Overexpression of the Csk homologous kinase (Chk tyrosine kinase) induces multinucleation: a possible role for chromosome-associated Chk in chromosome dynamics

Protein phosphorylation is well known to regulate cell cycle progression through activation and inactivation of cyclin-dependent kinases. Multiple protein kinases and phosphatases, most of which are serine/threonine-specific, are involved in spindle assembly and chromosome segregation (Nigg et al., 1996).

The Src family of non-receptor-type protein-tyrosine kinases consists of proto-oncogene products and related proteins, and is involved in a variety of signal transduction events underlying cell activation and growth (Thomas and Brugge, 1997). The members of the Src family include c-Src, c-Yes, Fyn, Lyn, c-Fgr, Hck, Lck and Blk. Some Src family members are expressed in a variety of tissues, while others tend to be restricted to specific cell types of hematopoietic origin. The tyrosine kinase activity of Src family kinases is tightly regulated by tyrosine phosphorylation and dephosphorylation (Cooper, 1990). The non-receptor-type tyrosine kinase Csk has been shown to function as a ubiquitous negative-regulator of Src family kinases (Okada et al., 1991; Cooper and Howell, 1993).

The expression of a second member of the Csk family, the Csk homologous kinase Chk, is restricted in hematopoietic and neuronal cells. Both Csk and Chk have Src homology 3 (SH3) and SH2 domains and lack the consensus tyrosine phosphorylation and myristoylation sites found in Src family kinases. Like Csk, Chk is shown to suppress the activity of Src family kinases (e.g. Lck, Fyn, c-Src and Lyn) in vitro or in a yeast co-expression system by phosphorylating their C-terminal negative regulatory tyrosine residues (Klages et al., 1994; Chow et al., 1994; Avraham et al., 1995; Davidson et al., 1997; Hirao et al., 1997). Chk was also found to selectively suppress the kinase activity of Lyn but not c-Src in platelets and megakaryocytic Dami cells in vivo (Hirao et al., 1997; Hirao et al., 1998).

There is increasing evidence that Src family kinases are involved in mitotic functions. The Src family kinases present in fibroblasts (c-Src, Fyn and c-Yes) are shown to become activated at the G₂/M phase as a consequence of dephosphorylation of their C-terminal negative regulatory tyrosine residues (Taylor and Shalloway, 1993; Roche et al., 1995). In addition, different Src family kinases are located in various compartments of cells. For instance, c-Src is found mainly in perinuclear membranes, including endosomes and the microtubule organizing center, in fibroblasts (David-Pfeuty and Nouvian-Dooghe, 1990; Kaplan et al., 1992). In T lymphocytes, whereas Lck is localized to the plasma membrane, Fyn is associated with the centrosomes and the mitotic spindles and poles (Ley et al., 1994). The distinct subcellular localization of each enzyme may be critical for its function. However, the roles of Src family kinases in mitosis remain elusive.

In this study, we show that Lyn and Chk associate with mitotic chromosomes and spindles in immature myeloid KMT-2 cells. Chk overexpression results in mitotic aberration accompanied by changes in tyrosine phosphorylation of proteins associated with mitotic chromosomes and spindles. Our findings lead to the hypothesis that Chk-dependent perturbation of tyrosine phosphorylation states, possibly via Lyn in mitotic chromosomes and spindles, may affect chromosome dynamics.

