Kruppel-like factor 4 mediates p53-dependent G1/S cell cycle arrest in response to DNA damage.

The tumor suppressor p53 is required for the maintenance of genomic integrity following DNA damage. One mechanism by which p53 functions is to induce a block in the transition between the G(1) and S phase of the cell cycle. Previous studies indicate that the Krüppel-like factor 4 (KLF4) gene is activated following DNA damage and that such activation depends on p53. In addition, enforced expression of KLF4 causes G(1)/S arrest. The present study examines the requirement of KLF4 in mediating the p53-dependent cell cycle arrest process in response to DNA damage. We show that the G(1) population of a colon cancer cell line, HCT116, that is null for the p53 alleles (-/-) was abolished following gamma irradiation compared with cells with wild-type p53 (+/+). Conditional expression of KLF4 in irradiated HCT116 p53-/- cells restored the G(1) cell population to a level similar to that seen in irradiated HCT116 p53+/+ cells. Conversely, treatment of HCT116 p53+/+ cells with small interfering RNA (siRNA) specific for KLF4 significantly reduced the number of cells in the G(1) phase following gamma irradiation compared with the untreated control or those treated with a nonspecific siRNA. In each case the increase or decrease in KLF4 level because of conditional induction or siRNA inhibition, respectively, was accompanied by an increase or decrease in the level of p21(WAF1/CIP1). Results of our study indicate that KLF4 is an essential mediator of p53 in controlling G(1)/S progression of the cell cycle following DNA damage.

The tumor suppressor p53 is required for the maintenance of genomic integrity following DNA damage. One mechanism by which p53 functions is to induce a block in the transition between the G 1 and S phase of the cell cycle. Previous studies indicate that the Krü ppel-like factor 4 (KLF4) gene is activated following DNA damage and that such activation depends on p53. In addition, enforced expression of KLF4 causes G 1 /S arrest. The present study examines the requirement of KLF4 in mediating the p53-dependent cell cycle arrest process in response to DNA damage. We show that the G 1 population of a colon cancer cell line, HCT116, that is null for the p53 alleles (؊/؊) was abolished following ␥ irradiation compared with cells with wild-type p53 (؉/؉). Conditional expression of KLF4 in irradiated HCT116 p53؊/؊ cells restored the G 1 cell population to a level similar to that seen in irradiated HCT116 p53؉/؉ cells. Conversely, treatment of HCT116 p53؉/؉ cells with small interfering RNA (siRNA) specific for KLF4 significantly reduced the number of cells in the G 1 phase following ␥ irradiation compared with the untreated control or those treated with a nonspecific siRNA. In each case the increase or decrease in KLF4 level because of conditional induction or siRNA inhibition, respectively, was accompanied by an increase or decrease in the level of p21 WAF1/CIP1 . Results of our study indicate that KLF4 is an essential mediator of p53 in controlling G 1 /S progression of the cell cycle following DNA damage.
The mammalian cell cycle is operationally divided into five distinct phases: gap 1 (G 1 ), DNA synthesis (S), gap 2 (G 2 ), mitosis (M), and growth arrest phase (G 0 ), also called quiescence (1). Complex networks of control mechanisms called "checkpoints" are responsible for the orderly progression of these events within the cell cycle. Defects in checkpoint control increase genetic instability, thereby contributing to uncontrolled proliferation (2). For example, damage to the DNA elicits a series of signal transduction pathways that result in an arrest of the cell cycle at various checkpoints (3). Much of the DNA damage-induced signals are funneled through p53, which directs further downstream actions that lead to inhibition of G 1 to S and G 2 to M transitions, among other events such as apoptosis (4). Therefore, it is not surprising that p53 is the most frequently mutated tumor suppressor gene in human cancers (5).
The arrest in the transition between the G 1 and S phase of the cell cycle elicited by p53 requires in part the transcriptional activation of the gene encoding the cyclin-dependent kinase (Cdk) 1 inhibitor p21 WAF1/CIP1 (6,7). p21 WAF1/CIP1 binds to several G 1 cyclin-Cdk complexes and inhibits phosphorylation of the retinoblastoma susceptibility gene product Rb (8), a step required for the onset of DNA synthesis (9). Recent evidence suggests that p21 WAF1/CIP1 is also required to sustain G 2 arrest after DNA damage (10). Here, p21 WAF1/CIP1 mediates the function of p53 in response to DNA damage by inhibiting Cdc2 (11), a Cdk required for entry into mitosis (12). The proportion of cells that arrests in G 1 /S or G 2 /M depends on the cell type and status of checkpoint controls in each cell (13).
