A monoclonal antibody, raised against mammalian centrosomes and screened by recognition of plant microtubule organizing centers, identifies a pericentriolar component in different cell types

Microtubules are filamentous structures in the cytoplasm of eucaryotic cells where they perform a remarkable range of functions (for reviews see Dustin,1984). They play a central role in the organization of the cell interior and in the building of the mitotic apparatus (for reviews see Alberts et al. 1989).

Microtubule organizing centers (MTOCs) are required for the assembly, proper organization, and polarity of microtubule networks (Bergen et al. 1980; McIntosh, 1983; Bornens and Karsenti, 1984; Tucker, 1984; Brinkley, 1985).

In animal cells, the major MTOC is the centrosome. Centrosomes have a dual structure. They comprise highly organized and specific elements, namely the centrioles, and a poorly defined pericentriolar material that shows extraordinary flexibility in its organization, location and functions (Mazia, 1984; Karsenti and Maro, 1986; Vorobjev and Nadezhdina, 1987; Bornens,1991). It shifts from a diffuse intracellular distribution to aggregated forms that appear at appropriate places and possibly induce the formation of centrosomes or of other forms of microtubule organizing centers (Rieder and Borisy, 1982; Maro et al. 1985; Tassin et al. 1985; Karsenti and Maro, 1986; Bré et al. 1990). Cytoplasmic microtubules are not nucleated directly on centrioles. They originate from the pericentriolar material (Buendia et al. 1992; for reviews see Cande, 1990; Huang, 1990).

The explanation of the function of centrosomes in molecular terms relies to a large extent on the biochemical and structural characterization of their components. Several centrosomal proteins have been identified (Frasch et al. 1986; Gosti et al. 1987; Buendia et al. 1990; Moudjou et al. 1991) and for a number of these the cDNA has been cloned and characterized (Snyder and Davis, 1988; Whitfield et al. 1988; Joswig et al. 1991), including γ-tubulin (Horio et al. 1991; Steams et al. 1991; Zheng et al. 1991). However, highly purified centrosomes show complex protein profiles (Bornens,1991) and the majority of their components remain to be characterized.

We have recently described preparative procedures yielding relatively high amounts of purified centrosomes from calf thymus (Komesli et al. 1989). Here, we have used these preparations to generate monoclonal antibodies directed against elements of the pericentriolar material.

We report the identification of a protein that in animal cells is specifically and tightly associated with the pericentriolar material and also combines with kineto-chores and the midplate during mitosis. In plant cells, a related antigen is found at the expected location of microtubule organizing centers.

Ingredients used for the construction of phosphate buffered saline (PBS: 150 mM NaCl, 10 mM sodium phosphate, pH 7.4), bovine serum albumin (BSA), TNM buffer (10 mM Tris HCl, pH 7.2, 10 mM NaCl, 5 mM MgCl₂), Pipes buffer (10 mM Pipes, pH 7.2, with KOH), MCM buffer (25 mM Mes, 8 mM CaCl₂, 600 mM mannitol, pH 5.5) and Hoechst dye 33258 were from Sigma Chemical Co., St Louis, MO, USA.

Urea, 2-mercaptoethanol, uranyl acetate and tannic acid were from Merck, Darmstadt, RFA; osmium tetroxide from Fluka, Mulhouse, France; Epon from TAAB, St Germain-en-Laye, France; SDS and Aquamount from BDH, Chassieu, France; Tween 20 (EIA Grade) from Bio-Rad, Ivry, France; ECL Kit from Amersham, les Ulis, France; PVDF Immobilon from Millipore Corporation, Bedford, MA, USA; glutaraldehyde from Polysciences Inc., Warrington, PA, USA.

Culture media were from Boehringer Mannheim Diagnostic, Inc, Houston, TX, USA.

HeLa cells, SP2O/Agl4 (non-secreting mouse myeloma) and Raji cells (Burkitt lymphoma, human) were kindly provided by Dr R. Berthier. 2017 cells (bovine foetal calf spleen cells) were a generous gift from Dr B. Guillemin.

Endosperm cells were isolated from young ovules of Haementhus katherinae bak. and prepared. as previously described (Bajer and Molé-Bajer, 1986).

