserved epitope. Introduction centrosomal antigen recognized by a monoclonal antibody (CTR2611) raised against a preparation of human lymphoid cell centrosomes (Bornens et al. 1987). In KE37 lymphocytes, the cell line from which the centrosomes were purified to prepare the antibody, the centrosomal region is strongly labeled both in interphase and at mitosis. Interestingly, in MDCK cells, where microtubules do not originate from the centrioles (Br6 et al. 1987), the CTR2611 antibody does not stain the pericentriolar region. Instead, it stains intermediate filaments. Yet, during mitosis, the same antibody stains the minus-end domain of the microtubules of the spindle. As indicated by different immunoblotting techniques, the CTR2611 antibody recognizes mainly a 74 x 103 MT polypeptide in all cell lines and in centrosome preparations. In addition, it stains polypeptides at 120 and 170xl0 3 M r but their detection is less obvious. The function of this centrosomal antigen is still unclear although its localization suggests that it plays a role in the interaction between different elements of the cytoskeleton, in the nucleation of microtubules or the localization of the nucleating material during cell differentiation and the cell cycle. In fibroblasts or motile cells, most of the interphasic microtubule network is organized by the centrosome. This microtubule-organizing center (MTOC) is made of centrioles surrounded by a fibrogranular material that can assume various shapes. It is often organized into radial arms surrounding the distal end of the centriolar cylinder and nucleates microtubules (Berns and Richardson, 1977; Bergen et al. 1980; Mclntosh, 1983; Karsenti and Maro, 1986). However, in some differentiated animal cells (Tucker, 1979; Tassin et al. 1985), in all plant cells as well as in the oocytes of many species, centrioles are absent (Pickett-Heaps, 1974; Karsenti and Maro, 1986) and in some differentiated cells, although centrioles are present, they nucleate only a fraction of the microtubules (Bre et al. 1987). In all these cases, microtubules are thought to be nucleated by pericentriolar material that has lost its association with centrioles (Tassin et al. 1985; Karsenti and Maro, 1986; Bre et al. 1987). We still know very little about this material, how its interaction with centrioles is regulated, how it is localized in the cell and how nucleation actually works. Nevertheless, the number of identified centrosomal proteins is growing (Gosti-Testu et al. 1986; Salisbury et al. 1986; Baron and Salisbury, 1988; Toriyama et al. 1988; Keryer et al. 1989) and this should help to shed some light on these questions. In this paper, we report on the characterization of a Journal of Cell Science 97, 259-271 (1990) Printed in Great Britain © The Company of Biologists Limited 1990 Key words: centrosome, intermediate filaments, mitotic spindle. Materials and methods Cells The KE37 cell line of T lymphoblastic origin (Mayer et al. 1982) 259 was cultivated in suspension in RPMI1640 containing 10 % fetal calf serum, in a humidified atmosphere, equilibrated with 5% CO2 in air at 37 °C. Madin-Darby canine kidney epithelial cells (MDCK) strain II were grown in Eagle's minimal essential medium with Eagle's salts supplemented with 10 mM Hepes, pH7.3, 2mM L-glutamine, 5% FCS, penicillin (110 units ml" 1 ) and streptomycin (lOO/igml"1). The cells were seeded either on plastic Petri dishes for biochemical analysis or on glass coverslips for immunofluorescence studies, and incubated in a humidified atmosphere, equilibrated with 5% CO2 in air at 37 °C. Drug treatments Nocodazole (Sigma Chemical GmbH, Deisenhofen, Federal Republic of Germany) was used at a final concentration of 33 /avi. Taxol was used at a final concentration of 5 f(M. Antibodies The rabbit anti-tubulin was a gift from Jan De Mey, the monoclonal CTR2611 (IgM) was obtained by immunization of mice with human centrosomes purified from the KE37 cell line according to Bornens et al. (1987). Secondary antibodies for immunofluorescence experiments were Affinipure antibodies purchased from Diannova (Hamburg, West Germany). The 9 nm protein A-gold and the 10 nm gold-goat anti-mouse IgM used for the electron microscopy were obtained from Jannssen (B-2430, Olen; Belgium). Immunofluorescence KE37 cells were washed in PBS (phosphate-buffered saline) and then sedimented onto 12 mm round coverslips previously coated with polylysine. The cells attached to the coverslips were then fixed with glutaraldehyde and double labeled for CTR2611 and tubulin as described below for the MDCK cells. Simple staining for CTR2611 or double staining for CTR2611 and tubulin or for CTR2611 and actin were carried out on cells fixed in glutaraldehyde as follows: after a brief rinse (2 s) in PBS at 37 °C, the coverslips were either quickly pre-extracted (twice, 2 s) in 80 mM K-Pipes, pH 6.8, 5 mM EGTA, 1 mM MgCl2,0.5 % Triton X-100 and then fixed in the same buffer plus 0.3 % glutaraldehyde or directly fixed and post-extracted with detergent (without any significant difference in the staining). The cells were rinsed briefly in PBS and incubated twice (lOmin) at room temperature in NaBH4 (1 mgml" 1 in PBS). The cells were then washed once in PBS and for 5min in PBS containing 0.1% Triton X-100. The CTR2611 ascites fluid was delipidated, dialysed against PBS containing 50 % glycerol and used diluted 3500-fold in PBS. The rabbit antitubulin was diluted 100-fold. Coverslips were incubated for 20 min with the antibodies. After two washes of 5 min each in PBS, the secondary antibodies were added for 15 min at room temperature (fluorescein-labeled goat anti-mouse (1/100) plus a Texas Red-labeled goat anti-rabbit (1/150) or Texas Red-labeled goat anti-mouse (1/100) and a fluorescein-labeled