MCB
Home Help [Feedback] [For Subscribers] [Archive] [Search] [Contents]
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Oikawa, T.
Right arrow Articles by Fukasawa, K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Oikawa, T.
Right arrow Articles by Fukasawa, K.

 Previous Article  |  Next Article 

Molecular and Cellular Biology, May 2005, p. 4046-4061, Vol. 25, No. 10
0270-7306/05/$08.00+0     doi:10.1128/MCB.25.10.4046-4061.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.

Transcriptional Control of BubR1 by p53 and Suppression of Centrosome Amplification by BubR1

Tatsuo Oikawa,1,2 Masaru Okuda,1* Zhiyong Ma,2 Rakesh Goorha,3 Hajime Tsujimoto,4 Hisashi Inokuma,1 and Kenji Fukasawa2

Laboratory of Veterinary Internal Medicine, Faculty of Agriculture, Yamaguchi University, Yamaguchi 753-8515,1 Department of Veterinary Internal Medicine, Graduate Schools of Agricultural and Life Sciences, University of Tokyo, Tokyo 113-8657, Japan,4 Department of Cell Biology, Neurobiology and Anatomy, University of Cincinnati College of Medicine, Cincinnati, Ohio 45267-0521,2 Department of Virology and Molecular Biology, St. Jude Children's Research Hospital, Memphis, Tennessee 381053

Received 18 January 2005/ Returned for modification 27 January 2005/ Accepted 31 January 2005


   ABSTRACT
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Elimination of the regulatory mechanism underlying numeral homeostasis of centrosomes, as seen in cells lacking p53, results in abnormal amplification of centrosomes, which increases the frequency of chromosome segregation errors, and thus contributes to the chromosome instability frequently observed in cancer cells. We have previously reported that p53–/– mouse cells in prolonged culture undergo genomic convergence similar to that observed during tumor progression; early-passage p53–/– cells are karyotypically heterogeneous due to extensive chromosome instability associated with centrosome amplification, while late-passage p53–/– cells are aneuploid yet karyotypically homogeneous and chromosomally stable. Moreover, they contain numerically normal centrosomes. Through the microarray analysis of early- and late-passage p53–/– cells, we identified the BubR1 spindle checkpoint protein, which plays a critical role in suppression of centrosome amplification and stabilization of chromosomes in late-passage p53–/– cells. Up-regulation of BubR1 augments the checkpoint function, which effectively senses the spindle/chromosome aberrations associated with centrosome amplification. We further found that BubR1 transcription is largely controlled by p53. In early-passage p53–/– cells, BubR1 expression is low and the checkpoint function in response to microtubule toxin is considerably compromised. In late-passage cells, however, regaining of BubR1 expression restores the checkpoint function to mitotic aberrations caused by microtubule toxin. Our studies demonstrate the molecular aspect of genomic convergence in cultured cells, providing critical information for understanding the stepwise progression of tumors.


   INTRODUCTION
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
The centrosome, a major microtubule-organizing center of an animal cell, consists of a pair of centrioles and a number of surrounding proteins (pericentriolar material). During mitosis, centrosomes direct the formation of bipolar spindles, which is an essential event for accurate chromosome segregation (14, 17). Since each daughter cell receives only one centrosome upon cytokinesis, centrosome must duplicate once during each cell cycle. Elimination of the regulatory mechanisms underlying centrosome duplication and numeral homeostasis of centrosomes result in the presence of more than two centrosomes (centrosome amplification), leading to aberrant mitoses and chromosome transmission errors (2, 14, 17).

Loss or mutational inactivation of p53 is known to induce centrosome amplification (16). It has been proposed that loss of p53 allows cells proceeding through the cell cycle upon cytokinesis failure, resulting in doubling of both genome and centrosome numbers (32). However, examination of p53–/– murine embryonic fibroblasts (MEFs) through passaging in culture has revealed the complexity of the causal mechanism of centrosome amplification by loss of p53. In early-passage (passage 1 [p1] to p5) p53–/– MEFs, genome doubling (multinucleation/multiploidization) was infrequent, yet many diploid or near-diploid cells were found to possess amplified centrosomes, indicating that these amplified centrosomes were generated not from multiploidization but from deregulated centrosome duplication (39). When p53–/– MEFs were passaged further (p10 to p15), an increased frequency of multinucleation was detected. However, multinucleation becomes much less frequent when cells are further passaged (~p25) in spite of the persistent numeral abnormality of centrosomes. Thus, in the absence of p53, centrosome amplification can result from both deregulated centrosome duplication and multiploidization, and their contributions differ depending on the passages in culture. When p53–/– MEFs are further cultured (~p50), centrosome amplification often disappears, accompanied by chromosome stabilization (K. Fukasawa, unpublished observation). A similar observation has been made previously in mouse primary epithelial cells lacking p53 (p53–/– mouse embryonic epithelial cells [MEEs]) (7). Centrosome amplification and chromosome instability (CIN), observed at high frequencies in early to mid passage (p1 to p30), were suppressed in late passages (>p50), which may be equated to the genomic convergence observed during tumor progression (19). In early passages in culture, p53–/– cells are karyotypically heterogeneous due to CIN associated with centrosome amplification. However, at a certain time point, one or a few cells acquire a karyotype that renders optimal growth properties and gradually dominate the culture. For this cell population, maintenance of this particular karyotype becomes a priority, forcing the selection of its progenies with a mutation(s) that suppresses the cause of CIN (i.e., centrosome amplification). Thus, it is reasonable to predict that late-passage p53–/– MEEs with normal centrosome profiles have acquired a mutation(s) that either directly or indirectly suppresses centrosome amplification. We performed a microarray analysis of early- and late-passage p53–/– MEEs. The gene for BubR1 (Bub1b) kinase was one of the genes with increased mRNA levels in late-passage cells. BubR1 was initially identified as being encoded by a gene mutated in colon carcinomas that exhibited a CIN phenotype (3), as a protein homologous to yeast protein Bub1 and Mad3 (3, 33, 40), and as a protein that interacts with CENP-E kinetochore motor protein (4). Subsequent studies have shown that BubR1 is a key component of the mitotic spindle checkpoint machinery (18, 20), which restrains cells from entering anaphase until all chromosomes are properly attached to bipolar spindles. This checkpoint function is activated by kinetochores unattached or inadequately attached to the spindles, as well as by improper spindle tension on kinetochores (8, 26, 30, 34). The spindle checkpoint inhibits the ability of Cdc20 to activate anaphase-promoting complex (APC)/cyclosome, a multisubunit E3 ubiquitin ligase, which promotes proteolysis of several key proteins responsible for progression to anaphase (43). Here we show that introduction of BubR1 into early-passage p53–/– MEEs efficiently eliminates cells with amplified centrosomes by inducing mitotic arrest/cell death via execution of the spindle checkpoint function in response to mitotic aberrations associated with amplified centrosomes. Moreover, silencing BubR1 in late-passage p53–/– MEEs induced reappearance of cells with amplified centrosomes. Thus, up-regulation of BubR1 is responsible for suppression of centrosome amplification in late-passage p53–/– cells, which is achieved via specifically targeting and eliminating cells with amplified centrosomes. The up-regulation of BubR1 in late-passage p53–/– cells was found to be more general, since BubR1 expression was also up-regulated in p53–/– adult mouse skin fibroblasts (MSFs) that had undergone genomic convergence (chromosome stabilization and suppression of centrosome amplification) after prolonged passages. We further found that BubR1 transcription and expression are largely controlled by p53 and thus significantly down-regulated in primary (early-passage) p53–/– cells. Moreover, introduction of BubR1 into early-passage p53–/– cells restored the checkpoint function in response to nocodazole (microtubule toxin) exposure initially described by Cross et al. (9), strongly suggesting that BubR1 is a downstream target of p53 in its nocodazole-responsive checkpoint activity. To our knowledge, this is the first demonstration of a molecular mechanism of genomic convergence in cultured cells, which can be equated with genomic convergence during tumor progression.


   MATERIALS AND METHODS
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Cells, plasmids, and transfection. Isolation and characterization of p53–/– MEEs were described previously (7). p53+/+ and p53–/– MSFs were generated from abdominal skins of 8-week-old male p53+/+ and p53–/– mice, which were littermates of p53+/– mouse crossing. Both MEEs and MSFs were maintained in complete medium (Dulbecco modified Eagle medium [DMEM] supplemented with 10% fetal bovine serum [FBS] with penicillin [100 U/ml] and streptomycin [100 µg/ml]) and grown in an atmosphere containing 10% CO2.