Balb/c mice were obtained from IFF A CREDO, L’Ar-bresle, France. Vinblastine sulfate and nocodazole were purchased from Aldrich, Strasbourg, France. Pectolyase Y23 was from Seishin Pharmaceutical Co., Tokyo, Japan; Macero-zyme R-10 from Serva, Heidelberg, RFA, and Caylase from Cayla, Toulouse, France.

mAb YL1/2, a monoclonal antibody specific to tyrosinated tubulin (Wehland et al. 1983) was a generous gift from Dr J. V. Kilmartin. A polyclonal antibody specific for detyrosinated tubulin was kindly provided by Dr J. C. Bulinski (Gundersen et al. 1984) and CREST serum was a generous gift from Dr B. Roussel.

Fluorescein-conjugated sheep anti-mouse antibody was from Bioart, Biochem, Nanterre, France; rhodamine-conjugated goat anti-rabbit and horseradish peroxidase goat antimouse antibodies were obtained from Cappel, West Chester, PA, USA.

Fluorescein-conjugated goat anti-rat, fluorescein-conjugated rabbit anti-human antibodies and rhodamine-conjugated rabbit anti-mouse Ig M antibody were obtained from Jackson, West Grove, PA, USA.

Rabbit anti-mouse IgG antibody conjugated with 5 nm gold particles was purchased from Janssen Pharmaceutica, Berrse, Belgium.

Balb/c mice were immunized intraperitoneally with native calf thymus centrosomes prepared as described by Komesli et al. (1989). Fifty micrograms of centrosomal proteins were used per injection, emulsified in complete Freund’s adjuvant. Three subsequent boosts of 50 μg of centrosomes in incomplete Freund’s adjuvant were given at 2-week intervals by intraperitoneal injection. Spleens were removed from immunized mice 3 days after the final injection for hybridoma production. Spleen cells were fused with mouse myeloma SP2O/Agl4 cells using standard procedures as described by Kohler and Milstein (1975). Supernatants from hybridoma cultures were screened by immunofluorescence for the presence of antibodies against isolated calf thymus centrosomes (for more details see section on immunofluorescence, below). Hybridomas found to secrete the desired antibody were cloned three times by limiting dilution. Fifteen positive clones were obtained. The corresponding antibodies were further tested for their immunostaining specificity on calf thymocytes (see Immunofluorescence). Here we describe one of these antibodies, designated mAb 6C6. This mAb is an IgM. Ascites fluid containing a high concentration of mAb 6C6 was prepared by injecting 4×10⁶ cloned hybridoma cells into the peritoneal cavity of Balb/c mice and harvesting the ascites tumor 20-25 days later.

HeLa cells, Raji cells and 2017 cells were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated bovine fetal calf serum, 2 mM glutamine, 1 mM sodium pyruvate, 100 i.u./ml penicillin and 100 μg/ml streptomycin at 37°C with 6.5% CO₂.

Raji cells were blocked with 10 Mg/ml of vinblastine sulfate for 12 h. Chromosomes were further isolated as described by Mitchison and Kirschner (1985).

In all immunofluorescence experiments, antibodies were diluted in PBS containing 1% BSA. Dilution factors are indicated in parenthesis.

Isolated centrosomes (Komesli et al. 1989) were sedimented onto 14 mm round coverslips at 24,000 g for 15 min and fixed in methanol at −20°C for 6 min. In single-labelling experiments, coverslips were incubated with mAb 6C6 ascites fluid (1/50,000) for 30 min. After reaction with the primary antibody, centrosomes on coverslips were washed three times in PBS, 0.1% Tween 20, incubated with a fluorescein-conjugated sheep anti-mouse antibody (1/100) for 30 min and washed again three times in PBS, 0.1% Tween 20.

In double-labelling 6C6/anti-tubulin experiments, centrosomes were sequentially incubated with mAb 6C6 ascites fluid (1/50,000), fluorescein-conjugated sheep anti-mouse antibody (1/100), anti-detyrosinated tubulin antibody (1/50) and rhodamine-conjugated goat anti-rabbit antibody (1/200).

