Negative regulation of Chk2 expression by p53 is dependent on the CCAAT-binding transcription factor NF-Y.

The kinase Chk2 and tumor suppressor p53 participate in an ill defined regulatory interaction in mammalian cells. The abundance of Chk2 mRNA and protein has now been shown to be decreased by the induction of p53 in Saos2 cells. Ionizing radiation also triggered the phosphorylation and subsequent down-regulation of Chk2 in human colorectal HCT116 (p53(+/+)) cancer cells; irradiation of its isogenic mutant HCT116 (p53(-/-)) cells, which lack functional p53, induced Chk2 phosphorylation but not its down-regulation. In addition, HCT116 (p53(+/+)) cells constitutively expressing a dominant negative p53 (V143A) failed to suppress Chk2 expression after irradiation. Reporter gene assays in HCT116 (p53(+/+)) cells revealed that wild-type p53 repressed, whereas a dominant negative p53 mutant increased, the activity of the human Chk2 gene promoter. Mutational analysis showed that a CCAAT box located between nucleotides -152 and -138 of the promoter was responsible for its negative regulation by p53. Electrophoretic mobility shift assays demonstrated that the transcription factor NF-Y binds to this CCAAT sequence. A dominant negative mutant of NF-YA abolished the effect of p53 on Chk2 promoter activity. These results suggest that p53 negatively regulates Chk2 gene transcription through modulation of NF-Y function and that this regulation may be important for reentry of cells into the cell cycle after DNA damage is repaired.

The kinase Chk2 and tumor suppressor p53 participate in an ill defined regulatory interaction in mammalian cells. The abundance of Chk2 mRNA and protein has now been shown to be decreased by the induction of p53 in Saos2 cells. Ionizing radiation also triggered the phosphorylation and subsequent down-regulation of Chk2 in human colorectal HCT116 (p53 ؉/؉ ) cancer cells; irradiation of its isogenic mutant HCT116 (p53 ؊/؊ ) cells, which lack functional p53, induced Chk2 phosphorylation but not its down-regulation. In addition, HCT116 (p53 ؉/؉ ) cells constitutively expressing a dominant negative p53 (V143A) failed to suppress Chk2 expression after irradiation. Reporter gene assays in HCT116 (p53 ؉/؉ ) cells revealed that wild-type p53 repressed, whereas a dominant negative p53 mutant increased, the activity of the human Chk2 gene promoter. Mutational analysis showed that a CCAAT box located between nucleotides ؊152 and ؊138 of the promoter was responsible for its negative regulation by p53. Electrophoretic mobility shift assays demonstrated that the transcription factor NF-Y binds to this CCAAT sequence. A dominant negative mutant of NF-YA abolished the effect of p53 on Chk2 promoter activity. These results suggest that p53 negatively regulates Chk2 gene transcription through modulation of NF-Y function and that this regulation may be important for reentry of cells into the cell cycle after DNA damage is repaired.
The fidelity of chromosome segregation is achieved by checkpoint controls that prevent cell cycle progression when damage to DNA is detected (1,2). Cell cycle arrest in response to DNA damage in mammals is mediated predominantly through the transcriptional activation of specific genes by the tumor suppressor protein p53 (3,4), the gene for which is commonly mutated in diverse types of human cancer (5). This nuclear phosphoprotein directly activates a subset of genes important for cell cycle arrest (3,4) or apoptosis (6 -9) by binding to specific nucleotide sequences located within the promoter region or first intron of these genes (10). In addition, p53 functions as a negative regulator of various genes, although the promoters of most of these genes do not contain a typical consensus binding site for this protein (11). The mechanism by which p53 exerts this inhibition is thought to involve physical interaction either with basal components of the transcription machinery (12,13) or with other unidentified factors (14,15), but its molecular basis has remained unclear.
Both the abundance and transactivation activity of p53 are regulated primarily by posttranslational modifications such as phosphorylation, acetylation, SUMO conjugation, and ubiquitylation (16 -18). The phosphorylation of p53 at several sites within its transactivation domain is thus rapidly induced on exposure of human cells to genotoxic stresses such as ionizing radiation (19). Both the kinases ataxia telangiectasia mutated (ATM) and ATM-and Rad3-related phosphorylate human p53 on Ser 15 , and this reaction appears to stabilize p53 by inhibiting its interaction with Hdm2 (the human ortholog of mouse Mdm2) (20,21). If DNA damage is too severe for repair, p53 is phosphorylated on Ser 46 by homeodomein-interacting protein kinase 2, resulting in specific activation of the gene for p53AIP1 and subsequent apoptosis (8). Thus, although the consequences of p53 phosphorylation on specific residues are not fully characterized, such modifications probably not only regulate the transactivation activity of p53 but also determine the target gene specificity of this activity.
