Protein Kinase CK2 Is Involved in G2 Arrest and Apoptosis Following Spindle Damage in Epithelial Cells

Protein kinase CK2 is involved in G2 arrest and apoptosis following spindle damage in epithelial cells

Mohamed Sayed1,2, Steven Pelech2, Connie Wong3, Anthony Marotta1 and Baljinder Salh*,1

1The Department of Experimental Medicine, Jack Bell Research Center, University of British Columbia, 2660 Oak Street, Vancouver, British Columbia V6H 3Z6, Canada; 2The Department of Medicine, Koerner Pavilion, University of British Columbia, Vancouver, British Columbia V6T 1Z3, Canada; 3Memorial Sloan Kettering Cancer Center, 1275 York Avenue, New York, New York, NY 10021, USA p53 undergoes phosphorylation on several residues in Introduction response to cellular stresses that include UV and ionizing radiation, however the in¯uence of spindle damage on The tumor suppressor protein p53 possesses a number this parameter is relatively unclear. Consequently, the of anti-oncogenic properties, which include the induc- eect of nocodazole on serine 392 phosphorylation was tion of growth arrest and apoptosis, as well as its examined in two epithelial cell lines. We show that this ability to block transformation by oncogenes (Finlay et process is dependent upon the stepwise activation of p38 al., 1989; Baker et al., 1990; Yonish-Rouach et al., mitogen-activated protein kinase (p38 MAPK) and 1991; Dulic et al., 1994; Miyashita and Reed, 1995; and protein kinase casein kinase 2 (CK2). Furthermore, this reviewed in Levine, 1997). Additionally, p53 was activation correlated with the biochemical regulation of proposed to be a component of the spindle assembly the maturation-promoting factor (MPF, cdc2/cyclin B), checkpoint (Cross et al., 1995), however, subsequent as both DRB and antisense depletion of CK2, as well as work indicates a role in the post-mitotic checkpoint SB203580 were associated with an inhibition of its (Lanni and Jacks, 1998). Both pieces of evidence activation in response to nocodazole. Strikingly, when suggest an important role for p53 in maintenance of the cell cycle characteristics of nocodazole treated cells diploidy and DNA integrity. were examined, we observed that depletion or inhibition The structure of p53 comprises four domains of the catalytic subunit of CK2, in the presence of (reviewed in Ko and Prives, 1996). The transcriptional microtubule inhibitors, resulted in a compromise of the activation domain is situated between amino acids 1 ± G2 arrest (spindle checkpoint). Furthermore, CK2- 43. The sequence-speci®c DNA-binding domain depleted, nocodazole treated cells demonstrated a stretches from amino acid 100 to 300, and is the site dramatic reduction in the apoptotic cell fraction, most frequently mutated in human tumors. The con®rming that these cells had been endowed with oligomerization and regulatory domains are located oncogenic properties. These changes were observed in at the C-terminus between 300 ± 393. The mechanism both HeLa cells and HCT116 cells. We also show that whereby cellular stress is communicated to p53, so that this eect is dependent on the presence of functional it is functionally activated remains incompletely under- wild-type p53, as this phenomenon is not apparent in stood, but likely involves p53 modi®cation by processes HCT116 p537/7 cells. Collectively, our results indicate such as phosphorylation and acetylation, which two novel roles for CK2 in the spindle checkpoint arrest, stabilize and activate the molecule (Hupp et al., 1992; in concert with p53. Firstly, to maintain increased Gu and Roeder, 1997; Sakaguchi et al., 1998). A range cyclinB/cdc2 kinase activity, as a component of G2 of dierent protein kinases, that includes c-Jun N- arrest, and secondly, a role in p53-mediated apoptosis. terminal kinase (JNK), p38 MAPK, protein kinase C These ®ndings may have implications for an improved (PKC), casein kinase 1 (CK1), ATM kinase and DNA- understanding of abnormalities of the spindle checkpoint dependent protein kinase (DNA-PK), may mediate p53 in human cancers, which is a prerequisite for de®ning phosphorylation (reviewed in Prives and Hall, 1999). future therapies. Oncogene (2001) 20, 6994 ± 7005. The exact roles of some of these molecules remain unclear. DNA-PK phosphorylates residues 15 and 37 Keywords: CK2; p53; spindle checkpoint; apoptosis of p53, and this appears to aect transactivation and growth arrest in some cells (Lees-Miller et al., 1992; Fiscella et al., 1993). However, these processes occur normally in cells lacking DNA-PK also, indicating the presence of possible compensatory mechanisms (Huang et al., 1996). Another important rami®cation of serine *Correspondence: B Salh, Department of Medicine, University of 15 phosphorylation by DNA-PK is the reduced British Columbia, 100-2647 Willow Street, Vancouver, BC, Canada, interaction of p53 with its negative regulator MDM2 V5Z 3P1; E-mail: [email protected] Received 9 May 2001; revised 17 July 2001; accepted 2 August (Haupt et al., 1997; Kubbutat et al., 1997; Shieh et al., 2001 1997). The same net eect occurs in response to Involvement of CK2 in spindle damage M Sayed et al 6995 MDM2 phosphorylation by DNA-PK and ATM the cell cycle, by the addition of serum-containing (Mayo et al., 1997; Khosravi et al., 1999). Notably, medium. Nocodazole, a drug that disrupts the mitotic spindle disruption also eects increased serine 15 spindle by inhibiting microtubule polymerization, was phosphorylation but this appeared to be