Calf thymocytes and Raji cells were centrifuged on coverslips at 300g for 5 min, in PBS. HeLa and 2017 cells were grown on coverslips and washed in PBS before further processing. All cells were fixed for 6 min in methanol at −20°C. The preparations were subsequently washed in PBS. For single mAb 6C6 labelling experiments, they were sequentially incubated with mAb 6C6 ascites fluid (1/50,000) for 30 min, then with fluorescein-conjugated sheep anti-mouse antibody (1/100).

In anti-tubulin/6C6 double-labelling experiments, the cells were stained by sequential incubation with mAb 6C6 ascites fluid (1/50,000), rabbit anti-mouse IgM antibody conjugated with rhodamine (1/250), mAb YL 1/2 (1/100) and fluorescein-conjugated goat anti-rat IgG antibody (1/250).

Double-labelling of centrosomes and kinetochores used sequential incubation with mAb 6C6 ascites fluid (1/50,000) and human CREST serum (1/100). The secondary antibodies were rabbit anti-mouse IgM antibody conjugated with rhodamine (1/250) and a rabbit anti-human IgG antibody conjugated with fluorescein (1/250).

Nuclei were stained using Hoechst 33258 at 10 μg/ml in PBS.

Isolated chromosomes were sedimented onto coverslips at 16,000g for 10 min at 4°C. The coverslips were removed, fixed in methanol for 6 min at −20°C and stained with mAb 6C6 as described for isolated centrosomes. Chromatin was stained using Hoechst.

Coverslips were mounted in aquamount and observed using a Zeiss Axiovert 35 M microscope (Zeiss, Oberkochen, RFA). Pictures were taken with a Nikon F-601 (Charenton-le-Pont, France) using Kodak Asa 400 film.

Negative control experiments, performed with the omission of primary antibodies, revealed only diffuse, low-level background staining.

Endosperm cell preparations were fixed for 10 min in methanol at 4°C. In single-labelling experiments, preparations were incubated with mAb 6C6 ascites fluid (1/50,000) for 2 h, washed three times in PBS, 0.1% Tween 20 and incubated with a rabbit anti-mouse IgM antibody conjugated with rhodamine (1/250) for 1 h. After three washes, nuclei were stained using Hoechst for 5 min at 0.25 mg/ml in PBS.

In anti-tubulin/6C6 double-labelling experiments, cells were sequentially stained with mAb 6C6 ascites fluid (1/50,000), mAb YL1/2 (1/100), rhodamine-conjugated rabbit anti-mouse IgM (1/250) and fluorescein-conjugated goat antirat IgG antibodies (1/250).

Seeds of maize were sprouted in the dark at 24°C. After three days, root tips were exposed, for 1 h at 32°C, to an enzyme solution containing 0.05% Pectolyase Y23, 1% Macerozyme R-10, 4% Caylase in MCM buffer, while gently shaken in the dark. Then root tips were washed twice for 10 min with the same buffer devoid of enzymes.

Cells were spread on coverslips coated with poly-L-lysine (1 mg/ml) and fixed for 10 min in methanol at —20°C. Cell preparations were stained with mAb 6C6 as described above for endosperm cells.

Calf thymus centrosomes were centrifuged onto coverslips and fixed in methanol for 6 min at —20°C. The coverslips were incubated with mAb 6C6 ascites fluid (1/50,000) for 30 min. After three washes in PBS, 0.1% Tween 20, they were reacted with anti-mouse antibody conjugated to 5 nm colloidal gold (1/100) for 30 min. The preparations were washed as above, post-fixed for 5 min in aqueous 1% glutaraldehyde and washed with PBS. They were subsequently incubated with 0.5% tannic acid, post-fixed with osmium tetroxide, dehydrated in ethanol and embedded in Epon. Sections parallel to the coverslips were observed by electron microscopy (Hitashi H600) after staining with uranyl acetate.

Thymocytes were fractionated as described by Gosti-Testu et al. (1986). For this, a thymus was mechanically dissociated in 450 ml PBS by pressing it through a 0.4 mm mesh stainless steel sieve. The cell suspension of thymocytes was filtered through a 0.25 mm mesh sieve and centrifuged at 32 g for 5 min to remove heavy contaminants. The supernatant was again filtered through a nylon mesh and centrifuged at 310 g for 8 min. Cells were resuspended in Nonidet-P40, 0.5% sodium deoxycholate in TNM buffer and vortexed briefly (2 min). The detergent-insoluble cell fraction, which contains nuclei and perinuclear cytoskeleton, and the detergentsoluble fraction were recovered after centrifugation at 300 g for 7 min. Protein determination of each fraction was done according to Lowry et al. (1951).

