Microinjection of a monoclonal antibody against SPN antigen, now identified by peptide sequences as the NuMA protein, induces micronuclei in PtK2 cells

NuMA protein was first described as a predominantly nuclear protein which relocated at mitosis to the pole regions. This protein with a molecular mass originally reported as 300 kDa was present in several human cell types but not in cells from other species (Lydersen and Pettijohn, 1980). It appeared to be autoantigen since NuMA antibodies were found in the sera of patients with rheumatic diseases (Price et al., 1984). Other high molecular mass proteins which relocate from the interphase nucleus to the poles of the mitotic spindle have also been characterized by their reactivity with monoclonal antibodies or with human autoimmune sera. These include centrophilin (Tousson et al., 1991), SP-H antigen (Maekawa et al., 1991), SPN antigen (Kallajoki et al., 1991) and the 1H1 and 1F1 antigens (Compton et al., 1991). Recently the cDNA for NuMA has been cloned from expression libraries by screening either with a human autoimmune NuMA serum (Yang et al., 1992) or with the monoclonal antibody 1F1 (Compton et al., 1992). NuMA protein expressed in bacteria encoded by the cDNA clone contained the epitopes for the 1H1 antibody, for the NuMA antibody 2E4 and for the centrophilin antibody 2D3 (Compton et al., 1992). Thus in spite of differences in the reported molecular mass, in staining patterns and in other properties reported in the literature, NuMA, centrophilin, 1F1 and 1H1 seem to represent different names for the same protein.

SPN antigen was originally characterized using monoclonal antibodies raised against a urea extract of nuclei isolated from the human adrenal cortex carcinoma cell line SW13 (Kallajoki et al., 1991). The SPN polypeptide had a reported molecular mass of 210 kDa and was a nuclear matrix component during interphase. During mitosis it relocated to the centrosome at prophase and accumulated at the spindle poles in metaphase and anaphase. During telophase it relocated to the reassembling nucleus. SPN antigen extracted from mitotic HeLa cells bound microtubules stabilized in vitro with taxol (Kallajoki et al., 1992), a property it shared with SP-H antigen (Maekawa et al., 1991). Treatment of mitotic cells with the microtubule disrupting drug nocodazole resulted in the dispersal of SPN antigen into many foci, which acted as microtubule organizing centers during recovery from the block (Kallajoki et al., 1991). A similar result was found for centrophilin by Tousson et al. (1991). In taxol-treated mitotic cells, SPN antigen was found at the center of the multiple microtubular asters induced by the drug (Kallajoki et al., 1992) and similar results have been reported with SP-H antigen (Maekawa et al., 1992). When five SPN antibodies were microinjected into the cytoplasm of HeLa cells, one antibody SPN-3 -caused a block in mitosis, spindle aberrations and resulted in cells with micronuclei (Kallajoki et al., 1991). These results suggest that SPN antigen acts as a microtubule minus-end organizer in mitotic cells and also plays a critical role in mitosis.

The question of whether SPN antigen and the NuMA antigen are identical is particularly important because the experiments just described show that SPN antigen plays an essential role in mitosis. Here we show that immunoprecipitated SPN antigen was recognized by a NuMA-type human autoimmune serum. In addition, immunoprecipitation with SPN-3 antibody was used to isolate SPN antigen for cyanogen bromide cleavage, while sequence analysis of three of the resulting peptides covering more than 50 residues showed a perfect fit to the protein sequence predicted for the human NuMA protein. These results show that the SPN antigen of HeLa cells is in fact NuMA. We have also extended the microinjection experiments with SPN-3 antibody on HeLa cells to PtK2 cells, so that we could inject cells at identifiable mitotic stages. Injection into the cytoplasm of interphase cells resulted after 24 hours in >50% cells with micronuclei. Injection of mitotic cells led to the formation of two micronucleated daughter cells in a stage-specific manner. Thus injection of prophase, prometaphase or metaphase cells resulted in the formation of two daughter cells with micronuclei whereas injection after the onset of anaphase resulted in the formation of daughter cells with normal nuclei. When metaphase cells were injected with SPN-3 and fixed after different times, daughter cells with micronuclei were first observed after 90 minutes. The finding that SPN and NuMA antigens are identical proteins, together with the results of the microinjection experiments, argue for a functional role of the NuMA protein in mitosis.