dependent added before cells entered the mitotic (M) phase. After on the agent used, and occurred in epithelial cells, but 9 ± 10 h, microscopic examination of transfected cells not in primary cultures of human ®broblasts (Stewart with either S oligonucleotide or buer alone revealed et al., 2001). that many of them became rounded, indicative of cells Protein kinase CK2 (formerly known as casein arrested in mitosis (Figure 1a). In contrast, cells kinase II) was ®rst shown to phosphorylate murine transfected with AS showed fewer rounded cells. We p53 at serine 389 (Meek et al., 1990; Herrmann et al., then measured the percentage of cells in mitosis 1991). Phosphorylation by CK2 has been shown to (mitotic index, MI) after nocodazole treatment. The enhance p53 sequence-speci®c DNA binding in vitro average MI of cells with either S-treated or buer was (Hupp et al., 1992), which in turn contributes to its 63% (1008 of 1600 cells), whereas the MI of cells transactivation function in vitro (Mundt et al., 1997). transfected with AS was 11% (176 of 1600) (Figure In vivo, the phosphorylation of p53 protein at the CK2 1b). These results indicated that CK2 depletion was site has been reported to regulate p53-speci®c tran- responsible for a reduction in the mitotic cell fraction. scriptional repression (Hall et al., 1996). More recent However they did not establish whether this occurred evidence indicates that rodent cells exposed to UV before or after cells had arrested in G2/M. Similar radiation but not ionizing radiation or etoposide also results were observed using the CK2 speci®c inhibitor exhibit increased serine 389 phosphorylation, and that DRB that was added 2 h prior to nocodazole exposure CK2 could phosphorylate the serine 389 site in vitro (data not shown). (Kapoor and Lozano, 1998; Lu et al., 1998). Further work using a dominant negative mutant of p38 MAPK CK2 functions as a component of the spindle assembly and the speci®c p38 MAPK inhibitor, SB202190, checkpoint indicates that this site is phosphorylated by p38 MAPK (Huang et al., 1999). In contrast to genotoxic stress, it The spindle assembly checkpoint is the mechanism that is currently unknown whether the CK2 site is involved prevents the onset of anaphase in cells that have in the response to spindle damage. defects in spindle assembly or chromosome alignment Protein kinase CK2 is a ubiquitous, second on the spindle (Li and Nicklas, 1995). Recently, it has messenger-independent protein serine/threonine kinase been demonstrated that inhibition of p38 MAPK which is composed of two catalytic a-ora'-subunits activity, in response to spindle disruption, compromises and two regulatory b-subunits (Allende and Allende, the spindle assembly checkpoint (Takenaka et al., 1995; Pinna and Meggio, 1997). Although its precise 1998). Based upon our earlier observations that p38 biochemical regulation remains unknown, we have MAPK and CK2 interact directly, we therefore tested demonstrated that CK2 may be activated by p38 whether CK2 might also participate in the spindle MAPK in response to stress, through a direct protein ± assembly checkpoint of the somatic cell cycle. protein interaction (Sayed et al., 2000). Here we report Transfected HeLa cells with S, AS, or left untreated that CK2 is involved in two novel functions in (control) were synchronized at G1/S phase boundary response to spindle damage, both in maintaining G2 by a double thymidine block and released into the cell arrest and in the subsequent apoptosis process. cycle. Cells were then either treated with nocodazole for 10 h or released without treatment for 9 h during which the mitotic cells were collected by mechanical shake o. Dramatic increases of maturation-promoting Results factor (MPF) and p34cdc2 activities were observed in nocodazole-treated cells, indicative of mitotic arrest Depletion or inhibition of CK2 compromises mitotic arrest (Figure 2a,b). These activities were substantially To investigate whether CK2 is required for mitotic suppressed in AS-transfected cells after nocodazole checkpoint function, we employed antisense to the exposure. Thus, the activation of a key component of alpha subunit to deplete the protein kinase. We the mitotic spindle assembly checkpoint induced by the transfected HeLa cells with either CK2 (a/a' subunits) microtubule poison nocodazole was compromised by antisense (AS) or sense (S) oligodeoxyribonucleotide. depletion of CK2. Similar results were obtained after The expected depletion of CK2 upon AS oligonucleo- nocodazole exposure in cells pre-treated with the p38 tide treatment was con®rmed by assaying the intracel- MAPK speci®c inhibitor SB203580. lular protein level of CK2 a/a' subunits by We next examined whether CK2 was activated in immunoblotting (Figure 1c). Previous studies, using nocodazole-treated cells or in normal M phase. Immune the same antisense oligonucleotide, have demonstrated complex kinase assay for CK2 revealed that the activity not only complete depletion of the catalytic CK2 a/a' of CK2 from nocodazole-arrested mitotic cell lysates was subunits (Ulloa et al., 1993), but also inhibition of four times greater than that from interphase cells sorting of the regulatory b subunit, preventing its (control). Little to no activity of CK2 was observed in translocation to the nuclei (Ulloa et al., 1994). Cells mitotic cells (shake-o) and the activity of CK2 by were then synchronized at G0 phase and released into nocodazole was attenuated by inhibition of p38 MAPK