goat anti-rabbit (1/100)). The coverslips were then washed in PBS and mounted in Moewiol (Hoechst, FRG). When necessary, the chromosomes were labeled with propidium iodide according to a procedure adapted from Krishan (1975). This staining was mainly used for the confocal microscope study. In this case, after immunostaining and the final washes, the cells were treated with RNase A (1 mgml" 1 in PBS) for 15 min at room temperature and the propidium iodide was added at O.l^gml" 1 in TBS (10 mM Tris, 150 mM NaCl) for 10 min at room temperature. The coverslips were then washed in PBS and mounted for the observation. Confocal fluorescence microscopy Confocal fluorescence microscopy was performed with the confocal scanning laser beam fluorescence microscope developed at the European Molecular Biology Laboratory. A description of the design and operating principles of the confocal fluorescence microscope has been published previously (Stelzer et al. 1989). Two wavelengths (476 and 514.5 /an) produced by an argon laser (2020-05; Spectra-Physics, Inc., Mountain View, CA) were used to excite FITC and Texas Red, respectively (see also Bacallao et al. 1989). Serial optical sections of 0.4/an were made. The stereo 260 B. Buendia et al. images shown in this paper (Figs 2, 4 and 7) are preferably viewed with stereoscopes or stereoglasses. Electron microscopy KE37 cells were pelleted and resuspended in a PHEM buffer (60 mM Pipes, 25 mM Hepes, 2 mM MgCl2 and 10 mM EGTA) plus 0.5% Triton X-100 before 0.3% glutaraldehyde prefixation (in PHEM buffer containing 0.5% Triton) for 10 min at 20 °C. Cells were then sedimented onto 12 mm round coverslips previously coated with polylysine. A further 10 min extraction was then performed with the PHEM buffer containing 0.5 % Triton X-100. The free aldehyde groups were blocked by two successive incubations of lOmin in NaBH4 (lmgml" 1 in PHEM). The coverslips were then incubated with the CTR2611 antibody (1/300) for l h , washed four times (10 min each) and incubated with gold-labeled goat anti-mouse IgM (10 nm size). The coverslips were again rinsed five times for a total of 1 h. The cells were postfixed in 2 % glutaraldehyde, 0.2 % tannic acid in Sorensen buffer (Langanger and De Mey, 1989). The samples were then treated with 0.5% OsO4 in Sorensen buffer on ice for 10 min followed by 0.5 % uranyl acetate and 1 % phosphotungstic acid in 70 % ethanol for 30 min at room temperature. After embedding in Epon, sections were cut and observed in a Philips 400 electron microscope. MDCK cells were grown on glass coverslips and treated according to a procedure adapted from the method described by Langanger and De Mey (1989) in order to achieve good preservation of the cytoskeleton. Cells were rinsed for I s in PHEM buffer, dipped for 2 s in PHEM containing 1% Triton, prefixed for 2 min in PHEM containing 0.5% Triton X-100 and 0.3 % glutaraldehyde. They were then fixed in PHEM containing 0.3% glutaraldehyde without detergent for 10 min and further permeabilized with 0.5% Triton X-100 in PHEM for 30 min. The free aldehyde groups were blocked by two successive 10 min incubations in NaBH4 (1 mgml""1 in PHEM). The coverslips were then incubated with the CTR2611 antibody (1/300 or 1/500) for 1 or 2 h, washed four times (10 min each) and incubated with goldlabeled goat anti-mouse IgM (10 nm size). In some experiments, the first antibody was revealed by the successive addition of a goat anti-mouse IgM (lh) followed after some washes (lh) by goldlabeled protein A (9 nm size). The samples were then treated as described for the KE37 lymphocytes. Preparation of cell extracts MDCK cells were grown on plastic dishes. Whole-cell extracts were obtained by direct resuspension of cells in Laemmli sample buffer (0.062 M Tris-HCl, pH6.8, 2% SDS, 2 to 5% betamercaptoethanol, 10 % glycerol). Electrophoresis and immunoblotting Proteins were separated electrophoretically using either 5% or 7.5% SDS-polyacrylamide gels (Laemmli, 1970). The proteins were transferred from unstained gels to 0.45/on nitrocellulose paper in a semi-dry blotting apparatus (Sartorius) for 1 h 30 min at 150 mA. The transfer buffer contained 48 mM Tris-HCl, pH 9.2, 39 mM glycine, 1.3 mM SDS and 20 % methanol. Subsequently the quality of the transfer was visualized by labeling the proteins with a solution of Ponceau Red. The nitrocellulose was then processed according to either the immunogold-silver staining procedure (Janssen) or the Enhanced ChemiLuminescence (ECL) gene detection system (Amersham). For the immunogold-silver staining, after saturation with 5% ovalbumin (OVA) in a Tris buffer (20 mM Tris-HCl, pH8.2, 0.9% NaCl), the nitrocellulose was washed three times for 5 min in a 0.1 % OVA-Tris buffer. The nitrocellulose was incubated with the CTR2611 Ab (1/7000 in the same buffer plus 1 % normal goat serum) for 1 h, then washed three times for 5 min in 0.1% OVA-Tris buffer. The second antibody, a gold-labeled goat anti-mouse IgM (1/30 in the same buffer plus 1/20 gelatin) was added for l h and subsequently washed with 0.1% OVA-Tris. The immunogold-stained transfer membrane was then incubated with a silver enhancement kit (Janssen, Olen, Belgium). For the ECL technique, the nitrocellulose was saturated with 5 % (w/v) milk in Tris buffer saline (TBS) containing 0.1 % Tween for 30 min at 37 °C. The nitrocellulose was then washed once in 5 % milk-TBS and twice in 3 % ovalbumin, 0.1% (w/v) gelatin, 0.1% Tween in PBS. The nitrocellulose was incubated with the CTR2611 Ab (1/1000 in the same buffer) for 1 h, then washed three times for 5 min in 5 % milk in Tris buffer saline. The second antibody, a goat anti-mouse IgM-biotinylated (1/200 in the same buffer) was added for l h and subsequently washed and then incubated with Streptavidin-biotinylated