Requirement of Krüppel-like factor 4 in preventing entry into mitosis following DNA damage.

Previous studies indicate that Krüppel-like factor 4 (KLF4 or GKLF) controls the G1/S cell cycle checkpoint upon DNA damage. We present evidence for an equally important role of KLF4 in maintaining the integrity of the G2/M checkpoint following DNA damage. HCT116, a colon cancer cell line with wild type p53 alleles, underwent sustained G2 arrest up to 4 days after gamma-irradiation. In contrast, HCT116 cells null for p53 were able to enter mitosis following irradiation. Western blot analyses of irradiated HCT116 cells showed increased levels of p53, KLF4, and p21WAF1/CIP1 and decreased levels of cyclin B1 when compared with unirradiated controls. In contrast, the levels of cyclin B1 increased in irradiated HCT116 p53-/- cells, in which KLF4 failed to increase due to the absence of p53. When KLF4 was inhibited by small interfering RNA, irradiated HCT116 cells exhibited increased mitotic indices and a rise in cyclin B1 levels. Conversely, irradiated HCT116 p53-/- cells that were infected with KLF4-expressing adenoviruses demonstrated a concurrent reduction in mitotic indices and cyclin B1 levels. In each case, Cdc2 kinase measurements showed an inverse correlation between Cdc2 kinase activities and KLF4 levels. Co-transfection experiments showed that KLF4 repressed the cyclin B1 promoter through a specific GC-rich element. Moreover, chromatin immunoprecipitation experiments demonstrated that both KLF4 and HDAC were associated with the cyclin B1 promoter in irradiated HCT116 cells. We conclude that KLF4 is essential in preventing mitotic entry following gamma-irradiation and does so by inhibiting cyclin B1 expression.

The eukaryotic cell cycle is divided into several phases: gap (G 1 ), DNA replication (S), gap 2 (G 2 ), mitosis (M), and a resting or quiescent phase (G 0 ) (1). The events in the cell cycle are ordered into dependent pathways so that late events cannot begin until early events are completed. For example, mitosis is dependent on the completion of DNA replication. Control mechanisms that enforce the dependence of the cell cycle are called checkpoints (2). Elimination of the checkpoints may result in cell death, infidelity in the distribution of chromosomes, or increasing susceptibility to environmental perturbations such as DNA-damaging agents. There are three major checkpoints in the cell cycle: G 1 /S, G 2 /M, and mitotic or spindle checkpoints (3)(4)(5). Controlling the checkpoints are three families of proteins: cyclins, cyclin-dependent kinases (Cdk), 1 and inhibitors of Cdk (Cki) (6 -9). Cyclins and Cdk form heterodimeric complexes that phosphorylate downstream targets to drive the cell cycle. Cki are inhibitors of the cyclin-Cdk complexes and function by halting the cell cycle at checkpoints.
In response to DNA damage such as that caused by ionizing radiation and chemotherapeutic drugs, cells are arrested at the transition from G 1 to S phase and G 2 to M phase (10 -13). Arrest in these checkpoints prevents DNA replication and mitosis in the presence of chromosomal alterations. The tumor suppressor, p53, plays a pivotal role in controlling both the G 1 /S and G 2 /M transitions following DNA damage (14 -16). Upon its activation by upstream mediators, p53 transcriptionally induces a Cki, p21 WAF1/CIP1 , which in turn inhibits G 1 cyclin-Cdk complexes and is mainly responsible for the resulting G 1 /S arrest (17,18). Both p53 and p21 WAF1/CIP1 are also required for sustained arrest in the G 2 phase following ␥-irradiation (19). The mechanisms by which p53 accomplishes this task are thought to be multiple and include its ability to transcriptionally suppress cdc2 and cyclin B1 (20), which are required for entry into mitosis (21). A second mechanism is through the induction of 14-3-3 expression (22,23), which sequesters cyclin B1-Cdc2 complex in the cytoplasm (24). Third, p21 WAF1/CIP1 inhibits cyclin B1-Cdc2 and prevents entry into mitosis (25,26). Last, 14-3-3 and p21 WAF1/CIP1 have also been shown to exert a cooperative effect in controlling the G 2 /M checkpoint (27).