Samples (10 /d) of calf thymus centrosomes (50 μg in 50 μl Pipes buffer) were diluted to a final volume of 50 μl in Pipes buffer containing urea at final concentrations of 0, 2, 4, 6 and 8 M. After a 30 min incubation at 4°C, 25 μl samples were centrifuged on coverslips for 20 min at 100,000 g and doublelabelled with mAb 6C6 and polyclonal anti-tubulin antibody as described above.

The remaining sample was centrifuged for 20 min at 100,000 g. The pellet and the supernatant were boiled in SDS, 25 mM Tris, 10% glycerol, 1% 2-mercaptoethanol for 3 min and analysed on immunoblots.

Gel electrophoresis analysis of each protein fraction was performed in one dimension according to Laemmli (1970).

Protein were transferred onto PVDF Immobilon membranes according to Towbin et al. (1979) in the presence of 0.035% SDS. The membrane was saturated in PBS, 0.1% Tween 20 at room temperature for 1 h and all subsequent washes and antibody dilutions were performed in this buffer. The membrane was incubated with mAb 6C6 (1/10,000) for 1 h. After three washes, the membrane was incubated for 1 h with anti-mouse antibody labelled with horseradish peroxidase (1/10,000), washed three times as above and developed using the Amersham ECL Kit.

Isolated centrosomes were double-stained with an antitubulin antibody and mAb 6C6. As previously shown (Komesli et al. 1989) calf thymus centrosomes stained with anti-tubulin antibodies appear as pairs of bright spots in close proximity to each other, which correspond to centrioles (Fig. 1a). The same organelles are recognized by mAb 6C6 (as shown by comparison of Fig. 1 a and b), but the staining pattern is more diffuse and the two centrioles are, in most cases, no longer distinct (Fig. 1b). This difference in shape is not due to an artefact of the double-labelling procedure. The same characteristic is observed with mAb 6C6 alone (Fig. 1c).

The most likely explanation is that whilst anti-tubulin antibodies react with centrioles, mAb 6C6 marks the pericentriolar material, and this was confirmed by further experiments.

Immunoelectron microscopy localization of the mAb 6C6 antigen on isolated centrosomes showed that the epitope lay in electron-dense material at the periphery of the organelle (Fig. 2). No staining was observed in the inner part of the organelle under conditions that made the centriolar tubulin accessible to various monoclonal or polyclonal antibodies. However, none of these tubulin antibodies were IgMs.

It has been shown that centrosomes can be biochemically dissected by sequential extraction with urea solutions of increasing concentration (Klotz et al. 1990). Calf thymus centrosomes were extracted with urea solutions of 2, 4, 6 and 8 M urea and the insoluble fractions were double-stained with anti-tubulin and mAb 6C6 antibodies. The tubulin staining was completely abolished after treatment with 4 M urea solutions; and this corresponds to the solubilization of centriolar structures (Klotz et al. 1990; and Fig. 3 g and h). The mAb 6C6 signal persisted in these conditions. It progressively disappeared at urea concentrations of 6 M and 8 M (Fig. 3 d and e); this corresponds to the behaviour of the components of the pericentriolar material (Klotz et al. 1990). These results are consistent with the microscopy data and confirm that the mAb 6C6 antigen is a component of the pericentriolar material.

Immunoblots of purified centrosomal proteins using mAb 6C6 as a primary antibody, showed a major band of 180 kDa (Fig. 4B, lane 4). This band corresponds to a protein that is strongly enriched during centrosome purification. When whole thymocyte proteins were immunoblotted, results were negative (Fig. 4B, lane 1).

Centrosome preparation involves thymocyte lysis at low ionic strength (Komesli et al. 1989). When the crude protein fractions corresponding to the soluble and the insoluble parts of the cells were immunoblotted, results were also negative (Fig. 4B, lanes 2, 3).