horseradish peroxidase complex (1/100 in 5 % milk-TBS) for 1 h. After three washes the immunoblot was incubated with the ECL gene detection solution (Amersham, Germany) for 3 min and then submitted to autoradiography overnight. Aster formation in vitro using mitotic frog egg extracts A metaphase II frog egg extract was prepared as described by Felix et al. (1989). An ATP-regenerating system was added to the first 10000g lysate, which was then recentrifuged for l h at 100 000 ^ a t 4°C. Extracts contained about 40 mg ml" 1 of proteins. A 15^1 sample was incubated at room temperature in the presence of 0.1 /JM taxol for desired times, and fixed by dilution in 1 ml of 0.25 % (w/v) glutaraldehyde in RG1 (80 mM K-Pipes, 1 mM EGTA, lmM MgCl2, lmM GTP, pH6.8, with KOH). The suspension was layered on a 5 ml cushion of 25 % glycerol (v/v) in RG2 (RG1 without GTP) in 15 ml corex tubes and centrifuged on a coverslip as described by Evans et al. (1985). The coverslips were Fig. 1. In KE37 lymphocytes the CTR2611 antigen is localized in the pericentriolar material and on intermediate filaments. KE37 lymphocytes were either fixed directly (A-D) or first treated with Nocodazole for 2 h (E-H). After fixation with glutaraldehyde they were subjected to double immunofluorescence using a rabbit anti-tubulin antibody (A,C,E,G) and the CTR2611 antibody (B,D,F,H) as described in Materials and methods. Observations were made on a Zeiss Axiophot photomicroscope. Bar, 10 /.an. Immunogold electron microscopy of thin sections of interphasic KE37 lymphocytes pre-stained with the CTR2611 antibody shows that the antigen is localized on the pericentriolar material next to the minus end of microtubules (arrows) and on patches of material associated with intermediate filaments (arrowhead). Bar, 1.5/on. Centrosomal antigen, intermediate filaments and mitotic spindle 261 postfixed in methanol at -20°C for 5min, and immunofluorescence was performed with a polyclonal rabbit anti-tubulin and the CTR2611 monoclonal antibody as described for the immunostaining of MDCK cells. Results In KE37 lymphocytes, the CTR2611 antigen is localized in the pericentriolar material and on intermediate filaments In interphasic lymphocytes, microtubules originate from a single MTOC containing the centrioles (Fig. 1A). The CTR2611 monoclonal antibody brightly stained a broad region around the MTOC with some additional granular labeling (Fig. IB). During mitosis (Fig. 1C-D), the antibody strongly labeled the poles of the mitotic spindle and gave a faint labeling of the cortex of the cell (Fig. ID). After microtubule depolymerization by Nocodazole, the interphasic centrosomes were still labeled by the CTR2611 antibody (Fig. 1H) but a brightly stained ribbon-like structure appeared. This was similar to the reorganization of intermediate filaments in cells treated by Nocodazole for some hours (Osborn et al. 1980) (Fig. 1F-H). By electron microscopy, in interphasic untreated cells, the CTR2611 antigen was found associated with a granular material surrounding the centrioles in the area of microtubule nucleation (Fig. II, arrows) and on intermediate filaments (Fig. II, arrowhead). No staining was observed when the CTR2611 monoclonal was omitted. Similar controls were made for all immunofluorescence and electron microscope studies and negative results were always obtained (data not shown). Extensive quenching using sodium borohydrate (two successive treatments with a fresh solution each time) after glutaraldehyde fixation is essential to obtain clean controls (see Materials and methods). In MDCK cells, the CTR2611 antigen is associated with intermediate filaments In subconfluent MDCK cells, microtubules do not orig- Fig. 2. Staining pattern obtained with the CTR2611 antibody in interphasic MDCK cells. The cells were fixed with glutaraldehyde and subjected to double immunofluorescence using a rabbit anti-tubulin antibody and the CTR2611 antibody as described in Materials and methods. Confocal stereo views are shown of a given field for the CTR2611 antigen (A) and tubulin (B). Bar, 5 fan. 262 B. Buendia et al. Fig. 3. In interphasic MDCK cells, the CTR2611 antigen is associated with intermediate filaments. Cells were fixed and processed for electron microscopy as described in Materials and methods. (A) Junction area (j) between two adjacent cells. The gold particles decorate intermediate filaments (if), (n) nucleus, (des) desmosome. (B) Peripheral area of the cell; (mt) microtubules, (if) intermediate filaments. (C) Microvilli are not labeled by the CTR2611 antibody (act) to actin. Bar, 0.5 ;im. Centrosomal antigen, intermediate filaments and mitotic spindle 263 inate from a clear MTOC during interphase (Br6 et al. 1987; Bacallao et al. 1989; Buendia et al. 1990). In these cells, the CTR2611 antibody did not stain structures colocalizing with centrioles. Instead, it clearly labeled a fibrous network (Fig. 2A). The observation of optical sections collected by confocal microscopy through interphasic cells showed that the CTR2611 antigen was present in fibers that extended from the upper cortex into the cytoplasm as well as on the basal side of the cells (Fig. 2A). This network did not coincide with the microtubule network of the same cells (Fig. 2B). Following disruption of the microtubule network by Nocodazole, the fibers stained by the CTR2611 antibody collapsed towards the nuclear periphery (not shown) as vimentin-type filaments do (Osborn et al. 1980; see review, Traub, 1985). This strongly suggested that in MDCK cells the antigen was essentially associated with intermediate filaments. In order to examine this possibility more closely, we investigated the localization of the CTR2611 antigen at the ultrastructural level. As shown in Fig. 3A,B, in