Transient transfection of wild-type (wt) and mutant BubR1 mutants was performed using a FuGENE 6 transfection system (Roche). Cells were transfected with a mammalian expression plasmid [pcDNA3.1(+); Invitrogen] encoding FLAG epitope-tagged mBubR1 (or mBubR1 mutant) together with a plasmid containing a puromycin resistance gene (pBabePuro) at a 20:1 molar ratio. After incubation for 12 h, cells were washed twice with medium and fed with fresh complete medium. Puromycin (2.5 µg/ml) was added to the medium 16 h later. Successfully transfected cells enriched by 48 h of puromycin treatment were seeded on coverslips and cultured for 24 h in complete medium. In another experiment, cells on coverslips were transiently transfected with a pcDNA3.1 plasmid containing mBubR1 together with an expression plasmid containing H2B-GFP at a 15:1 molar ratio. After incubation for 12 h, cells were washed and fed with fresh complete medium.

The short interfering RNA (siRNA) sequence used for silencing of mBubR1 corresponds to the coding sequence 211 to 230 (relative to the start codon). Annealing and cloning were performed using the pSUPER RNAi system (OliogoEngine) to generate an siRNA plasmid for BubR1 (pSUPER [BubR1/RNAi]). Cells were plated subconfluently 24 h before transfection and were fed with fresh medium 4 h before transfection. Cells were then transfected with either a pSUPER BubR1/RNAi plasmid or a pSUPER vector together with a plasmid containing a hygromycin resistance gene as a selection marker at a 20:1 molar ratio. After incubation for 12 h, cells were washed and fed with fresh complete medium. Twenty-four hours later, hygromycin (100 µg/ml) was added to the medium and the hygromycin-resistant colonies formed during 2 weeks of hygromycin treatment were subcloned.

Microarray analysis. Microarray analysis of total RNA prepared from PE and PL MEEs was performed using BD Atlas Glass Mouse 1.0 Microarrays according to the manufacturer's instructions (BD Biosciences).

Immunostaining and fluorescence microscopy. For all the immunostaining experiments except for coimmunostaining of {gamma}-tubulin and centrin, cells grown on coverslips were washed twice with phosphate-buffered saline (PBS) and fixed with 10% formalin for 20 min at room temperature. Cells were then washed with PBS and permeabilized with 1% NP-40 in PBS for 5 min. Cells were first incubated with blocking solution (10% normal goat serum in PBS) for 1 h and then probed with either rabbit anti-{gamma}-tubulin polyclonal antibody or a mixture of mouse anti-{alpha}-tubulin (DM1A; Sigma) and anti-ß-tubulin (Tub 2.1; Sigma) monoclonal antibodies for 1 h at room temperature. The antibody-antigen complexes were detected with either Alexa Fluor 546-conjugated goat anti-mouse immunoglobulin G (IgG) antibody for {alpha}- and ß-tubulin and Alexa Fluor 546-conjugated goat anti-rabbit IgG antibody for {gamma}-tubulin (Molecular Probes) by incubation for 1 h at room temperature. The samples were washed with Tris-buffered saline (TBS) for 30 min after each incubation and then counterstained with 4',6'-diamidino-2-phenylindole (DAPI) DNA dye.

For coimmunostaining of {gamma}-tubulin and centrin, cells were first placed on ice for 20 min to destabilize microtubules nucleated at the centrosomes. The cold-treated cells were subjected to brief extraction (~10 s) with cold extraction buffer [0.75% Triton X-100, 5 mM piperazine-N,N'-bis(2-ethanesulfonic acid) (PIPES), 2 mM EGTA (pH 6.7)], washed in cold PBS, and fixed with 10% formalin-10% methanol. Cells were then permeabilized and incubated in blocking solution as described above, followed by incubation with rabbit anti-{gamma}-tubulin polyclonal and anticentrin monoclonal antibodies. The antibody-antigen complexes were detected with Alexa Fluor 488-conjugated goat anti-rabbit IgG antibody for {gamma}-tubulin and Alexa Fluor 546-conjugated goat anti-mouse IgG antibody for centrin by incubation for 1 h at room temperature. The samples were washed with TBS for 30 min after each incubation and then counterstained with DAPI. Immunostained cells were examined under a fluorescence microscope (Zeiss Axioplan 2 Imaging, 40x and 60x objective lenses), and images were captured by a Hamamatsu ORCA-ER digital camera.

Immunoprecipitation and immunoblot analysis. For detection of endogenous BubR1 levels using commercially available anti-BubR1 antibody (C-20; Santa Cruz Biotech) by immunoblot analysis, mitotic cells enriched by mitotic shake-off of exponentially growing cells were lysed in NP-40 lysis buffer (1% NP-40, 50 mM Tris [pH 8.0], 150 mM NaCl, 4 mM Pefabloc SC [Boehringer Mannheim], 2 µg/ml leupeptin, 2 µg/ml aprotinin). The lysates were cleared by centrifugation at 20,000 x g for 10 min at 4°C. The lysates (200 µg of total proteins) that contained similar number of mitotic cells determined by immunoblot analysis with anti-phosphohistone H3 antibody (Cell Signaling Technology) were incubated with goat anti-BubR1 polyclonal antibody (C-20) for 4 h at 4°C. Immunocomplexes were collected by protein G-agarose beads. After several washes with ice-cold NP-40 lysis buffer, the samples were denatured in sodium dodecyl sulfate (SDS) sample buffer (2% SDS, 10% glycerol, 60 mM Tris [pH 6.8], 5% ß-mercaptoethanol, 0.01% bromophenol blue), fractionated by 8% SDS-polyacrylamide gel electrophoresis (PAGE), and blotted onto Immobilon (Millipore) sheets. The blots were incubated in blocking buffer (5% [wt/vol] nonfat dry milk in TBS plus Tween 20 [TBS-T]) for 1 h and then probed with goat anti-BubR1 antibody (C-20) overnight at 4°C. The blots were then rinsed in TBS-T and incubated with horseradish peroxidase-conjugated rabbit anti-goat IgG antibody for 1 h at room temperature. The blots were then rinsed in TBS-T, and the antibody-antigen complex was visualized by ECL chemiluminescence (Amersham). In other experiments, the lysates prepared from exponentially growing cells were immunoblotted with rabbit anti-BubR1 antibody (a gift from W. Dai), which showed a higher enough sensitivity to detect hyperphosphorylated BubR1.

For detection of ectopically expressed FLAG-tagged BubR1 or BubR1 mutants, it was not necessary to enrich mitotic cells or enrich proteins by immunoprecipitation for the commercial anti-BubR1 antibody (C-20). Total lysates were directly subjected to immunoblot analysis using rabbit anti-FLAG polyclonal antibody (Affinity Bioreagents) as described above. Antigen-antibody complexes were detected by alkaline phosphatase-conjugated goat anti-rabbit IgG antibody (Boehringer Manheim), visualized by use of alkaline phosphatase color developing reagents (Bio-Rad).

Northern blot analysis. Total RNAs were prepared from passage 3 p53+/+ and p53–/– MSFs using Trizol RNA isolation reagent (Invitrogen) and subjected to Northern blot analysis. The cDNA sequence (576 to 945) of mBubR1 was used as a probe for detection of BubR1 mRNA.

Flow cytometry. Cells were collected by trypsinization, washed once with PBS, and fixed in 70% ethanol for 30 min. Fixed cells were briefly pelleted, resuspended in PBS, and treated with RNase A (100 µg/ml) for 30 min at 37°C. Propidium iodide (10 µg/ml) was added to the cell suspension prior to the analysis.

In vitro histone H1 kinase assay. Cells were lysed in Triton X lysis buffer (1% Triton X-100, 50 mM Tris [pH 8.0], 4 mM Pefabloc SC, 2 µg/ml leupeptin, 2 µg/ml aprotinin), and cell extracts were cleared by centrifugation for 15 min at 20,000 x g and 4°C. The supernatants were preincubated with protein A-agarose. The precleared lysates (100 µg of total proteins) were immunoprecipitated with anti-CDK1 antibody, as well as anti-cyclin B antibody (Santa Cruz Biotech). The antibody-antigen complexes were collected with protein A-agarose and tested for histone H1 kinase activity as described previously (34). 32P incorporation of histone H1 was quantitated by scanning with Fuji Phosphoimager 1000.


   RESULTS
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Increased expression of BubR1 in late-passage p53–/– MEEs. We performed a comparative DNA microarray analysis of early-passage (passage 5) (PE cells) and late-passage (passage 60) p53–/– MEEs (PL cells) under optimal growth conditions. Approximately 35% of the PE cells and <2% of the PL cells used for the microarray analysis contained amplified centrosomes (n ≥ 3) (data not shown; see reference 7). We compared the expression patterns of 1,101 genes between PE and PL cells, which revealed 31 genes with altered fluorescence ratios of >2.0-fold. Of these, nine genes were induced in PL cells (Fig. 1A). Among them, because centrosome amplification results in improper spindle formation, as well as chromosome aberration, we explored the potential involvement of BubR1 spindle checkpoint kinase in suppression of centrosome amplification in late-passage p53–/– MEEs.