Although the induction of p21 WAF1/CIP1 expression has been shown to be a consequence of direct binding of p53 to its promoter, evidence implicates many other transcription factors in regulating p21 WAF1/CIP1 transcription (28). Among these is Krü ppel-like factor 4 (KLF4; also known as gut-enriched Krü ppel-like factor or GKLF), a member of the mammalian KLF family of transcription regulators (29 -31). KLF4 was initially identified as an epithelially enriched gene with preferential expression in the terminally differentiated, postmitotic epithelial cells of the intestine and epidermis (32,33). In cultured cells, expression of KLF4 is associated with conditions that lead to growth arrest such as serum deprivation or contact inhibition (32). Consistent with these findings, constitutive expression of KLF4 inhibits DNA synthesis and reduced cell proliferation (32,34,35). This is in part due to cell cycle arrest at the G 1 /S boundary as a result of the ability of KLF4 to transcriptionally activate expression of p21 WAF1/CIP1 (36 -38).
In support of a checkpoint function for KLF4, we recently showed that its expression is activated in a p53-dependent fashion upon DNA damage by agents such as methyl methane sulfonate and ␥-irradiation (37,39). This induction is correlated with an increase in the levels of p21 WAF1/CIP1 with consequent G 1 /S cell cycle arrest in cells with wild type p53 (39). Importantly, inhibition of KLF4 expression in such cells after ␥-irradiation results in abrogation of the G 1 arrest in a manner similar to the cell cycle profile seen in irradiated cells that are null for p53 (39). Conversely, conditional expression of KLF4 in irradiated cells null for p53 restored G 1 arrest as if the they were wild type for p53 (39). These findings indicate that KLF4 is a necessary and sufficient mediator of p53 for the G 1 /S cell cycle arrest resulting from DNA damage and does so by activating p21 WAF1/CIP1 expression. Since p21 WAF1/CIP1 has also been shown to be required for sustained G 2 arrest following ␥-irradiation (19), we sought to determine in the present study whether KLF4 may also be involved in controlling the G 2 /M checkpoint after DNA damage.

EXPERIMENTAL PROCEDURES
Cell Lines-The colon cancer cell lines wild type and null for p53, HCT116 p53ϩ/ϩ, and HCT116 p53Ϫ/Ϫ, respectively, were generous gifts from Dr. Bert Vogelstein (Johns Hopkins University) (19). The cells were cultured in McCoy's medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. EcR116 p53Ϫ/Ϫ cells were established by stably transfecting pVgRXR (36), which contains VgEcR and RXR that form a receptor for the insect hormone, ecdysone, into the parental HCT116 cell line and selected with 100 g/ml Zeocin (Invitrogen). The level of RXR expression was determined by Western blot analysis.
␥-Irradiation-␥-Irradiation of cultured cells was performed using a 137 Cs ␥-irradiator at 0.8 Gy/min for 15 min, for a total of 12 Gy. Cells were harvested at 0, 24, 48, 72, and 96 h after ␥-irradiation for subsequent assays. Medium was changed at the time of collection for the remaining plates.
FACS Analysis-Cell cycle analysis was performed as previously described (39). Cells were rinsed in Dulbecco's phosphate-buffered saline (DPBS; Mediatech, Inc.), treated with trypsin, and resuspended in McCoy's medium containing 10% FBS. Cells were then collected by centrifugation, washed with DPBS, collected again by centrifugation, resuspended in 70% ethanol, and fixed at Ϫ20°C overnight. Cells were pelleted once again by centrifugation and resuspended in a solution containing 50 g/ml propidium iodide, 50 g/ml RNase A, 0.1% Triton X-100, and 0.1 mM EDTA at room temperature for 30 min. Flow cytometry was performed on a FACSCalibur cytometer (Becton Dickinson).
Measurement of Mitotic Indices-At each time point, cells were fixed in 3% formaldehyde for 15 min. Cold 100% methanol was then added, and cells were incubated at room temperature for 20 min. Cells were then rinsed three times with DPBS. A Hoechst 33258 solution (10 g/ml) was added to each dish to a final concentration of 0.2 g/ml, which was incubated at room temperature for 15 min. After the incubation, cells were rinsed five times with DPBS, and nuclei were visualized by fluorescence microscopy (Nikon). A minimum of 400 cells were examined per experiment. Mitotic figures were scored for cells with condensed chromosomes.