interphasic MDCK cells, the antigen was indeed associated with intermediate filaments both at the periphery and in the cytoplasm of the cell. No gold particles were found associated with microtubules, free in the cytoplasm or around the centrioles (not shown). The small actincontaining microvilli present in MDCK cells were usually unlabeled although a few gold particles were occasionally associated with them (Fig. 3C). Localization of the CTR2611 antigen during the cell cycle In lymphocytes, the antigen was clearly associated with the poles of the mitotic spindle (Fig. ID). In MDCK cells, although the antigen was not associated with the centrosome during interphase, it was relocalized, at least partially, onto the poles of the mitotic spindle during mitosis (Fig. 4). The cells shown in this figure were stained for CTR2611 antigen and either tubulin or chromosomes in order to determine precisely the different stages of mitosis and the observations were made by confocal microscopy. We show the stereo pairs of the CTR2611 staining only, to save space. In early prophase, the staining became more diffuse although some fibers were still visible near the zones of cell contacts (Fig. 4A) and a faint labeling of the center of the two prophase asters became apparent (Fig. 4A, arrows). At the time of nuclear envelope breakdown, while the interphasic microtubule network collapsed and the two dense asters of short microtubules assembled, the CTR2611 antibody brightly labeled the center of each aster. The antigen seemed to be organized on a ring formed of aggregated material (Fig. 4B). Dots were actually aligned on some microtubules (Fig. 4B, arrows). At the onset of metaphase, the CTR2611 antigen became localized at the poles of each half spindle, though staining localized on fibers at the periphery of the cell persisted (Fig. 4C). During anatelophase, the CTR2611 antigen began to spread out and finally seemed to migrate towards the midbody in late telophase (Fig. 4 D, E). The midbody itself was not stained. These confocal studies suggested that the antigen was interacting with the pericentriolar region, but the staining appeared to be more spread into the spindle than simply associated with the centrosome itself (Fig. 4C). This was confirmed in the electron microscope: the antigen was mostly localized along the minus end of the microtubules of the mitotic spindle (Fig. 5A). Microtubules had granular material associated with them but it was 264 B. Buendia et al. Fig. 4. Redistribution of CTR2611 antigen during the cell cycle in MDCK cells. The cells were fixed with glutaraldehyde and labeled with the CTR2611 antibody. The stages of mitosis were determined by labeling either the chromosomes with propidium iodide or the microtubules with an anti-tubulin antibody (not shown). Confocal stereo view of 5 different cells corresponding to: (A) early prophase (black arrows indicate the faint staining in the areas corresponding to the center of the two asters of microtubules in this cell); (B) late prophase (white arrows indicate staining aligned on some microtubules); (C) metaphase; (D) late anaphase; and (E) late telophase. Bar, difficult to determine if the staining was associated with it or directly to the microtubule walls. The centrioles were not labeled (Fig. 5B). During mitosis, few intermediate filaments were found in the cytoplasm. By contrast, the cortex of the cells contained many that were strongly labeled by the CTR2611 antibody (Fig. 5C). Interestingly, in telophase, there was still some labeling associated with the microtubules originating from the centrioles but strong labeling was found along intermediate filaments clustered next to the centrioles (Fig. 6). Following microtubule disruption by Nocodazole in mitotic cells, all the antigen was redistributed onto a ring of intermediate filaments surrounding dispersed chromosomes (not shown). In mitosis, the CTR2611 antigen is associated with the center of taxol-induced microtubule asters In mitotic cells, taxol treatment induced mitotic spindle disruption and the assembly of many microtubule asters as originally described by DeBrabander et al. (1986). In the cell shown in Fig. 7, there were two large microtubule asters, which probably contained the centrioles as well as many small asters that had assembled at the cell periphery. There was a large accumulation of CTR2611 antigen in the center of each aster. Some labeling organized in fibers was still visible in the cell cortex, however. In interphasic cells, microtubule stabilization by taxol did not affect strongly the distribution of the CTR2611 antigen that remained associated with the intermediate filaments, despite the fact that dense bundles of microtubules had assembled. This suggested that, in MDCK cells, the affinity of CTR2611 antigen for intermediate filaments was sufficiently high in interphase to prevent its interaction with microtubules and that it was reduced during mitosis permitting its redistribution onto microtubules. This was interesting, and we thought that it would be important to study the interaction of this antigen with microtubules in an in vitro system. We have recently developed concentrated frog egg extracts that can be prepared from metaphase II-arrested eggs or from interphase eggs and that retain the metabolic properties of each cell cycle state upon incubation at room temperature (Felix et al. 1989; Verde et al. 1990). In mitotic extracts taxol induced first the assembly of short microtubule bundles that quickly reorganized into asters similar to those observed in mitotic MDCK cells (Fig. 8C). In microtubules assembled shortly after addition of taxol (Fig. 8A), the CTR2611 antigen seemed to be distributed all along the microtubules in small patches (Fig. 8B). After 30 min of