Roles of BCCIP in Chromosome Stability and Cytokinesis

XMeng1, J Fan2 and Z Shen1,2

1Department of Molecular Genetics and Microbiology, University of New Mexico School of Medicine, Albuquerque, NM, USA and 2Department of Radiation Oncology, The Cancer Institute of New Jersey, Robert Wood Johnson Medical School, New Brunswick, NJ, USA

The BRCA2 gene is involved in recombinational DNA a failure of the immediate G1/S checkpoint activation repair and cytokinesis. BRCA2 defects are associated (or tetraploid checkpoint) (Margolis et al., 2003). with chromosomal abnormalities, which is a hallmark of The tumor suppressor gene BRCA2 is involved in genomic instability that contributes to tumorigenesis. homologous recombinational DNA repair that contri- Here, we show that downregulation of a BRCA2 butes to structural chromosome stability (Moynahan interacting protein (BCCIP) in HT1080 cells leads to et al., 2001; Venkitaraman, 2002; Powell and Kachnic, chromosomal polyploidization, centrosome amplification 2003). BRCA2 also participates in the regulation of and abnormal mitotic spindle formation. The BCCIP mitosis and cytokinesis that contribute to numerical knockdown cells can enter mitosis and retain spindle chromosomal stability (Daniels et al., 2004; Rudkin and checkpoint, but fail to complete cytokinesis. Our data Foulkes, 2005). BRCA2 interacting protein a (BCCIPa) suggest an essential role of BCCIP in the maintenance of is a BRCA2 and CDKN1A (p21, Waf1 and Cip1) genomic integrity. Interacting Protein (Ono et al., 2000; Liu et al., 2001). Oncogene (2007) 26, 6253–6260; doi:10.1038/sj.onc.1210460; Owing to alternative splicing, a second isoform BCCIPb published online 23 April 2007 is expressed in human cells (Meng et al., 2003). BCCIPa and BCCIPb share identical N-terminal 258 amino Keywords: cytokinesis; chromosome instability; BRCA2; acids, but each has a unique C-terminal sequence (Meng BCCIP; polyploidy et al., 2003). Additional studies have shown that BCCIPb interacts with BRCA2 and p21 (Meng et al., 2004a; Lu et al., 2005). The chromatin-bound fraction of BCCIPa/b colocalizes with BRCA2 and contributes to BRCA2 and RAD51 nuclear focus formation (Lu et al., B Introduction 2005). A moderate knockdown ( 50% of downregulation) of both or either BCCIP isoform significantly Human cancers are characteristic of numerical and reduces DNA double-strand break-induced homologous structural chromosomal instability. Structural chromo- recombination, impairs G1/S checkpoint activation, some instability, including reciprocal and nonreciprocal abrogates p53 transactivation activities and downregu- translocations, regional chromosomal amplifications/ lates p21 expression (Meng et al., 2004a, b, 2007; Lu insertions/deletions, may arise from defects in cell cycle et al., 2005). In this sudy, we show that severely checkpoints, DNA recombination, DNA repair and downregulated BCCIP expression results in mitosis telomere maintenance. Numerical chromosomal instability defect and chromosome instability in HT1080 cells, (gain or loss of chromosomes) arises from chromosome suggesting BCCIP as a critical player in maintaining segregation errors during mitosis, or failure to complete chromosome stability. cell separation at the end of mitosis (cytokinesis). Aneuploidy generally results from chromosome segregation errors, which may be a consequence of misregula- Results tion of microtubule dynamics, centrosome replication, chromosome condensation, kinetochore assembly, chro- BCCIP knockdown causes chromosome instability mosome cohesion and spindle checkpoint control In a study when BCCIPa and/or BCCIPb were partially (Lengauer et al., 1998). A second type of numerical downregulated by RNAi, we observed an increase in chromosomal abnormality is polyploidization, which polyploid cells after an extended culture of cells with may result from failed completion of mitosis followed by moderate BCCIP knockdown (Supplementary Figure S1), suggesting a potential role of BCCIP in chromosomal instability. To investigate this, we established cell lines Correspondence: Dr Z Shen, Department of Radiation Oncology, The with severe BCCIPa/b knockdown (>95% downregula- Cancer Institute of New Jersey, Robert Wood Johnson Medical tion) in HT1080 cells by combining two short hairpin School, 195 Little Albany Street, New Brunswick, NJ 08903, USA. E-mail: [email protected] RNA (shRNA) targeted at two independent regions of Received 28 September 2006; revised 5 February 2007; accepted 9 March the BCCIPa/b mRNA (Figure 1a). Although the 2007; published online 23 April 2007 growths of these cells are compromised in later stage BCCIP, chromosome instability and cytokinesis X Meng et al 6254 of culture, the cells can be maintained in culture for a subsequent DNA replication) after a failed cell division few passages. In these cells, we observed an increase in in mitosis. To distinguish the potential mechanisms by polyploid cells between passages 2 and 5 (Figure 1b). which BCCIP knockdown induces polyploidization, Consistent with this observation, severe BCCIP knock- HT1080 cells were incubated with nocodazole to be down induces cells with large or multiple nuclei blocked at metaphase, and then stained with an (Figure 1c). In addition, we used a chromosome 12- specific centromeric DNA probe to quantify chromosome numbers by fluorescent in situ hybridization (FISH). As shown in Figure 2a and d, control HT1080 cells are mostly diploid. However, the BCCIP knockdown cells displayed a significant increase in cells with more than two copies of chromosome 12 (Figure 2b–d). These data strongly suggest chromosome instability in cells with severely downregulated BCCIP.