The human immature myeloid cell line KMT-2 (Tamura et al., 1990) was cultured in Iscove’s modified Dulbecco’s medium containing 10% fetal bovine serum in the presence of 2 ng/ml granulocyte-macrophage colony-stimulating factor (a gift from Nippon Hoechst Co., Ltd). KMT-2 cells were transfected by electroporation with the pMKITneo vector (a gift from Drs K. Maruyama and T. Yamamoto) or the pMKITneo vector encoding human Chk or Chk tagged with the FLAG epitope (Chk-FLAG) as described (Hirao et al., 1997), and stably transfected cell clones were selected in 400 μg/ml G418. Monoclonal antibodies to Chk (13G2) (Hirao et al., 1997), Csk (#52, Transduction Laboratories), phosphotyrosine (4G10, Upstate Biotechnology Inc.), FLAG (M2, Kodak), c-Src (clone 327, Oncogene Research Products), Lyn and Yes (Lyn8, Lyn9 and 1B7, Wako Chemicals, Osaka, Japan), DNA topoisomerase II (#2010-1, TopoGEN), histone H1 (AE-4, Leinco Technologies), CD34 (4A1, Nichirei, Tokyo, Japan), α-tubulin (YOL1/34 and YL1/2, Serotec) and a control antibody (MOPC21, Sigma) were used. Polyclonal antibodies to Lyn (Lyn44), Yes (Yes3), Fyn (FYN3), c-Fgr (c-Fgr:N-47), Blk (Blk:K-23), Hck (Hck:N-30) and cytoplasmic phospholipase A2 (cPLA2:N-216) were obtained from Santa Cruz Biotechnology. Antibodies to Lck (Lck:NT) and SHP-1 (SH-PTP1) were obtained from Upstate Biotechnology Inc. Horseradish peroxidase-conjugated F(ab′)₂ fragments of anti-mouse Ig, anti-rat Ig and anti-rabbit Ig were purchased from Amersham. FITC-conjugated F(ab′)₂ fragments of anti-rat IgG or anti-mouse IgG were from BioSource.

COS-1 cells were transiently transfected with pMKITneo vector alone or pMKITneo vector encoding human Chk-FLAG using TransIT Transfection Reagent (Mirus Corporation), according to the manufacturer’s instructions.

Cells were cultured with or without 0.4 μg/ml nocodazole for 12 hours. Cell lysates were recovered from unsynchronized and nocodazole-arrested cells solubilized at 4°C in Triton lysis buffer [50 mM Hepes, pH 7.4, 10% glycerol, 1% Triton X-100, 0.25% deoxycholate, 100 mM NaF, 10 mM EDTA, protease inhibitors (50 μg/ml aprotinin, 100 μM leupeptin, 25 μM pepstatin A, 2 mM benzamidine, 100 μM bestatin, 100 μM chymostatin, 1 mM phenylmethylsulfonyl fluoride) and 1 mM Na₃VO₄].

Cells were swollen in hypotonic buffer (30 mM Hepes, pH 7.4, 1 mM MgSO₄ and protease inhibitors), followed by mechanical rupture through a 26-gauge needle at 4°C. After adjusting the salt concentration to 150 mM NaCl, the supernatants were recovered by centrifugation at 2,500 g for 2 minutes and separated into soluble (S100) and particulate (P100) fractions by ultracentrifugation at 100,000 g for 30 minutes at 4°C.

(1) Unsynchronized cells were swollen in swelling buffer (7.5 mM Tris-HCl, pH 7.4, 40 mM KCl, 1 mM EDTA-KOH, 0.1 mM spermine, 0.25 mM spermidine, protease inhibitors, and 1 mM Na₃VO₄), as described (Gasser and Laemmli, 1987; Taagepera et al., 1995). The cell pellets were solubilized at 4°C in twice-concentrated swelling buffer containing 1% CHAPS and nuclei were collected by centrifugation at 200 g for 3 minutes. (2) Cells were exposed to 1 μg/ml aphidicolin for approx. 22 hours, washed free of aphidicolin, and then cultured for approx. 18 hours in 0.1 μg/ml nocodazole. The cells were swollen and solubilized as above. Lysates were centrifuged at 200 g for 3 minutes repeatedly until no contamination of nuclei and cell debris was detected. The resultant supernatants were loaded on glycerol gradients, which consisted of 40% and 80% glycerol layers, and centrifuged at 4°C for 5 minutes at 1000 g and 20 minutes at 4000 g. Mitotic chromosomes were collected from the 40/80% interface and in the 80% glycerol. Mitotic chromosomes were solubilized at 4°C with Empigen lysis buffer [50 mM Tris-HCl, pH 9.0, 25 mM KCl, 4 mM EDTA-KOH, 10% glycerol, protease inhibitors, 1 mM Na₃VO₄ and 1% Empigen BB (Calbiochem)], as described for nuclear matrices (Staufenbiel and Deppert, 1984). (3) Mitotic chromosomes were treated with 80 μg/ml DNase I in DNase buffer (20 mM Tris-HCl, pH 7.4, 3 mM MgCl₂, 3 mM CaCl₂ and protease inhibitors), as described (Gasser and Laemmli, 1987; Taagepera et al., 1995). The supernatants (S-1) were separated and the pellets were incubated for 20 minutes at 4°C in histone buffer (10 mM Tris-HCl, pH 9.0, 2 M NaCl, 10 mM EDTA, 0.1% CHAPS and protease inhibitors). After the supernatants (S-2) were collected, the residual pellets were recovered as scaffolds. The S-1 and S-2 fractions were combined as nonscaffolds.