Mitotic trapping experiments were performed by adding nocodazole to the culture media to a final concentration of 0.2 g/ml. Media containing nocodazole were replaced every 24 h. Mitotic figures were examined following nuclear staining.
Adenovirus Infection-The recombinant adenoviruses containing GFP and KLF4 (AdEGI-KLF4) or GFP alone (AdEGI) were described previously (36,40). EcR116 p53Ϫ/Ϫ cells were grown to 40% confluence in 10-cm dishes and replenished with fresh media containing 2% FBS followed by the addition of 10 8 plaque-forming units of recombinant virus per dish. Infected cells were incubated at 37°C for 6 h, at which time cells were ␥-irradiated, and the media were changed. Cells were treated with 5 M ponasterone A (Invitrogen) for 0, 24, 48, 72, and 96 h and then collected for further analysis. Medium was changed at the time of collection for the remaining plates.
Preparation of Small Interfering RNA (siRNA) and Transfection-The KLF4-specific siRNA was described previously (39). A nonspecific, double-stranded siRNA with identical length was also generated based on the sequence of an unrelated protein and used as a control.
HCT116 p53ϩ/ϩ cells were grown to 40% confluence in 10-cm dishes, ␥-irradiated for a total of 12 Gy, and transfected with annealed siRNA using DMRIE-C reagent (Invitrogen) for 6 h as recommended by the manufacturer. McCoy's media containing 20% FBS and 2% penicillin/ streptomycin were added to each dish to a final concentration of 10% FBS and 1% penicillin/streptomycin. Cells were harvested at 0, 24, 48, 72, and 96 h for further assays. Medium was changed at the time of collection for the remaining plates.
Chromatin Immunoprecipitation (ChIP) Assay-ChIP assays were performed based on a published protocol (43). HCT116 p53ϩ/ϩ cells were irradiated or not for the amount of 12 Gy and maintained in culture for an additional 48 h. Cells were washed twice with PBS, followed by the addition of 1% formaldehyde in PBS, and incubated for 15 min at room temperature. At the end of the incubation, the crosslinking was terminated by the addition of glycine to a final concentration of 0.125 M. Cells were then washed twice with PBS and lysed with 1 ml of radioimmune precipitation assay buffer that contained 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 5 mM EDTA, 1% Nonidet P-40, 0.5% deoxycholate, and protease inhibitors. Cell lysates were then sonicated to yield chromatin fragments of ϳ600 bp, as assessed by agarose gel electrophoresis. Insoluble materials were removed by centrifugation, at 13,000 ϫ g for 10 min, and the supernatant was transferred to a new tube. The supernatant was diluted 10-fold with dilution buffer (20 mM Tris-HCl, pH 8.1, 150 mM NaCl, 2 mM EDTA, and 1% Triton X-100), and the diluted extracts were precleared by incubation with 45 l of protein G-Sepharose beads (Sigma) and sheared salmon sperm DNA. After centrifugation for 5 min at 7,000 ϫ g, the supernatant was transferred to a fresh tube. Immunoprecipitation was performed by rocking overnight at 4°C the precleared cell lysates with the appropriate antibodies. Immune complexes were then precipitated with protein G-Sepharose beads and sheared salmon sperm DNA. The beads were collected by centrifugation and washed twice sequentially with radioimmune pre-cipitation assay wash buffer I (20 mM Tris-HCl, pH 8.1, 150 mM NaCl, 0.1% SDS, 1% Triton X-100, and 2 mM EDTA), wash buffer II (20 mM Tris-HCl, pH 8.1, 500 mM NaCl, 0.1% SDS, 1% Triton X-100, and 2 mM EDTA), wash buffer III (10 mM Tris-HCl, pH 8.1, 0.25 M LiCl, 1% Nonidet P-40, 1% deoxycholate, and 1 mM EDTA), and wash buffer IV (10 mM Tris-HCl, pH 8.1, and 1 mM EDTA). The immunoprecipitates were eluted by adding 200 l of elution buffer (1% SDS, 0.1 M NaHCO 3 ) to the washed beads followed by incubation at 65°C overnight. After centrifugation for 5 min at 7,000 x g, DNA was extracted from the supernatant using a Qiagen PCR purification kit. A 3-l DNA sample was then subjected to amplification using primers encompassing nt Ϫ287 and ϩ23 of the cyclin B1 promoter (5Ј-TCTTGCCCGGCTAAC-CTTTCCAGG-3Ј and 5Ј-TTCCGCCGCAGCACGCCGAGAAGA-3Ј).