Although earlier studies indicate that expression of p21 WAF1րCIP1 is the result of direct binding of p53 to its promoter (14), it is now evident that a myriad of transcription factors under various physiologic conditions can also lead to the transcriptional activation of p21 WAF1րCIP1 (15). Among these is the zinc finger-containing transcription factor, Krü ppel-like factor 4 (KLF4), also called gut-enriched Krü ppel-like factor or GKLF (16,17). KLF4 is a member of a rapidly expanding family of mammalian Krü ppel-like factors that exhibit homology to the Drosophila protein Krü ppel (18). Expression of KLF4 is highly enriched in the postmitotic terminally differentiated epithelial cells of the intestine and epidermis (19,20). In cultured cells expression of KLF4 is associated with growth arrest as a result of serum deprivation or contact inhibition (19,21). Conversely, enforced expression of KLF4 inhibits DNA synthesis and results in decreased cell proliferation (19,22,23). These studies suggest that KLF4 is a negative regulator of cell growth.
Recently, it was demonstrated that expression of KLF4 is also induced by DNA damage and that such induction is dependent on p53 (24). Importantly, KLF4 was shown to physically interact with p53, resulting in a synergistic activation of the p21 WAF1րCIP1 promoter. Moreover, antisense inhibition of KLF4 leads to a decreased level of p21 WAF1/CIP1 in response to DNA damage (24), suggesting that KLF is a potentially important mediator of p53-induced growth arrest. Indeed, recent studies using an inducible system for KLF4 indicate that its induction leads to arrest in the G 1 /S transition of the cell cycle (25). In the present study, we further characterize the role of KLF4 in mediating p53-dependent cell cycle arrest. By manip-* This work was supported in part by Grants DK52230 and CA84197 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
¶ ulating KLF4 expression, we show that KLF4 is essential for the G 1 /S cell cycle arrest that results from 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 of Johns Hopkins University (10). The cells were cultured in McCoy's medium supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin. EcR116 p53Ϫ/Ϫ cells were established by stably transfecting pVgRXR (25), which contains VgEcR and retinoid X receptor that form a receptor for the insect hormone ecdysone, into the parental HCT116 cell line and selecting with 100 g/ml Zeocin (Invitrogen). The level of retinoid X receptor 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 24 h after ␥ irradiation for subsequent assays.
Adenovirus Infection-The recombinant adenovirus containing green fluorescence protein and KLF4 (AdEGI-KLF4) or green fluorescence protein alone (AdEGI) were described previously (25,26). EcR116 p53Ϫ/Ϫ cells were grown to 40% confluence in 10-cm dishes and replenished with fresh media containing 2% fetal bovine serum 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 medium was changed. Cells were treated with 5 M ponasterone A (Invitrogen) for 24 h and then collected for further analysis.
Preparation of siRNA and Transfection-23-nucleotide singlestranded RNAs were produced by Integrated DNA Technologies (Coralville, IA). The small interfering RNA (siRNA) sequences targeting KLF4 (GenBank TM accession number XM_047517) correspond to the coding region between nucleotides 121-141 from the translation initiation site. The complementary single-stranded RNAs were dissolved in 10 mM Tris-HCl and 1 mM EDTA (pH 7.0) and annealed in 25 mM KoAc, 10 mM Tris-HCl, and 1 mM EDTA (pH 7.0) by briefly heating to 70°C, then incubating for 20 min each at 37 and 23°C. A nonspecific doublestranded 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 medium containing 20% fetal bovine serum and 2% penicillin-streptomycin was added to each dish to a final concentration of 10% fetal bovine serum and 1% penicillin-streptomycin. Cells were harvested 24 h later for further assays.