View larger version (31K):
[in this window]
[in a new window]
  		FIG. 1. Increased BubR1 expression in late-passage p53–/– cells at both transcriptional and protein levels. (A) Microarray analysis of RNA prepared from early- and late-passage p53–/– MEEs. The analysis was performed twice, and the graph shows the nine genes whose expression was increased more than twofold in late-passage (PL) cells compared with early-passage (PE) cells in both analyses. Abbreviations: GJB4, gap junction beta 4 protein; SIX4, sine oculis-related homeobox protein 4 homolog; MAGI, membrane-associated guanylate kinase inverted protein 1; SDF4, stromal cell-derived factor 4; DGCR6, DiGeorge syndrome chromosome region 6 protein. (B) The lysates prepared from PE and PL p53–/– MEEs enriched for mitotic cells by mechanical shake-off were first characterized by immunoblot analysis using anti-phosphohistone (P-histone) H3 antibody (a, upper part). The mitotic lysates with equivalent anti-phosphohistone H3 antibody reactivity were subjected to immunoprecipitation using goat anti-BubR1 antibody (C-20; Santa Cruz Biotech). The immunoprecipitates were resolved by 8% SDS-PAGE and subjected to immunoblot analysis using the same antibody (a, lower part). The BubR1 protein level was increased more than eightfold in PL cells compared with PE cells. The lysates prepared from exponentially growing PE and PL cells were run on 8% SDS-PAGE for a longer duration to maximize the separation of hyperphosphorylated and hypophosphorylated BubR1 and immunoblotted with rabbit anti-BubR1 antibody (a gift from W. Dai) (b, top part). The same lysates were also probed for Mad2 (C-19; Santa Cruz Biotech) (b, middle part) and ß-tubulin (b, bottom part). (C) p53–/– MSFs were cultured for 40 passages. The early (PE, passage 4) and late (PL, passage 40) cells were immunostained with anti-{gamma}-tubulin antibody, and the number of centrosomes per cell was determined by fluorescence microscopy (part a). For each experiment, >300 cells were examined. The results are shown as the average ± standard error from three experiments. The lysates prepared from exponentially growing PE and PL p53–/– MSFs were subjected to immunoblot analysis and probed for BubR1 (b, top part), Mad2 (b, second part), phosphohistone H3 (b, third part), and ß-tubulin (b, bottom part).

 
We first tested whether the different levels of BubR1 mRNA between PE and PL cells detected by the microarray analysis would be proportionally translated to the protein levels. The mitotic cells enriched by mechanical shake-off of PE and PL cells were lysed, and the lysates containing equivalent mitotic cells, determined by immunoblotting for anti-phosphohistone H3 antibody (Fig. 1B, part a, top), were subjected to immunoprecipitation followed by immunoblot analysis using anti-BubR1 antibody (C-20, Santa Cruz Biotech) (Fig. 1B, part a, bottom). The level of BubR1 in mitotic PE cells was barely detectable, while mitotic PL cells showed readily detectable levels of BubR1. It has been shown that BubR1 is hyperphosphorylated when the spindle checkpoint is activated (10, 27). To examine the extent of phosphorylation of BubR1 in PL cells, the lysates from PE and PL cells under an optimal growth condition were immunoblotted using the anti-BubR1 antibody that can efficiently recognize both hyperphosphorylated and nonphosphorylated BubR1 (Fig. 1B, part b, top). Approximately 10% of total BubR1 in PL cells were found to be hyperphosphorylated. We also immunoblotted the cell lysates with antibody against Mad2, another major spindle checkpoint protein (30) (Fig. 1B, part b, middle). Although Mad2 was slightly up-regulated in PL cells, the difference in the Mad2 levels between PE and PL cells is much less significant when compared with that of BubR1. Thus, BubR1 is rather specifically up-regulated in both mRNA and protein levels in late-passage cells, indicating that p53–/– MEEs had acquired a mutation(s)/genetic modification(s) during passaging which resulted in increased expression of BubR1.

We next tested whether the up-regulation of BubR1 in the genomically converged p53–/– cells after prolonged culture is a phenomenon specific to the particular cell type (MEEs) or to this particular culture (PL cells generated after prolonged passages). To this end, we prepared skin fibroblasts from abdominal skin samples from p53–/– adult mice (MSF) and subjected them to prolonged passaging. Similar to embryonic cells (MEFs and MEEs), a high frequency of centrosome amplification was observed in the early-passage p53–/– MSFs. However, at passage 40, centrosome amplification was significantly suppressed (Fig. 1C, part a) and showed genomic convergence (karyotypic homogeneity and chromosome stabilization) (data not shown). Immunoblot analysis revealed that BubR1 is up-regulated in late-passage p53–/– MSFs (Fig. 1C, part b), similar to p53–/– MEEs. Thus, up-regulation of BubR1 in p53–/– cells with numerically restored centrosome profiles after prolonged culture appears to be rather a general phenomenon.

Suppression of centrosome amplification in early-passage p53–/– MEEs by exogenously introduced BubR1. To address the potential role of BubR1 in suppression of centrosome amplification in late-passage p53–/– MEEs, we tested whether introduction of BubR1 in PE cells would restore a normal centrosome profile. A plasmid encoding wt murine BubR1 (mBubR1) (11) was transiently transfected into PE cells together with a plasmid containing a puromycin resistance gene as a rapid selection marker. The vector was transfected as a control. The successfully transfected cells were enriched by puromycin treatment for 48 h, and the surviving cells were further cultured in fresh medium for an additional 24 h. Cells were first examined for the expression of BubR1 by immunoblot analysis. The mBubR1-transfected cells showed readily detectable levels of BubR1 comparable to the levels up-regulated in PL cells, while endogenous BubR1 (vector control cells) was barely detectable (Fig. 2A). The cells were then examined for centrosomes by immunostaining for {gamma}-tubulin, a major component of centrosomes (Fig. 2B and C). Among the vector-transfected cells, ~35% of cells contained amplified (n ≥ 3) centrosomes (representative images are shown in Fig. 2B, parts a to c). In contrast, less than 3% of the mBubR1-transfected cells were found to contain amplified centrosomes (Fig. 2B, parts d to f). In one experiment, cells were examined for integrity of centrosomes by coimmunostaining of {gamma}-tubulin and centrin. Immunostaining of centrin can identify individual centrioles within centrosomes at a high magnification (35) (Fig. 2B, parts g to l). The majority of dot signals detected by anti-{gamma}-tubulin antibody overlapped with two anticentrin antibody-reactive dots that represent a pair of centrioles. Thus, ectopic expression of BubR1 in early-passage p53–/– MEEs efficiently and selectively eliminated cells with amplified centrosomes from culture.



View larger version (47K):
[in this window]
[in a new window]
  		FIG. 2. Ectopic expression of wt mBubR1 in PE cells suppresses centrosome amplification. wt mBubR1 was transfected into PE cells together with a plasmid containing a puromycin resistance gene (20:1 molar ratio). As a control, the vector was transfected. Puromycin was added to the medium 16 h after transfection. The puromycin-resistant cells at 48 h of puromycin treatment were replated in fresh medium and cultured for additional 24 h. (A) The total cell lysates prepared from the transfectants and PL cells were subjected to immunoblot analysis with anti-BubR1 antibody. (B) The mBubR1- and vector-transfected cells were subjected to cold treatment and brief extraction (see Materials and Methods) and coimmunostained with rabbit anti-{gamma}-tubulin antibody (centrosomes, green) and mouse anticentrin antibody (centrioles, red). Cells were also counterstained with DAPI (parts b and e). The {gamma}-tubulin immunostaining of the vector-transfected PE cells shows extensive centrosome amplification (part a, cells with amplified centrosomes are indicated by arrowheads), while amplified centrosomes were rarely detected in the mBubR1 transfectants (part d). Parts c and f show the overlay images. Parts g to i and j to l show the magnified images of the areas indicated by arrows; {gamma}-tubulin (parts g and j) and centrin immunostaining (parts h and k) is shown. Parts i and l show the overlay images of {gamma}-tubulin and centrin immunostaining. Scale bar, 20 µm. The vector- and mBubR1-transfected PE cells immunostained for {gamma}-tubulin (and centrin) were statistically analyzed for the number of centrosomes per cell, and the results are shown in panel C. For each experiment, >300 cells were examined. The results are shown as the average ± standard error from three experiments.