Trichostatin A (TSA) Treatment-HCT116 p53ϩ/ϩ cells were irradiated or not at day 0 and fed with media containing 2 M TSA or vehicle alone for 2 days. Cell extracts were then harvested for Western blot analysis of cyclin B1 and ␤-actin content.

p53 Is Required for Preventing Mitotic Entry in HCT116
Cells following ␥-Irradiation-The tumor suppressor p53 exerts critical checkpoint functions. To demonstrate a dependence of sustained G 2 arrest on p53 following DNA damage, we irradiated HCT116 p53ϩ/ϩ and HCT116 p53Ϫ/Ϫ with a 12-Gy ␥-ray and examined daily the cell cycle profiles up to 4 days following irradiation. Consistent with previous findings (19,39), parental cells (p53ϩ/ϩ) arrested in either G 1 with 2n DNA content or G 2 with 4n DNA content from day 1 to 4 after irradiation, as expected for cells with intact checkpoint functions (results not shown). Cells with deleted p53 alleles (p53Ϫ/Ϫ) were arrested mostly in G 2 for the same duration. However, although a substantial number of cells were arrested in G 2 for both genotypes, morphological examination shows a steady increase in the number of HCT116 p53Ϫ/Ϫ cells that entered mitosis beginning at day 2 following irradiation ( Fig.  1B). In contrast, irradiated HCT116 p53ϩ/ϩ cells had a persistent decrease in the percentage of cells in mitosis from days 1 to 4 when compared with that before irradiation (Fig. 1B). Control unirradiated cells displayed a steady and equal level of mitotic indices for both genotypes (Fig. 1A). Both HCT116 p53ϩ/ϩ and p53Ϫ/Ϫ cells had an intact mitotic apparatus as evidenced by their ability to be become trapped in mitosis temporarily by the microtubule-disrupting agent, nocodazole (19) (Fig. 1C). However, only HCT116 p53Ϫ/Ϫ cells were able to enter mitosis following ␥-irradiation in the presence of nocodazole (Fig. 1D). These results indicate that p53 is required for sustaining G 2 arrest and preventing mitotic entry in HCT116 cells following DNA damage.
Levels of G 2 /M Markers in HCT116 p53ϩ/ϩ Cells Are Decreased after ␥-Irradiation-To determine the mechanism by which p53 sustains G 2 arrest and prevents mitosis in irradiated HCT116 cells, we performed Western blot analysis on several markers of the G 2 /M checkpoint. In particular, Cdc2 and cyclin B1 were selected, since they have been shown to be required for entry into mitosis. As seen in Fig. 2, there was a persistent increase in the levels of p53 in irradiated HCT116 p53ϩ/ϩ cells from day 1 to 4 following irradiation when compared with unirradiated cells. The levels of KLF4 and p21 WAF1/CIP1 were concomitantly increased in irradiated cells. In contrast, there was a significant and sustained decrease in the levels of cyclin B1 in irradiated cells. The levels of Cdc2 were also decreased following ␥-irradiation but to a lesser extent than that of cyclin B1. These results suggest that the reduced levels of Cdc2 and cyclin B1 contributed to the sustained G 2 arrest observed in irradiated HCT116 p53ϩ/ϩ cells.