Knockdown of BCCIP does not prevent the entry of G2 cells into M phase Polyploidy can be induced by endoreduplication, by which the cells do not enter mitosis but reenter S-phase after the previous round of DNA replication. Polyploidy can also result from reentry into interphase (and

pPUR/U6-GFP + ++ Figure 2 Chromosome number abnormality in BCCIP knock- pPUR/U6-αβ311 + + + down cells. A chromosome 12-speciﬁc centromeric DNA probe was pSilencer-scramble ++ used for FISH analysis and to index chromosomal stability of pSilencer-αβ633 ++ control and BCCIP knockdown cells (passage 5) as reported by others (Jallepalli et al., 2001). DNA was stained with DAPI (blue). αβ ++ pSilencer- 730 (a) represents control cells, (b and c) represent BCCIP knockdown cells and (d) illustrates the distribution of cells with different BCCIPα number of chromosome 12. More than 300 cells were scored for BCCIPβ each group.

β-actin Figure 1 BCCIP knockdown by shRNA induces polyploidization 1 23456 of HT1080 cells. (a) Knockdown of BCCIP by shRNA. Three common regions between BCCIPa and BCCIPb mRNA at locations ab311 bp, ab 633 bp and ab 730 bp were selected for shRNA targeting, because they have no significant homology with bcany other human expressed sequence tags (EST) sequences based on a basic local alignment search tool (BLAST) search (see Materials and methods section for the details of nucleotide sequences and vector construction). The efficacy of combining Control 2.7% two shRNAs to severely knockdown BCCIP expression is shown. Combining two shRNAs is feasible, because pPUR/U6 and pSilencer use different selection markers (see Materials and methods section for details). Cells transfected with indicated shRNA vectors were selected with puromycin and hygromycin. Immunoblots of whole cell extract were carried out with anti- (passage 2) BCCIP (top panel), or anti-b-actin blot (lower panel). As shown here (lanes 2–4), targeting by a single BCCIP shRNA creates a 28.6% moderate knockdown (B50%) condition for BCCIP. By combining two shRNAs (lanes 5 and 6), shRNA-ab311/shRNA-ab633 or shRNA-ab311/shRNA-ab730, a severe knockdown condition with B BCCIP- 95% BCCIPa/b downregulation was created. These two cell lines (lanes 5 and 6) have identical phenotypes and were used for shRNA subsequent experiments unless stated otherwise. (b) Polyploidiza- (passage 5) tion in BCCIP knockdown cells measured by DNA content analysis. The DNA content of BCCIP knockdown cells were 46.4% analysed by flow cytometry at passages 2 and 5. The percentages of cells with more than 4N DNA content are indicated. (c) Formation of cells with large nucleus after BCCIP knockdown. Representative cell morphology after the cells were stained with anti-a-tubulin and 2N 4N 8N DAPI to contrast the nuclei from the cytoplasm. Arrows indicate BCCIP knockdown cells with large or multiple nuclei.