 
BubR1 expression in early-passage p53–/– MEEs selectively eliminates cells with amplified centrosomes. We next dissected the mechanism underlying the rapid disappearance of cells with amplified centrosomes in the BubR1-transfected PE cells. PE cells were transfected with a plasmid encoding mBubR1 along with a plasmid encoding a histone-green fluorescent protein (GFP) fusion protein (H2B-GFP) as a tracking marker of the transfected cells (22). At every 24 h after transfection (total, 96 h), cells were fixed and counterstained with DAPI, and the percentage of the transfected (GFP-positive) cells over the total cell number was determined (Fig. 3A; representative images are shown in Fig. 3B). In vector-transfected cells, 15 to 20% of the total cell population was GFP positive throughout the 96-h chase period, indicating that expression of H2B-GFP is not cytotoxic and the transfected cells undergo cell division at similar rates with untransfected cells. In contrast, in BubR1-transfected cells, there was a small reduction in the percentage of GFP-positive cells (~15%) at 24 h posttransfection. Moreover, the percentage of GFP-positive cells progressively declined during the chase period, and only ~5% of the cells were GFP positive at 72 h. However, the percentage of GFP-positive cells became stable after 72 h. In parallel, we also tested the effect of ectopically expressed BubR1 in PL cells. We observed no noticeable toxicity of BubR1 introduction in PL cells (data not shown). These results suggest that introduction of BubR1 confers cellular toxicity in PE cells, which appears to target a particular population of PE cells.



View larger version (41K):
[in this window]
[in a new window]
  		FIG. 3. Cellular toxicity imposed by introduction of mBubR1 in PE cells. PE cells seeded on coverslips were transfected with either a vector or a wt mBubR1 plasmid together with a plasmid encoding H2B-GFP (15:1 molar ratio). After transfection, cells were fed with fresh medium. At every 24 h after transfection for a total of 96 h, cells were examined for the percentage of GFP-positive cells over the total number of cells, and the results are shown in panel A as the average ± standard error from three experiments. Representative fluorescence microscopic images are shown in panel B.

 
The high frequency of centrosome amplification in PE cells raises the possibility that aberrant mitoses due to amplified centrosomes might be targeted by the spindle checkpoint function augmented by ectopically expressed BubR1, resulting in cell cycle arrest and/or cell death. To test this, we immunostained microtubules of the BubR1/H2B-GFP-transfected cells using antibodies against {alpha}- and ß-tubulin. In vector-transfected cells at 24 h posttransfection, the mitotic index of the GFP-positive cells is ~5%, and a similar mitotic index was observed at 48 and 72 h (Fig. 4A). In contrast, in BubR1-transfected cells, the mitotic index of the GFP-positive cells was >20% at 24 h, >50% at 48 h, and >40% at 72 h (Fig. 4A). Thus, either mitotic delay or arrest was induced by ectopic expression of BubR1. Moreover, >75% of those GFP-positive mitotic cells show aberrant metaphase spindles organized by multiple spindle poles (Fig. 4B, part I, arrow), many of which showed readily observable misaligned or broken chromosomes (part II). In BubR1-transfected cells, we also detected a significant increase in the frequency of GFP-positive cells that were dead or dying (~5% at 24 h, ~10% at 48 h, and ~15% at 72 h) (Fig. 4A), which show either nuclear fragmentation (Fig. 4C, parts a and a') or nuclear diffusion (parts b and b').



View larger version (47K):
[in this window]
[in a new window]
  		FIG. 4. Exogenously introduced mBubR1 efficiently and selectively eliminates cells with amplified centrosomes. Vector- and mBubR1-transfected cells as in Fig. 3 were subjected to immunostaining with anti-{alpha}- and ß-tubulin monoclonal antibodies and DAPI staining. A significantly higher number of GFP-positive mBubR1-transfected cells were found in mitotic phase (identified by nuclear membrane breakdown, spindle formation, and condensed chromosomes) than GFP-positive vector-transfected cells (A). Representative images of increased frequency of GFP-positive mitotic cells in mBubR1-transfected cells at 48 h posttransfection are shown in panel B. Parts I and II are the overlay magnified images of the {alpha}- and ß-tubulin staining and DAPI staining of the areas indicated in part a. Many GFP-positive mitotic mBubR1-transfected cells show abnormalities including multipolar spindles (part I, arrow) and the presence of misaligned, fragmented, or unattached chromosomes (part II). In the GFP-positive mBubR1-transfected cells, dying cells were frequently detected. The increased frequency of cell death among the GFP-positive mBubR1-transfected cells became more evident at later time points (A). The representative images of dying cells in the GFP-positive mBubR1-transfected cells at 72 h posttransfection are shown in panel C, in which parts a and b show GFP-H2B and parts a' and b' show DAPI staining. Parts a and a' show a cell with nuclear fragmentation, while parts b and b' show a cell with cytoplasmic diffusion of DNA. Scale bar, 20 µm. The percentages of interphase, mitotic, and dead/dying cells among GFP-positive vector- and mBubR1-transfected cells shown in panel A were determined from two independent transfections, which gave nearly identical results, and the results shown in panel A are from one of the experiments. For each experiment, >200 GFP-positive cells were examined. (D) Vector and mBubR1 transfectants at 24, 48, and 72 h were immunostained for {gamma}-tubulin and counterstained with DAPI, and the surviving GFP-positive interphase cells were examined for the number of centrosomes per cell. The results shown are from one of two independent experiments that gave nearly identical results. For each experiment, >200GFP-positive interphase cells were examined. Centrosome amplification is observed in the vector-transfected, GFP-positive interphase cells at 72 h posttransfection at a similar frequency with those at 24 h posttransfection. In contrast, most of the GFP-positive, mBubR1-transfected interphase cells at 72 h contain one or two centrosomes. The surviving GFP-positive cells in both vector and mBubR1 transfectants were also examined for BubR1 expression by immunostaining. All surviving GFP-positive cells in the mBubR1 transfectants show higher immunostaining signals than the vector transfectants, indicating that the surviving GFP-positive cells overexpress BubR1 (data not shown). (E) A plasmid encoding mMad2 was cotransfected with a plasmid encoding H2B-GFP as described for mBubR1 transfection. Vector and mMad2 transfectants at 24, 48, and 72 h were immunostained with anti-{gamma}-tubulin antibody and counterstained with DAPI, and the surviving GFP-positive interphase cells were examined for the number of centrosomes per cell. For each experiment, >200 GFP-positive cells were examined. (F) Vector- and mBubR1-transfected early-passage (PE) p53–/– MSFs at 24, 48, and 72 h were immunostained for {gamma}-tubulin and counterstained with DAPI, and the surviving GFP-positive interphase cells were examined for the number of centrosomes per cell. The results shown are from one of two independent experiments that gave nearly identical results. For each experiment, >200 GFP-positive interphase cells were examined.

 
We then examined the number of centrosomes per cell in the surviving GFP-positive cells (Fig. 4D). Among the vector-transfected GFP-positive surviving cells, ~65% of the cells contained either one or two centrosomes and ~35% contained amplified centrosomes, and there was no significant change in the ratio of one/two centrosomes versus amplified centrosomes throughout the chase period. Thus, in the vector-transfected cells, there was no selection of cells with one/two centrosomes or amplified centrosomes. In contrast, in the BubR1-transfected GFP-positive surviving cells, a significant reduction in the percentage of cells with amplified centrosomes (~20%) was detected as early as 24 h posttransfection. The percentage of GFP-positive surviving cells with amplified centrosomes continued to decline, and they were nearly absent at 72 h. Thus, there was a clear selection of cells with one/two centrosomes in the BubR1-transfected cells.

To corroborate the finding that the augmented spindle checkpoint function may be responsible for selective elimination of cells with amplified centrosomes, we transiently cotransfected a plasmid encoding murine Mad2 (mMad2), another key spindle checkpoint protein (30), together with an H2B-GFP plasmid. Cells transfected with mMad2 (GFP-positive cells) showed progressive elimination of cells with amplified centrosomes similar to mBubR1-transfection (Fig. 4E). However, the effectiveness of eliminating cells with amplified centrosomes was less pronounced in the mMad2 transfectants than the mBubR1 transfectants. This perhaps reflects the fact that Mad2 and BubR1 functionally interact to elicit the overall spindle checkpoint (30), and PE cells express low levels of BubR1 (described later in detail), which limits the effect of Mad2 overexpression on the spindle checkpoint. Nevertheless, this result further supports the notion that spindle checkpoint activity plays a critical role in rapid and selective elimination of cells with amplified centrosomes.

We also examined whether elimination of cells with amplified centrosomes by ectopic expression of BubR1 could occur in early-passage p53–/– MSFs. The early-passage p53–/– MSFs were transfected with a mBubR1 (or a vector as a control) plasmid along with a H2B-GFP plasmid as described above (Fig. 3), and the surviving GFP-positive cells were immunostained for {gamma}-tubulin (Fig. 4F). Similar to p53–/– MEEs (PE), there was a significant reduction in the number of cells with amplified centrosomes in the mBubR1 transfectants. Thus, selective elimination of cells with amplified centrosomes by introduction of BubR1 appears to be not cell type specific. Taken together, these results demonstrate that exogenously introduced BubR1 specifically targets PE cells with amplified centrosomes via executing the checkpoint function against the mitotic defects associated with centrosome amplification, leading to mitotic arrest and cell death.