PtK2 cells were cultured in Minimal Essential Medium with Earle’s salts, glutamine, non-essential amino acids, 1 mM sodium pyruvate and 10% fetal calf serum. Large scale cultures (2 liters) of HeLa S3 cells were grown in spinner flasks in Joklik’s MEM supplemented with 10% fetal calf serum or 10% normal calf serum.

To obtain synchronous mitotic cells, cells were first grown in 2.5 mM thymidine for 18-22 h and then centrifuged for 5 min at 1500 revs/min in a Sorvall RC-5 centrifuge using the GSA rotor. The cells were resuspended in fresh medium with 0.06 μg/ml of colcemid and cultured for a further 16-20 h.

The mitotic index was determined from 10 ml of cells. Cells were pelleted by centrifugation for 10 min at 800 revs/min, suspended in 0.075 M KCl, and incubated for 10 min at 37°C. Then cells were again centrifuged for 10 min at 800 revs/min and fixed in methanol:acetic acid (3:1) for 10 min. The suspension was centrifuged for 10 min at 800 revs/min and the pellet was suspended in 0.5 to 1.0 ml of fixative. Drops of this suspension were smeared on microscope slides, air dried, and stained with Hoechst 33258 as described below. The mitotic index was >95%.

HeLa S3 cells, grown in suspension and synchronized as described above were harvested at a concentration between 3 × 10⁵ and 5 × 10⁵ cells/ml and washed three times in PBS (137 mM NaCl, 7 mM Na₂HPO₄, 1.5 mM KH₂PO₄, 2.7 mM KCl, pH 7.1) by centrifugation at 1000 revs/min for 5 min. The final pellet was suspended in ice-cold 0.5% Triton X-100 in PBS supplemented with 1 mM EGTA, 1 mM PMSF, 10 μg/ml aprotinin, 10 μM leupeptin, 1 μM pepstatin and 10 μM E-64 at a cell concentration of 2 × 10⁷/ml. Typically 30-50 ml were obtained from 2 liters of HeLa cells. Cells were incubated on ice for 5 min and then homogenized with a Potter-Elvehjem homogenizer. Cell disruption was confirmed by phase contrast microscopy. The suspension was centrifuged for 5 min at 13,000 g. The mitotic cell supernatant was aliquoted and stored at −70°C.

For large-scale immunoprecipitation of the SPN antigen, 10 ml aliquots of the mitotic HeLa cell extract were thawed and clarified by centrifugation in a Beckman ultracentrifuge TL-100 using the TLA-100.3 fixed angle rotor at 200,000 g for 20 min. Four ml of SPN-3 antibody (Kallajoki et al., 1991) supernatant from overgrown hybridoma cultures were added per 10 ml of cell extract. The mixture was incubated on ice for 1 h. 120 μl of affinity-purified rabbit antibody against mouse immunoglobulins (1.9 mg/ml; Dako, Glostrup, Denmark) was added and incubation continued for 30 min. Then 1 ml of pre-swollen Protein A-Sepharose beads (Pharmacia, Uppsala, Sweden) was added and the suspension was incubated for 1 h at 4°C with rocking. Beads were harvested by centrifugation at 1000 revs/min for 5 min and washed three times with ice-cold 0.5% Triton X-100 in PBS. The final pellet was suspended in 600 μl of 2× concentrated SDS-PAGE sample buffer, heated to 85°C for 10 min and centrifuged at 13,000 g for 5 min. The SPN supernatants were frozen at −20°C, and later used to isolate SPN antigen (see below).

Proteins were separated by SDS-PAGE on 7.5% acrylamide 0.5 mm thick minigels. Each 4 mm well was loaded with 5 μl of immunoprecipitated material or with 5 μl of supernatant fraction before or after collection of the immunocomplexes. Gels were either stained with Coomassie brilliant blue or used for immunoblotting. Proteins were transferred electrophoretically to nitrocellulose (0.2 μm pore size, Schleicher and Schuell Co., Dassel, FRG) in transfer buffer containing 25 mM Tris, 192 mM glycine, 0.01% SDS and 20% methanol at 250 mA constant current for 16 h. Protein transfer was controlled by Ponceau S staining. The sheets were blocked with 4% bovine serum albumin (BSA) in Tris-buffered saline (TBS: 20 mM Tris-HCl, pH 7.4, 0.15 M NaCl) overnight. Nitrocellulose sheets were incubated without primary antibody or with SPN-3 as undiluted culture supernatant, or with a NuMA-type human autoimmune serum (a kind gift from Dr. H. Ponstingl, German Cancer Research Center, Heidelberg, FRG) diluted 1:500 into 1% BSA, 0.2% Tween 20 in TBS. Incubation was for 2 h at 37°C. Washes were with 0.2% Tween 20 in TBS. The second antibody for SPN-3 was rabbit anti-mouse immunoglobulins conjugated to peroxidase (Dako Immunochemicals, Klostrup, Denmark) diluted 1:200. To detect the human NuMA autoantibody, an affinity-purified, peroxidaseconjugated, sheep anti-human immunoglobulins diluted 1:500 (Amersham, UK)was used. After 1 h incubation the nitrocellulose sheets were washed with TBS-Tween and the peroxidase was detected using 4-chloronaphthol as the chromogen.