Oncogene BCCIP, chromosome instability and cytokinesis X Meng et al 6255 antibody to Serine-10 phosphorylated histone H3, knockdown causes polyploidization, which has likely ph(Ser10)H3, which marks mitotic cells (Hans and caused a failure of a critical step after the cells Dimitrov, 2001). The specificity of this antibody to have passed the metaphase (Figure 3), we investigated mitotic cells was confirmed by immunostaining of whether BCCIP is involved in cytokinesis. We HT1080 cells (Supplementary Figure S2). After the found that BCCIP knockdown significantly increased DNA was co-stained with propidium iodide, the the percentage of cells with failed cytokinesis ph(Ser10)H3 positive cells were scored by flow cytome- (Figure 4), suggesting that BCCIP knockdown try. In control cells, ph(Ser10)H3 positive cells accumu- induces chromosomal polyploidization by impairing late in diploid population but little in tetraploid after cytokinesis, although we cannot exclude the possibility nocodazole block (Figure 3a, b and e). However, in that a small portion of the cells may adopt endoredu- BCCIP knockdown cells, ph(Ser10)H3 positive cells plication or abort mitosis at a stage earlier than accumulate in both the diploid and tetraploid popula- cytokinesis. tions after nocodazole block (Figure 3c–e). The accu- It has been suggested that polyploidization may cause mulation kinetics of ph(Ser10)H3 positive tetraploid centrosome amplification in p53-deficient cells (Meraldi BCCIP knockdown cells is approximately the same as et al., 2002). We further investigated whether the that of the diploid control cells. These data suggest that knockdown of BCCIP can affect centrosomes. As the BCCIP knockdown cells indeed enter mitosis. They shown in Figure 5a, we indeed observed an increase of also suggest that BCCIP knockdown cells have normal cells with abnormal number of centrosomes in BCCIP spindle checkpoint activation as nocodazole effectively knockdown cells. In control cells, about 1.2% cells have blocks cells at metaphase. Therefore, the polyploidiza- more than two centrosomes. However, we observed tion in BCCIP knockdown cells is likely due to a failure B8% of the BCCIP knockdown cells with more than of cell division after passing the metaphase (see below), two centrosomes (Figures 5b). Centrosome amplifica- but unlikely due to an endoreduplication of DNA in S tion was also observed in cells with single BCCIP phase. isoform knockdown (Supplementary Figure S3). In addition, downregulation of BCCIP induces formation BCCIP knockdown causes cytokinesis failure and of multipolar mitotic spindles (Figure 6). Centrosome centrosome amplification amplification and multipolar mitotic spindles are further BRCA2 is involved in cytokinesis (Daniels et al., evidence of genome instability in BCCIP knockdown 2004), and interacts with BCCIP. Because BCCIP cells.

a e 2.52% 0.28% 9.57% 0.78% 17.07% 1.43% 20

15 Control b 10 5.60% 2.27% 2.27%

5 Phosphorylated Histone-3

c of Cells with Ser-10 Percentage 0.76% 1.24% 4.29% 6.51% 6.05% 14.45% 0 0 48 Hours of Nocodazole Block BCCIP- shRNA Control (diploid) Control (tetraploid) d 55.53% shRNA (diploid) shRNA (tetraploid) 26.69% 26.69%

0 hour 4 hours 8 hours Figure 3 BCCIP knockdown does not affect the entry into mitosis. Control or BCCIP knockdown HT1080 cells (passage 3) were incubated with nocodazole for 4 or 8 h to block cells at metaphase. After being ﬁxed with ethanol, cells were double stained with Serine-10 phosphorylated histone-3 antibody (ph(Ser10)H3) (green) to mark mitotic cells and propidium iodide for DNA (red). In (a–d), the horizontal axis represents the DNA content. In (b and d), the vertical axis represents the number of cells. In (a and c), the vertical axis represents the ph(Ser10)H3 level, and regions R3 and R4 represent the diploid and tetraploid mitotic cells, respectively. (e) shows the percentage of ph(Ser10)H3 positive diploid and tetraploid cells in control and BCCIP knockdown HT1080 cells.

Oncogene BCCIP, chromosome instability and cytokinesis X Meng et al 6256

Figure 5 Centrosome ampliﬁcation in BCCIP knockdown cells. (a) shows centrosome staining in control cells and in BCCIP knockdown cells (passage 5) with more than two centrosomes (bottom two panels). Centrosomes (red) were stained with anti- pericentrin, and nuclei with DAPI (blue). (b) shows the percentage of cells with more than two centrosomes per nucleus. Presented are averages of two independent cell lines (lanes 5 and 6 in Figure 1a). More than 300 cells were scored from each cell line.

Figure 4 BCCIP knockdown induces cytokinesis failure. Cells were stained with anti-a-tubulin. (a–c) represent typical BCCIP knockdown cells with cytokinesis failure. (d) shows the percentage of cells with morphology consistent with cytokinesis failure. Shown are the average results of two cell lines at passage 5 (lanes 5 and 6 in Figure 1a). A minimum of 300 cells was scored in each cell line.