Cell Cycle Analysis-Cells were rinsed in Dulbecco's phosphate-buffered saline (Mediatech), trypsinized, resuspended in McCoy's medium containing 10% fetal bovine serum and 1% penicillin-streptomycin, collected by centrifugation, washed with Dulbecco's phosphate-buffered saline, again collected by centrifugation, resuspended in 70% ethanol, and fixed at Ϫ20°C overnight. Cells were pelleted again by centrifugation and re-suspended in a staining solution containing 50 g/ml propidium iodide, 50 g/ml RNase A, 0.1% Triton X-100, and 0.1 mM EDTA for 30 min. Flow cytometry was performed on a FACSCalibur (BD Biosciences) cytometer.

RESULTS
G 1 /S Arrest Depends on p53 in HCT116 Cells following ␥ Irradiation-Both HCT116 p53ϩ/ϩ and p53Ϫ/Ϫ cells exhibited comparable cell cycle profiles before irradiation (Fig. 1, A,  B, and E). Following 12 Gy of ␥ irradiation, HCT116 p53ϩ/ϩ cells demonstrated a normal cell cycle arrest pattern, with ϳ15% of cells in G 1 , ϳ80% cells in G 2 , and a significantly reduced S population (Fig. 1, C and F). However, HCT116 p53Ϫ/Ϫ cells exhibited an abnormal cell cycle pattern after ␥ irradiation, with ϳ90% of the cells in G 2 and few remaining in either G 1 or S (Fig. 1, D and F). Consistent with the effect of ␥ irradiation on the cell cycle of HCT116 p53ϩ/ϩ cells, protein levels of KLF4 and p21 WAF1/CIP1 were both significantly increased in response to an increase in p53 protein levels (Fig. 2,   FIG. 2. Western blot analysis of p53, KLF4, and p21 WAF1/CIP1 in response to ␥ irradiation. The levels of p53, KLF, and p21 WAF1/CIP1 were determined by Western blot analysis in HCT116 p53ϩ/ϩ (lanes 1 and 2) and HCT116 p53Ϫ/Ϫ (lanes 3 and 4) without irradiation (lanes 1 and 3) and 24 h after 12 Gy of ␥ irradiation (lanes 2 and 4). 1. The effect of ␥ irradiation on HCT116 p53؉/؉ and HCT116 p53؊/؊ cells. In panels A through D flow cytometric analyses of HCT116 p53ϩ/ϩ and HCT116 p53Ϫ/Ϫ were performed 24 h after 0 or 12 Gy of ␥ irradiation. Cells were stained with propidium iodide, and DNA content was analyzed by flow cytometry. The DNA content of haploid and diploid cells is designated 2n and 4n, respectively. Panels E and F show the means and standard deviations of percentage of the G 1 , S, and G 2 /M populations from five independent experiments in non-irradiated and irradiated cells, respectively. Open bars represent HCT116 p53ϩ/ϩ, and closed bars represent HCT116 p53Ϫ/Ϫ. 15,000 cells were analyzed in each experiment. *, p Ͻ 0.05 compared with HCT116 p53ϩ/ϩ cells. lanes 1 and 2). In contrast, no induction in the level of either KLF4 or p21 WAF1/CIP1 was observed following ␥ irradiation in HCT116 p53Ϫ/Ϫ cells (Fig. 2, lanes 3 and 4). These results suggest that at least part of the cell cycle arrest caused by ␥ irradiation is a result of p53-dependent activation of KLF4 and p21 WAF1/CIP1 .