Cells were exposed to 2 mM thymidine for 1 day, washed free of thymidine, and cultured for 5 hours. The cells were incubated with 0.1 μg/ml nocodazole for 22 hours, washed to remove nocodazole, and then incubated for 45 minutes to proceed through mitosis. Enriched mitotic cells were cultured for a further 15 minutes in 10 μg/ml taxol to stabilize mitotic spindles in vivo. The cell pellets were washed with distilled water, and solubilized in isolation buffer (2 mM Pipes, pH 6.9, 0.25% Nonidet P-40, protease inhibitors and 10 μg/ml taxol), as described (Zieve and Solomon, 1982; Kuriyama et al., 1984). To remove any contaminants, the mitotic spindles devoid of chromosomes were sequentially washed with isolation buffer and RIPA buffer (50 mM Hepes, pH 7.4, 1% Triton X-100, 1% deoxycholate, 0.1% SDS, protease inhibitors and 1 mM Na₃VO₄).

Triton X-100 cell lysates were subjected to SDS-polyacrylamide gel electrophoresis and electrotransferred onto polyvinylidene difluoride membranes (Millipore). Immunodetection was performed by enhanced chemiluminescence (Amersham) as described (Hirao et al., 1997). For immunoprecipitation, Triton X-100 cell lysates were incubated at 4°C with protein G-Sepharose beads (Pharmacia) precoated with the appropriate mAb. The immune complexes were analyzed by SDS-polyacrylamide gel electrophoresis and western blotting. Sequential reprobing of membranes with a variety of antibodies were performed after the complete removal of primary and secondary antibodies from membranes in stripping buffer, according to the manufacturer’s instructions.

After washing immune complexes with modified RIPA buffer (50 mM Hepes, pH 7.4, 1% Triton X-100, 0.25% SDS, 4 mM EDTA, 1 mM Na₃VO₄ and 2 mM phenylmethylsulfonyl fluoride) and Triton X-100 lysis buffer containing 2 M NaCl, a portion of each immunoprecipitate was subjected to an in vitro kinase assay with 2.5 μM [γ-³²P]ATP and an equal portion was applied to a quantitative immunoblot, as described previously (Hirao et al., 1997; Hirao et al., 1998). Acid-denatured enolase was used as an exogenous substrate for Src family kinases. The samples were separated on SDS-polyacrylamide gels. The gels were treated with 1 N KOH at 56°C for 2 hours and subjected to a BAS 2000 BioImage Analyzer (Fuji Photo Film Co., Tokyo, Japan).

Cells were cultured in the presence of 0.05 μg/ml colcemid for approx. 20 hours, and washed with phosphate-buffered saline. After incubation for 30 minutes in 75 mM KCl, the cells were fixed in methanol:acetic acid (3:1), allowed to air-dry on glass slides, and stained with Giemsa.

Exponentially growing cells were washed with phosphate-buffered saline, and fixed with 70% ethanol at 4°C for 60 minutes. After washing with phosphate-buffered saline, the cells were treated with 100 μg/ml RNaseA followed by addition of 100 μg/ml propidium iodide for 15 minutes to stain DNA. The propidium iodide fluorescence of nuclei in individual cells was measured using a FACScan (Becton Dickinson).