These results suggest that the 180 kDa protein is specifically associated with purified centrosomes.

The 180 kDa polypeptide is the only one to be consistently detected on immunoblots. However, fainter variable bands are also visible. We used biochemical dissection of centrosomes, as described above, to purify further the mAb 6C6-specific antigen. Soluble and insoluble centrosomal protein fractions generated by urea extraction were immunoblotted. The 180 kDa protein remained associated with the insoluble material at 4 M urea concentration (Fig. 5, lane 4) and was progressively solubilized at higher urea concentrations (Fig. 5, lanes 7,9). This paralleled the behavior of the fluorescent signal (Fig. 3) and strongly suggested that this signal was causally related to the presence of the 180 kDa protein.

Lower molecular weight polypeptides remained associated with urea-insoluble material. Like the 180 kDa polypeptide, they copurified with the centrosomes and did not react with other IgMs. They might represent insoluble degradation products of the 180 kDa polypeptide.

In interphase cells, mAb 6C6 specifically and uniquely immunostains centrosomes (Fig. 6). This was observed in bovine thymocytes (Fig. 6a), bovine spleen cells (Fig. 6b), human lymphoblastic cells (Fig. 6c) and HeLa cells (Fig. 6d).

The centrosome staining persists during the entire cell cycle. However, from prometaphase to telophase, mAb 6C6 also shows sequential staining with other mitotic organelles (Fig. 7).

During prometaphase, a punctate staining pattern appears (Fig. 7 a, b and c). The punctate elements are not as intensely stained as centrosomes but they are always visible in prometaphase cells and are not observed with IgMs unrelated to mAb 6C6.

Late in metaphase, the spots gather in the metaphase plate (Fig. 7 d, e, and f) and this suggests that they are associated with kinetochores. This possibility was further substantiated by double-staining experiments with mAb 6C6 and a CREST antibody, showing a global co-localization of the spots stained by both antibodies (Fig. 8 a and b). Definitive evidence arose when isolated metaphase chromosomes were stained with mAb 6C6. Clear images of double spots located in the kinetochore region were observed (Fig. 8 d).

At the onset of anaphase, or immediately after, the mAb 6C6 antigen apparently dissociates from kinetochores (Fig. 7 g, h and i) and subsequently associates with the mid plate (Fig. 7 j, k and 1), finally concentrating in the midbody (Fig. 7 m, n and o). Even within the limits of resolution of optical microscopy the signal in the midzone region is clearly restricted to the central part of the furrow, which is not stained by antitubulin antibodies (Fig. 7 h), and this is the location of a recently described mitotic organelle, the telophase disc (Andreassen et al. 1991).

Previous studies have shown that phosphorylated epitopes can exhibit cyclic appearance and disappearance on mitotic organelles (Murray and Kirschner, 1989; Vandré and Borisy, 1989) and these observations raise the possibility that the mAb 6C6 antigenic determinant was phosphorylated.

Fixed cells and isolated centrosomes were extensively dephosphorylated using alkaline phosphatase. This did not modify the staining pattern (not shown). Similarly, dephosphorylation of purified centrosomal protein prior to immunoblotting (Davis et al. 1983) did not abolish the reactivity of the 180 kDa polypeptide (not shown). These results strongly argue against the possibility that mAb 6C6 reacts with a phosphorylated epitope.

We have also investigated the microtubule dependence of the centrosomal and kinetochore staining. As shown above, the kinetochore signal remains associated with isolated chromosomes.

Nocodazole treatment of cells, under conditions in which microtubules are fully disassembled, does not modify the centrosomal or kinetochore staining (not shown). In particular, the kinetochore staining is present in all nocodazole-treated cells showing condensed chromosomes. Obviously some of these cells entered mitosis after the dissolution of the microtubule network and this result shows that the association of the mAb 6C6 antigen with kinetochores is not microtubuledependent.

In higher plant cells, the functional equivalent of the mammalian pericentriolar material that functions as a MTOC, is distributed at the surface of the nucleus (Lambert, 1980; Vantard et al. 1990). Monoclonal antibody 6C6 staining of different plant cell types showed a series of closely adjacent spots around the nuclei (Fig. 9).