Restoration of the spindle checkpoint in late-passage p53–/– cells. To further test the restoration of spindle checkpoint and normal centrosome profiles in PE cells by ectopic expression of BubR1, we first examined mitotic early (PE)- and late (PL)-passage p53–/– MEEs and MSFs in more details (Fig. 5). Under normal growth conditions, there was no significant difference in mitotic indices determined by chromosome condensation and rounded-up morphology between PE and PL cells (Fig. 5A). However, we noticed the frequent presence of lagging chromosomes and chromosome bridges in the mitotic PE cells, but not in mitotic PL cells. The presence of lagging chromosomes during anaphase and formation of chromosome bridges during telophase/cytokinesis have been shown to reflect the impaired spindle checkpoint (13, 21, 23). We, thus, statistically determined the frequencies of lagging chromosomes and chromosome bridges in PE and PL cells. In PE cells (both p53–/– MEEs and MSFs), we detected significantly higher frequencies of lagging chromosomes and chromosome bridges compared with PL cells (Fig. 5A; representative images of lagging chromosomes and chromosome bridges in PE cells are shown in Fig. 5B), suggesting that the spindle checkpoint may be compromised in PE cells but restored in PL cells. However, alternatively this result may merely reflect the decrease in the frequency of centrosome amplification in PL cells. To further test whether the spindle checkpoint in PE p53–/– cells is compromised and is restored in PL p53–/– cells, we exposed PE and PL p53–/– MEEs to nocodazole (a potent microtubule-destabilizing agent) for 30 h. Since cyclin B (CDK1/cyclin B activity) is one of the primary targets of APC, and spindle checkpoint suppresses activation of APC (18, 30), we first prepared cell lysates from nocodazole-treated cells at every 6 h and subjected them to immunoprecipitation with anti-Cdk1, as well as anti-cyclin B, antibody. The immunoprecipitates were then tested in the in vitro kinase assay using histone H1 as a substrate (Fig. 5C). We detected significant CDK1/cyclin B activity in PL cells but almost no activity in PE cells at 30 h. Thus, the spindle checkpoint is activated more effectively in PL cells than in PE cells, resulting in delayed inactivation of CDK1/cyclin B in PL cells. In parallel, nocodazole-treated PE and PL cells were fixed and immunostained with antibody against MPM2 (a known M-phase-specific phosphoantigen) (12) and counterstained with DAPI DNA dye. Contrary to our prediction, we did not see a significant difference in mitotic index (red bar) between PE and PL cells throughout the nocodazole treatment period, even at 30 h (Fig. 5D). However, we noticed significantly more dead/dying cells in PL cells compared with PE cells (yellow bar). Another phenomenon to be noted, which occurred more frequently in late hours of nocodazole treatment (i.e., 30 h), is the formation of multiple "mininuclei" (green bar) (Fig. 5F shows morphology). This results from nuclear membrane formation around the dispersed chromosomes that have failed to congress at the metaphase plate. There were more cells with mininuclei in PE cells than in PL cells at 30 h. However, the more striking difference in the cells with mininuclei between PE and PL cells is the reactivity to anti-MPM2 antibody: the majority (60 to 80%) of those in PL cells were detected by anti-MPM2 antibody as efficiently as mitotic cells, while most (>90%) of those in PE cells were reactive to this antibody at a much lesser degree, similar to interphase cells (Fig. 5E, representative images are shown in 5F). Since it has been shown that CDK1/cyclin B is primarily responsible for generation of MPM2 epitopes (24), this may explain the conflicting data on CDK1/cyclin B activities and mitotic index in the nocodazole-treated PE and PL cells shown above. CDK1/cylin B remained active in PL cells that had phenotypically progressed (or regressed) to interphase (i.e., nuclear membrane formation and flattened morphology), while the activity of CDK1/cyclin B was suppressed in those of PE cells. Thus, although the numbers of mitotic cells (in terms of morphology) were similar between nocodazole-treated PE and PL cells, especially at 30 h, we could detect extended CDK1/cyclin B activity in PL cells. That is, the checkpoint against mitotic defects caused by nocodazole is more effectively activated in PL cells than in PE cells, resulting in suppression of APC and delayed inactivation of CDK1/cyclin B. Together with the finding that PE cells show significantly higher incidence of lagging chromosomes and formation of chromosome bridges than PL cells under normal growth conditions (Fig. 5 A and B), it is likely that the spindle checkpoint function is compromised in PE cells and is restored in PL cells. However, the nocodazole-treated PL cells did not show considerable mitotic delay/arrest compared with PE cells, but morphologically transformed to interphase cells that retained mitotic activities, suggesting that the up-regulation of BubR1 in PL cells may elicit unconventional checkpoint activity against the defects associated with centrosome amplification, as well as microtubule toxins (see Discussion).



View larger version (38K):
[in this window]
[in a new window]
  		FIG. 5. Compromised spindle checkpoint in early-passage PE cells and restoration of spindle checkpoint in PL cells. (A) PE and PL p53–/– MEEs and MSFs under normal growth conditions were stained with DAPI, and the mitotic index and the frequencies of lagging chromosomes and chromosome bridge formation were determined. Representative images of lagging chromosomes (part a) and chromosome bridge formation (part b) in PE cells are shown in panel B, which includes magnified images of the indicated areas. (C) PE and PL p53–/– MEEs were exposed to nocodazole for 30 h. At 6, 12, 24, and 30 h, lysates were prepared and subjected to immunoprecipitation using anti-CDK1 antibody. The immunoprecipitates were assayed for in vitro kinase activities using histone H1 as a substrate. The kinase activities were quantitated by a PhosphorImager, and the values (arbitrary units) are presented in the graph. We also performed the in vitro kinase reaction of the anti-cyclin B antibody immunoprecipitates and obtained almost identical results (data not shown). (D) Nocodazole-treated PE and PL p53–/– MEEs at 12, 24, and 30 h were immunostained with anti-MPM2 antibody and counterstained with DAPI. The percentages of mitotic (determined by rounded-up morphology and condensed chromosomes), interphase, dead/dying (cf. Fig. 4C), and postmitotic (mininuclei) cells were determined by fluorescence microscopy. (E) The postmitotic (mininuclei) cells in the nocodazole-treated PE and PL p53–/– MEEs at 12, 24, and 30 h were statistically analyzed for anti-MPM2 antibody reactivity (positive, similar intensity to mitotic cells; negative, similar intensity to interphase cells). Most of the PE cells with mininuclei are negative, while the majority of PL cells with mininuclei (60 to 80%) are positive for MPM2 reactivity. Representative images of PE and PL cells with mininuclei are shown in panel F. MN, cells with mininuclei; M, mitotic cells; INT, interphase cells.

 
The kinase activity of BubR1 is not essential but required for maximal suppression of centrosome amplification. mBubR1 has a kinase domain at the C-terminal region (amino acids [aa] 740 to 1020 in mBubR1) (11). To determine whether the kinase activity of BubR1 is required for suppression of centrosome amplification, FLAG epitope-tagged wt mBubR1 and four deletion mutants (Fig. 6A) were transiently transfected into PE cells along with a plasmid containing a puromycin resistance gene as described for Fig. 2. As a control, the vector was transfected. The transfectants enriched by puromycin treatment for 48 h were replated and cultured for an additional 24 h in fresh medium. We first examined the expression of the transfected wt and deletion mutant mBubR1 proteins by immunoblot analysis using anti-FLAG antibody (Fig. 6B). All of the transfectants expressed comparable levels of transfected mBubR1 with expected molecular weights. The transfectants were also examined for centrosomes by immunostaining for {gamma}-tubulin (Fig. 6C). The mBubR1/KD mutant lacks 9 aa (residues 784 to 792) within the conserved catalytic core, and deletion of this sequence has been shown to inactivate its kinase activity (5). Similarly, the mBubR1(1-739) mutant lacks the entire kinase domain. Both kinase-dead mutants retain the Bub3-binding domain (aa 385 to 426), which is critical for kinetochore localization of BubR1 (40). Expression of these kinase-dead mutants resulted in significant suppression of centrosome amplification, although not as effective as wt mBubR1. The mBubR1(450-1052) mutant, which retains the kinase domain but lacks the N-terminal region (aa 1 to 449) that includes the Bub3-binding domain, failed to suppress centrosome amplification. Likewise, the mBubR1(450-739) mutant, which lacks both the kinase domain and the N-terminal domains, was unable to suppress centrosome amplification. Thus, the kinase activity of BubR1 is dispensable yet appeared to be required for maximal checkpoint activity. Moreover, our studies show that the N-terminal region (Bub3-binding domain) is an absolute requirement for BubR1 to exert its checkpoint function against the defects associated with centrosome amplification.