Cells were placed in a chamber on a glass slide, and maintained above 30°C in a warmed room. Nomarski differential-interference-contrast images were monitored using a Fluoview laser scanning microscope (Olympus, Tokyo).

Cells were cytocentrifuged and fixed with 3.7% paraformaldehyde for 20 minutes, permeabilized with phosphate-buffered saline containing 5% bovine serum albumin and 0.1% saponin for 30 minutes as described (Yamaguchi and Fukuda, 1995), and incubated for 1 hour with either the rat anti-tubulin antibody YOL1/34, the mouse anti-Chk antibody 13G2 or the mouse anti-FLAG antibody. The cells were washed, incubated with FITC-conjugated anti-rat IgG or FITC-conjugated anti-mouse IgG for 1 hour, and treated with 100 μg/ml RNaseA plus 100 μg/ml propidium iodide for 15 minutes.

Mitotic chromosomes were cytocentrifuged, washed with HMEV (50 mM Hepes, pH 7.4, 5 mM MgCl₂, 10 mM EGTA, 10 mM Na₃VO₄) containing 0.1% CHAPS, and fixed with 1% paraformaldehyde in HMEV for 10 minutes. The chromosomes were blocked with HBSV (50 mM Hepes, pH 7.4, 150 mM NaCl, 10 mM Na₃VO₄, 0.1% saponin) containing 5% bovine serum albumin for 30 minutes, and stained with either the anti-Chk antibody 13G2 or the isotype-matched control MOPC21 antibody in conjunction with FITC-anti-mouse IgG, followed by staining with 20 μg/ml propidium iodide for 25 minutes. Confocal images were obtained using a Fluoview laser scanning microscope or an LSM510 laser scanning microscope (Zeiss), as described (Tada et al., 1999). Care was taken to ensure that detection sensitivity was kept constant throughout the study and that there was no bleed-through from the fluorescein signal into the red channel. Emission signals were detected at between 505 and 530 nm for fluorescein and more than 650 nm for propidium iodide (Mera et al., 1999). All Z-series sections at 0.3-0.5 μm intervals were merged.

To examine the effect of Chk on cell growth, we transiently expressed the FLAG-epitope tagged version of Chk (Chk-FLAG) at large amounts in COS-1 cells. Fig. 1A shows that ectopically expressed Chk-FLAG was clearly visualized in nuclei (arrows) and the cytoplasm of COS-1 cells by immunofluorescence staining with anti-FLAG. Some of the transfected cells were stained in both nuclei and the cytoplasm, and the others were preferentially stained in nuclei. Expression of Chk-FLAG was observed in about 50% of the total cells. Using anti-Chk, similar results were obtained in COS-1 cells transiently expressed with intact Chk (data not shown).

We examined whether ectopic expression of Chk affected proliferation of COS-1 cells. Fig. 1B shows typical proliferation curves of vector- and Chk-transfected cells, indicating that expression of Chk did inhibit proliferation of COS-1 cells. Given that Chk expression was detected in at most 50% of the total number of cells, the degree of inhibition of cell proliferation could correspond to the efficiency of Chk transfection.