The SPN supernatants in SDS-PAGE sample buffer described above were unfrozen and immediately boiled. They were then subjected to preparative SDS-PAGE using 7.5% polyacrylamide gels 1 mm in thickness. Gels were stained with Coomassie brilliant blue for 5 min and then destained in 7.5% acetic acid and 5% methanol for approximately 5 min. As soon as the SPN band became visible, the protein band was cut out. Gel pieces were suspended in equilibration buffer (0.125 M Tris-HCl, pH 6.8, 0.1% SDS, 1 mM EDTA). After 1 h the buffer was changed and equilibration continued for 1 h on a rocking table. The buffer was then removed and the gel pieces were stored at −20°C. When sufficient pieces had been collected, the SPN protein was electrophoretically eluted into dialysis tubing which had been washed with electrophoresis buffer. Electrophoresis was in 25 mM Tris, 192 mM glycine, 0.1% SDS at 150 V for 48 h. The solution (1 to 1.5 ml) was removed from the tubing and concentrated to about 400 μl in the Speed Vac (Savant Instruments). The protein was recovered by chloroform/methanol precipitation (Wessel and Flügge, 1984) and suspended in trifluoroacetic acid (TFA). Water was added to reach 70% TFA. Treatment with CNBr was in the dark at room temperature for approximately 16 h. The reaction was terminated by drying in the Speed Vac. The residue was extracted with 100 μl of 50% TFA followed by centrifugation. The soluble fragments recovered after drying were dissolved in 0.1 M Tris-HCl, pH 8.5, 6 M guanidine-HCl, 1 mM DTT and subjected to reversed phase HPLC using a Vydac 214 TP52 column. Solvent A was 0.1% TFA and solvent B was 0.08% TFA, 70% acetonitrile. Gradient elution was from 10% to 90% solvent over 90 min. Peak fractions were subjected to sequence analysis using an Applied Biosystems gas phase sequenator (model A470) or a Knauer sequenator (model 810). Both instruments were equipped with an on line PTH-amino acid analyzer.

These were grown on round 12 mm coverslips. Shortly before microinjection the medium was changed to the same medium without sodium bicarbonate but with 20 mM Hepes, pH 7.05. Small colonies usually containing 10-15 cells were injected using a semiautomatic microinjection apparatus (Eppendorf, Hamburg, FRG). Usually 30-40 cells were injected in each experiment in a marked area.The coverslips were placed in fresh medium without Hepes and the cells were kept at 37°C for 24 or 25.5 h before being fixed and subjected to immunofluorescence analysis.

For microinjection of mitotic cells, PtK2 cells were grown on 25 mm × 25 mm coverslips with etched grid markings (Bellco, Vinland, NJ, USA). In a single grid square a particular mitotic cell was identified and the stage of mitosis determined by phase contrast microscopy. This cell was then immediately microinjected. On each coverslip 10-20 mitotic cells were localized and injected within 20-30 min. After 3 h at 37°C cells were fixed and subjected to immunofluorescence analysis. In a second series of experiments involving metaphase PtK2 cells, 10 metaphase PtK2 cells were microinjected with SPN-3 antibody. Only a single metaphase cell was injected per coverslip. Cells were either fixed immediately after microinjection or after 2, 5, 10, 20, 30, 60, 90, 120 or 180 min at 37°C.

The SPN-3 antibody and the SPN-5 control antibody were purified from ascites fluid with a Protein G column (Mab Trap, Pharmacia).

HeLa cells were fixed in −10°C methanol for 10 min. HeLa cells were incubated with the human NuMA-type autoimmune serum diluted 1:500 for 1 h. The second antibody was FITC-conjugated goat anti-human IgGs (Miles Laboratories Inc., Kankaler, USA).