Oncogene Involvement of CK2 in spindle damage M Sayed et al 6996

Figure 1 Function of CK2 as a mitotic checkpoint protein. (a) HeLa cells, treated with CK2 sense, antisense or left untreated, were deprived of serum and released into the cell cycle by addition of serum. After 18 h, cells were either not treated (control) or exposed to nocodazole (400 ng/ml) for 12 h before being photographed. (b) Failure of HeLa cells transfected with antisense to undergo mitotic arrest. The shaded bars and error bars represent the means and standard deviations, respectively, determined from at least three independent assessments. (c) Depletion of the endogenous CK2a in HeLa cells transfected with CK2a-a' antisense. Equal amounts of protein extracts were subjected to protein immunoblot analysis with monoclonal anti-CK2a antibody. Lower

using SB203580 (Figure 2c). Collectively, with the Nocodazole-induced p53 serine 392 phosphorylation is previous studies (Sayed et al., 2000; Takenaka et al., mediated by CK2 and is SB203580 sensitive 1998), we conclude that nocodazole treatment appears to activate p38 MAPK in association with CK2 by In our previous study we showed that the phosphor- disrupting spindle formation. ylation of p53 at serine 392 induced by stress, was

Oncogene Involvement of CK2 in spindle damage M Sayed et al 6997

Figure 2 (a) Suppression of nocodazole-induced MPF activity by either CK2 antisense or a speci®c inhibitor of p38. HeLa cells were synchronized at G1-S by double thymidine block and either collected (control) or released into cell cycle. After release, the cells were incubated with or without nocodazole (400 ng/ml). In some experiments SB203580 (20 mM) or DRB (20 mM) was added 2 h prior to nocodazole treatment. Cell lysates were then prepared and histone H1 phosphotransferase activities were measured. (b) cdc2 activity is downstream to CK2 and p38 MAPK. Proteins were immunoprecipitated with anti-cdc2 antibody and its phosphotransferase activity was assayed toward histone H1. (c) Activation of CK2 in nocodazole-arrested cells is dependent on p38 MAP kinase. CK2 protein in cell lysates, prepared from the above treatment, was immunoprecipitated with monoclonal anti- CK2a antibody and assayed for phosphotransferase activity towards the CK2 speci®c substrate RRADDSDDDDD. The lower panel shows equal amounts of CK2 protein were immunoprecipitated. (d) CK2 is the major kinase phosphorylating p53 at serine 392, Cell lysates from the above treatment were immunoblotted with phospho antibodies to p53 (serine 392) and p38a.(e) p38 protein was immunoprecipitated with anti-p38a antibody and assayed for phosphotransferase activity toward activating transcription factor 2 (ATF2). The lower panel shows equal amount of p38 protein was immunoprecipitated. All of these

mediated by CK2 as a result of direct interaction with shows that the hyperphosphorylation of p53 at serine p38 MAPK (Sayed et al., 2000). We reexamined this 392 due to nocodazole, was completely suppressed in ®nding in the context of spindle disruption. Figure 2d cells either transfected with AS or treated with p38

Oncogene Involvement of CK2 in spindle damage M Sayed et al 6998 speci®c inhibitor SB 203580. (Whereas a suppression of cells treated with nocodazole appeared to exhibit an p53 phosphorylation was observed using DRB, no override of the mitotic assembly checkpoint and an change was found with CK2 sense oligonucleotide, apparent reentry into S phase. This was also manifested data not shown). This eect was independent of by an alteration of cyclin B expression (Figure 3c) and changes in p53 protein expression (lower panel). The cdc2 inactivation (Figure 3b). depletion of CK2 did not aect the phosphorylation or To further investigate if the expression of cyclin B the activation of p38 upon nocodazole treatment was correlated with phosphorylation-dependent inacti- (Figure 2d,e), con®rming its role downstream of p38 vation of cdc2, the electrophoretic mobility of cdc2 was MAPK activation. Thus, these results indicate that used to assess whether it was phosphorylated on CK2 phosphorylates p53 at serine 392, due to the threonine 14, tyrosine 15 or both. Figure 3d shows nocodazole-induced activation of the spindle assembly the phosphorylation stages of cdc2 upon nocodazole checkpoint, and p38 MAPK is its upstream regulator. stimulation in the presence and absence of CK2 It should be noted that HeLa cells are known to antisense. The most gel-retarded form of cdc2, species harbor human papilloma virus (HPV), which has been `a', is phosphorylated on both threonine 14 and shown to