Chk is restrictedly expressed in hematopoietic and neuronal cells. To investigate the role of Chk in hematopoietic cell growth, we stably overexpressed Chk in human immature myeloid KMT-2 cells that express a significant amount of endogenous Chk, and established independent cell clones (Fig. 2A). Overexpression of either intact Chk or Chk-FLAG brought about enlargement of cells in individual clones (Fig. 2B, KMH and KMF clones). Cell proliferation assays showed that Chk overexpression prolonged the average doubling time of approx. 1.5 days in control cells to approx. 2 days (Fig. 2C). Chk-overexpressing cells were mainly hypotetraploid and some possessed more than 110 chromosomes (Fig. 2Db,c), although vector-transfected cells (KMV-1 clone) showed 45 chromosomes per cell, as observed in parental KMT-2 cells (Fig. 2Da; Tamura et al., 1990). Noticeably, FACS analysis showed that the ratio of Chk-overexpressing cells (KMH-15 and KMF-6 clones) in the G₂/M phase (8N positions) to those in the G₁ phase (4N positions) was larger than that of control cells (KMV-1 clone) in the G₂/M phase (4N position) to those in the G₁ phase (2N position) (Fig. 2Dd-f), and that only a small fraction of apoptotic cells with reduced DNA contents (<4N) was observed in Chk overexpression (Fig. 2De,f). Fig. 3 shows that Chk-overexpressing cells displayed high proportions of polymorphonuclear shape and multinucleation (C-F, short arrows). The aberrant phenotypes included the appearance of various sizes of aligned chromosome masses (long arrows). These results suggest that Chk overexpression augments the number of cells in the G₂/M phase. Inhibition of cell proliferation by Chk overexpression may be primarily explained by a blockade in mitosis, not by induction of apoptosis.

We examined mitotic states of living cells in a time-lapse manner under a differential-interference-contrast microscope. Fig. 4 shows representative mitotic phenotypes in control (Fig. 4A) and Chk-overexpressing cells (Fig. 4B-D). While mitotic chromosomes in control cells were properly aligned and segregated (Fig. 4A), mitotic chromosomes in Chk-overexpressing cells were loosely aligned and separated into two masses of mitotic chromosomes without complete alignment (Fig. 4B). A Chk-overexpressing cell with four masses of mitotic chromosomes divided into two multinucleated daughter cells through a pseudo-multidivision (Fig. 4C). The appearance of more than four masses of mitotic chromosomes gave rise to multinucleation (Fig. 4D). These results suggest that Chk overexpression may affect mitotic chromosome dynamics.

Next, we double-stained cells with anti-α-tubulin for microtubules and propidium iodide for DNA. Confocal images were analyzed and detection sensitivity was carefully kept constant. As shown in Fig. 5, Chk-overexpressing cells showed sparse mitotic spindles in the metaphase (Fig. 5E,G,H) and anaphase (Fig. 5F,I), whereas control cells properly formed mitotic spindles and aligned chromosomes in the metaphase (Fig. 5B) and anaphase (Fig. 5C). Control (Fig. 5A) and Chk-overexpressing cells (Fig. 5D) displayed similar staining levels of microtubule fibers in the interphase. These results suggest that Chk overexpression may disturb the formation of mitotic spindles.

Furthermore, we double-stained cells with anti-Chk and propidium iodide for DNA. Each confocal image of individual samples was analyzed in a single plane of focus. Chk was found to localize to interphase nuclei and mitotic chromosomes in Chk-overexpressing (Fig. 6C-E) and control cells (Fig. 6A,B), although considerable amounts of Chk were distributed in the cytoplasm of the cells. These results suggest that a fraction of Chk is associated with chromosomes throughout the cell cycle.

Since Chk is known to be a negative regulator of Src family kinases, we determined the repertoire of Src family kinases present in KMT-2 cells. Immunoprecipitation and in vitro kinase assays revealed that Lyn and c-Yes were expressed in KMT-2 cells (Fig. 7A). To determine whether Chk overexpression suppressed these kinase activities, Lyn and c-Yes were immunoprecipitated from Triton X-100 cell lysates and subjected to in vitro kinase assays with acid-denatured enolase as substrate. Fig. 7B shows that Lyn but not c-Yes immunoprecipitated from Chk-overexpressing cells exhibited reductions in autophosphorylation and tyrosine phosphorylation of enolase, compared to Lyn immunoprecipitated from control cells. Upon Chk overexpression, the apparent net dephosphorylation of Lyn detected with the anti-phosphotyrosine antibody 4G10 resulted in reduction of Lyn kinase activities. Although phosphorylation of Lyn at the C-terminal tyrosine residue is offset by reduced autophosphorylation due to catalytic inactivation, it is possible that the antibody is preferentially able to recognize the phosphorylated tyrosine residue in the catalytic domain of Lyn rather than that present in the C-terminal regulatory domain. These results suggest that Chk negatively regulates Lyn but not c-Yes in KMT-2 cells. This is consistent with previous observations that Chk selectively suppresses Lyn activity in hematopoietic cells in vivo (Hirao et al., 1997; Hirao et al., 1998).