To identify microinjected PtK2 cells the cells were fixed in −10°C methanol for 10 min and incubated for 1 h with rhodamineconjugated, affinity-purified goat anti-mouse IgGs (Dianova, Hamburg, FRG). When detergent treatment was used, cells were extracted in microtubule stabilizing buffer containing 0.5% Triton X-100 for 15 s (see Kallajoki et al., 1991) prior to fixation with methanol. For double labelling of SPN antigen and tubulin, cells were fixed and incubated first with an affinity-purified rabbit-tubulin antibody at 20 μg/ml (Osborn et al., 1978). Then samples were incubated simultaneously with affinity-purified, fluorescein-conjugated sheep anti-mouse IgGs (Amersham) to reveal the SPN antibody distribution and with affinity-purified rhodamine-conjugated goat anti-rabbit IgGs (Dianova, Hamburg) to reveal the microtubular distribution. Cells were then stained for 1 min with Hoechst 33258 (20 μg/ml in 25% ethanol/75% PBS) and embedded in Mowiol 4.88 (Hoechst AG, Frankfurt, FRG). Microscopy was with a Zeiss Axiophot microscope. Micrographs were taken with Kodak T-Max 400 film processed at 1600 ASA.

SPN present in mitotic HeLa cell extracts was purified by immunoprecipitation using the monoclonal murine SPN-3 antibody followed by affinity-purified rabbit anti-mouse immunoglobulins. The resulting immune complexes were harvested by Protein A-Sepharose beads, extensively washed, and dissolved in hot SDS-sample buffer. SDS-gels of SPN antigen purified by immunoprecipitation were either stained with Coomassie brilliant blue or electrophoretically blotted on to nitrocellulose. Fig. 1A, lane 3, shows a Coomassie brilliant blue-stained gel of the SPN-3 precipitated material. A band corresponding to a molecular mass of 210 kDa was clearly visible and this band reacted when tested in immunoblotting with SPN-3 antibody as shown in Fig. 1B, lane 2. A faint band at ∼180 kDa was also visible in the immunoblots and is probably a breakdown product of the SPN antigen. Fig. 1A also documents the complex protein composition of the mitotic HeLa extract used for immunoprecipitation (lane 1) and the resulting supernatant fraction devoid of SPN antigen (lane 2). The SPN antigen was clearly a minor component of mitotic cells since it was not visible in lane 1.

Fig. 1B, lanes 3 and 4, shows that the polypeptide precipitated by the SPN serum was strongly decorated by a human NuMA-type autoimmune serum. Again weak staining of a 180 kDa polypeptide was detected.

When the SPN antibody and the NuMA antibody were used in immunofluorescence microscopy on HeLa cells, identical staining patterns were seen (compare Fig. 2A and C). Thus interphase nuclei were positive, and in addition strong staining was seen at the pole regions of metaphase cells.

Although the SPN polypeptide could be electrophoretically transferred from SDS-PAGE to nitrocellulose or polyvinyldifluoride membranes and subsequently detected by SPN antibodies, the efficiency of transfer was poor. Several variations of the electroblotting procedure of the SPN polypeptide present in the immunoprecipitate did not yield enough protein on the blot to allow subsequent microsequencing procedures. Therefore the immunoprecipitate was processed by preparative SDS-PAGE. After a short staining and destaining of the gel the 210 kDa band was excised. Gel fragments were collected and subjected to electroelution using the procedure described in Materials and methods. The SPN polypeptide was recovered and treated with cyanogen bromide (CNBr). Soluble fragments were separated on a C4 reversed phase HPLC column.

Although the elution profile obtained for the CNBr fragments was very complex, three fragments were pure upon sequence analysis. These were peptide 1-GDILQTPQFQ-, peptide 2-GNELERLXAALMESQ- and peptide 3-LKKAHGLLAEENRGLGERANLGRQF LEVE. When the cDNA sequence of human NuMA became subsequently available (Yang et al., 1992; Compton et al., 1992), the three peptide sequences we obtained on HeLa SPN showed a perfect fit with the predicted protein sequence of human NuMA protein. Peptide 1 occupies residues 206 to 215 and lies directly before the putative helical domain. Peptide 2 covers residues 978 to 992 located in the helical domain. The unassigned residue X is an arginine in both NuMA sequences. Peptide 2 is an overlapping fragment which arose due to incomplete cleavage at methionine 989. Peptide 3 spans residues 1440 to 1468 and is located towards the C-terminal end of the helical domain. In agreement with the known specificity of CNBr cleavage, all three peptide sequences are preceded by a methionine in the proposed protein sequence of NuMA. Our peptide sequences covering a total of more than 50 amino acid residues show that the SPN antigen of HeLa cells is NuMA.