Cytokinesis failure and centrosome amplification are immediate events following BCCIP knockdown To determine whether cytokinesis failure and centrosome amplification are the primary consequences of BCCIP knockdown, we transfected a mixture of two independent small interference RNAs (siRNAs) targeted at regions ab311 and ab633 into the cells. The level of BCCIP starts to decline 24 h after transfection, and reaches the lowest level by day 4. However, at day 5 after the transient transfection, the BCCIP level starts to recover (Figure 7c), which is common when siRNA is transiently transfected into the cells. This may be due to Figure 6 Multipolar spindle formations in BCCIP knockdown the transient nature of transfected siRNA, or due to a cells. Centrosomes were stained with anti-pericentrin (red), mitotic potential fast growth of a non-transfected subpopula- spindle (green) with anti-a-tubulin and DNA (blue) with DAPI. (a) tion of cells. In these cells with acute BCCIP knock- shows normal spindle in control and multiple polar spindles in down, we observed an increase of cytokinesis failure and BCCIP knockdown cells. (b) shows the percentage of mitotic cells with multipolar spindles, which was obtained by counting 100 centrosome amplification as early as 2 days after mitotic cells from two independent BCCIP knockdown cell lines at transfection of BCCIP siRNA (Figure 7a and b), passage 5 (lanes 5 and 6 of Figure 1a). suggesting that the cytokinesis failure and centrosome amplification are early effects of BCCIP downregulation. These data also show that the kinetics of centrosome amplification and cytokinesis failure (thus centrosome amplification may also be caused by tetraploidization) are almost identical in HT1080. unbalanced mitotic division. Among the cells with more than two centrosomes, In addition, the method described by Fukasawa et al. about 48% have three or five centrosomes, and 52% (1996) was modified to identify whether centrosome have four or six centrosomes, suggesting that the amplification and cytokinesis failure occur during the

Oncogene BCCIP, chromosome instability and cytokinesis X Meng et al 6257 Centrosome amplification and cytokinesis failure caused by BCCIP knockdown can be prevented with expression of exogenous BCCIP To further support the role of BCCIP in cytokinesis and centrosome amplification, and to rule out the possibility that the cytokinesis failure and centrosome amplification in BCCIP knockdown cells are caused by off-target effects, exogenous flag-BCCIPa or flag-BCCIPb were constitutively expressed in HT1080 cells. Then the expressions of endogenous BCCIPa or BCCIPb were downregulated by siRNAs targeted at the 30-untranslated regions (30-UTR) of BCCIPa or BCCIPb mRNA (30-BCCIP siRNA). Expressions of the exogenous flag- BCCIPa or flag-BCCIPb were not affected by the 30-BCCIP siRNA (Figure 8c) because the flag-BCCIP vectors contain no 30-UTR. As shown in Figure 8a and b (columns 3 and 5), targeting BCCIPa or BCCIPb 30-untranslated regions alone increased the percentage of cells with cytokinesis failure and centrosome amplification. However, expression of flag-BCCIPa or flag- BCCIPb in these cells significantly reduced cytokinesis failure and centrosome amplification (columns 4 and 6, Figure 8a and b). These data further support the role of BCCIP in cytokinesis and centrosome amplification and ruled out off-target effect of BCCIP RNAi on centrosome amplification and cytokinesis failure.