To examine subcellular localization of Lyn, c-Yes, Csk and Chk, we prepared cytosol and particulate (membrane) fractions by subcellular fractionations. Fig. 8A shows that significant proportions of endogenous Chk and Chk-FLAG were present in membranes whereas Csk was abundant in the cytosol (Fig. 8A, lanes 1 and 2), consistent with previous observations in platelets and megakaryocytic cells (Hirao et al., 1997; Hirao et al., 1998). Lyn and c-Yes were localized to membranes but not cytosol due to lipid modification.

When we isolated interphase nuclei and mitotic chromosomes, the tyrosine kinases Lyn, c-Yes, Csk and Chk were all present in nuclei and chromosomes (Fig. 8A, lanes 3 and 4). DNA topoisomerase II (Topo II), a major chromosome-associated protein, localized to nuclei and chromosomes where neither the cytosolic phospholipase A2 (cPLA2) nor the cell-surface transmembrane protein CD34 was present, indicating that these fractions are free of contamination by cytosolic and membrane fractions. After histone H1 was extracted into the nonscaffold fraction by treatment with DNase I and high salt, the residual chromosome scaffolds retained Lyn, c-Yes, Csk and Chk (Fig. 8A, lanes 5,6). These results suggest that each enzyme may firmly interact with nuclear matrices and chromosome scaffolds throughout the cell cycle.

To visualize chromosomal localization of Chk, we performed immunofluorescence studies. Fig. 8B shows that Chk could be visualized in the entire region of the chromosomes, supporting the biochemical results shown in Fig. 8A. However, with our present antibody reagents, we have been unable to visualize Lyn, c-Yes and Csk.

To examine the effect of Chk on the metaphase, we compared tyrosine phosphorylation of proteins between the interphase and metaphase. Fig. 9 shows that unsynchronized (Fig. 9, lanes 1-3) and nocodazole-arrested metaphase cells (Fig. 9, lanes 4-6) exhibited similar patterns of tyrosine phosphorylation of Triton X-100-solubilized proteins. Upon Chk overexpression no discernible metaphase-specific tyrosine-phosphorylated proteins appeared to be present. Note that Chk overexpression decreased autophosphorylation of Lyn but not c-Yes in the metaphase (Fig. 9, lanes 4-6), as shown in unsynchronized cells (Fig. 9, lanes 1-3).

Since Triton X-100 is incapable of solubilizing nuclei and chromosomes, we purified chromosomes from nocodazole-arrested cells (Fig. 9, lanes 7-9). The patterns of tyrosine phosphorylation of proteins in chromosomes were different from those observed in Triton X-100-solubilized cell lysates (compare Fig. 9 lanes 7-9 with 1-6). Several tyrosine-phosphorylated proteins, including SHP-1 protein-tyrosine phosphatase, pp64 and pp80, were detected in chromosomes but not in Triton X-100 cell lysates. Upon Chk overexpression, tyrosine phosphorylation of chromosome-associated Lyn and pp80 decreased and that of pp64 increased (Fig. 9, lanes 7-9).

To test the hypothesis that interaction of Chk with mitotic spindles somehow affects tyrosine phosphorylation in mitotic spindles, we attempted to purify mitotic spindles. Nocodazole-arrested cells were washed free of drug and allowed to form mitotic spindles. Taxol-stabilized mitotic spindles were isolated and adjusted to approximately equal amounts of tubulins per lane for western blotting, because the yields of the spindles from Chk-overexpressing cells were roughly less than half of those from control cells (see also Fig. 5). Fig. 9 shows that mitotic spindles, like chromosomes, were found to contain Chk, Csk, Lyn, c-Yes, SHP-1 and pp64 (Fig. 9, lanes 10-12). The patterns of tyrosine phosphorylation in mitotic spindles were similar to those in chromosomes, although pp80 was not detected. These results suggest that Chk overexpression perturbs the levels of tyrosine phosphorylation of proteins that associate with chromosomes and mitotic spindles.