Since rat kangaroo PtK2 cells stay relatively flat during mitosis, it is possible to inject mitotic cells. Here we document the results of injecting SPN-3 antibody (a) into interphase PtK2 cells, and (b) into mitotic PtK2 cells at defined mitotic stages. All experiments used SPN-3 antibody at a concentration of ∼5 mg/ml. As a control SPN-5 antibody was used at the same concentration, since this antibody does not affect mitosis of HeLa cells (Kallajoki et al., 1991).

In three experiments, SPN-3 antibody was injected into interphase cells and the cells were fixed and stained after 24 or 25.5 hours (Table 1). 58-77% of the antibody positive cells had developed micronuclei (Fig. 3 a-b). The other antibody positive cells showed only a single nucleus. Rounded cells, arrested in mitosis in a prometaphase-like state as reported for HeLa cells (Kallajoki et al., 1991), were not seen when of PtK2 cells were injected with SPN-3 antibody. In two control experiments, 75 interphase PtK2 cells were injected with SPN-5 antibody, and when examined one day later only a single cell had micronuclei.

Single cells in different stages of mitosis were injected with SPN-3 antibody. The stage of mitosis of a particular cell was determined by phase contrast microscopy and the cell was immediately injected with SPN-3 antibody. The positions of injected cells on the grid marked coverslip were noted. 10-20 such cells were injected on each coverslip and the cells were incubated at 37°C for 3 hours before fixation and immunofluorescence analysis. Results are summarized in Table 2 and Figure 4. In PtK2 cells, injection of SPN-3 antibody did not seem to cause mitotic arrest, but instead resulted in daughter cells that had micronuclei. The results depended on whether cells were injected before or after the onset of anaphase. Thus injection of prophase (Fig. 4a, b), prometaphase (Fig. 4c, d) or metaphase (Fig. 4e, f) cells resulted in formation of micronucleated daughter cells for 90%, 65% and 61% of the cells injected in each of these mitotic stages (Table 2, Fig. 4a-f). When injection was performed after the onset of anaphase, only 5% of the resulting daughter cells had micronuclei, while almost all other cells yielded two daughter cells with apparently normal nuclei (e.g. Fig. 4g, h). Very occasionally, injection of SPN-3 led to defective cytokinesis, resulting in one or three daughter cells with one, two or multiple nuclei (see Table 2). As a control, 20 prophase, prometaphase and metaphase cells were injected with SPN-5 antibody at an equivalent concentration (Fig. 5). This led to defective cytokinesis in one cell, resulting in a single cell with two nuclei. Two normal daughter cells were formed from each of the other 19 injected cells, regardless of the stage of mitosis used for injection. Examples are shown for cells injected with the control SPN-5 antibody in prophase (Fig. 5a, b), prometaphase (Fig. 5c, d) and metaphase (Fig. 5e, f).

In a second set of experiments, metaphase cells were identified and injected with SPN-3 antibody. They were incubated for different times, and then fixed and stained to visualize the injected SPN-3 antibody, the mitotic spindle and the chromosomes. Cells fixed immediately (Fig. 6a-c), and at 2, or 5 (Fig. 6d-f) or 10 minutes after injection showed normal mitotic spindles and normal chromosome arrangements as judged by staining with tubulin antibody and with Hoechst. The injected SPN-3 antibody was distributed diffusely throughout the cell with spindle pole regions detectable above the background of free antibody (Fig. 6a, d), demonstrating that the antibody recognized its native antigen inside the living cell. Cells injected with SPN-3 antibody in metaphase appeared to proceed through anaphase normally as judged by cells fixed 20 minutes after injection (Fig. 6g-i). In telophase the injected cells showed normal organization of microtubules in the intercellular bridge (Fig. 6k, n), but the decondensing chromosomes appeared strangely clumped in cells fixed 60 minutes after injection (Fig. 6j-l). Micronucleated cells were first seen 90 minutes after injection of metaphase cells, and showed a normal organization of the interphase microtubule network (Fig. 6n).