The kinase Chk2, the mammalian ortholog of Saccharomyces cerevisiae Rad53 (22,23) and Schizosaccharomyces pombe Cds1 (24), is implicated in the DNA damage checkpoint operative at the G 2 phase of the cell cycle (25)(26)(27). Chk2 is thus rapidly activated as a result of its phosphorylation by ATM in response to the induction of DNA double strand breaks by ionizing radiation, and the activated enzyme is able to phosphorylate Cdc25A, Cdc25B, and Cdc25C in vitro (25)(26)(27)(28)(29)(30). In addition, a potential role for Chk2 in the G 1 cell cycle checkpoint has been suggested by the observation that overexpression of Chk2 induced G 1 arrest in immortalized or transformed cells (31). Chk2 phosphorylates human p53 at multiple sites, including Ser 20 , in vitro (31,32). However, phosphorylation of Ser 20 does not inhibit the interaction of human p53 with Hdm2 in vitro, and substitution of the equivalent residue (Ser 23 ) of murine p53 with Ala failed to prevent the accumulation or activation of p53 in embryonic stem cells in response to ionizing radiation (33). The physiological relevance of phosphorylation of human p53 on Ser 20 has therefore remained unclear.
Recent observations with Chk2 knockout (Chk2 Ϫ/Ϫ ) mice have implicated Chk2 in p53-dependent apoptosis (34 -36). Both cell cycle arrest at G 1 and p53-dependent transcriptional activation in response to ionizing radiation are impaired in embryonic fibroblasts from Chk2 Ϫ/Ϫ mice. These observations thus suggested that Chk2 is an upstream regulator of p53 function. Furthermore, the abundance of Chk2 mRNA was shown to be inversely related to p53 status both in cell lines (27) and in specimens of human gastric cancer (37), suggesting the possibility that Chk2 is also a downstream target of p53. We have now investigated the regulatory interaction between Chk2 and p53. We show here that p53 negatively regulates Chk2 expression at the level of Chk2 gene transcription and that the transcription factor NF-Y appears to play an important role in this regulation.
Northern and Immunoblot Analyses-For Northern blot analysis, total RNA was isolated from Saos2 cells with the use of Isogen (Nippon Gene). The RNA (15 g) was denatured, fractionated by electrophoresis on a 1% agarose-formaldehyde gel, and transferred to a nylon membrane. Chk2, p53, and glyceraldehyde-3-phosphate dehydrogenase cDNAs were labeled with [␣-32 P]dCTP with the use of a random primer labeling kit (Amersham Biosciences). The membranes were then subjected to hybridization with the labeled cDNAs, and hybridization signals were detected by autoradiography.
Flow Cytometry-Cells were harvested by exposure to trypsin, washed with phosphate-buffered saline, and fixed with 70% ethanol at Ϫ20°C. They were then stained for 30 min in the dark with phosphatebuffered saline containing propidium iodide (30 g/ml) and RNase A (5 g/ml) before analysis of DNA content with a FACScan flow cytometer and Cell Quest software (Becton Dickinson).
Cloning of the 5Ј-Flanking Region of Human CHK2-A 6.0-kb fragment of the 5Ј regulatory region of the human Chk2 gene (CHK2) was generated by PCR from genomic DNA of HeLa cells. The PCR primers, 5Ј-ACTGAATATAGTCTAACGGAGAACCCTTGGA-3Ј (forward) and 5Ј-AAGCGAAGCTCAGGAGACTCCGTTCGCACA-3Ј (reverse), were designed on the basis of sequence information deposited in GenBank TM (accession number AL117330). The PCR product was subcloned into the pGEM-T easy vector (Promega), yielding pGEM-T Chk2 6.0, and was sequenced for verification.