target p53 for degradation (Schener et al., tyrosine 15 whereas the intermediate form, species `b', 1993). However, it has been reported that cervical is phosphorylated on threonine 14 or tyrosine 15, but carcinoma cells, including HeLa, express functional not both. The fastest migrating form of cdc2 (species wild-type p53, which is sucient to activate the `c') is not phosphorylated on either threonine 14 or transcription of cellular target genes (Butz et al., tyrosine 15, and represents either the active form of the 1995, 1999). The level of p53 expression observed in kinase (phosphorylated on threonine 161 and bound to our studies is consistent with this (Figure 2d). cyclin B), or cdc2 that is not bound to cyclin Furthermore, we have measured the expression of (monomeric, inactive cdc2) (Atherton-Fessler et al., p53-stimulated p21WAF1 protein in HeLa cells treated 1994; Liu et al., 1997). These results were con®rmed by with dierent genotoxic stresses to con®rm that it is using the phospho-speci®c anti-cdc2 tyrosine 15 anti- indeed functional (results not shown). body (results not shown). Figure 3d demonstrates that the phosphorylation state of cdc2 remains unchanged in AS-transfected cells, whereas cyclin B turnover is Depletion of CK2 compromises mitotic arrest following altered (Figure 3c). Collectively, these results indicate spindle damage that the inhibition of cdc2 activation in CK2-depleted Having demonstrated an eect of CK2 depletion on cells upon nocodazole exposure, was not correlated mitotic arrest (Figure 1), cyclin B/cdc2 kinase activity with changes in the phosphorylation of cdc2, but was and p53 phosphorylation (Figure 2), we wanted to likely due to the altered cyclin B kinetics. Similar determine the cell cycle characteristics of these cells in results were also obtained by using the CK2 inhibitor, more detail. Therefore, exponentially growing HeLa DRB (data not shown). cells were transfected with CK2 AS, S or buer alone. Cells were then synchronized in G1 phase by serum Mitotic spindle checkpoint compromise requires p53 deprivation (0.5%) for 30 h and then released into the protein cell cycle by the addition of 10% serum containing medium. Nocodazole was added before the cells entered p53 has been shown to be involved in G2/M phase to mitosis for the indicated times shown (Figure 3a). The maintain cell diploidy (reviewed in Ko and Prives, FACS analysis data for synchronized cells showed DNA 1996). Several studies have shown that p537/7 cells content of 75% of cells with 2N and 9% accumulated accumulate 4N, 8N and 16N populations upon 4N populations (data not shown). After 6 h following exposure to microtubule inhibitors for 448 h (Cross release into the cell cycle, cells progressed through S et al., 1995; Di Leonardo et al., 1997; Lanni and Jacks, phase and entered G2 and mitosis with similar kinetics 1998). From our ®ndings thus far, we were unable to (Figure 3a). However, from 12 h, the behavior of these directly implicate p53 in the altered spindle checkpoint populations was dramatically dierent. AS-treated cells arrest observed following CK2 depletion. Therefore, we (control), S- or buer-treated cells continued to investigated this hypothesis in the colorectal MIN accumulate in mitosis for the next 12 h with DNA (microsatellite instability) HCT116 cancer cell line, content of 4N. The time-dependent increase of cdc2 because it has an intact spindle-dependent checkpoint activity (Figure 3b), as well as corresponding cyclin B (Bunz et al., 1998). As expected, HCT116 cells with expression (Figure 3c), and mitotic index (Figure 4b) by either wild-type (wt) p53+/+ or with disrupted p537/7 18 h, all con®rm that these cells remained blocked in alleles, were shown to arrest in M phase upon mitosis upon exposure to nocodazole. Signi®cantly, cells nocodazole exposure (Figure 4a). However, p53+/+ devoid of CK2 appeared to progress into the cell cycle, cells transfected with CK2 AS, in common with HeLa compromising the mitotic block (Figure 3a), with a large cells similarly treated, also appeared to show a population of cells progressing into a G1-like state with dramatically compromised mitotic spindle checkpoint DNA content of 2N populations by 24 h. The mitotic with 32% of the cells in G1/G0 compared with only index (Figure 4b) of the cycling cells indicates that 3% for p537/7 cells, after 24 h. Furthermore approxi- 530% of cells entered mitosis by 24 h, despite mately 30% of the cells had reached mitosis by 24 h. In disruption of mitotic assembly. Thus, CK2-depleted striking contrast, p537/7 cells transfected with CK2 AS