In this study, ectopic and high expression of Chk in COS-1 cells displays nuclear localization and leads to suppression of cell growth. Overexpression of Chk in myeloid KMT-2 cells results in growth retardation and multinucleation. Several tyrosine kinases, including Chk and Lyn, associate with mitotic chromosome scaffolds and mitotic spindles in myeloid KMT-2 cells, leading to a hypothesis that Chk- and Lyn-mediated signal transduction in mitotic chromosomes and spindles may participate in metaphase chromosome dynamics.

Although Src family kinases are believed to localize to the plasma membranes due to lipid modification, c-Src is reported to localize to spindle poles in mitotic 3T3 fibroblasts (David-Pfeuty and Nouvian-Dooghe, 1990; Kaplan et al., 1992) and Fyn localizes to mitotic spindles and mitotic poles in T lymphocytes (Ley et al., 1994). However, protein-tyrosine kinases thus far have not been reported to associate with mitotic chromosomes. Our results show that Lyn and c-Yes and their regulators are all localized to interphase nuclei, metaphase chromosomes and mitotic spindles. The localization of these enzymes to nuclei in KMT-2 cells supports previous observations that Lyn associates with nuclear matrices in interphase HeLa cells and Rat2 fibroblasts (Radha et al., 1996). Additionally, we observed that a significant amount of Lyn localizes to nuclei in COS-1 cells (data not shown). Taken together, these results suggest that Lyn continues to associate with nuclear matrices and chromosome scaffolds throughout the cell cycle.

The mitotic chromosome scaffold, which is thought to play important roles in chromatid shape, chromosome condensation and alignment, and sister chromatid separation, contains abundant components such as Topo II, the SMC family, centromere-associated proteins, cytoskeletal proteins, and a number of unidentified proteins expressed at lower levels (Lewis and Laemmli, 1982; Gasser and Laemmli, 1987; Yen et al., 1991; Saitoh et al., 1994; Strunnikov et al., 1995; Wood et al., 1997). Components of the mitotic chromosome scaffolds are still resistant to extraction by Triton X-100 even after DNase I and high salt treatment. Empigen BB partly solubilized Chk, Csk and tyrosine-phosphorylated proteins including Lyn, c-Yes, SHP-1 and pp64, whereas these proteins were not solubilized by 2 M NaCl, 1% CHAPS or RIPA buffer containing a mixture of 1% deoxycholate, 0.1% SDS and 1% Triton X-100. Empigen BB can also solubilize nuclear matrix proteins (Staufenbiel and Deppert, 1984). These results suggest a novel phospholipid-based association of proteins with chromosome scaffolds and nuclear matrices, because the structure of Empigen BB is similar to that of phosphatidylcholine (Allen and Humphries, 1975).

Overexpression of a truncated protein-tyrosine phosphatase induces multinucleation and asynchronous nuclear division in hamster BHK cells (Cool et al., 1992). Tyrosine kinase inhibitors, genistein and 6-dimethylaminopurine, cause unusual chromosome movement in grasshopper spermatocytes and rat kangaroo PtK1 cells (Nicklas et al., 1993). Tyrosine kinase and phosphatase activities associate with chromosomes in murine lymphoid P388D1 cells and PtK1 cells, and levels of tyrosine phosphorylation fluctuate during mitosis (Taagepera et al., 1995). These data imply that disturbance of the equilibrium of tyrosine phosphorylation deregulates various cellular processes during mitosis.

We found that Chk overexpression changed the tyrosine phosphorylation states of proteins associated with mitotic chromosomes and spindles, i.e. a decrease in tyrosine phosphorylation of Lyn and an increase in that of pp64. Chk appears to suppress Lyn activity in mitotic spindles and chromosome scaffolds, as observed in Triton X-100 lysates of KMT-2 cells (Fig. 7B) and other hematopoietic cells (Hirao et al., 1997; Hirao et al., 1998). However, because of difficulties in obtaining sufficient amounts of mitotic chromosomes and mitotic spindles, we have been unable to immunoprecipitate these proteins from appropriate fractions. These data suggest that negative regulation of Lyn, possibly by Chk, might indirectly induce an increase in tyrosine phosphorylation of pp64. Another possibility is that Chk might directly phosphorylate pp64 besides Src family kinases.