When SPN-3 antibody is injected into interphase PtK2 cells, after 24 hours 58-77% of the cells had micronuclei, while the remaining cells had apparently normal nuclei (Table 1; Fig. 3). In contrast, when the same experiment was performed with HeLa, approximately half of the cells had micronuclei while most of the remaining cells were rounded and arrested in a prometaphase-like state. These cells showed distorted mitotic spindles (Kallajoki et al., 1991).

Micronuclei can also be induced by mitotic inhibitors such as colchicine, colcemid, vinblastin and griseofulvin as well as by X-irradiation (Ringertz and Savage, 1976). The percentage of cells that develop micronuclei depended on the mitotic inhibitor, the dose of the drug and the time of exposure. Cells from different species also displayed different sensitivities to the same mitotic inhibitor. Drugs such as colcemid also induced mitotic arrest.

We therefore compared the response of interphase PtK2 and HeLa cells to colcemid and taxol treatment. A striking difference in response of the two cell types to the drugs was observed. After 24 hours in either colcemid (0.5 μg/ml) or in taxol (20 μg/ml), ∼80% of the PtK2 cells had micronuclei while the others were flat, still attached to the coverslip and had only a single nucleus (Fig. 7a-c). In contrast, after 24 hours in either drug the vast majority of HeLa cells (>95%) were rounded, and arrested in a prometaphase-like state (Fig. 7b, d). The majority of flat spread cells attached to the coverslips had normal nuclei while a minority had micronuclei.

Thus PtK2 cells respond to both SPN-3 antibody injection, and to drug treatment, by preferentially forming micronucleated cells (∼70% micronucleated after SPN-3 injection vs 80% after drug treatment). In contrast HeLa cells respond to SPN-3 antibody injection and to drug treatment in a qualitatively similar but quantitatively different manner. Thus 50% of the cells are micronucleated after SPN-3 injection vs 1% after drug treatment, and ∼50% of cells are arrested in a prometaphase-like state after SPN-3 injection vs 95% after drug treatment.

Our results show that SPN antigen is in fact NuMA protein. Immunoprecipitated SPN antigen is recognized on immunoblots by a human autoimmune NuMA antiserum (Fig. 1B). In addition, sequence analysis of three peptides derived from HeLa SPN show identity with the predicted amino acid sequence of human NuMA protein (see Results). Antibodies 1F1 and 1H1 (Compton et al., 1991), the centrophilin antibody 2D3 (Tousson et al., 1991) and the NuMA antibody 2E4 (Lydersen and Pettijohn, 1980) all react with recombinant NuMA protein expressed in E. coli (Compton et al., 1992). Thus it is clear that at least NuMA protein, the 1F1 and the 1H1 antigens, centrophilin and SPN antigen are the same protein. SP-H antigen (Maekawa et al., 1991), characterized with a human autoantibody, shows very similar behaviour during the cell cycle and a similar molecular mass to the antigens listed above and also shares other properties with SPN and the other antigens (Kallajoki et al., 1992; Compton et al., 1992). Thus it seems probable that all six different antigens represent different names for the same protein. Alternatively, since a functional role for SPN antigen has been directly demonstrated by microinjection experiments, the same function can now be assumed to have been shown for NuMA/centrophilin/1F1/1H1 and probably also for SP-H antigen.

A total of 108 cells in different stages of mitosis were injected. Injection of SPN-3 antibody into PtK2 cells in prophase, prometaphase or metaphase resulted in abnormal mitosis in 90%, 78% and 77% of cells, respectively. Usually each injected cell yielded two daughter cells with micronuclei (Table 2 and Fig. 4). In contrast, injection into PtK2 cells in anaphase resulted in abnormal mitosis in only 16% of the injected cells while the other daughter cells had normal nuclei. Anaphase and/or telophase are not particularly short when compared to other mitotic stages. Thus for PtK 2 cells prometaphase is about 12 minutes (range 5-15 minutes), metaphase about 16 minutes (range 7.5-18 minutes), anaphase about 7 minutes (range 5-10 minutes) and telophase 4-7 minutes (De Brabander et al., 1986). Thus it seems unlikely that the results with anaphase cells could be explained by the rate of SPN-3 antibody binding. Instead our results suggest a central role for SPN/NuMA beginning in early prophase, and ending in early anaphase. Thus the critical time interval, in which SPN/NuMA has to be functional, starts with the relocation of SPN/NuMA to the centrosomes at the beginning of prophase. In taxol-treated cells the striking redistribution of SPN (and SP-H) antigen into multiple foci which act as organizing centers for the multiple microtubule asters also occurs during the same time interval (Maekawa et al., 1991; Kallajoki et al., 1992). That injection of SPN-3 antibody into PtK2 cells does not cause abnormal spindles but does prevent taxol-induced aster formation in cells injected in early mitotic stages (Kallajoki et al., 1992) may perhaps be explained by the different structures of the spindle pole and the taxol-induced asters. For example taxol-induced asters contain neither centules nor pericentular components such as the 5051 antigen. Thus taxol-induced asters may be more labile than spindle poles in PtK2 cells