Oncogene Involvement of CK2 in spindle damage M Sayed et al 6999

Figure 3 Mitotic arrest is compromised in the absence of CK2. (a) Synchronized HeLa cells (in G0), treated with CK2 antisense or sense, were released into cell cycle and collected for ¯ow cytometric analysis in the presence and absence of nocodazole (200 ng/ml) at the indicated time. Lysates from the above treated cells were immunoprecipitated with anti-cdc2 antibody and assayed for cdc2 phosphotransferase activity (b) or either subjected to immunoblotting with anti-cyclin B (c) or cdc2 (d) protein. This experiment was maintained a high mitotic index comparable with responsible for the failure to transgress the spindle nocodazole treated p537/7 cells (Figure 4b), accumulat- checkpoint. The p53+/+ cells exhibited a down-regulation ing 4N with diminished 2N populations (Figure 4a). of cyclin B expression in response to antisense CK2 Thus, it appeared that the normal checkpoint status of depletion in keeping with the HeLa cell data. MIN cells following CK2 depletion and spindle Another important ®nding to emerge from investiga- damage, is converted to that characteristic of CIN tion of the HCT116 cells was that antisense depletion (chromosomal instability) cells (Cahill et al., 1998). of CK2 did not in¯uence the ability of the cells to We evaluated whether cyclin B expression was aected reach the mitotic checkpoint. This is exempli®ed by the in p537/7 cells. Figure 4c shows that cyclin B was not fact that both at 18 h and at 24 h, there are equivalent aected by depletion of CK2. These ®ndings argue fractions of cells at the G2/M stage for both against an impairment of cyclin B expression, as being nocodazole, and nocodazole+antisense treated p537/7

Oncogene Involvement of CK2 in spindle damage M Sayed et al 7000

Oncogene Involvement of CK2 in spindle damage M Sayed et al 7001 cells (60 and 61% at 18 h, and 67 and 65% at 24 h expected, these cells exhibited a much lower frequency respectively). Moreover, this reduces the possibility of apoptosis in response to nocodazole in comparison that AS treated HeLa cells underwent a G1 or S phase to p53+/+ cells (10 vs 32.2%). In further support of co- delay. Furthermore, our data indicated that neither the operativity between CK2 and p53, it is apparent from activity nor the protein levels of CK2 were aected in this data that p537/7 cells do not exhibit a reduction of HCT116 p537/7 cells (results not shown). Therefore, apoptotic cell fraction upon CK2 depletion. these studies seem to indicate that abrogation of the mitotic assembly checkpoint and cyclin B degradation are dependent on p53 protein, and in addition, are Discussion regulated by CK2. In this study we have demonstrated for the ®rst time, a functional co-operation between protein kinase CK2 Cells transfected with CK2 antisense escaped and p53, in modulation of cdc2 kinase activity and the nocodazole-induced cell death mitotic checkpoint arrest following nocodazole-induced The tumor suppressor p53 has been shown to mediate spindle damage. This eect was in¯uenced by the p38 apoptosis in several cell types. Loss of p53 prevents MAPK, as both the biochemical events and the cell- thymocytes from undergoing apoptosis after gamma cycle changes could be reversed using the p38 MAPK irradiation, and decreases the ecacy of chemotherapy speci®c inhibitor, SB203580. Nocodazole-induced CK2 in vitro, and in vivo (Clarke et al., 1993; Lowe et al., activation correlated with increased levels of cyclin B 1993; reviewed in Ko and Prives, 1996). To examine expression and cell cycle arrest, whereas the compro- cell death in cells transfected with CK2 AS, we mised checkpoint was associated with a reduction in analysed and quanti®ed the phenotypic changes levels of cyclin B and consequently a reduction in cdc2 characteristic of apoptotic cells by double staining activation. The latter event was shown to be dependent HeLa cells with annexin V and propidium iodide (PI). upon p53. The reduction in apoptosis exhibited by CK2 Exponentially growing control or AS-transfected cells depleted cells, accompanied by the signi®cantly reduced were exposed to nocodazole, and then harvested every mitotic index provides the strongest support for the 6 h. Figure 5a shows that 22% of control cells had notion that the mitotic checkpoint arrest was breached. undergone apoptosis (lower right panel) after 24 h of The identical cell cycle distribution pattern observed for exposure with nocodazole. These observations are both nocodazole-treated and AS CK2-depleted p537/7 consistent with previous studies in certain cell lines, cells (Figure 4) provides the most compelling evidence which undergo apoptosis following persistent exposure against an alternative cell cycle modulating role for to nocodazole (Minn et al., 1996; Taylor and McKeon, depletion of CK2 in mediating the above responses. 1997). In contrast, cells transfected with CK2 AS The C-terminal domain of p53 is a target for CK2 showed a signi®cant, reproducible 50% reduction in phosphorylation. It has been shown that CK2 and p53 apoptosis. An even greater reduction in apoptosis associate in vivo via the binding of the regulatory b (72%) consequent upon CK2 depletion was observed subunit of CK2 to the C-terminal domain of p53 using HCT116 p53+/+ cells (Figure 5b). However, (Appel et al., 1995; Gotz et al., 1999). The eect of this HCT116 p537/7 cells transfected with CK2 AS and phosphorylation upon the growth suppressing and exposed to nocodazole, demonstrated no change in the transactivating activities of p53 is controversial (Hupp percentage of apoptotic cell fractions (Figure 5c). As et al., 1992; Mundt et al., 1997; Rolley and Milner,