SHP-1 activates Src family kinases in certain hematopoietic cells by dephosphorylating their C-terminal tyrosine residues, although SHP-1 is involved in inhibitory regulation of various signal transduction pathways (Falet et al., 1996; Somani et al., 1997). We can consider the possibility that Src family kinases associated with mitotic spindles and chromosomes may be regulated positively by SHP-1 and negatively by Chk and Csk to maintain their activities properly during the metaphase and anaphase. Additionally, pp80, which was mostly detected in chromosome scaffolds, might play a specific role in chromosomes.

Chromosome movement appears to be affected by insufficient formation of mitotic spindles, whereas microtubule formation in the interphase was not inhibited in Chk-overexpressing cells (Figs 3, 4). Although the mechanism by which Chk overexpression interferes with chromosome dynamics is not known, we hypothesize that Chk overexpression could induce perturbation of chromosome movement through impairment of mitotic spindle formation, perhaps giving rise to multinucleation. The presence of Lyn, c-Yes, Chk, Csk, SHP-1 and pp64 in both chromosomes and mitotic spindles (Fig. 9) also supports the hypothesis that chromosomes play an important role in mitotic spindle assembly (Vernos and Karsenti, 1995). However, our studies do not preclude the possibility that Chk overexpression in KMT-2 cells might affect cytokinesis as a whole. In addition, we observed that HeLa cells underwent multinucleation upon expression of the truncated form of Lyn lacking the tyrosine kinase domain (data not shown), suggesting that blockade of Lyn-mediated signal transduction may perturb normal processes of cell division in a variety of cell types that express Lyn.

Our observations of cells having multiple masses of the aligned chromosomes show that multinucleation of cells occurs through a complicated pseudo-multidivision (Fig. 4). Our results might provide a clue as to how cells, such as bone marrow megakaryocytes, undergo multinucleation. In addition, our data are partly consistent with the model that polyploidizing megakaryocytes have a unique mechanism in the anaphase of the cell cycle (Nagata et al., 1997). It would be of interest to examine whether Chk overexpression might affect multiplication of centrosomes.

Finally, inhibition of Src family kinases by microinjection of antibodies suppresses passage through the G₂ phase of the cell cycle in NIH3T3 fibroblasts (Roche et al., 1995). We therefore speculate that, as well as at steps downstream of cell-surface receptors in the G₀/G₁ and G₂ phases, the signal transduction mediated by Src family kinases may take place during spindle assembly and chromosome movement in the metaphase and anaphase. Further studies are required to clarify precise roles of Lyn, Chk, SHP-1 and the tyrosine-phosphorylated proteins in the metaphase and to compare metaphase-specific tyrosine phosphorylation in chromosomes and spindles among hematopoietic cells and other types of cells.

We are grateful to Drs Kazuo Maruyama (Tokyo Medical and Dental University, Tokyo, Japan) and Tadashi Yamamoto (The Institute of Medical Science, The University of Tokyo, Tokyo, Japan) for providing the pMKITneo vector, to Nippon Hoechst Co., Ltd. (Tokyo, Japan) for providing granulocyte-macrophage colony-stimulating factor, and to the technical support staff of Fuji Photo Film Co., Ltd (Tokyo, Japan) and Dr Takashi Saito (Chiba University Graduate School of Medicine, Chiba, Japan) for helping with preparation of the color prints. We thank Ms Yuko Takahashi (School of Pharmaceutical Sciences, University of Shizuoka, Shizuoka, Japan) for technical assistance and Dr Robert L. Margolis (Institut de Structurale, Grenoble, France) for critical reading of the early version of the manuscript. This work was supported in part by Grants-in-aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology, Japan.