Discussion

In this study, we identified a functional role of BCCIP in numerical chromosome stability in HT1080 cells. We show that BCCIP downregulation causes failure of cytokinesis, centrosome amplification and abnormal Figure 7 Induction of cytokinesis failure and centrosome ampli- spindle formation. These data strongly support a role fication after acute BCCIP knockdown with siRNA. Various times after HT1080 cells transfected with control luciferase siRNA or a of BCCIP in the maintenance of genome stability. mixture of BCCIP siRNA ab311 and ab633 siRNA, cells were Because chromosome instability is associated with stained with anti-a-tubulin to score cytokinesis failure and with tumor progression, these findings indicate that loss of anti-pericentrin to score centrosome amplification. For the purpose BCCIP may be associated with tumor aggression. The of counting centrosomes, cells with cytokinesis failure but more human BCCIP gene is located at 10q26 (Liu et al., than one nucleus were counted as a single cell. The percentages of cells with cytokinesis failure (a) or with more than two centrosomes 2001), which has been implicated in many forms of (b), and the average and standard deviation from three independent human tumors, including astrocytic brain tumors experiments are shown. (c) shows the reduced expression of BCCIP (Rasheed et al., 1992; Maier et al., 1998; Merlo, 2003; in BCCIP knockdown cells used in (a) and (b), and anti-actin blot Ohgaki et al., 2004). We have previously shown that was used as a loading control. BCCIP expression is downregulated in kidney cancer (Meng et al., 2003), BCCIPa expression was not detectable in astrocytic brain cancer cell line A172 (Liu et al., 2001), and BCCIP deletion has been first cycle of cell division following release from G0 suggested in a brain tumor cell line (Roversi et al., arrest. Briefly, 1 day after BCCIP siRNA transfection, 2005). Recently, we found loss of BCCIP expression in a cell cycle progression was arrested by 0.5% serum significant portion of aggressive astrocytic brain tumors starvation for 60 h and then released by changing (manuscript in preparation), further supporting the role medium back with 10% serum. Sixteen hours after the of BCCIP in genomic instability and tumor aggression. release from the G0 arrest, we observed 0.5% of control Cells that prematurely exit mitosis must bypass the cells and 8.9% of BCCIP knockdown cells with subsequent G1/S checkpoint (sometimes referred as cytokinesis failure, and 0.7% of control cells and 7.7% tetraploidy G1/S checkpoint) before the next DNA of BCCIP knockdown cells with more than two replication to become polyploidy (Margolis et al., 2003). centrosomes. These data suggest that the observed It has been suggested that this G1/S tetraploidy cytokinesis failure and centrosome amplification occur checkpoint is dependent on p53 function (Andreassen during the immediate round of cell cycle after release et al., 2001). Thus, the p53 functions may have also been from G0 phase. impaired in BCCIP knockdown HT1080 cells. We

Oncogene BCCIP, chromosome instability and cytokinesis X Meng et al 6258 previously reported that BCCIP knockdown inactivates the G1/S checkpoint (Meng et al., 2004a, b). Recently, we have shown that the transactivation activity of wild- type p53 is ablated by BCCIP downregulation (Meng et al., 2007). Therefore, the BCCIP knockdown cells would be able to silence the tetraploidy checkpoint after failed cytokinesis, to bypass G1/S checkpoint and to become susceptible for the next round of DNA replication, all of which leads to polyploidization. Both aneuploidy and polyploidy are forms of chromosomal instability. The mechanisms by which aneuploidy and polyploidy occur are subtly