Figure 4 p53 is required for CK2 to override the mitotic arrest. (a) Flow cytometric analysis of exponentially growing HCT116 (p53+/+) and HCT116 (p537/7) cell lines transfected with CK2 antisense or left untreated (control) in the presence and absence of nocodazole (200 ng/ml) for the indicated times (cells were synchronized as for Figure 3). (b) Mitotic indices of HCT116 and HeLa cells are shown at the indicated times after exposure to nocodazole. (c) HCT116 (p537/7) and p53+/+ (d) cell lysates were immunoblotted with anti-cyclin B. A control immunoblot using a-GAPDH antibody is shown for the p53+/+ cells. Similar results

Oncogene Involvement of CK2 in spindle damage M Sayed et al 7002

Figure 5 Suppression of apoptosis in the absence of CK2. Quantitative analysis of the apoptotic cells using annexin-V FITC in exponentially growing HeLa cells (a), HCT116 p53+/+ (b), or HCT116 p537/7 (c), transfected with CK2 antisense or left untreated (control) in the presence and absence of nocodazole (100 ng/ml) for 24 h. The ¯ow cytometric analysis was performed on 100 000 cells and the percentage of apoptotic (A), live (L) and dead (D) cells was measured. The left lower quadrant represents live cells, the right lower quadrant apoptotic cells, and the top right quadrant dead cells. This experiment was repeated with similar results on at least three occasions

Oncogene Involvement of CK2 in spindle damage M Sayed et al 7003 1994). Most of the studies support a growth suppres- essential element of the mitotic checkpoint which sing eect that is dependent on DNA binding prohibits re-duplication of DNA (Notterman et al., (reviewed in Prives and Hall, 1999). An intriguing 1998). It has been reported that p53 possesses checkpoint study by Hao et al. (1996) observed that whilst wild- regulatory activity following nocodazole treatment type p53 is inactive in G1-arrested cells, it became (Lanni and Jacks, 1998), speci®cally, to prevent S phase activated by mutation of the murine CK2 site to re-entry following adaptation from mitotic arrest by glutamic acid. Mutation of the same site to alanine led inducing p53-dependent arrest in 4N adapted cells. to a restricted ability of p53 to repress the SV40 Signi®cantly, our observations in epithelial cells of two promotor, whilst leaving unaected the p21 (and RGC) dierent origins, extend these ®ndings by implicating for promotors. In our study, using two dierent tumor cell the ®rst time, CK2 in this response. We therefore lines, HeLa and colon cancer HCT116 cells, it was propose that the depletion of CK2, which correlated shown that p53's function during M phase, was with suppression of p53 protein phosphorylation at dependent on the direct eect of CK2 on p53, serine 392, may confer upon p53 the ability to following spindle damage. Previously, p53 has been compromise the spindle checkpoint arrest, as a result demonstrated to prohibit re-duplication of DNA of facilitating cyclin B degradation. Previous work has (Notterman et al., 1998) without completion of mitosis. shown that mutation of p53 protein at the CK2 Collectively, taken in context with previous observa- phosphorylation site (serine 392) inhibited the ability of tions, our results suggest a novel role of CK2 in p53 to cause growth repression but had no eect on its concert with p53 protein to prohibit further DNA re- transactivation activity (Hall et al., 1996). It has also duplication in response to spindle damage. been suggested that p53 can exist in a con®guration The question remained as to whether the compromised which might be inhibited for DNA binding but might mitotic arrest, observed in the absence of CK2, might be remain active for other functions (reviewed in Ko and due to p53-mediated, abrogated expression of cyclin B. Prives, 1996). This is further supported by the well- p53 regulates cell cycle checkpoints in response to accepted role for p53 transrepression in the induction of genotoxic stress by targeting the transcription of speci®c apoptosis (reviewed in Vogt Sionov and Haupt, 1999). gene that inhibit cyclin, cyclin-dependent kinase com- Thus, our results support a model in which p53 may lose plexes. For example, at the G2/M checkpoint, there are the ability to maintain mitotic integrity, in the absence or several such genes, which include 14-3-3s (Hermeking et dysfunction of CK2, by a loss of its ability to al., 1997; Chan et al., 1999), GADD45 (Zhan et al., 1999) appropriately regulate cyclin B levels, as well as an and B99 (Utrera et al., 1998). In the example of 14-3-3s, inability to execute apoptosis. G2 arrest occurs through the binding of cdc25C Protein kinase CK2 activity is elevated in a multi- phosphatase and the resulting inhibition of activation tude of human cancer cells including leukemias and of cdc2. This explanation is clearly inadequate, as cells solid tumors (Pinna and Meggio, 1997; Xu et al., de®cient in p53 are equally able to enter into G2 arrest 1999). Although the precise signi®cance of this remains following DNA damage. Additionally, it has also been unknown, Landesman-Bollag et al. (1998) have shown that p53 can directly regulate cyclin B expression demonstrated that misexpression of CK2a contributes by lowering the intracellular levels of cyclin B protein to the loss of p53 alleles. We hypothesize that (Innocente et al., 1999). Therefore, we determined cdc2/ misexpression of CK2a, a common observation in cyclin B activity in a cell line de®cient for p53 (p537/7). most tumors (Allende and Allende, 1995) may have a Indeed, we found that neither phosphorylation of cdc2 role in compromising genomic stability, by impairing (data not shown) nor degradation of cyclin B (Figure 4c) the appropriate response to spindle damage. were aected in the HCT116 p537/7 cell line (in contrast In conclusion, we propose a mechanistic signaling to p53+/+ cells), in the absence or inhibition of CK2, pathway that may regulate p53 in response to spindle during persistent activation of the mitotic spindle damage. Upon microtubule disruption, p38 MAPK is checkpoint by nocodazole. This supports our hypothesis, activated and associates with CK2, regulating CK2's that p53 functions to repress cyclin B expression. Similar activity. In addition, it also modulates the phosphor- ®ndings have been reported in normal human foreskin ylation of p53, modifying p53's function to ensure ®broblasts (HFFs), where p53 depletion (by transduc- maintenance of genomic mitotic arrest. This is shown tion with human papillomavirus type E6) was associated to be due at least in part through regulation of cyclin with prolonged G2 arrest, in association with high levels B/cdc2 kinase activity levels and subsequent induction of cyclin B/cdc2 kinase activity (Passalaris et al., 1999). of apoptosis. Furthermore, HeLa cells expressing a constitutively active cdc2 mutant (CDC2AF), and also exhibiting Materials and methods increased levels of cyclin B/cdc2 kinase activity, were also able to arrest in G2 (Jin et al., 1996). In direct agreement Cell culture and synchronization with these observations, our cell cycle analysis data HeLa and HCT116 (p53+/+ and p537/7, kindly donated by showed that cells lacking functional p53 underwent Dr B Vogelstein) were grown for 3 days until near con¯uency mitotic delay for up to 24 h (Figure 4). Although it has 2 at 378C and 5% CO2 on 175 cm ¯asks in 10% fetal calf serum been demonstrated that p53 does not aect the initiation, (FCS) supplemented DMEM and McCoy's medium (GIBCO) duration or magnitude of the mitotic delay following respectively. Cells were split at 56105 cells /10 cm2 dish 24 h disruption of the mitotic spindle, in ®broblasts, it is an prior treatment. For cell synchronization at G1/S phase, HeLa