Inducible Expression of KLF4 in HCT116 p53Ϫ/Ϫ Cells Restores G 1 Peak-The failure of ␥ irradiation to induce expression of KLF4 and p21 WAF1րCIP1 in HCT116 p53Ϫ/Ϫ cells correlated with the reduction in G 1 and S populations (Figs. 1 and 2). This suggests that activation of KLF4, with consequent activation of p21 WAF1րCIP1 , may be necessary for the accumulation of cells in G 1 . To test this hypothesis, we established a stable HCT116 p53Ϫ/Ϫ cell line that expressed the receptors for the insect nuclear hormone, ecdysone, and its partner, retinoid X receptor (25). This cell line, called EcR116 p53Ϫ/Ϫ, was infected with the recombinant adenovirus AdEGI or AdEGI-KLF4 (25) that contained enhanced green fluorescence protein as a control or enhanced green fluorescence protein plus KLF4, respectively. Following infection, cells were ␥-irradiated or not and then treated with the inducer, ponasterone A, or vehicle alone for 24 h before being harvested for cell cycle analysis. As seen in Fig. 3, treatment of AdEGI-KLF4-infected cells with ponasterone A without irradiation resulted in a statistically significant increase in the G 1 population and a decrease in the G 2 /M population (Fig. 3, C, D, and J), whereas AdEGI-infected cells without irradiation and treated with ponasterone A had no effect on the cell cycle when compared with untreated cells (Fig. 3, A, B, and I). Cells infected with AdEGI followed by irradiation showed G 2 /M arrest in the absence or presence of ponasterone A (Fig. 3, E, F, and K) as did cells infected with AdEGI-KLF4 and irradiated without any ponasterone A treatment (Fig. 3, G and L). In contrast, upon the addition of ponasterone A, AdEGI-KLF4-infected and irradiated cells had a statistically significant increase in the G 1 population (Fig. 3, H  and L). This finding is reminiscent of the G 1 /S arrest seen in HCT116 p53ϩ/ϩ cells following irradiation (compare Figs. 3H  and 1C). Fig. 4 shows that only cells infected by AdEGI-KLF4 and induced with ponasterone A (lanes 11 and 12) had appreciable amounts of KLF4. The increase in the KLF4 level correlated with an increase in the p21 WAF1/CIP1 level, a finding consistent with our previous observation that KLF4 is an activator of p21 WAF1րCIP1 expression (25). The combined results of Figs. 3 and 4 indicate that the inducible expression of KLF4 in irradiated cells lacking p53 restores the characteristic G 1 /S arrest in cells with wild type p53 following irradiation. This finding indicates that KLF4 is necessary and sufficient in mediating the G 1 cell cycle effect of p53 following DNA damage.

DISCUSSION
Cell cycle progression is regulated by checkpoint controls, which function to safeguard the integrity of the genome. Activation of DNA integrity checkpoints occurs through the detection of damaged or unreplicated DNA and is in effect until DNA damage has been repaired (30). The checkpoint that arises after DNA damage can activate during G 1 , S, or G 2 (3,31). Arrest in G 1 permits repair prior to replication, whereas arrest in S or G 2 permits repair of the genome before mitotic segregation. The p53 tumor suppressor has been shown to be integral to both the G 1 (32,33) and G 2 (34, 35) DNA damage machinery. This was supported by the results in Fig. 1, which showed that HCT116 p53ϩ/ϩ cells arrested at either G 1 or G 2 /M after ␥ irradiation as expected for cells with intact checkpoint function. The resultant activation of p53 because of ␥ irradiation was accompanied by a significant increase in the level of KLF4 and p21 WAF1/CIP1 (Fig. 2) in a manner similar to the previously observed response of fibroblasts subjected to DNA damage caused by methyl methanesulfonate (24). HCT116 p53Ϫ/Ϫ cells, in contrast, showed no induction of either KLF4 or p21 WAF1/CIP1 by ␥ irradiation and arrested only in G 2 /M (Fig. 1). The latter result was consistent with that from a previous study, which also demonstrated that p53 was necessary to sustain G 2 arrest (10).
Recent studies indicate that the G 1 checkpoint control after DNA damage consisted of two steps (31). The first step is a rapid and p53-independent induction of the G 1 checkpoint. It is a result of rapid redistribution of p21 WAF1/CIP1 from cyclin D1-Cdk4/6 complexes to cyclin E-Cdk2 complexes, which are inhibited by p21 WAF1/CIP1 (36,37). The second step involves the post-translational modifications of p53 by upstream protein kinases, including ataxia telangiectasia mutated/ataxia telangiectasia and Rad3 related and Chk1/Chk2 (38,39), which results in p53 activation and subsequent transcriptional induction of p21 WAF1րCIP1 (31). Several lines of evidence suggest that KLF4 is involved in the p53-dependent induction of p21 WAF1րCIP1 . First, p53 mediates the transcriptional induction of KLF4 in response to DNA damage (24). Second, the induction in KLF4 precedes that in p21 WAF1րCIP1 following DNA damage (24). Third, KLF4 binds to a specific cis-element in the proximal promoter of the p21 WAF1րCIP1 gene and activates the promoter (24). Fourth, p53 and KLF4 physically interact and cause a synergistic induction in p21 WAF1րCIP1 gene expression (24). The importance of KLF4 in mediating the transcriptional induction of p21 WAF1րCIP1 is further demonstrated by the observation that p53 fails to activate the p21 WAF1րCIP1 promoter if the KLF4 response element in the promoter is mutated (24).