Inhibition of KLF4 Expression in HCT116 p53ϩ/ϩ Cells Results in Abrogation of G 2 /M Arrest following ␥-Irradiation-Previous studies show that p21 WAF1/CIP1 , the expression of which is p53-dependent, is required for the sustained G 2 arrest following ␥-irradiation (19). Since p21 WAF1/CIP1 is a target of KLF4 (36 -38), we examined the role of KLF4 in sustaining G 2 arrest after DNA damage. To accomplish this goal, we used siRNA to inhibit KLF4 expression (39). Fig. 3 shows that HCT116 p53ϩ/ϩ cells treated with control, nonspecific siRNA had low levels of mitotic index throughout the 4 days following ␥-irradiation in a manner similar to that seen for untreated cells (Fig. 1B). In contrast, cells treated with KLF4-specific siRNA entered mitosis beginning on day 2 following irradiation (Fig. 3). This pattern is similar to that observed for irradiated HCT116 p53Ϫ/Ϫ cells (Fig. 1B). These results indicate that KLF4 is involved in preventing mitotic entry in HCT116 p53ϩ/ϩ cells following ␥-irradiation.
We also performed Western blot analysis for G 2 /M markers in irradiated HCT116 p53ϩ/ϩ cells that had been treated with KLF4-specific or control siRNA. Fig. 4 shows that the levels of p53 were not significantly altered in cells receiving either siRNA, as expected. The levels of KLF4 were persistently decreased in cells treated with KLF4-specific siRNA (ϩ lanes) as compared with those treated with control siRNA (Ϫ lanes). There was a corresponding decrease in the levels of p21 WAF1/CIP1 in cells treated with KLF4-specific siRNA. Importantly, the levels of cyclin B1 were increased in cells treated with KLF4specific siRNA from day 2 to 4 when compared with those treated with control siRNA. In contrast, the levels of Cdc2 did not alter in KLF4-siRNA-treated cells. These results suggest that the increased cyclin B1 in KLF4-siRNA-treated HCT116 p53ϩ/ϩ cells may be the reason for the increased mitotic indices following irradiation as observed in Fig. 3.

Overexpression of KLF4 Suppresses Mitosis in ␥-Irradiated
HCT116 p53Ϫ/Ϫ Cells-We also attempted to determine the reason why HCT116 p53Ϫ/Ϫ cells were able to enter mitosis following ␥-irradiation as noted in Fig. 1B. This was accomplished by analyzing protein levels of G 2 /M markers by Western blot analysis as demonstrated in Fig. 5. As expected, p53 was completely absent from these cells. KLF4 was also absent in either irradiated or unirradiated cells because of the lack of p53. The levels of Cdc2 were similar between irradiated and unirradiated cells. However, cyclin B1 levels were increased in irradiated cells, which is the opposite of that observed for HCT116 p53ϩ/ϩ cells (Fig. 2). These results suggest that p53 suppresses cyclin B1 expression in irradiated HCT116 p53ϩ/ϩ cells and that in the absence of p53, cyclin B1 is no longer inhibited and contributes to the increased mitosis seen in the irradiated HCT116 p53Ϫ/Ϫ cells.
To determine the effect of re-expressing KLF4 in irradiated HCT116 p53Ϫ/Ϫ cells, we used a conditional expression system involving adenoviruses as previously reported (36,39). EcR116 p53Ϫ/Ϫ cells, a derivative of HCT116 p53Ϫ/Ϫ containing the receptor for the insect hormone, ecdysone, were infected with either a control adenovirus, AdEGI, or an adenovirus containing an ecdysone-inducible KLF4, AdEGI-KLF4 (36,39). Following infection and irradiation, cells were treated with the inducer, ponasterone A, to activate expression of the transgene. As seen in Fig. 6, EcR116 p53Ϫ/Ϫ cells infected with the control virus, AdEGI, continued to enter mitosis following ␥-irradiation. In contrast, the mitotic indices after irradiation were reduced in those cells infected with AdEGI-KLF4 and upon induction by ponasterone A. Importantly, induced expression of KLF4 in irradiated HCT116 p53Ϫ/Ϫ cells was accompanied by  2. Western blot analysis of p53, KLF4, p21 WAF1/CIP1 , and G 2 markers in HCT116 p53؉/؉ cells following ␥-irradiation. HCT116 p53ϩ/ϩ cells were irradiated on day 0, and Western blot analysis was performed for the stated proteins daily up to 4 days following irradiation (ϩ lanes). Controls were unirradiated cells (Ϫ lanes). ␤-Actin was used as a loading control. The experiments were repeated three times, and shown is the result of a representative experiment. a reduction in cyclin B1 levels and an increase in p21 WAF1/CIP1 levels (Fig. 7). The levels of Cdc2 were not influenced by KLF4 (Fig. 7). These results show that induced expression of KLF4 in the absence of p53 is capable of suppressing entry into mitosis in HCT116 p53Ϫ/Ϫ cells following ␥-irradiation. This effect is probably accomplished by a combination of suppression of cyclin B1 and activation of p21 WAF1/CIP1 expression by KLF4.