Oncogene Involvement of CK2 in spindle damage M Sayed et al 7004 cells were treated with double thymidine block (two 16 h loaded on sodium dodecyl sulfate (SDS) 712%, 37.5 : 1 periods of exposure to 2 mM thymidine separated by a 10 h polyacrylamide gel; 15%, 150 : 1 polyacrylamide gel was used exposure without thymidine). The cells were then rinsed twice for gel shift assay. Western blot analyses were performed with with phosphate buer saline (PBS) and cultured in complete mouse monoclonal; anti-human cyclin B1 (Pharmingen), anti- growth medium. Nocodazole (100 ± 400 ng/ml) or was added p53 (StressGen), phospho-p53 ser-392, phospho-cdc2 (New to either exponentially growing cells for the indicated times or England Biolabs), and CK2a (kindly donated by Dr Olaf- to G1/S synchronized cells for 10 h. In some experiments Georg Issinger). Rabbit polyclonal anti-p34 cdc2 (Kinetek HeLa cells were cultured in DMEM containing FCS (0.5%) Pharmaceuticals), phospho-p38a, (New England Biolabs) and for 30 h to synchronize cells in G0 phase, fresh DMEM p38a (Santa-Cruz) antibodies, were also used. supplemented with 10% FCS was then added to induce the cells to re-enter the cell cycle. After 18 h, nocodazole (400 ng/ Protein kinase assays ml) was added before cells entered M phase, as described (Takenaka et al., 1998). Mitotic cells were collected by Immunoprecipitations were performed by incubating 500 mg mechanical shake o, 9 h after release from thymidine block of whole lysate with 4 mg with anti-CK2a, or cdc2, or p38 (G1/S) or 22 ± 26 h from G0 release (Takenaka et al., 1998). In MAPK and 20 ml of protein A-Sepharose (preblocked with some experiments, the potent and speci®c CK2 or p38 MAPK 1% BSA) for 3 ± 4 h at 48C. The beads were washed four inhibitors; 5,6-dichloro-1-b-D-ribofuranosylbenzimidazole times with 20 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% NP- (DRB) (20 mM) and SB203580 (20 mM) (Calbiochem) respec- 40 and the above protease and phosphatase inhibitors. For tively, were used 2 h prior to the addition of nocodazole. immune-complex kinase assays, the immunoprecipitated beads were washed one additional time with buer A (12 m MOPS (pH 7.2) and 15 mM MgCl ), and incubated Oligonucleotide transfection M 2 with either 5 mg of the speci®c CK2 substrate Cells were grown in medium containing FCS (0.5%) for 24 h RRADDSDDDDD (Calbiochem) or with H1-histone (for prior to transfection, washed twice with PBS and supple- cdc2), or ATF2 (for p38 MAPK), in a ®nal volume of 25 ml mented with fresh medium containing 10% FCS. CK2a-a' with 5 or 10 mCi of 100 mM [g-32P]ATP in buer A, for antisense or sense oligonucleotide (2 mg/ml) was transfected 10 min at 308C. Reactions were stopped by either spotting into cells using Eectene (Qiagen) according to the 20 ml onto 1.5 cm2 piece of Whatman P-81 phosphocellulose manufacturer's protocol. CK2a-a' antisense and sense were paper and the incorporated radioactivity was measured in designed as previously described (Ulloa et al., 1993), with a Wallec (LKB) scintillation counter, or by the addition of modi®cation in which the entire oligonucleotides were SDS ± PAGE sample buer. After boiling, the samples were phosphorothioated. Sequence: antisense (5' to 3'): GTA subjected to SDS ± PAGE for autoradiography with X-ray ATC ATC TTG ATT ACC CCA or sense (5' to 3'): TGG ®lm or by immunoblotting analysis. GGT AAT CAA GAT GAT TAC. Cells were incubated for 18 ± 24 h prior to any treatment. Mitotic index Cells were collected by incubation with trypsin containing Flow cytometry EDTA, centrifuged, counted for normalization and then ®xed Cells were trypsinized, washed with PBS and ®xed with 70% in a solution containing 3.7% formaldehyde, 0.5% NP-40 ethanol at the indicated time points then stored at 7208C. and Hoescht 33258 (10 mg/ml) in PBS. Nuclei were visualized The ®xed cells were then stained with propidium iodide by ¯uorescent microscopy. Nuclei with condensed, evenly (50 mg/ml) with RNase (10 mg/ml). The stained cells were stained chromosomes were scored as mitotic. At least 600 analysed for DNA content by ¯uorescence-activated cell cells were counted for each determination. sorting (FACS) in a FACScan (Becton Dickinson Instru- ments). The forward and side scatter gates were set to exclude Apoptosis assay any dead cells from the analysis; 10 000 events within this gate were acquired per sample. Cells were collected at the indicated time points by incubation with trypsin-EDTA, centrifuged and washed twice with ice-cold PBS, then stained with Annexin-V FITC and Western blotting propidium iodide (PI) according to the manufacturer's Adherent cells were washed twice with ice-cold PBS, scraped, protocol (Pharmingen). Cells were analysed using the then lysed in 50 mM HEPES (pH 7.2), 100 mM NaCl, 1 mM CellQuest software on a Becton Dickinson FACScan and EGTA, 20 mM NaF, 1% Non-idet P-40, 1 mM sodium 10 000 events were acquired per sample. orthovanadate, 1% aprotinin, 1 mM phenylmethylsulfonyl- ¯uoride, 1 mM pepstatin, 0.5 mg/ml leupeptin and 10 mg/ml soybean trypsin inhibitor. Lysates were subjected to brief Acknowledgments sonication for 2610 s, centrifuged at 15 000 r.p.m. for 5 min, This work was supported by grants from the Cancer and the supernatants were collected and stored frozen prior Research Society, and the Northwestern Society for Intest- to use. The protein concentration was determined by the inal Research to B Salh, and the National Cancer Institute of method of Bradford using bovine serum albumin as the Canada to S Pelech. M Sayed is a recipient of a Medical standard. Between 20 ± 50 mg of total cell lysate protein was Research Council of Canada Doctoral Research Award.