different. Aneuploidy often results from chromosome segregation errors during mitosis, while polyploidy may be a consequence of failed cytokinesis, DNA endoreduplication or endomitosis as in megakaryocyte differentiation (Geddis and Kaushansky, 2004). Our data have shown that cytokinesis failure is the major consequence of BCCIP knockdown, which is accompanied by centrosome amplification in the first cell cycle after BCCIP knockdown (Figure 7). We also observed that some BCCIP knockdown cells have three copies of chromosome (Figure 2d), suggesting aneuploidy. Although we cannot rule out that the modest-level aneuploidy is a result of chromosome segregation errors, it is unlikely caused by spindle checkpoint defect as nocodazole effectively blocked BCCIP knockdown cells at metaphase (Figure 3). It is possible that the aneuploidy is a result of abnormal mitotic spindle formations, which are consequence of the centrosome amplification and cytokinesis failure. In addition, some BCCIP knockdown cells have more than two and unpaired (such as three or five) centrosomes. This would add further mechanism for aneuploidy formation. BRCA2 is involved in cytokinesis and homologous recombination (Moynahan et al., 2001; Venkitaraman, 2002; Powell and Kachnic, 2003; Daniels et al., 2004). We have previously reported that chromatin-bound fraction of BCCIP colocalizes with BRCA2, and a modest knockdown of BCCIP (B50% of BCCIP reduction) reduces DNA double-strand break-induced homologous recombination (Lu et al., 2005). In this study, we show that BCCIP is involved in cytokinesis and chromosomal Figure 8 Prevention of cytokinesis failure and centrosome stability. Therefore, BCCIP may regulate the mainte- amplification in BCCIP knockdown cells by exogenous BCCIP nance of genome stability through multiple pathways: expression. Flag-BCCIPa or flag-BCCIPb was expressed in homologous recombinational repair (Lu et al., 2005), G1/ HT1080 cells. Control cells were transfected with an empty vector. Then these cells were transfected with various siRNAs as defined: S checkpoint (Meng et al., 2004a, b) and cytokinesis. It lane (1) control cells transfected with luciferase siRNA; lane (2) remains to be determined whether the interaction control cells transfected with BCCIP siRNA targeted at common between BRCA2 and BCCIP plays a direct role in regions 311 and 633; lane (3) control cells tranfected with 30-siRNA cytokinesis. In addition, it has been reported that p53 0 against BCCIPa (3 -siRNA-a); lane (4) flag-BCCIPa expressing protein has both transcription-dependent and -indepen- cells transfected with siRNA targeted at the 30 untranslated region of BCCIPa mRNA (30-siRNA-a); lane (5) control cells transfected dent role in preventing centrosome amplification (Tar- with siRNA targeted at the 30untranslated region of BCCIPb apore et al., 2001; Tarapore and Fukasawa, 2002; mRNA (30-siRNA-b) and lane (6) flag-BCCIPb expressing cells Shinmura et al., 2006), and the transcriptional-dependent transfected with siRNA targeted at the 30-untranslated region of function of p53 in centrosome amplification is likely BCCIPb mRNA (30-siRNA-b). After the siRNA transfections, cells with cytokinesis failure or more than two centrosomes per nucleus mediated by p21 (Tarapore et al., 2001) and BCCIP were counted. (a) shows the percentage of cells with cytokinesis interacts with p21 (Meng et al., 2004a, b) and BCCIP failure, (b) shows the percentage of cells with centrosome down-regulation abrogates the transcriptional activity of amplification, (c) shows the expression of BCCIP as blotted by p53 (Meng et al., 2007). It is possible that BCCIP may anti-BCCIP, anti-flag and anti-actin in these cells. regulate centrosome stability through p53 and p21 functions.