incubation in the presence of taxol, the CTR2611 antigen was concentrated near the center of each aster, often forming a ring (Fig. 8D). This was strikingly similar to what was observed in MDCK mitotic cells treated with taxol. Observation of these structures in the Centrosomal antigen, intermediate filaments and mitotic spindle 265 Fig. 5. During mitosis, the CTR2611 antigen is associated with the minus end of spindle microtubules. MDCK cells were fixed and processed for electron microscopy as described in Materials and methods. (A) Labeling of the minus end of mitotic spindle microtubules; ki, kinetochore; chr, chromosomes. (B) The centriole and pericentriolar area are not labeled. (C) Intermediate filaments in the cortex of the cell are labeled. Bar, 0.5 /an. electron microscope confirmed that they indeed had a center containing some amorphous material but no centrioles. The CTR2611 antigen was not in the center, but rather on' the microtubules next to the center (not shown), just as in the mitotic spindle of intact cells. In interphasic extracts, taxol produced disorganized bundles of microtubules that were diffusely stained by the CTR2611 antibody (not shown). Therefore, this in vitro system can be used to study further how the CTR2611 antigen interacts with microtubules. CTR2611 antibody recognizes a 74xlO3Mr polypeptide in several species As shown in Fig. 9A and B, the antibody recognized one polypeptide of 74xlO 3 M r in all cell lines tested: MDCK cells, HeLa cells, KE37 lymphocytes, PtK 2 cells and frog egg extracts (interphasic or mitotic) as well as in the preparation of pure centrosomes isolated from KE37 lymphocytes. This staining was obtained using two 266 B. Buendia et al.. entirely different detection methods (immunogold and silver enhancement or immunoperoxidase and Enhanced ChemiLuminescence (ECL)). In addition, a band was revealed at 170xl0 3 M r using the immunogold method, whereas a very faint band around 120xl0 3 M r was detected by ECL. Discussion Interaction of the CTR2611 antigen with the centrosome and intermediate filaments In human lymphocytes during interphase, the CTR2611 antigen is mainly localized in the pericentriolar material. However, it is also associated with finely granular material localized on intermediate filaments as determined by electron microscopy. Following microtubule depolymerization, the intermediate filaments reorganize into a coil (Geiger and Singer, 1980; Osborn et al. 1980; Ball and Singer, 1981; Geuens et al. 1983; for review, •%' T* m ^ Fig. 6. Relocalization of CTR2611 antigen onto intermediate filaments during telophase. Cells werefixedand processed for electron microscopy as described in Materials and methods. The CTR2611 antigen is still present on microtubules (mts) close to the chromosomes (chr), which are close to the centriole (c). Intermediate filaments (if) also located close to the centriole are decorated with gold particles. Bar, 0.5 |(m. Traub, 1985), which is brightly stained by the CTR2611 antibody by immunofluorescence microscopy. Under these conditions, the centrosome is still stained. It is striking that in interphasic MDCK cells, the CTR2611 antigen is localized exclusively on intermediate filaments, because in these cells most microtubules do not originate from the pericentriolar region. In fact, there is little material around the centrioles in MDCK cells (Bre et al. 1987; Buendia et al. 1990). It may be that most of the pericentriolar material is redistributed onto intermediate filaments and the microtubules grow from these sites. We tried to find such events through an extensive electronmicroscope survey of microtubule nucleation in MDCK cells. However, this turned out to be difficult to interpret because nucleation may occur only over short distances. We do see interactions between microtubules and intermediate filaments, but we have no way of deciding if this is fortuitous or not (C. Antony, unpublished). Recently, McBeath and Fujiwara (1989) reported a tight interaction between intermediate filaments and microtubules in collagen-secreting cells. It was not clear, however, if some of the microtubules were nucleated from regions on intermediate filaments or if these filaments acted only as guides to align centrosome nucleated microtubules. Schatten et al. (1987) have reported that a monoclonal antibody raised against Drosophila intermediate filament proteins stained centrosomes in sea-urchin eggs. These results and the present study with the CTR2611 antibody suggest that there must be some relationship between intermediate filament proteins and centrosomal antigens. The class of intermediate filaments with which the CTR2611 antigen is associated in MDCK cells is probably of the vimentin type. Indeed, lymphocytes possess only vimentin and the redistribution of the antigen observed both in MDCK and in lymphocytes after Nocodazole treatment is typical of vimentin filaments (for review, Traub, 1985). Possible function of CTR2611 antigen association with microtubule minus end domain in the mitotic spindle In both MDCK cells and lymphocytes the CTR2611 antigen is redistributed onto the minus-end domain of microtubules in the mitotic spindle during the interphase-metaphase transition. The redistribution of the CTR2611 antigen during the cell cycle is particularly striking in MDCK cells because in interphase the antigen is localized only on intermediate filaments. However, in all cells tested so far (from starfish (A. Picard, unpublished) to human), the CTR2611 antibody stains a broad domain around the poles of the mitotic spindle. It is widely accepted that the microtubules in the mitotic spindle are nucleated by the pericentriolar material (Mclntosh, 1983; Mitchison and Kirschner, 1984). It is also assumed that the