One day after injection of SPN-3 antibody into the cytoplasm of interphase PtK2 cells, 58-77% of cells had micronuclei. Presumably the antibody gains access to the SPN antigen as the nuclear membrane breaks down at the onset of prophase. Daughter cells with micronuclei would then be formed in a manner analogous to that discussed above for cells injected in the early mitotic stages. Injection of the same antibody into interphase HeLa cells also resulted in 38-56% of the cells being micronucleated one day after microinjection (Kallajoki et al., 1991).

In HeLa cells it was relatively easy to demonstrate that injection of SPN-3 results not only in micronuclei formation but also in abnormal spindle formation, since a substantial fraction (up to 60%) of the cells were found arrested in a prometaphase-like state. As judged by tubulin staining, mitotic spindle microtubules were present, but the spindles were abnormal and multipolar, and the chromosomes were widely scattered (Kallajoki et al., 1991). In the current study with PtK2, cells arrested in prometaphase with abnormal spindles were not seen. In addition, again at the light microscope level, abnormalities in spindle structure were not seen by tubulin staining when metaphase cells were injected with SPN-3 antibody and followed for different times (Fig. 6). Our results show that interphase HeLa and PtK2 cells also behave differently when treated for long times with mitotic drugs such as taxol and colcemid. Thus at 24 hours the vast majority of HeLa cells became arrested in prometaphase. Mitotic spindle microtubules were not formed, and only a small minority of the cells contained micronuclei. In contrast, under the same conditions, ∼80% of PtK2 cells had micronuclei, and cells arrested in mitosis were not found. Thus the differences seen after microinjection of SPN-3 antibodies into the two cell types is paralleled by the differences seen after drug treatment of the two cell types.

Several functions have been proposed for NuMA protein. Originally it was suggested that NuMA plays a role in postmitotic nuclear assembly (Lyderson and Pettijohn, 1980; Pettijohn et al., 1984; Price and Pettijohn, 1986). NuMA proteins relocate to the nucleus before nuclear lamins, again suggesting a role in nuclear assembly (Yang et al., 1992). Although the cDNA sequence of NuMA protein did not show significant homology to any other protein in the data bank, the central domain of NuMA is similar to coiled-coil forming regions of structural proteins such as myosin and intermediate filament proteins (Yang et al., 1992; Compton et al., 1992). Thus like these proteins NuMA may form filaments. Recently a nuclear skeleton with 10 nm diameter filaments and with a 23 nm axial repeat has been described (He et al., 1990; Jackson and Cook, 1988). These observations and the association of NuMA (Lyderson and Pettijohn, 1980) and SPN antigen (Kallajoki et al., 1991) with the nuclear matrix suggest that NuMA may be a structural component of interphase nuclei.

A role for NuMA in spindle function is supported by the localization of NuMA proteins to mitotic spindle pole regions (Lyderson and Pettijohn, 1980; Tousson et al., 1991; Kallajoki et al., 1991, 1992; Maekawa et al., 1991). NuMA proteins also located to sites of microtubule nucleation during recovery or after treatment with drugs such as nocadazole and taxol (Tousson et al., 1991; Maekawa and Kuriyama, 1991; Kallajoki et al., 1991, 1992). In cells treated with drugs in prophase or prometaphase, the NuMA protein is invariably found at the centers of multiple microtubule asters (5-20/cell) (Tousson et al., 1991; Maekawa and Kuriyama, 1991; Kallajoki et al., 1991, 1992). Injection of SPN-3 antibody into taxol-treated PtK2 cells prevented the formation of such asters (Kallajoki et al., 1992). In addition SPN or SP-H antigen from mitotic HeLa cells bound to microtubules in vitro (Maekawa and Kuriyama, 1991; Kallajoki et al., 1992), suggesting that the NuMA protein may function as a microtubule minus end organizer in vivo. However we note, as did Yang et al. (1992) and Compton et al. (1992), that the sequence of the NuMA protein does not reveal known microtubule binding motifs.