Cdc2 Kinase Activities Correlate with Mitotic Entry and Are Inhibited by KLF4 following DNA Damage-Cdc2 and cyclin B1 are two components of Cdc2 kinase required for mitotic entry. Previous studies indicate that p53 inhibits Cdc2 kinase (24). To determine the involvement of KLF4 in regulating Cdc2 kinase activity, we performed kinase assays under the various experimental conditions described above. As shown in Fig. 8A, Cdc2 kinase activity was reduced (lane 2) at day 3 following ␥-irradiation of HCT116 p53ϩ/ϩ cells when compared with unirradiated control (lane 1). In contrast, there was an increase in Cdc2 kinase activity (lane 4) at day 3 following irradiation of HCT116 p53Ϫ/Ϫ cells comparing to the unirradiated control (lane 3). The activities of Cdc2 kinase correlated with the mitotic indices of cells in these conditions as revealed in Fig. 1, A and B. When irradiated HCT116 p53ϩ/ϩ cells were treated with KLF4-specific siRNA (Fig. 8B, lane 2), Cdc2 kinase activity was increased compared with cells treated with control siRNA (Fig. 8B, lane 1). Conversely, Cdc2 kinase activity was reduced in irradiated HCT116 p53Ϫ/Ϫ cells that were infected with the AdEGI-KLF4 virus and induced with ponasterone A (Fig. 8C, lane 2) as compared with those infected with the control AdEGI virus (Fig. 8C, lane 1). In each case, Cdc2 kinase activities correlated with the extent of mitotic indices as documented in earlier figures. These results indicate that KLF4 inhibits Cdc2 kinase activity.
KLF4 Suppresses the Cyclin B1 Promoter Activity-Previous studies indicate that G 2 arrest caused by overexpression of p53 is in part due to the ability of p53 to transcriptionally suppress cyclin B1 expression (20). The correlative changes in the levels of KLF4 and p53 (Figs. 2, 4, and 5) and the inverse relationship between the levels of KLF4 and cyclin B1 (Figs. 2, 4, 5, and 7) suggest that KLF4 may mediate the effect of p53 in suppressing cyclin B1 activity. We therefore performed co-transfection experiments to determine whether KLF4 may transcriptionally suppress cyclin B1 expression. As seen in Fig. 9A, KLF4 suppressed cyclin B1 promoter activity by 90% (compare lanes 1  and 2). The suppressive effect of KLF4 was mediated in part by a GC-rich element between nt Ϫ137 and Ϫ142 in the cyclin B1 promoter, since mutation in this element abrogated the sup-pressive effect of KLF4 (Fig. 9A, compare lanes 3 and 4). The association of KLF4 with the cyclin B1 promoter was further demonstrated by in vivo ChIP experiments. As seen in Fig. 9B, an antibody directed against KLF4 specifically precipitated a DNA fragment containing the cyclin B1 promoter (lane 4) in irradiated (ϩ ␥) but not in unirradiated (Ϫ ␥) HCT116 p53ϩ/ϩ cells. An antibody directed against the histone deacetylase  1 and 3) for the next 2 days before subject to Western blot analysis for cyclin B1. ␤-Actin was used as a loading control. The experiments were repeated three times, and shown is the result of a representative experiment.