microtubules have their minus end 'capped' by this material. Although this is true in vitro for centrosomenucleated microtubules (Mitchison and Kirschner, 1984; Centrosomal antigen, intermediate filaments and mitotic spindle 267 Fig. 7. In mitotic MDCK cells, the CTR2611 antigen accumulated in the center of taxol-induced asters. Cells treated with taxol (5 ;<M) for 2 h were fixed and processed for double immunofluorescence as described in Materials and methods. Confocal stereo views of a mitotic cell labeled for: A, tubulin; and B, CTR2611 antigen. Bar, Toriyama et al. 1988), this is not at all clear in vivo in the mitotic spindle. Microtubule polymerization could also occur by local stabilization of their minus ends through the binding of pericentriolar material along the microtubule wall. This was proposed previously by DeBrabander et al. (1985) and recently Mitchison (1989) has produced evidence for a flux of tubulin subunits from the kinetochore towards the poles of the mitotic spindle. This implies that the minus end should not be capped. Also, using a human auto-antibody directed against pericentriolar material (the 5051 antibody), LaClaire and Goddard (1989) have described an accumulation of reactive material along the minus end of algal mitotic spindle microtubules. Moreover, we have shown that the microtubule asters that form at the expense of the mitotic spindle in taxol-treated mitotic cells contain the CTR2611 antigen arranged as a ring around their center. More strikingly, similar asters can be generated in vitro in mitotic frog egg extracts. This process occurs through the rearrangement of microtubules and, apparently, the CTR2611 antigen binds along 268 B. Buendia et al. microtubules and relocates at the center of the asters upon their formation. These observations suggest that 'centrosomal' material can bind to microtubules and migrate towards their minus end or accumulate there in some unknown fashion. This is important because there are many instances in which an 'organized' centrosome existing prior to the assembly of the poles of the mitotic spindle is not obvious. This is the case in plants (PickettHeaps, 1974; De Mey et al. 1982; Vantard et al. 1985) but also during meiosis in many species (Karsenti and Maro, 1986). In these instances a mechanism by which the centrosomal material could bind along microtubules and migrate towards their minus end would help to assemble the spindle poles. This will be discussed more extensively elsewhere (Verde et al. unpublished data). CTR2611 antigen and other centrosomal proteins By immunoblotting using the immunogold detection method, the CTR2611 antibody revealed two polypeptides: one at 74xlO 3 M r and another at 170xl03Afr (Fig. 9A). Fig. 8. The CTR2611 antigen accumulates in the center of taxol-induced microtubule asters assembled in mitotic frog egg extracts. The extracts were incubated with 0.1 f(M taxol at room temperature, fixed with glutaraldehyde and processed for immunofluorescence as described in Materials and methods. Observations were made on a Zeiss Axiophot photomicroscope. (A and C) Tubulin staining. (B and D) CTR2611 staining. (A-B) Microtubules fixed lmin after taxol addition; (C-D) microtubules were fixed 15min after taxol addition, Bar, 8/on. When more stringent conditions were used (ECL technique in the presence of milk and Tween), the 74xlO 3 M r polypeptide remained stained whereas the 170xl0 3 M r protein was not detected any more and a faint band appeared at 120xl0 3 M r (Fig. 9B). The relationship between these polypeptides and the immunofluorescence staining is not entirely clear. Given the fact that the 74xlO3Afr polypeptide is detected in all conditions and in all species tested so far, we think that this protein is the antigen actually detected by immunofluorescence. Several 'centrosomal' proteins have been identified that are specifically localized at the poles of the mitotic spindle (Brady et al. 1986; Sager et al. 1986; Salisbury et al. 1986; Kotani et al. 1987; Baron and Salisbury, 1988; Neighbors et al. 1988; Toryama et al. 1988; Keryer et al. 1989; Kuryama, 1989; Rao et al. 1989). Recently, Keryer et al. (1989), have characterized an antibody prepared in the same way as the CTR2611 antibody. This CTR532 antibody labels the mitotic spindle poles in different mammalian cell types as well as in the ciliate Paramecium tetraurelia. It recognizes a 170xl0 3 M r protein in Paramecium extracts. This antigen is clearly different from that recognized by the CTR2611 antibody. Indeed, apart from the fact that the main polypeptides recognized by both antibodies are different, staining at the poles of the mitotic spindle by CTR2611 is abolished following microtubule depolymerization whereas this is not the case for the CTR532. Toriyama et al. (1988) have characterized molecules with microtubule-nucleating activity isolated from the mitotic apparatus of sea-urchin eggs. They obtained a 51xlO 3 M r major component. However, when they solubilized the microtubule-nucleating activity from whole egg homogenates, they eluted a fraction from a phosphocellulose column containing the 51 x 103 MT protein as well as proteins of 100xl0 3 M r , 75xlO 3 M r and 60xl0 3 M r . The proteins recognized by the CTR2611 antibody are highly conserved, since they are detected in human lymphocytes, in canine kidney epithelial cells, in frog egg extracts and in sea-urchin eggs (Picard et al. unpublished). It would be interesting to know if the 180 xlO 3 and the 75xlO 3 M r proteins coeluted with the aster-forming material in sea-urchin eggs have any relationship to the proteins recognized by the CTR2611 antibody. A striking outcome of this investigation is that a monoclonal raised against centrosomes recognizes an antigen present in pericentriolar material as well as along microtubule minus end in mitosis and intermediate filaments. Interactions between microtubules and inter- Centrosomal antigen, intermediate filaments and mitotic spi/idle 269 A { 200- typical 'centrosomal' staining. Yet, electron-microscopic observation has revealed that it was not. Instead, the antigen was associated with the minus ends of the spindle microtubules. Therefore, the use of antibodies and immunofluorescence to characterize the behavior of the centrosome during the cell cycle and cell differentiation should be done cautiously. 