Our microinjection experiments with the SPN-3 antibody (Kallajoki et al., 1991, 1992, and this paper) are currently the strongest argument for a direct role of NuMA in mitosis. Clearly the target in such injection experiments is the SPN/NuMA protein. Thus inactivation of this protein can lead to micronuclei in both HeLa and PtK2 cells. Mitotically arrested cells with abnormal spindles have been demonstrated after injection of HeLa, but not of PtK2 cells. Of particular interest is the unexpected similarity in the effects of antibody injection and of treatment with antimitotic drugs such as colchicine or taxol. These drugs bind directly to tubulin or microtubules. Drug treatment led to micronuclei in low numbers in HeLa cells and in large numbers in PtK2 cells. Treatment of HeLa cells with colcemid resulted in large numbers of cells which are mitotically arrested and which do not contain microtubules in the mitotic spindle. Thus although SPN-3 antibody injection and colcemid target different proteins, the net result in HeLa cells was the same, i.e. micronucleation and mitotically arrested cells due to an inactive spindle in SPN-3 antibody injected cells or to a non-existent spindle in colcemidtreated cells. In PtK2 cells, both SPN-3 antibody injection and drug treatment led to micronuclei formation. The mechanism proposed by Ringertz and Savage (1976) for micronuclei formation supposes that the drug initially blocks cells in prometaphase, resulting in cells with widely scattered chromosomes, which cannot form a metaphase array because of microtubule disturbances. Micronuclei are formed as the nuclear envelope re-forms around individual, or groups of chromosomes. Thus the difference between HeLa and PtK2 cells with both SPN antibody injection and drug treatment may be that while many HeLa cells are blocked in prometaphase, PtK2 cells are able to progress further through mitosis and form micronuclei. Major differences in the response of human and rodent cell lines to drugs such as colcemid have been noted (Kung et al., 1990; Rieder and Palazzo, 1992). Kung et al. (1990) have shown that differences exist in the ability of different cell lines to progress into the next cell cycle in the absence of mitosis. They have further suggested that such differences may be linked to different levels of cyclin B and cdc2 kinase found in arrested HeLa and CHO cells.

Antibodies to other spindle components have also been injected into cells, and their effects can be compared and contrasted to those of the SPN-3 antibody. When injected in concentrations higher than 6 mg/ml into interphase cells a rat monoclonal antibody YL 1/2 reacting specifically with the tyrosylated form of alpha-tubulin, prevented disassembly of the cytoplasmic microtubule complex and therefore spindle formation, so cells were arrested in mitosis (Wehland and Willingham, 1983). 80% of the PtK2 cells injected in prophase with YL 1/2 did not divide and showed abnormal spindle structures after 2 hours. Polyclonal antibodies with high affinity for beta-tubulin were found to disrupt cytoplasmic microtubules after microinjection of interphase PtK2 cells, whereas mitotic microtubules were resistant to even high antibody concentrations and cells were able to proceed through mitosis and divide normally (Füchtbauer et al., 1985). The effects of injecting antibodies against a variety of non-tubulin spindle associated components have been summarized by McIntosh and Koonce (1989). Antibodies to CENP-A, -B, -C and -E (Bernat et al., 1990; Yen et al., 1991) and to the CHO-1 antigen (Nislow et al., 1990) have recently also been injected into cells. Looking at the effect of injection of antibodies to such proteins only antibodies to p13suc1 are reported to cause mitotic abnormalities and micronuclei formation (Riabowoi et al., 1989). The effects seen when p13suc1 antibodies are injected into fibroblasts are very similar to those seen with the SPN-3 antibody. However the distribution of SPN/NuMA antigen and p13suc1 during the cell cycle seem different.

Finally, NuMA/SPN antigen could also have dual or multiple functions. Thus it could have a structural role in the interphase nuclei and it could act as a microtubule minus end organizer during mitosis. Further experiments, e.g. with in vitro microtubule nucleation model systems, with in vitro systems for studying nuclear assembly, or with transfected cells are necessary to get a deeper insight to the function of the NuMA protein.

The data in this paper were presented at the International Cell Biology meeting in Madrid in July 1992. We thank Uwe Plessmann, Susanne Isenberg and Monika Dietrich for expert technical assistance, and Claudia Hake and Jaoko Liippo for their photographic expertise. M.K. was supported by a long-term EMBO fellowship. This work was supported in part by a grant from the Deutsche Forschungsgemeinschaft to M.O. (Os 70/2-1).