View larger version (32K):
[in this window]
[in a new window]
  		FIG. 6. The kinase activity of BubR1 is dispensable but required for maximal suppression of centrosome amplification. (A) Schematic representation of mBubR1 deletion mutants. The FLAG epitope-tagged wt BubR1 and four deletion mutants were generated by PCR-based mutagenesis. The KD mutant lacks a 9-aa conserved sequence (784 to 792) within the catalytic core of BubR1 (5). The BubR1(1-739) mutant retains the Bub3-binding domain but lacks the kinase activity. The BubR1(450-1052) mutant retains the kinase activity but lacks the Bub3-binding domain. The BubR1(450-739) mutant lacks both the kinase activity and the Bub3-binding domain. FLAG-tagged wt BubR1 and the mutants described in panel A were transiently transfected into PE cells together with a plasmid containing a puromycin resistance gene at a 20:1 molar ratio. As a control, the vector was transfected. Puromycin was added to the medium 16 h after transfection. Successfully transfected cells enriched by puromycin treatment for 48 h were replated in fresh medium and cultured for an additional 24 h. The cell lysates were prepared and probed with anti-FLAG polyclonal antibody (B). The transfectants were examined for suppression of centrosome amplification as described in the legend to Fig. 2B and C. Three independent experiments were performed. One experiment was performed with coimmunostaining of {gamma}-tubulin and centrin, and two experiments were performed with immunostaining of {gamma}-tubulin. These two methods gave similar results. The results presented in panel C were the percentage of cells with amplified (three or more) centrosomes as the average ± standard error from three experiments. For each experiment, >300 cells were examined.

 
Silencing BubR1 in late-passage p53–/– MEEs results in reemergence of amplified centrosomes. If the increased expression of BubR1 is responsible for suppression of centrosome amplification in late-passage p53–/– MEEs, silencing BubR1 expression should result in reemergence of PL cells with amplified centrosomes. To test this, we generated the PL cell lines whose BubR1 expression was silenced by siRNA. A pSuper siRNA vector containing a short sequence specific for mBubR1 (pSuper/BubR1), which provides intracellular expression of siRNA mBubR1 to achieve long-term silencing of endogenous mBubR1, was stably transfected into PL cells together with a plasmid containing a hygromycin resistance gene as a selection marker. As a control, the vector was transfected. Among hygromycin-resistant colonies formed in the pSuper/BubR1-transfected cells, two subcloned cell lines (BubR1/RNAi-1 and -2) showed significantly reduced levels of BubR1 compared with two randomly subcloned vector-transfected cell lines (Vector-1 and -2) (Fig. 7A). These cell lines, under optimal growth conditions, were examined for centrosomes. Virtually no centrosome amplification was detected in the vector-transfected cells (Fig. 7B and C, part a), similar to the untransfected PL cells. In contrast, 10 to 15% of BubR1/RNAi cells contained amplified centrosomes (Fig. 7B and C, part b). It should be noted that the increase in the number of cells with amplified centrosomes in PL cells by silencing of BubR1 was not due to cytokinesis failure, since the BubR1/RNAi-1 and -2 cells with bipolar spindles progress normally through mitosis to complete cytokinesis, and there was no significant increase in multiploidization compared with the vector-transfected cells as examined both microscopically and flow cytometrically (data not shown). These results demonstrate that the increased expression of BubR1 plays an important role in the suppression of centrosome amplification in PL cells. Although the increase in the frequency of centrosome amplification in PL cells by silencing BubR1 is significant, it is clearly less than that observed in PE cells, in which 30 to 35% of cells contain amplified centrosomes. This may reflect incomplete silencing of BubR1 by siRNA (Fig. 7A), and thus PL cells transfected with BubR1 siRNA may retain partial capability of executing spindle checkpoint function. Alternatively, suppression of centrosome amplification in PL cells may be achieved by multiple mechanisms, and increased expression of BubR1 may be one such mechanism.



View larger version (38K):
[in this window]
[in a new window]
  		FIG. 7. siRNA-mediated silencing of mBubR1 in PL cells induces reemergence of cells with amplified centrosomes. PL cells were transfected with either a pSUPER(BubR1/RNAi) plasmid, which stably expresses inhibitory RNA duplex molecules of BubR1 (211 to 230 of the coding sequence) or a pSUPER vector together with a plasmid containing a hygromycin resistance gene (20:1 molar ratio). The colonies formed during a 2-week hygromycin treatment were subcloned and examined for BubR1 expression by immunoprecipitation/immunoblot analysis of mitotic cells using anti-BubR1 antibody as described in Fig. 1B (A). Among the subcloned cell lines, two vector-transfected cell lines (Vector-1 and -2) and two pSUPER(BubR1/RNAi)-transfected cell lines (BubR1/RNAi-1 and -2), both of which show decreased expression of BubR1, were maintained for further experimentation. Exponentially growing cells were immunostained for centrosomes as described in the legend to Fig. 2B and C, and the centrosome profiles were determined. The results shown in panel B are presented as the average ± standard error from three independent immunostainings. For each immunostaining, >300 cells were examined. Representative immunostaining images are shown in panel C, in which part a shows the overlay image of {gamma}-tubulin immunostaining and DAPI staining of Vector-1 cells and part b shows that of BubR1/RNAi-1. Scale bar, 20 µm. Arrows, cells with amplified centrosomes. (D) Vector-1 and BubR1/RNAi-1 cells were subjected to the nocodazole-responsive checkpoint assay (9). Cells were exposed to nocodazole (0.12 µg/ml) for 44 h and examined by flow cytometry. Cell cycle progression was halted in the Vector-1 cells in response to nocodazole treatment, indicating that these cells are equipped with an intact checkpoint function. In contrast, BubR1/RNAi-1 cells continued cycling without cytokinesis, resulting in generation of a cell population with 8N DNA content (black arrow). The open arrow indicates the position of 8N DNA content.

 
If the reduction of the checkpoint function is responsible for reappearance of PL cells with amplified centrosomes when BubR1 expression is silenced, PL cells silenced for BubR1 expression should show reduced checkpoint function against the defects caused by exposure to microtubule toxins. To test this, Vector-1 and BubR1/RNAi-1 were exposed to nocodazole for 44 h and subjected to flow cytometric analysis (Fig. 7D). The nocodazole-responsive checkpoint was intact in Vector-1 cells, resulting in accumulation of a G2/M cell population. In contrast, octaploidization (8N), which indicates progression through mitosis, was readily observed in nocodazole-treated BubR1/RNAi-1 cells, showing that the nocodazole-responsive checkpoint was significantly impaired in BubR1/RNAi cells.

p53 controls BubR1 transcription. The observation that BubR1 mRNA and BubR1 protein are present at significantly reduced levels in early-passage p53–/– MEEs raised the possibility that BubR1 expression may be controlled by p53. To test this, MSFs were prepared from adult male p53–/– mice and p53+/+ male littermates from two independent p53+/– mouse crosses. These cells were designated p53–/– MSF1, p53+/+ MSF1, p53–/– MSF2, and p53+/+ MSF2, respectively. We performed Northern blot analysis of total RNA prepared from these cells (passage 2) under exponential growth using mBubR1 as a probe (Fig. 8A). Both p53–/– MSF1 and p53–/– MSF2 showed barely detectable levels of BubR1 mRNA. In contrast, both p53+/+ MSF1 and p53+/+ MSF2 showed 8- to 10-fold higher levels of BubR1 mRNA, demonstrating that BubR1 transcription largely depends on the presence of p53.



View larger version (55K):
[in this window]
[in a new window]
  		FIG. 8. p53 controls BubR1 mRNA expression. (A) Total RNA of MSFs (passage 2) prepared from p53+/+ and p53–/– littermate mice from two independent p53+/– mouse mating were subjected to Northern blot analysis using mBubR1 as a probe (see Materials and Methods) (top part). As a loading control, the same blot was reprobed with the glyceraldehyde-3-phosphate dehydrogenase gene (GAPDH) (bottom part). (B) p53–/– MSFs (passage 2) were transfected with vector, wt p53, or Val135 mutant p53 along with a plasmid containing a hygromycin resistance gene for a selection marker (20:1 molar ratio). After hygromycin treatment, colonies were pooled and examined for p53 expression by immunoblot analysis using anti-p53 monoclonal antibody (Pab421) (top part). As a loading control, the same blot was reprobed with anti-ß-tubulin monoclonal antibody (bottom part). (C) The pooled cell lines generated in panel B were examined for BubR1 mRNA levels by Northern blot analysis using mBubR1 as a probe (top part). As a loading control, the same blot was reprobed with GAPDH (bottom part).