200- t 116- 11697-- 97- We thank E. Stelzer for help with the confocal microscope, T. Kreis for many stimulating discussions, as well as J. DeMey and A. Merdes for useful suggestions. This work was supported by an EEC grant no. ST2000411 to B.B. and by the CNRS (M.B.). 3 B 4 References BACALLAO, R., ANTONY, G., DOTTI, C , KARSENTI, E. AND SIMONS, K. (1989). The subcellular organization of MDCK cells during the formation of a polarized epithelium. J. Cell Biol. 109, 2817-2832. BALL, E. H. AND SINGER, S. J. (1981). Association of microtubules and intermediate filaments in normal fibroblasts and its disruption upon transformation by a temperature-sensitive mutant of Rous sarcoma virus. Proc. natn. Acad. Sci. U.S.A. 78, 6986-6990. BARON, A. T. AND SALISBURY, J. L. (1988). Identification and localization of a novel cytoskeletal, centrosome associated protein in PtK2 cells. J. Cell Biol. 107, 2669-2678. 200- 116- BERGEN, L. G., KURIYAMA, R. AND BORISY, G. (1980). Polarity of 97- microtubules nucleated by centrosomes and chromosomes of Chinese hamster ovary cells in vitro. J. Cell Biol. 84, 151-159. BERNS, M. W. AND RICHARDSON, S. M. (1977). Continuation of mitosis after selective laser microbeam destruction of the centriolar region. J. Cell Biol. 75, 977-982. BLOOM, G. S. AND VALLEE, R. B. (1983). Association of microtubuleassociated protein 2 (MAP2) with microtubules and intermediate filaments in cultured brain cells. J. Cell Biol. 96, 1523-1531. 66- BORNENS, M., PAINTRAND, M., BERGES, J., MARTY, M.-C. AND KARSENTI, E. (1987). Structural and chemical characterization of isolated centrosomes. Cell Motil. Cytoskel. 8, 238-249. 42- BRADY, R. C, CABRAL, F. AND DEDMAN, J. R. (1986). Identification of a 52 KDa calmodulin binding protein associated with the mitotic spindle apparatus in mammalian cells. J. Cell Biol. 103, 1855-1861. 1 2 Fig. 9. Nature of the CTR2611 antigen. (A) Immunoblotting of different extracts analyzed by electrophoresis on a 5 % (lanes 1-2) and a 7.5% (lanes 3-4) polyacrylamide gel. After transfer to nitrocellulose the immunoblotting was performed with the CTR2611 antibody using immunogold-silver staining. Lane 1, MDCK cells; lane 2, PtK 2 cells; lane 3, mitotic frog egg extract; lane 4, interphasic frog eggs extract. Mr values are shown (xlO ). (B) Immunoblotting of cell extracts analyzed by electrophoresis on an 8% polyacrylamide gel. After transfer to nitrocellulose, the immunoblotting was performed with the CTR2611 antibody using the Enhanced ChemiLuminescence technique. This figure shows the autoradiogram after one night of exposure. Lane 1, KE37 human lymphocytes; lane 2, MDCK cells; lane 3, HeLa cells; and lane 4, pure centrosomes isolated from KE37 lymphocytes. BRE, M. H., KREIS, T. E. AND KARSENTI, E. (1987). Control of microtubule nucleation and stability in Madin-Darby Canine Kidney cells: the occurrence of noncentrosomal, stable detyrosinated microtubules. J. Cell Biol. 105, 1283-1296. BUENDIA, B., BRE, M. H., GRIFFITHS, G. AND KARSENTI, E. (1990). Cytoskeletal control of centrioles movement in MDCK cells forming a confluent monolayer. J. Cell Biol. 110, 1123-1135. CELIS, J. E., SMALL, J. V., LARSEN, P. M., FEY, S. J., DE MEY, J. AND CELIS, A. (1984). Intermediate filaments in monkey kidney TC7 cells: focal centers and interrelationship with other cytoskeletal systems. Proc. natn. Acad. Sci. U.S.A. 81, 1117-1121. DEBRABANDER, M., AERTS, F., DE MEY, J., GEUENS, G., MOEREMANS, M., NUYDENS, R. AND WILLEBRORDS, R. (1985). Microtubule dynamics and the mitotic cycle: a model. In Aneuploidy (ed. Dellarco, V. I., Voytek, P. E. and Hollaender, A.), pp. 269-277. Plenum, NY. DEBRABANDER, M., GEUENS, G., NUYDENS, R., WILLEBRORDS, R., AERTS, F. with the participation of J. R. MCINTOSH (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. DE MEY, J., LAMBERT, A. M., BAJER, A. S., MOEREMANS, M. AND mediate filaments have already been documented (Geiger and Singer, 1980; Geuens et al. 1983; Knapp et al. 1983; Celis et al. 1984) and MAP2, which possesses binding sites for both microtubules and intermediate filaments, seems to play a role in their interaction in brain cells (Bloom and Vallee, 1983; Flynn et al. 1987). Apparently, the CTR2611 antigen shares with MAP2 the property of interacting with both microtubules and intermediate filaments. However, the antibody may also recognize a specific epitope present on different proteins that are present, respectively, on microtubules and intermediate filaments. Finally, this study shows that one must be careful at interpreting immunofluorescence data. Indeed, the staining at the pole of the spindle looked very much like a 270 B. Buendia et al. DEBRABANDER, M. (1982). Visualization of microtubules in interphase and mitotic plant cells of Haemantus endosperm with the immunogold staining method. Proc. natn. Acad. Sci. U.S.A. 79, 1898-1902. EVANS, L., MITCHISON, T. J. AND KIRSCHNER, M. W. (1985). Influence of the centrosome on the structure of nucleated microtubules. J. Cell Biol. 100, 1185-1191. FELIX, M. A., PINES, J., HUNT, T. AND KARSENTI, E. (1989). A post- ribosomal from activated Xenopus eggs that displays posttranslational regulated oscillation of its CDC2 mitotic kinase activity. EMBO J. 8, 3059-3069. FLYNN, G., JOLY, J. C. AND PURICH, D. L. (1987). The 28000 Mr microtubule binding domain of microtubule associated protein 2 also contains a binding site for intermediate filaments. Biochem. biophys. Res. Commun. 