During mitosis, following nuclear envelope breakdown, the mAb 6C6 staining is no longer detected, but it reappears in late telophase when the nuclear envelope is reconstructed (Fig. 9c-e). No staining is observed either at the cell poles or at the kinetochores. We have verified that IgMs unrelated to mAb 6C6 do not stain plant cells.

Using double-labelling we experienced difficulties in ascertaining the correspondence between the foci stained by mAb 6C6 and the origins of interphasic microtubule clusters. A major problem was that microtubules were extremely abundant in the perinuclear region, and this and other factors made the localization of their origin ambiguous. However, as illustrated (Fig. 10), when the microtubule network was partially dissolved and clusters of microtubules with clearly defined origins were visible these origins appeared to correspond to spots stained by mAb 6C6.

The understanding of centrosomal functions in molecular terms depends to a large extent on the identification and structural characterization of centrosomal proteins. We still know little about the biochemical composition of centrosomes. Nevertheless the number of identified centrosomal proteins is growing. Amongst these proteins some are permanently and primarily associated with centrosomes during the cell cycle. These include a 350 kDa pericentriolar protein sharing antigenic determinants with lactate dehydrogenase (Gosti et al. 1987), a 74 kDa protein that shows analogy with intermediate filaments (Buendia et al. 1990), a 62–64 kDa pericentriolar protein involved in microtubule nucleation (Moudjou et al. 1991), a 185 kDa protein associated with the nucleus and centrosomes in Drosophila (Frasch et al. 1986; Whitfield et al. 1988), a 34.5 kDa protein, centrosomin A (Joswig et al. 1991) and a 59 kDa protein, SPA 1, which plays an important role in cell growth and chromosome segregation (Snyder and Davis, 1988). Finally y-tubulin, a pericentriolar component involved in microtubule nucleation, has recently been discovered by a molecular genetic approach (Horio et al. 1991; Steams et al. 1991; Zheng et al. 1991).

Some antibodies, primarily directed against MAPs or against non-centrosomal parts of the mitotic apparatus, also stain centrosomes in a cell cycle-dependent manner and/or specifically in the presence of microtubuleblocking agents (Kellog et al. 1989; Cande, 1990; Huang, 1990; Nislow et al. 1990; Baron et al. 1991; Compton et al. 1991; Leslie et al. 1991; Tousson et al. 1991).

The 180 kDa protein identified by mAb 6C6 differs from previously described proteins with regard to its apparent M_r and/or its intracellular location. It remains associated with centrosomes during the whole cell cycle in the presence or absence of microtubule blocking agents. Morphological and biochemical data converge to localize the mAb 6C6 antigen in the pericentriolar material.

Centrophilin, a pericentriolar protein of similar M_r, is not associated with centrosomes in interphase in the absence of microtubule-blocking agents, and shows a different nuclear distribution pattern (Tousson et al. 1991). Thus, mAb 6C6 most probably identifies a new pericentriolar component.

The pericentriolar material is thought to be derived from primitive microtubule organizing centers associated with DNA replication origins (Cavalier-Smith, 1987). In higher plant cells, which are acentriolar, microtubules are thought to be nucleated by pericentriolar material that has lost its association with centrioles and is distributed in distinct foci around the nuclear envelope (Vantard et al. 1990; Zhang et al.1990). According to this view, plant and animal MTOCs are expected to share common components.

In the past, human auto-antibodies (such as 5051 Ab) that recognize pericentriolar material have been found to stain the periphery of the nucleus and the spindle poles of higher plant cells (Clayton et al. 1985). However, the specificity of these antibodies has been questioned (Harper et al. 1989).

A clear example of similarities in the composition of animal and plant MTOCs has been provided by the identification of SPA 1, a 59 kDa protein associated with the pericentriolar material in animal cells, with spindle pole bodies in yeast, and with the nuclear envelope in interphase plant cells (Snyder and Davis,1988).

Our results, showing cross-reactivity of a pericentriolar component and of plant MTOCs with mAb 6C6, provide further evidence for a common origin for these structures and suggest that the mAb 6C6 antigen is a conserved element of MTOCs.