Oncogene BCCIP, chromosome instability and cytokinesis X Meng et al 6259 Materials and methods BCCIPb isoforms, several shRNA sequences targeted at the shared region of BCCIPa and BCCIPb were used, including: Cell culture shRNA-ab311 (50-GTG TGA TTA AGC AAA CGG AT HT1080 cells were cultured in aMEM (Gibco BRL, Grand G-30), shRNA-ab633 (50-GCC ATG TGG GAA GTG CTA Island, NY, USA) with 10% fetal bovine serum (Biowhittaker, C-30), and shRNA-ab730 (50-GCT GCG TTA ATG TTT Walkersville, MD, USA), 20 mM glutamine and 1% penicillin/ GCA AAT-30). We found that application of a single shRNA streptomycin (Gibco BRL). Plasmids were transfected into can only cause B50% downregulation of BCCIP (Figure 1a), cells using the Geneporter transfection kit (Gene Therapy which is referred as ‘moderate knockdown’ in this study. To Systems Inc., San Diego, CA, USA). Flag-BCCIP expression knockdown BCCIPa/b further, we combined two shRNAs: in HT1080 cells were mediated by the pLXSP vector-based ab311 and ab633 or ab311 and ab730, creating B95% retrovirus infection as reported previously (Liu et al., 2001). downregulation of BCCIP (Figure 1a), referred as ‘severe knockdown’ condition in this study. Briefly, HT1080 cells were Antibodies and Western blot transfected with pPUR/U6 (with puromycin as selection Rabbit anti-BCCIPa/b antibodies were reported previously marker) vectors expressing either control or BCCIP shRNA- (Liu et al., 2001). Other antibodies were purchased as: anti- ab311, and then selected in puromycin for 2 weeks. The puro- Ser-10 phosphorylated histone H3 (ph(Ser10)H3) antibodies mycin-resistant cells were further transfected with pSilencer2.1 from Upstate (Lake Placid, NY, USA); anti-b-actin and anti- (with hygomycin as the selection marker) control or BCCIP a-tubulin antibodies from Sigma (St Louis, MO, USA); and shRNA-ab633 or shRNA-ab730 vectors and selected by anti-pericentrin antibody from Covance Research Products hygromycin B. After 2 weeks’ selection, puromycin and Inc. (Berkeley, CA, USA). Protein extracts were prepared from hygromycin double-resistant HT1080 cells with BCCIP severe cells lysed with 50 mM 4-(2-hydroxyethyl)-1-piperazineethane- knockdown were established and designated as passage 1. Cells sulfonic acid, pH 7.6, 250 mM NaCl, 5 mM ethylenediaminete- were passed every 3 days and cultured in the medium traacetic acid and 0.1% Nonidet P-40. Proteins were separated containing both puromycin and hygromycin. by sodium dodecyl sulfate–polyacrylamide gel electrophoresis, and Western blotting was performed as described (Meng et al., Acute knockdown of BCCIP expression by siRNA 2004a, b). Anti-flag antibody was purchased from Sigma. The siRNA was synthesized by in vitro transcription according to the protocol of the manufacturer (Ambion Inc.). The firefly Immunocytochemistry luciferase control siRNA 50-CTT ACG CTG AGT ACT TCG Cells grown on coverslips were washed with phosphate- A-30 was used as negative control. The BCCIPa 30-terminus- buffered saline (PBS), and fixed with methanol. The fixed cells untranslated region-specific siRNA is 50-AAC ATC TCG were washed with Tris buffered saline with Tween-20 (TBS-T) GCA CCT AGT AAT-30. The BCCIPb 30-terminus untrans- (25 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% Tween-20) lated region specific siRNA is 50-AAC TCA GAC TTT ATT three times and blocked for 30 min with 3% bovine serum CAG ATT AA-30. siRNA was transfected by lipofectamine albumin (BSA) in TBS-T. Cells were incubated in rabbit anti- (Gibco BRL, Rockville, MD, USA). pericentrin (1:300), or mouse anti-a-tubulin (1:300) primary antibody diluted in 3% BSA blocking buffer at 41C overnight. Flow cytometry analysis After washing three times with TBS-T, cells on coverslips were DNA content analysis by flow cytometry has been described incubated with appropriate secondary antibodies conjugated previously (Meng et al., 2004a, b, c). Cells were fixed with 70% with fluorescein (FITC) or Texas Red (Jackson ImmunoR- ethanol overnight, washed with PBS and incubated with esearch, West Grove, PA, USA) for 1 h at room temperature in mouse anti-ph(Ser10)H3 antibody (1:200 dilution). After the dark, washed and mounted using mounting medium 0 washing, these cells were incubated with FITC-conjugated containing 4 ,6-diamidino-2-phenylindole (DAPI). The images anti-mouse immunoglobulin before analysed by flow cytome- were recorded using a Zeiss fluorescence microscopy with an try to score ph(Ser10)H3 positive cells. Axioskop digital camera. FISH analysis of chromosomes Knockdown of BCCIP expression by shRNA A chromosome 12-specific centromeric FISH probe (Vysis We used two vectors to express shRNA to knockdown BCCIP Inc., Downers Grove, IL, USA) was used to hybridize HT1080 expression. The plasmid pPUR/U6 uses a puromycin resis- cells according to the manufacturer’s protocol with high tance cassette, while the pSilencer2.1Hyg (Ambion Inc., stringent washing condition. Austin, TX, USA) uses hygromycin B as the selection marker. The control vectors express either a scrambled sequence 50- ACT ACC GTT GTT ATA GGT G-30 (Ambion Inc.) in Acknowledgements pSilencer2.1Hygomycin or a GFP cDNA target sequences 50-GGT TAT GTA CAG GAA CGC A-30 in pPUR/U6. This research was supported by National Institute of Health The isoform specific BCCIP silencing has been reported (Lu grants CA115488 and ES08353 and by the US Army Medical et al., 2005). Briefly, isoform-specific sequences targeting Research and Materiel Command grants DAMD17-02-1-0515 BCCIPa (50-GGG AAC CTT CAT GAC TGT TGG) or and DAMD17-03-1-0317. We thank the technical support BCCIPb (50-GGG AAG CAA ATG GTC TTT TGA) were from the flow cytometry and the fluorescence microscopy inserted into pPUR/U6. To knockdown both BCCIPa and facility of UNM Cancer Research and Treatment Center.