148, 1453-1459. GEIGER, B. AND SINGER, S. J. (1980). Association of microtubules and intermediate filaments in chicken gizzard cells as detected by double immunofluorescence. Proc. natn. Acad. Sci. U.S.A. 77, 4769-4773. GEUENS, G., DE BRABANDER, M., NUYDENS, R. AND DE MEY, J. (1983). The interaction between microtubules and intermediate filaments in cultured cells treated with taxol and Nocodazole. Cell Biol. Int. Rep. 7, 35-47. 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. KARSENTI, E. AND MARO, B. (1986). Centrosomes and the spatial distribution of microtubules in animal cells. Trends biochem. Sci. 11, 460-463. KERYER, G., GARREAU DE LOUBRESSE, N., BORDES, N. AND BORNENS, M. (1989). Identification of a spindle-associated protein in ciliate micronuclei. J. Cell Sci. 93, 287-298. KNAPP, L. W., O'GUIN AND SAWYER, R. H. (1983). Rearrangement of the keratin cytoskeleton after combined treatment with microtubule and microfilament inhibitors. J. Cell Biol. 97, 1788-1794. KOTANI, S., MUROFUSHI, H., MAEKAWA, S., AIZAWA, H., KAJI, Z. AND SAKAI, H. (1987). Identification of the 190 KDa microtubule-associated protein in cultured fibroblasts and its association with interphase and mitotic microtubules. Cell Struct. Fund. 12, 1-9. KRISHAN, A. (1975). Rapid flow cytofluorometric analysis of mammalian cells by propidium iodide staining. J. Cell Biol. 66, 188-193. KURIYAMA, R. (1989). 225 KDa phosphoprotein associated with mitotic centrosomes in sea urchin eggs. Cell Motil. Cytoskel. 12, 90-103. LACLAIRE, J. W. AND GODDARD, R. H. (1989). Immunolocalization of pericentriolar material in algal mitotic spindle. Cell Motil. Cytoskel. 13, 225-238. LAEMMLI, U. (1970). Cleavage of structural proteins during the assembly of the head of Bacteriophage T4. Nature 227, 680-685. LANGANGER, G. AND DE MEY, J. (1989). Detection of cytoskeletal proteins in cultured cells at the ultrastructural level. In Techniques in Immunochemistry, vol. 4 (ed. Bullock, G. R. and Petruoz, P.), pp. 47-65. Academic Press, London. MAYER, L., MANFU, S. AND KUNKEL, H. G. (1982). Human T cell hybridomas secreting factors for IgA-specific help, polyclonal B cell activation and B cell proliferation. J. exp. Med. 156, 1860-1865. MCBEATH, E. AND FUJIWARA, K. (1989). Coalignment of microtubules, cytokeratin intermediate filaments, and collagen fibrils in a collagensecreting system. Eur. J. Cell Biol. 50, 510-521. Me INTOSH, J. R. (1983). The centrosome as a organizer of the cytoskeleton. Modern Cell Biol. 2, 115-142. MITCHISON, T. J. (1989). Microtubule flux in the mitotic spindle: evidence from photoactivation offluorescence.J. Cell Biol. 109, 637-652. MITCHISON, T. J. AND KIRSCHNER, M. (1984). Microtubule assembly nucleated by isolated centrosomes. Nature 312, 232-237. intermediate filament systems in the same cell by double immunofluorescence microscopy. Expl Cell Res. 125, 37-46. PICKETT-HEAPS, J. D. (1974). Plant microtubules. In Dynamic Aspects of Plant Ultrastructure (ed. Robards, A. W.), pp. 219-255. London: McGraw-Hill. RAO, P. N., ZHAO, J. Y., GANJU, R. K. AND ASHORN, C. L. (1989). Monoclonal antibodies against the centrosome. J. Cell Sci. 93, 63-69. SAGER, P. R., ROTHFIELD, N. L., OLIVER, J. M. AND BERLIN, R. D. (1986). A novel mitotic spindle pole component that originates from the cytoplasm during prophase. J. Cell Biol. 103, 1863-1872. SALISBURY, J. L., BARON, A. T., COLING, D. E., MARTINDALE, V. E. AND SANDERS, M. A. (1986). Calcium-modulated contractile proteins associated with the eukaryotic centrosome. Cell Motil. Cytoskel. 6, 193-197. SCHATTEN, H., WALTER, M., MAZIA, D., BIESSMAN, H., PAWELETZ, N., COFFE, G. AND SCHATTEN, G. (1987). Centrosome detection in sea urchin eggs with a monoclonal antibody against Drosophila intermediate filament proteins: Characterization of stages of the division cycle of centrosomes. Proc. natn. Acad. Sci. U.S.A. 84, 8488-8492. STELZER, E. H. K., STRICKER, R., PICK, R., STORZ, C. AND HANNINEN, P. (1989). Confocal fluorescence microscopes for biological research. In Scanning Imagings. Proc. Soc. Photoopt. Instrum. Eng. Bellingham, WA(ed. Tony Wilson), pp. 146-151. Spie, Washington. TASSIN, A. M., MARO, B. AND BORNENS, M. (1985). Fate of microtubule organizing centers during in vitro myogenesis. J. Cell Biol. 100, 35-46. TORIYAMA, M., OHTA, K., ENDO, S. AND SAKAI, H. (1988). 51 KDa protein, a component of microtubule-organizing granules in the mitotic apparatus involved in aster formation in vitro. Cell Motil. Cytoskel. 9, 117-128. TRAUB, P. (1985). Intermediate Filaments. A Review, pp. 82-110. Springer-Verlag, Berlin, Heidelberg. TUCKER, J. B. (1979). Spatial organization of microtubules. In Microtubules (ed. Roberts, K. and Hyams, J. S.), pp. 315-358. New York: Academic Press. VANTARD, M., LAMBERT, A. M., DE MEY, J., PICQUOT, P. AND VAN ELDIK, L. J. (1985). Characterization and immunocytochemical distribution of calmodulin in higher plant endoderm cells: Localization in the mitotic apparatus. J. Cell Biol. 101, 488-499. VERDE, F., LABBS, J. C , DOR£E, M. AND KARSENTI, E. (1990). Regulation of microtubule dynamics by cdc2 protein kinase in cell free extracts of Xenopus eggs. Nature 343, 233-238. NEIGHBORS, B. W., WILLIAMS, R. C. JR AND Me INTOSH, J. R. (1988). Localization of kinesin in cultured cells. J. Cell Biol. 106, 1193-1204. OSBORN, M., FRANKE, W. W. AND WEBER, K. (1980). Direct demonstration of the presence of two immunologically distinct (.Received 16 May 1990 - Accepted 3 July 1990) Centrosomal antigen, intermediate filaments and mitotic spindle 271