 
We next examined whether p53 controls BubR1 transcription through its transactivation function. We first transfected p53–/– MSFs (passage 2) with a plasmid encoding either mouse wt p53 or transactivation-negative point mutant p53 (Ala to Val at aa 135) (29) together with a plasmid containing a hygromycin resistance gene. The vector was transfected as a control. The hygromycin-resistant colonies were pooled and examined for expression of transfected p53 by immunoblot analysis (Fig. 8B). The mutant p53 transfectants expressed high levels of p53, as expected from the stable nature of a transactivation-negative mutant. In contrast, wt p53 transfectants expressed low but detectable levels of p53, which was expected since wt p53 is unstable and high-level expression of wt p53 is known to be cytotoxic. The vector-, wt p53-, and mutant p53-transfected cells were then subjected to Northern blot analysis using the mBubR1 probe (Fig. 8C). The mutant p53 transfectants expressed a barely detectable level of mBubR1 mRNA, similar to the vector transfectant, while wt p53 transfectants expressed more-than-fivefold higher levels of mBubR1 mRNA. Thus, p53 positively controls BubR1 transcription through its transactivation function.

Introduction of BubR1 in early-passage p53–/– cells restores the checkpoint against the defects associated with centrosome amplification and microtubule toxins. p53 was implicated in the spindle checkpoint function: cells prepared from p53–/– mice proceed through mitosis in the presence of nocodazole-induced spindle damage (9). Although the involvement of p53 in the spindle checkpoint was later challenged (25), the finding that BubR1 transcription largely depends on the presence of p53 raises the possibility that p53 may be indirectly involved in the spindle checkpoint through modulating the levels of BubR1. To test this, early-passage p53–/– MSFs were transfected with mBubR1 together with a plasmid containing a hygromycin resistance gene as a selection marker. For a control, the vector was transfected. The hygromycin-resistant colonies were pooled and confirmed for the expression of BubR1 by immunoblot analysis (Fig. 9A). BubR1 transfectants (p53–/– MSF/BubR1) expressed three- to fourfold higher levels of BubR1 than the vector transfectants (p53–/– MSF/Vec).



View larger version (38K):
[in this window]
[in a new window]
  		FIG. 9. Introduction of BubR1 into early-passage p53–/– cells restores a nocodazole-responsive checkpoint function. Early-passage p53–/– MSFs were transfected with a plasmid encoding mBubR1 or a vector together with a hygromycin resistance gene-containing plasmid (20:1 molar ratio). The surviving cells after a 2-week hygromycin treatment were pooled and designated p53–/– MSF(Vec) and p53–/– MSF(BubR1), respectively. The lysates prepared from these cells were examined for BubR1 expression by immunoblot analysis using goat anti-BubR1 antibody (A). The sensitivity of this antibody is too low to detect hyperphosphorylated BubR1. BubR1 expression in the vector-transfected cells is virtually undetectable, while the BubR1-transfected cells show readily detectable levels of BubR1 (top part). The blot was also probed with anti-ß-tubulin antibody. (B) p53–/– MSF(Vec) and p53–/– MSF(BubR1) were immunostained for {gamma}-tubulin and counterstained for DAPI, and the frequencies of cells with amplified centrosomes were determined under a fluorescence microscope (graph a). The frequencies of lagging chromosomes and chromosome bridge (bridge ch.) formation in p53–/– MSF(Vec) and p53–/– MSF(BubR1) were also determined as described in the legend to Fig. 5. The results are presented as the average ± standard error from three experiments. For each experiment, >300 cells were examined. (C) p53–/– MSF(Vec) and p53–/– MSF(BubR1), as well as untransfected p53+/+ and p53–/– MSFs, were subjected to the nocodazole-responsive checkpoint assay as described in the legend to Fig. 7D. Cell cycle progression was halted in the untransfected p53+/+ MSFs, while the untransfected p53–/– MSFs continued cycling without cytokinesis, resulting in generation of a cell population with 8N DNA content (black arrow). p53–/– MSF(Vec) cells behaved similarly to the untransfected p53–/– MSFs, while the cycling of p53–/– MSF(BubR1) was halted, similar to the untransfected p53+/+ MSFs. Open arrows indicate the positions of 8N DNA content.

 
p53–/– MSF/BubR1 showed significant suppression of centrosome amplification (Fig. 9B, graph a), as well as frequencies of lagging chromosomes and formation of chromosome bridges (graph b) compared with p53–/– MSF/Vec cells. We further examined p53–/– MSF/BubR1 and p53–/– MSF/Vec cells, as well as untransfected p53+/+ and p53–/– MSFs, after 44 h of nocodazole treatment by flow cytometry (Fig. 9C). As previously shown (9), p53+/+ MSFs were arrested and octaploidization (8N) was not observed while octaploidization was observed in nocodazole-treated p53–/– MSFs. The control p53–/– MSF/Vec cells also showed octaploidization, as expected. In contrast, p53–/– MSF/BubR1 cells were arrested and virtually no octaploidization was observed. Thus, exogenously introduced BubR1 enables early-passage p53–/– cells to exert the checkpoint activity against mitotic defects associated with centrosome amplification and microtubule toxins, and thus, p53 may be involved in the spindle checkpoint activity indirectly via modulating the levels of BubR1.


   DISCUSSION
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
The genomic convergence accompanied by restoration of numerical integrity of centrosomes in prolonged culture of p53–/– MEEs (7) provided an opportunity to identify a protein(s) that suppresses centrosome amplification. Among several genes differentially expressed between early- and late-passage p53–/– MEEs, we exploited the role of BubR1, a key component of the spindle checkpoint machinery in suppression of centrosome amplification in late-passage p53–/– cells. Together with other proteins (i.e., Bub1, Bub3, Mad1, Mad2, Mad3, CENP-E, and Mps1), BubR1 senses the failure of spindle attachment or lack of spindle tension at the kinetochore and executes checkpoint function to prevent anaphase onset. Ectopic expression of BubR1 in early-passage p53–/– cells resulted in selective and rapid elimination of cells with amplified centrosomes. Moreover, silencing of BubR1 in late-passage p53–/– cells resulted in reemergence of cells with amplified centrosomes. From these observations, we concluded that the up-regulation of BubR1 is the primary cause of suppression of centrosome amplification in late-passage p53–/– cells. Moreover, up-regulation of BubR1 expression in prolonged culture in p53–/– cells appears to be a more general phenomenon, since we repeated the long-term culturing (genomic convergence) of p53–/– MSFs and observed suppression of centrosome amplification and increased expression of BubR1 in late-passage p53–/– MSFs similar to p53–/– MEEs. Centrosome amplification is the target of chromosome stabilization during genomic convergence in p53–/– cells. In this context, any mutations that suppress centrosome amplification can achieve the same goal, namely, stabilization of chromosomes. Why then is the up-regulation of BubR1 specifically targeted? The answer may be linked to our finding that BubR1 transcription is positively controlled by p53. In early-passage p53–/– cells, BubR1 expression is down-regulated and thus the checkpoint function against the mitotic aberrations associated with centrosome amplification is significantly compromised, allowing the cells with amplified centrosomes to persist in the culture. Thus, during genomic convergence, restoring the normal levels of BubR1 may be the most efficient way to suppress centrosome amplification, which in turn selectively eliminates cells with amplified centrosomes.

The important question to be addressed is whether the spindle checkpoint function is indeed impaired in early-passage p53–/– cells. The previous report has implicated p53 in the spindle checkpoint function (9); upon exposure to microtubule toxins (i.e., nocodazole), p53–/– MEFs continue cell cycling without cytokinesis, becoming cells with multiploidy (4N, 8N, 16N...). In contrast, p53+/+ MEFs become arrested with 4N DNA content. This study concluded that loss of p53 impairs the spindle checkpoint function, allowing continuous cell cycling without cytokinesis. However, the examination of early-passage p53+/+ and p53–/– MEFs in response to nocodazole subsequently revealed that both p53+/+ and p53–/– MEFs proceeded into G1 phase morphologically (from mitotic rounded-up cells to interphase flattened cells) without cytokinesis at similar kinetics, but p53+/+ MEFs became arrested in G1, while p53–/– MEFs continued cell cycling (25), arguing the previous conclusion of the involvement of p53 in the spindle checkpoint. Our present study shows that BubR1 transcription largely depends on p53 (or transactivation function of p53), and exogenously introduced BubR1 augments the checkpoint function against mitotic defects associated with centrosome amplification and microtubule toxins in early-passage p53–/– cells, suggesting that p53 may be indirectly (through controlling BubR1 expression) involved in the spindle checkpoint, and loss of p53 may at least partially impair the spindle checkpoint function. We further made a potentially important observation: in early-passage p53–/– (PE) cells, BubR1 expression is down-regulated, and the spindle checkpoint function is compromised when compared with the genomically converged late-passage p53–/– (PL) cells that express BubR1 at a similar level with p53+/+ cells. When nocodazole-treated PE and PL cells were examined at a molecular level, we found that there was a significant delay in CDK1/cyclin B inactivation in PL cells, although they show similar mitotic indices. Further analysis revealed that the postmitotic PL cells that show interphase phenotypes (i.e., flattened morphology) retain CDK1/cyclin B-dependent mitosis-specific phospho-antigens (MPM2 antigens) at the level equivalent to mitotic cells. In contrast, the postmitotic interphase PE cells show minimal MPM2 antigens, similar to normal interphase cells. Thus, PL cells appear to undergo interphase transformation (flattened morphology) in the presence of active CDK1/cyclin B. Since CDK1/cyclin B is one of the major targets of APC, these findings suggest that the spindle checkpoint is compromised in PE cells, resulting in APC-mediated inactivation of CDK1 activity and progression into interphase. In PL cells, due to up-regulation of BubR1, the spindle checkpoint is augmented, resulting in suppression of APC in the presence of nocodazole, which in turn delays inhibition of CDK1 activity. These PL cells appear to transform to interphase morphology and either become arrested or undergo cell death, suggesting that persistent mitotic activities/phenotypes due to activation of the spindle checkpoint may interfere with further cell cycle progression. However, the activated spindle checkpoint is known to halt (or delay) mitotic progression and sustain mitotic (rounded-up) morphology for a period of time. This is not the case for PL cells undergoing nocodazole treatment. Therefore, it remains possible that the up-regulation of BubR1 in PL cells may elicit the checkpoint activity against the mitotic abnormalities associated with centrosome amplification, as well as exposure to microtubule toxins, which is different from the conventional spindle checkpoint. Apparently, more studies are needed to clarify this issue.