References

Allende JE and Allende CC. (1995). FASEB J., 9, 313 ± 323. Atherton-Fessler S, Liu F, Gabrielli B, Lee MS, Peng CY Appel K, Wagner P, Boldyre B, Issinger OG and and Piwnica-Worms H. (1994). Mol. Biol. Cell, 5, 989 ± Montenarh M. (1995). Oncogene, 11, 1971 ± 1978. 1001.

Oncogene Involvement of CK2 in spindle damage M Sayed et al 7005 Baker SJ, Markowitz S, Fearon ER, Willson JK and Landesman-Bollag E, Channavajhala PL, Cardi RD and Vogelstein B. (1990). Science, 249, 912 ± 915. Seldin DC. (1998). Oncogene, 16, 2965 ± 2974. Bunz F, Dutriaux A, Lengauer C, Waldman T, Zhou S, Lanni JS and Jacks T. (1998). Mol. Cell. Biol., 18, 1055 ± Brown JP, Sedivy JM, Kinzler KW and Vogelstein B. 1064. (1998). Science, 282, 1497 ± 1501. Lees-Miller SP, Sakaguchi K, Ullrich SJ, Appella E and Butz K, Shahabeddin L, Geinsen C, Spitkovsky D, Ullmann Anderson CW. (1992). Mol. Cell. Biol., 12, 5041 ± 5049. A and Hoppe-Seyler F. (1995). Oncogene, 18, 2381 ± 2386. Levine AJ. (1997). Cell, 88, 323 ± 331. Butz KN, Whitaker N, Denk C, Ullmann A, Geisen C and Li X and Nicklas RB. (1995). Nature, 373, 630 ± 632. Hoppe-Seyler F. (1999). Oncogene, 18, 2381 ± 2386. Liu F, Stanton JJ, Wu Z and Piwnica-Worms H. (1997). Mol. Cahill DP, Lengauer C, Yu J, Riggins GJ, Willson JK, Cell. Biol., 17, 571 ± 583. Markowitz SD, Kinzler KW and Vogelstein B. (1998). Lowe SW, Ruley HE, Jacks T and Housman DE. (1993). Nature, 392, 300 ± 330. Cell, 74, 957 ± 967. Chan TA, Hermeking H, Lengauer C, Kinzler KW and Lu H, Taya Y, Ikeda M and Levine AJ. (1998). Proc. Natl. Vogelstein B. (1999). Nature, 401, 616 ± 620. Acad.Sci.USA,95, 6399 ± 6402. Clarke AR, Purdie CA, Harrison DJ, Morris RG., Bird CC, Mayo LD, Turchi JJ and Berberich SJ. (1997). Cancer Res., Hooper ML and Wyllie AH. (1993). Nature, 362, 849 ± 57, 5013 ± 5016. 852. Meek DW, Simon S, Kikkawa U and Eckhart W. (1990). Cross SM, Sanchez CA, Morgan CA, Schimke MK, Ramel EMBO J., 9, 3253 ± 3260. S, Idzerda RL, Raskind WH and Reid BJ. (1995). Science, Minn AJ, Boise LH and Thompson CB. (1996). Genes Dev., 267, 1353 ± 1356. 10, 2621 ± 2631. Di Leonardo A, Khan SH, Linke SP, Greco V, Seidita G and Miyashita T and Reed JC. (1995). Cell, 80, 293 ± 299. Wahl GM. (1997). Cancer Res., 57, 1013 ± 1019. MundtM,HuppT,FritscheM,MerkleC,HansenS,LaneD Dulic V, Kaufmann WK, Wilson SJ, Tlsty TD, Lees E, and Groner B. (1997). Oncogene, 15, 237 ± 244. Harper JW, Elledge SJ and Reed SI. (1994). Cell, 76, Notterman D, Young S, Wainger B and Levine AJ. (1998). 1013 ± 1023. Oncogene, 17, 2743 ± 2751. Finlay CA, Hinds PW and Levine AJ. (1989). Cell, 57, 1083 ± Passalaris TM, Benanti JA, Gewin L, Kiyono T and 1093. Galloway DA. (1999). Mol. Cell. Biol., 19, 5872 ± 5881. Fiscella M, Ullrich SJ, Zambrano N, Shields MT, Lin D, Pinna LA and Meggio F. (1997). Prog. Cell Cycle Res., 3, Lees-Miller SP, Anderson CW, Mercer WE and Appella 77 ± 97. E. (1993). Oncogene, 8, 1519 ± 1528. Prives C and Hall PA. (1999). J. Pathol., 187, 112 ± 126. Gotz C, Scholtes P, Prowald A, Schuster N, Nastainczyk W Rolley N and Milner J. (1994). Oncogene, 9, 3067 ± 3070. and Montenarh M. (1999). Mol. Cell. Biochem., 191, 111 ± SakaguchiK,HerreraJE,SaitoS,MikiT,BustinM, 120. Vassilev A, Anderson CW and Appella E. (1998). Genes Gu W and Roeder RG. (1997). Cell, 90, 595 ± 606. Dev., 12, 2831 ± 2841. Hall SR, Campbell LE and Meek DW. (1996). Nucleic Acids Sayed M, Kim SO, Salh BS, Issinger OG and Pelech SL. Res., 24, 1119 ± 1126. (2000). J. Biol. Chem., 275, 16569 ± 16573. Hao M, Lowy AM, Kapoor M, Dee A, Liu G and Lozano Schener M, Huibregtse JM, Vierstra RD and Howley PM. G. (1996). J. Biol. Chem., 271, 29380 ± 29385. (1993). Cell, 75, 495 ± 505. Haupt Y, Maya R, Kazaz A and Oren M. (1997). Nature, Shieh SY, Ikeda M, Taya Y and Prives C. (1997). Cell, 91, 387, 296 ± 299. 325 ± 334. Hermeking H, Lengauer C, Polyak K, He TC, Zhang L, Stewart ZA, Tang LJ and Pietenpol JA. (2001). Oncogene, Thiagalingam S, Kinzler KW and Vogelstein B. (1997). 20, 113 ± 124. Mol. Cell, 1, 3 ± 11. Takenaka K, Moriguchi T and Nishida E. (1998). Science, Herrmann CP, Kraiss S and Montenarh M. (1991). 280, 599 ± 602. Oncogene, 6, 877 ± 884. Taylor SS and McKeon F. (1997). Cell, 89, 727 ± 735. Huang C, Ma WY, Maxiner A, Sun Y and Dong Z. (1999). J. Ulloa L, Diaz-Nido J and Avila J. (1993). EMBO J., 12, Biol. Chem., 274, 12229 ± 12235. 1633 ± 1640. Huang LC, Clarkin KC and Wahl GM. (1996). Cancer Res., Ulloa L, Diaz-Nido J and Avila J. (1994). Cell. Mol. 56, 2940 ± 2944. Neurobiol., 14, 407 ± 414. Hupp TR, Meek DW, Midgley CA and Lane DP. (1992). Utrera R, Collavin L, Lazarevic D, Delia D and Schneider C. Cell, 71, 875 ± 886. (1998). EMBO J., 17, 5015 ± 5025. Innocente SA, Abrahamson JL, Cogswell JP and Lee JM. Vogt Sionov R and Haupt Y. (1999). Oncogene, 18, 6145 ± (1999). Proc. Natl. Acad. Sci. USA, 96, 2147 ± 2152. 6157. Jin P, Gu Y and Morgan DO. (1996). J. Cell. Biol., 134, 963 ± Xu X, Landesman-Bollag E, Channavajhala PL and Seldin 970. DC. (1999). Mol. Cell. Biochem., 191, 65 ± 74. Kapoor M and Lozano G. (1998). Proc. Natl. Acad. Sci. Yonish-Rouach E, Resnitzky D, Lotem J, Sachs L, Kimchi A USA, 95, 2834 ± 2837. and Oren M. (1991). Nature, 352, 345 ± 347. Khosravi R, Maya R, Gottlieb T, Oren M, Shiloh Y and Zhan Q, Antinore MJ, Wang XW, Carrier F, Smith ML, Shkedy D. (1999). Proc. Natl. Acad. Sci. USA, 96, 14973 ± Harris CC and Fornace Jr AJ. (1999). Oncogene, 18, 14977. 2892 ± 2900. Ko LJ and Prives C. (1996). Genes Dev., 10, 1054 ± 1072. Kubbutat MH, Jones SN and Vousden KH. (1997). Nature, 387, 299 ± 303.

Oncogene