(HDAC) was also able to precipitate the cyclin B1 promoter fragment but again only in irradiated cells (lane 5). This result suggests that HDAC may be part of the co-repressor complex that assists in the suppression of the cyclin B1 promoter by KLF4. To substantiate this finding, we treated HCT116 p53ϩ/ϩ cells with TSA, a HDAC inhibitor, following ␥-irradiation and compared the levels of cyclin B1 with untreated cells. As seen in Fig. 9C, the levels of cyclin B1 remained the same between TSA-treated (lane 2) and untreated (lane 1) cells that were not irradiated. In contrast, the levels of cyclin B1 were increased in irradiated cells that had been treated with TSA (lane 4) when compared with the untreated but irradiated cells (lane 3). These observations indicate that inhibition of HDAC results in a derepression of the cyclin B1 gene by KLF4. DISCUSSION The eukaryotic cell cycle is a carefully orchestrated event. A round of cell division requires accurate DNA duplication in the S phase and proper chromosomal segregation during the M phase. Checkpoint controls are critical for the progression of the cell cycle in the event of chromosomal alterations such as those induced by ionizing radiation. In particular, the onset of mitosis is carefully guarded at the G 2 /M boundary and is not permitted until a round of DNA replication is completed earlier in the cell cycle (21,44). It has become apparent that a central and rate-limiting function of the transition from G 2 to M is performed by a protein kinase activity earlier referred to as maturation-promoting factor (48,49). Maturation-promoting factor is a complex of two proteins in equimolar amounts (50). The first component is the catalytic subunit of the protein kinase Cdc2 (51,52). The second component is cyclin B1 (53,54), which is required for the full activity of Cdc2 by a mechanism that includes specification of substrates or subcellular localization of the maturation-promoting factor protein kinase dimer (55). Cdc2 is primarily regulated by phosphorylation, both inhibitory and stimulatory, whereas cyclin B1 is regulated by both phosphorylation and synthesis (56).
Much progress has been made in understanding the mechanisms that control the cell cycle checkpoints in response to DNA damage (10,57). Central to this process is the tumor suppressor, p53, whose function is required for the maintenance of genomic integrity (14 -16). A reason for the importance of p53 lies in its versatility in regulating numerous cellular processes during genotoxic responses including checkpoint controls at the G 1 /S and G 2 /M boundaries as well as controlling the spindle checkpoint and maintaining centrosome homeostasis (15). A major target of p53 activation is p21 WAF1/CIP1 , which is responsible for the consequent G 1 /S and G 2 /M cell cycle arrest that ensues following DNA damage (17)(18)(19). Previous work in our laboratory demonstrates that KLF4 is transcriptionally activated upon DNA damage-induced activation of p53 and that KLF4 transcriptionally activates p21 WAF1/CIP1 expression by binding to a specific cis-element in the proximal p21 WAF1/CIP1 promoter (37). Importantly, this cis-element is required for the stimulatory effect of p53 on the p21 WAF1/CIP1 promoter despite the fact that p53 binds further upstream in the promoter (37). Moreover, p53 and KLF4 form a heterodimer and synergistically activate the p21 WAF1/CIP1 promoter (37). These results provide strong biochemical evidence for a crucial role of KLF4 in mediating the transcriptional activating effect of p53 on p21 WAF1/CIP1 (37). Indeed, recent genetic evidence also supports an essential function for KLF4 in mediating the action of p53 in eliciting G 1 /S arrest following DNA damage (39). Thus, when KLF4 is inhibited, cells can no longer arrest at G 1 despite the presence of an intact p53 (39). Since p21 WAF1/CIP1 has also been shown to be required for sustained G 2 arrest following DNA damage (19), it is highly likely that the G 2 arrest effect of KLF4 seen in this study is in part contributed by the ability of KLF4 to transactivate p21 WAF1/CIP1 as previously documented (36 -39).