In addition to the biochemical evidence supporting a crucial role for KLF4 in mediating the transcriptional induction of p21 WAF1րCIP1 by p53, the present study provides the genetic evidence to further substantiate the significance of KLF4 in p53-mediated G 1 arrest caused by DNA damage. Specifically, the conditional induction of KLF4 in ␥-irradiated HCT116 p53Ϫ/Ϫ cells restored the G 1 population of cells that are normally present in irradiated HCT116 p53ϩ/ϩ cells (Fig. 3). Conversely, inhibition of KLF4 expression in irradiated HCT116 p53ϩ/ϩ cells resulted in an abolishment of the G 1 peak in a manner that resembles the consequence of ␥ irradiation of HCT116 p53Ϫ/Ϫ cells (Fig. 6). In each case, the induction or inhibition of KLF4 expression was accompanied by a corresponding increase or decrease, respectively, in the level of FIG. 5. The effect of KLF4 siRNA on protein levels of p53, KLF4, and p21 WAF1/CIP1 in HCT116 p53؉/؉ cells following ␥ irradiation. The levels of p53, KLF4, and p21 WAF1/CIP1 were determined in ␥-irradiated (even lanes) or non-irradiated (odd lanes) HCT116 p53ϩ/ϩ cells that were untransfected (lanes 1 and 2), mock-transfected (lanes 3 and 4), transfected with 2 (lanes 5 and 6) or 4 g (lanes 7 and 8) of KLF4-specific siRNA, or transfected with 2 (lanes 9 and 10) or 4 g (lanes 11 and 12) of nonspecific siRNA. Cells were harvested 24 h later for Western blot analysis of protein levels.
FIG. 6. The effect of KLF4 siRNA on the cell cycle in HCT116 p53؉/؉ cells following ␥ irradiation. Flow cytometric analyses were performed in HCT116 p53ϩ/ϩ cells that were non-irradiated (panels A-E) or irradiated (panels F-J) followed by mock transfection (panels A and F), transfection with 2 (panels B and G) or 4 g (panels C and H) of KLF4specific siRNA, or transfection with 2 (panels D and I) or 4 g (panels E and J) of nonspecific siRNA. Cells were studied for 24 h following irradiation. Panels K and L show the means and standard deviations of percent cells in G 1 phase for each of the treatments in five independent experiments. 15,000 cells were analyzed in each experiment. *, p Ͻ 0.05 compared with mock-transfected cells. p21 WAF1/CIP1 (Figs. 4 and 5). Coupled with the findings from our previous study, which demonstrated that inducible expression of KLF4 causes a G 1 /S cell cycle arrest (25), it is highly likely that KLF4 serves a pivotal role in mediating the G 1 checkpoint function of p53 in response to DNA damage.
In addition to its effect on the G 1 /S checkpoint, p53 also regulates the G 2 /M transition in response to DNA damage (11). Part of the mechanism by which p53 inhibits the G 2 checkpoint involves inhibition of Cdc2, the cyclin-dependent kinase required to enter mitosis (12). Binding of Cdc2 to cyclin B1 is required for its activity, and repression of the cyclin B1 gene by p53 contributes to the blocking of entry into mitosis (40,41). p53 also represses expression of the Cdc2 gene (42,43) to help ensure that cells do not escape from the initial block. Moreover, several of the transcriptional targets of p53 can inhibit Cdc2, including p21 WAF1/CIP1 , 14-3-3, and Gadd45 (44 -46). Therefore, it is of great interest to note that a recent analysis of KLF4 target genes by cDNA microarrays showed that KLF4 activates expression of 14-3-3, in addition to p21 WAF1րCIP1 , and represses expression of Cdc2 (47). Whether KLF4 is also involved in mediating the G 2 checkpoint function of p53 in response to DNA damage is currently being determined.