References

Andreassen PR, Lohez OD, Lacroix FB, Margolis RL. (2001). Daniels MJ, Wang Y, Lee M, Venkitaraman AR. (2004). Tetraploid state induces p53-dependent arrest of nontrans- Abnormal cytokinesis in cells deﬁcient in the breast cancer formed mammalian cells in G1. Mol Biol Cell 12: 1315–1328. susceptibility protein BRCA2. Science 306: 876–8769.

Oncogene BCCIP, chromosome instability and cytokinesis X Meng et al 6260 Fukasawa K, Choi T, Kuriyama R, Rulong S, Vande Woude Meraldi P, Honda R, Nigg EA. (2002). Aurora-A over- GF. (1996). Abnormal centrosome amplification in the expression reveals tetraploidization as a major route absence of p53. Science 271: 1744–1747. to centrosome amplification in p53-/- cells. Embo J 21: Geddis AE, Kaushansky K. (2004). Megakaryocytes express 483–492. functional Aurora-B kinase in endomitosis. Blood 104: Merlo A. (2003). Genes and pathways driving glioblastomas 1017–1024. in humans and murine disease models. Neurosurg Rev 26: Hans F, Dimitrov S. (2001). Histone H3 phosphorylation and 145–158. cell division. Oncogene 20: 3021–3027. Moynahan ME, Pierce AJ, Jasin M. (2001). BRCA2 is Jallepalli PV, Waizenegger IC, Bunz F, Langer S, Speicher required for homology-directed repair of chromosomal MR, Peters JM et al. (2001). Securin is required for breaks. Mol Cell 7: 263–272. chromosomal stability in human cells. Cell 105: 445–457. Ohgaki H, Dessen P, Jourde B, Horstmann S, Nishikawa T, Lengauer C, Kinzler KW, Vogelstein B. (1998). Genetic Di Patre PL et al. (2004). Genetic pathways to glioblastoma: instabilities in human cancers. Nature 396: 643–649. a population-based study. Cancer Res 64: 6892–6899. Liu J, Yuan Y, Huan J, Shen Z. (2001). Inhibition of breast Ono T, Kitaura H, Ugai H, Murata T, Yokoyama KK, and brain cancer cell growth by BCCIPalpha, an evolutio- Iguchi-Ariga SM et al. (2000). TOK-1, a novel p21Cip1- narily conserved nuclear protein that interacts with BRCA2. binding protein that cooperatively enhances p21-dependent Oncogene 20: 336–345. inhibitory activity toward CDK2 kinase. J Biol Chem 275: Lu H, Guo X, Meng X, Liu J, Allen C, Wray J et al. (2005). 31145–31154. The BRCA2-interacting protein BCCIP functions in RAD51 Powell SN, Kachnic LA. (2003). Roles of BRCA1 and BRCA2 and BRCA2 focus formation and homologous recombina- in homologous recombination, DNA replication fidelity and tional repair. Mol Cell Biol 25: 1949–1957. the cellular response to ionizing radiation. Oncogene 22: Maier D, Zhang Z, Taylor E, Hamou MF, Gratzl O, Van Meir 5784–5791. EG et al. (1998). Somatic deletion mapping on chromosome Rasheed BK, Fuller GN, Friedman AH, Bigner DD, Bigner 10 and sequence analysis of PTEN/MMAC1 point to the SH. (1992). Loss of heterozygosity for 10q loci in human 10q25-26 region as the primary target in low-grade and high- gliomas. Genes Chromosomes Cancer 5: 75–82. grade gliomas. Oncogene 16: 3331–3335. Roversi G, Pfundt R, Moroni RF, Magnani I, van Reijmersdal Margolis RL, Lohez OD, Andreassen PR. (2003). G1 S, Pollo B et al. (2005). Identification of novel genomic tetraploidy checkpoint and the suppression of tumorigenesis. markers related to progression to glioblastoma through J Cell Biochem 88: 673–683. genomic profiling of 25 primary glioma cell lines. Oncogene Meng X, Liu J, Shen Z. (2003). Genomic structure of the 25: 1571–1583. human BCCIP gene and its expression in cancer. Gene 302: Rudkin TM, Foulkes WD. (2005). BRCA2: breaks, mistakes 139–146. and failed separations. Trends Mol Med 11: 145–148. Meng X, Liu J, Shen Z. (2004a). Inhibition of G1 to S cell cycle Shinmura K, Bennett RA, Tarapore P, Fukasawa K. progression by BCCIP beta. Cell Cycle 3: 343–348. (2006). Direct evidence for the role of centrosomally Meng X, Lu H, Shen Z. (2004b). BCCIP functions through localized p53 in the regulation of centrosome duplication. p53 to regulate the expression of p21Waf1/Cip1. Cell Cycle Oncogene (E-pub 30 Oct 2006). 3: 1457–1462. Tarapore P, Fukasawa K. (2002). Loss of p53 and centrosome Meng X, Yuan Y, Maestas A, Shen Z. (2004c). Recovery from hyperamplification. Oncogene 21: 6234–6240. DNA damage-induced G2 arrest requires actin-binding Tarapore P, Horn HF, Tokuyama Y, Fukasawa K. protein filamin-A/actin-binding protein 280. J Biol Chem (2001). Direct regulation of the centrosome duplication 279: 6098–6105. cycle by the p53-p21Waf1/Cip1 pathway. Oncogene 20: Meng X, Yue J, Liu Z, Shen Z. (2007). Abrogation of the 3173–3184. transactivation activity of p53 by BCCIP down-regulation. Venkitaraman AR. (2002). Cancer susceptibility and the J Biol Chem 282: 1570–1576. functions of BRCA1 and BRCA2. Cell 108: 171–182.

Supplementary Information accompanies the paper on the Oncogene Web site (http://www.nature.com/onc).

Oncogene