p53 prevents entry into mitosis when cells enter G 2 with damaged DNA or when they arrest in the S phase due to depletion of the substrates required for DNA synthesis (24). A mechanism by which p53 accomplishes this effect is by inhibiting Cdc2. Cdc2 is simultaneously inhibited by three transcriptional targets of p53: p21 WAF1/CIP1 , 14-3-3, and Gadd45 (24). In addition, p53 is able to directly repress the cdc2 promoter (20), resulting in a decrease in the level of Cdc2 upon activation of p53, as seen in the current study (Fig. 2). 14-3-3, the expression of which is p53-dependent (22), functions to sequester Cdc2 and cyclin B1 in the cytoplasm, thus disabling Cdc2 from initiating mitosis (23). It is of interest to note that a recent cDNA microarray analysis of the target genes of KLF4 using the colon cancer cell line RKO harboring a conditionally expressed KLF4, identified 14-3-3 as an up-regulated and cdc2 as a down-regulated gene, in addition to p21 WAF1/CIP1 as an up-regulated gene upon KLF4 activation (38). However, we observed little change in the levels of 14-3-3 in HCT116 p53ϩ/ϩ cells subjected to ␥-irradiation in the present study (results not shown). In addition, the levels of Cdc2 varied little upon manipulation of KLF4 levels ( Figs. 4 and 7). The inhibitory effect of KLF4 on mitotic entry in HCT116 cells must therefore not depend on either Cdc2 or 14-3-3 to any large extent. Whether RKO cells may enter sustained G 2 arrest upon prolonged induction of KLF4 by affecting the levels of Cdc2 and 14-3-3 as previously observed (38) remains to be examined.
Results of our study demonstrate that the levels of cyclin B1 are influenced by manipulation of KLF4 levels in HCT116 cells (Figs. 4 and 7). Thus, there is an inverse relationship between the amounts of KLF4 and cyclin B1 whether KLF4 is inhibited by siRNA or overexpressed by adenovirus-mediated induction. These results suggest that KLF4 may be a repressor of cyclin B1 expression. Indeed, results of co-transfection experiments confirm a direct suppressive effect of KLF4 on the cyclin B1 promoter (Fig. 9A). The changes in cyclin B1 levels are probably the reason for the altered Cdc2 kinase activities under the various experimental conditions as observed in Fig. 8. Previous studies indicate that p53 decreases cyclin B1 levels and attenuates activity of the cyclin B1 promoter (58,59). This inhibitory effect of p53 is exerted at a transcriptional level on the cyclin B1 promoter through a cis-element between nucleotides Ϫ287 and Ϫ123 of the promoter (20). It should be noted that this region of the cyclin B1 promoter does not contain any sequence recognized by p53 (60,61). Instead, a GC-rich, Sp1-binding site is present between nt Ϫ137 and Ϫ142 of the promoter (42). Site-directed mutagenesis of this element abolished the suppressive effect of KLF4 on the cyclin B1 promoter (Fig. 9A), consistent with the ability of KLF4 to bind to GC-rich sequences (37,(62)(63)(64)(65). Thus, KLF4 may mediate the action of p53 in modulating expression of yet another key cell cycle checkpoint gene in a manner reminiscent of its previously documented effect on the p21 WAF1/CIP1 promoter (37). Together, these findings further strengthen the significance of KLF4 in checkpoint regulation. They also support our recent findings on a potential role for KLF4 as a tumor suppressor in colorectal cancer (66).
Our study also identified HDAC as part of the co-repressor complex in the suppression of the cyclin B1 promoter by KLF4. Thus, HDAC is associated with the cyclin B1 promoter in irradiated HCT116 p53ϩ/ϩ cells (Fig. 9B). Consistent with this finding, TSA, a HDAC inhibitor, was able to relieve the suppressive effect of KLF4 on cyclin B1 promoter (Fig. 9C). These observations indicate that chromatin remodeling by hi-stone deacetylase is important in the regulation of the cyclin B1 promoter by KLF4. They are also consistent with a previous report that transcriptional repression by p53 requires histone deacetylase (41).
In addition to a crucial role in controlling the G 1 /S and G 2 /M checkpoints, KLF4 may exert an effect on the spindle checkpoint. This is based on the previous observation that among the genes regulated by KLF4, many are components of the kinetochore that have checkpoint functions in spindle formation (38). It is also of interest to note that cells without p53 or p21 proceed into mitosis after ␥-irradiation but have a 4n DNA content (19) (this study) rather than the 2n DNA content expected for cells that had gone through mitosis. This is due to the fact that irradiated p53Ϫ/Ϫ and p21Ϫ/Ϫ cells never completed cytokinesis such that the majority of the cells contain abnormally shaped, multilobulated nuclei (19). In addition, p53Ϫ/Ϫ and p21Ϫ/Ϫ cells contain an abnormal number of centrosomes (19). Whether KLF4 is directly or indirectly involved in the spindle checkpoint or centrosome duplication process remains to be determined.