A survey of recent literature shows the emergence of a collection of antibodies that react with various mitotic organelles in a cell-cycle-dependent manner. The corresponding antigenic proteins have recently been termed “passenger proteins” (Earnshaw and Cooke,1991). The concept of “passenger proteins” relies on the assumption that the antibodies do not identify a covalent modification of proteins, which would occur in various places at different time points during the cell cycle, and that they do not react with different proteins on different organelles (Yen et al. 1991).

Sequential association with different mitotic organelles during mitosis has been described in the case of kinetochore proteins that subsequently become components of the midzone. This class of proteins includes INCENPs (Cooke et al. 1987; Earnshaw and Cooke, 1991), CENT E (Yen et al. 1991), TD 60 (Andreassen el al. 1991), class III antigens of Compton et al. (1991) and two polypeptides of M_r 140,000 and 155,000, described by Pankov et al. (1990).

Common features of these proteins are that they are located at the periphery of the kinetochore, that they tightly associate with kinetochores at prometaphase and that they suddenly leave this organelle at the onset of anaphase to be concentrated finally in the midbody. When detailed structural studies have been performed, they have revealed differences in their distribution in the midzone. For example, INCENPs are distributed along the midbody microtubule bundles (Earnshaw and Cooke, 1991) whilst TD 60 marks a disk that is not stained by anti-tubulin antibodies, and with which microtubules associate (Andreassen et al. 1991).

Recently, Compton et al. (1991) have described a monoclonal antibody (3G3), raised against kinetochores, reacting with centrosomes in metaphase, and with midbodies in telophase. The monoclonal antibody 6C6 provides a symmetrical example of an antibody raised against centrosomes and staining kinetochores and the midzone in a cell-cycle-dependent manner. The mAb 6C6 antigen behaves as a genuine centrosomal component during the entire cell cycle. It also associates with the kinetochores in metaphase. Subsequently, it shows the same remarkable behavior as the proteins described above, leaving kinetochores at the onset of anaphase apparently to become a component of the midbody, and most probably of the telophase disc (Andreassen et al. 1991).

The monoclonal antibody 6C6 can therefore be added to the growing list of antibodies raised against a mitotic organelle and identifying apparent “passenger proteins” (Earnshaw and Cooke, 1991). It suggests that exchanges among mitotic organelles during mitosis may concern many of their component elements.

The mAb 6C6 antigen association with kinetochores occurs whether or not cells are treated with microtubule-blocking agents, and this suggests that it is not transported from centrosomes to kinetochores along microtubules. It is therefore likely that it shifts from a poorly abundant, soluble pool to aggregated forms that coalesce at the location of mitotic organelles. This is the observed behavior of the pericentriolar material (Rieder and Borisy, 1982; Maro et al. 1985; Tassin et al. 1985; Karsenti and Maro, 1986; Bré et al. 1990) and probably applies to elements of the kinetochores and of the midbody as well. The association of the mAb 6C6 antigen with mitotic organelles may reflect the cellcycle-regulated appearance of specific binding sites during the organization of these organelles. Alternatively, the specific aggregation of specialized proteins in discrete locations might play a role in their organization (Compton et al. 1991; Earnshaw and Cooke, 1991; Tousson et al. 1991; Yen et al. 1991).

The potential significance of the apparent migration of proteins from the kinetochores to the mid-plate region is obvious. It provides an elegant explanation of how the position of the metaphase plate is related to the position of the cleavage plane in animal cells. The physiological significance of the variable association of the mAb 6C6 antigen with centrosomes, kinetochores and the telophase disc is less obvious. The most straightforward explanation is that it reflects a common feature of these organelles, which are all involved in interactions with microtubules. On the other hand, mitotic organelles are not merely microtubule-organizing centers. They share a large number of other properties related to their controlled duplication and/or to their involvement in cell organization. The appearance of common components on these organelles in a cell-cycle-dependent manner might well represent an important mechanism in the integration of cellular functions.

We are deeply indebted to Dr Michel Bornens for continuous advice and encouragement during this work and for communication of new results prior to publication. We are most grateful to Dr Rolande Berthier and Dr Annie Andrieux for their kind help during the early steps of monoclonal antibody production, and to Dr R.L. Margolis for reviewing our manuscript. We thank Mrs N. Milloz for her expert assistance during the preparation and typing of this manuscript.