Recent studies have provided critical information on the molecular aspects of how BubR1 works as a spindle checkpoint protein. BubR1 physically interacts with Cdc20 and APC (5, 41), and BubR1 effectively inhibits APC by sequestering Cdc20 in an in vitro APC/cyclosome inhibition assay system (15, 37, 38). It remains to be clarified whether the inhibition of APC by BubR1 operates independently from Mad2, another well-established Cdc20 inhibitor (36), or additively or synergistically with Mad2. In our system, similar to ectopic expression of BubR1, Mad2 overexpression restored the checkpoint function targeting the abnormalities associated with amplified centrosomes in early-passage p53–/– cells which express minimal levels of BubR1. Although this finding indicates that early-passage p53–/– cells lack the checkpoint function against the spindle and chromosome abnormalities associated with amplified centrosomes, it cannot address the functional interaction between BubR1 and Mad2. BubR1 is also known to physically interact with CENP-E (5, 42), which is a kinetochore-based kinesin-like motor protein, and is implicated in microtubule attachment to the kinetochore, generation of kinetochore tension, and checkpoint responses (1, 5, 42). It has recently been found that the association of CENP-E activates the kinase and spindle checkpoint activities of BubR1 (28).

The requirement of the kinase activity of BubR1 in its spindle checkpoint function has been not conclusively determined. In the in vitro APC inhibition assay, the kinase-dead mutants (point mutant and mutant lacking the entire kinase domain) were shown to be able to inhibit the activation of APC by Cdc20 as efficiently as wt BubR1 (6, 38). Similarly, depletion of ATP from the in vitro assay did not diminish the inhibitory effect of BubR1 (15). Although these studies do not directly address the requirement of the kinase activity for BubR1 in the spindle checkpoint, they suggest that the kinase activity may be dispensable. In contrast, the in vivo studies using HeLa cells showed that transfection of the kinase-dead mutant BubR1 (a similar small deletion mutant used in our study) resulted in elimination of spindle checkpoint (5). Moreover, it has been shown that BubR1 can directly phosphorylate Cdc20, suggesting the posttranslational control of Cdc20 activity by BubR1-mediated phosphorylation (41). In our study, the expression of BubR1 kinase-dead mutants in early-passage p53–/– cells showed partial suppression of centrosome amplification. This result is not inconsistent with the aforementioned study by Mao et al. showing that only a small fraction of the total BubR1 needs to be kinase competent, but sufficient levels of total BubR1 (either kinase competent or dead) are needed for the spindle checkpoint activity (28). Upon introduction of the kinase-dead mutant, the endogenous wt BubR1 (at least partially) satisfies the amount of the kinase-competent BubR1 needed for eliciting the spindle checkpoint in response to the defects associated with centrosome amplification. In addition, we found that the mutants lacking the N-terminal sequence (aa 1 to 449) of BubR1 failed to suppress centrosome amplification. Taylor et al. (40) have shown that the aa 392 to 433 sequence of human BubR1 (aa 385 to 426 in mBubR1) is important for physical interaction with Bub3, and this interaction is required for BubR1 to localize to kinetochores. Thus, kinetochore association of BubR1 is critical for its checkpoint response to mitotic defects due to centrosome amplification.

Our present study provides the first molecular mechanism underlying genomic convergence in cultured cells. Normally, centrosome amplification and consequential mitotic defects trigger checkpoint activity, leading to cell death. However, early-passage p53–/– cells have reduced levels of BubR1 and hence fail to execute an effective checkpoint against spindle/chromosome aberrations associated with amplified centrosomes, allowing anaphase progression. This in turn results in an increased frequency of chromosome transmission errors and thus exacerbates karyotypic heterogeneity. At a certain time point during prolonged culturing of p53–/– cells, one or a few cells acquire the chromosome composition that promises the most advantageous growth phenotypes and gradually dominate the culture. For these cells, maintenance of this particular karyotype becomes a primal selection pressure. Thus, restoration of numeral integrity of centrosomes becomes a target of this selection pressure. This can be achieved by acquisition of mutations that either correct the cause of centrosome amplification or specifically eliminate the cells with amplified centrosomes by inducing growth arrest and cell death. In p53–/– cells, a mutation(s) and/or a genetic modification(s) that results in p53-independent expression of BubR1 may be a common event, restoring/augmenting the checkpoint function, which in turn efficiently targets cells with amplified centrosomes. We thus provided the molecular mechanism of genomic convergence in cultured cells, which may well be equated with the genomic convergence observed during tumor progression.


   ACKNOWLEDGMENTS
 
We thank J. Salisbury for anticentrin antibody, G. Wahl for an H2B-GFP plasmid, M. Oren for mouse p53 plasmids, W. Dai for anti-BubR1 antibody, and P. Rao for anti-MPM2 antibody. We also thank Sandy Schwemberger and George Babcock for FACS analysis. We also thank A. Staubach, E. Hunter, and I. Wenker for technical assistance.

This research was supported by grants-in-aid from the Naito Foundation and the Ministry of Education, Sports and Culture, Japan (13760221 and 1590492 to M.O.), the Japan Society for the Promotion of Science (to T.O.), and the National Institutes of Health (CA90522 and CA95925 to K.F.).


   FOOTNOTES
 
* Corresponding author. Mailing address: Laboratory of Veterinary Internal Medicine, Faculty of Agriculture, Yamaguchi University, 1677-1 Yoshida, Yamaguchi 753-8515, Japan. Phone: 81-83-933-5893. Fax: 81-83-933-5893. E-mail: okudamu{at}yamaguchi-u.ac.jp. Back


   REFERENCES
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 
1. Abrieu, A., J. A. Kahana, K. W. Wood, and D. W. Cleveland. 2000. CENP-E as an essential component of the mitotic checkpoint in vitro. Cell 102:817-826.[CrossRef][Medline]

2. Brinkley, B. R., and T. M. Goepfert. 1998. Supernumerary centrosomes and cancer: Boveri's hypothesis resurrected. Cell Motil. Cytoskelet. 41:281-288.[CrossRef][Medline]

3. Cahill, D. P., C. Lengaurer, J. Yu, G. L. Riggins, J. K. Willson, S. D. Markowitz, K. W. Kinzler, and B. Vogelstein. 1998. Mutations of mitotic checkpoint genes in human cancers. Nature 392:300-303.[CrossRef][Medline]

4. Chan, G. K. T., B. T. Schaar, and T. J. Yen. 1998. Characterization of the kinetochore-binding domain of CENP-E reveals interactions with the kinetochore proteins CENP-F and hBUBR1. J. Cell Biol. 143:49-63.[Abstract/Free Full Text]

5. Chan, G. K. T., S. A. Jablonski, V. Sudakin, J. C. Hittle, and T. J. Yen. 1999. Human BUBR1 is a mitotic checkpoint kinase that monitors CENP-E functions at kinetochores and binds the cyclosome/APC. J. Cell Biol. 146:941-954.[Abstract/Free Full Text]

6. Chen, R.-H. 2002. BubR1 is essential for kinetochore localization of other spindle checkpoint proteins and its phosphorylation requires Mad1. J. Cell Biol. 158:487-496.[Abstract/Free Full Text]

7. Chiba, S., M. Okuda, J. G. Mussman, and K. Fukasawa. 2000. Genomic convergence and suppression of centrosome hyperamplification in primary p53–/– cells in prolonged culture. Exp. Cell Res. 258:310-321.[CrossRef][Medline]

8. Cleveland, D. W., Y. Mao, and K. F. Sullivan. 2003. Centromeres and kinetochores: from epigenetics to mitotic checkpoint signaling. Cell 112:407-421.[CrossRef][Medline]