RNase Protection Assay-A 274-bp DNA fragment was generated by PCR with the primers 5Ј-CTAATGTTGCTGATTGGCTG-3Ј (nucleotides Ϫ160 to Ϫ141) and 5Ј-ATATGACTCACCGCGTGAGC-3Ј (nucleotides 95-114) and with pGEM-T Chk2 6.0 as the template and was then cloned into pGEM-T easy. The resulting vector was linearized with NcoI and an antisense RNA probe (370 bp) was synthesized by SP6 RNA polymerase in the presence of [␣-32 P]UTP with the use of a MAXIscript in vitro transcription kit (Ambion). An RNase protection assay was performed with 10 g of total RNA from U937 cells and an RPAIII kit (Ambion). The size of protected fragments was determined by electrophoresis and comparison with the electrophoretic mobilities of molecular size markers and a sequencing ladder obtained from the single-stranded circular DNA of bacteriophage M13mp18, which was supplied with the Sequenase version 2.0 DNA sequencing kit (Upstate, Lake Placid, NY).
Transient Transfection and Reporter Assay-HCT116 cells were grown to 70 -80% confluence in 12-well plates and then transfected, with the use of the Trans IT Transfection Reagent (Mirus), with 0.1 g of human CHK2 promoter-luciferase reporter construct, 0.3 g of a ␤-galactosidase expression vector (pCMV␤-gal; an internal control for normalization of transfection efficiency), 0.6 g of pcDNA3 encoding either wild-type human p53, a dominant negative mutant, or a phosphorylation site mutant thereof, and, where indicated, 3 g of pNF-YA29, a dominant negative mutant of NF-YA. Twenty-four hours after transfection, the cells were harvested in Reporter lysis buffer (Promega) and assayed for luciferase and ␤-galactosidase activities with the use of a Bright-Glo luciferase assay system and ␤-galactosidase enzyme assay system (Promega). All assays were performed at least three times in duplicate, and representative data are presented.
Electrophoretic Mobility Shift Assay Analysis-Electrophoretic mobility shift assays were performed as described (15). In brief, doublestranded oligonucleotides containing either wild-type or mutant CCAAT (nucleotides Ϫ158 to Ϫ134) or C/EBP 1 (Ϫ177 to Ϫ153) motifs of the human CHK2 promoter region were synthesized in vitro. The sequences of the oligonucleotides were as follows: wild-type CCAAT motif, 5Ј-AATGTTGCTGATTGGCTGGGGAGTC-3Ј; mutant CCAAT motif, 5Ј-AATGTTGCTGCAGCTCTGGGGAGTC-3Ј; and wild-type C/EBP motif, 5Ј-GAGAGCGTCTAACCAGACTAATGTT-3Ј. The wild-type CCAAT oligonucleotide was labeled with [␥-32 P]ATP by T4 polynucleotide kinase. Nuclear extracts of HeLa cells were prepared as described (40), and 5 g of extract protein were incubated for 30 min on ice with 1 g of poly(dI-dC) (Amersham Biosciences), in the absence or presence of an unlabeled competitor oligonucleotide, in a final volume of 25 l of a solution containing 25 mM HEPES-NaOH (pH 7.9), 20 mM KCl, 30 mM NaCl, 0.5 mM EDTA, 0.25 mM dithiothreitol, and 10% glycerol. The labeled probe (ϳ25,000 cpm) was added to the mixture, which was then incubated for an additional 20 min at room temperature. DNA-protein complexes were separated by electrophoresis on a 6% polyacrylamide gel. For "supershift" analysis, 2 g of antibodies specific for the A or B subunits of NF-Y (Rockland) or for C/EBP␣,␤,␦ or C/EBP␤ (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) were incubated with the DNA/ protein mixtures.

RESULTS
Negative Regulation of Chk2 Expression by p53-We have previously shown that the abundance of Chk2 mRNA in cancer cells was inversely related to their p53 status (27). Furthermore, transformation of normal cells by SV40 or adenovirus E6 up-regulated the amount of Chk2 mRNA and reintroduction of wild-type p53 reversed this up-regulation (27), suggesting that wild-type p53 inhibits the expression of Chk2. We therefore investigated this possibility further with the use of Tet-on Saos2 cells in which the expression of p53 is inducible by doxycyclin. Cells were lysed at various times after incubation with doxycyclin, and the amounts of Chk2 and p53 mRNA and protein were determined by Northern blot and immunoblot analyses. As expected, the abundance of both p53 mRNA and p53 protein increased with time of exposure to doxycyclin, with this effect being apparent as early as 2 h and maximal at 48 h (Fig. 1A). In contrast, the abundance of both Chk2 mRNA and protein decreased with a time course similar to that apparent for up-regulation of p53 expression, with Ͻ10% of the original amounts remaining after 48 h. As controls, the amounts of glyceraldehyde-3-phosphate dehydrogenase mRNA and ␤-actin were not affected by doxycyclin treatment.
The expression of Chk2 has been shown to be regulated in a cell cycle-dependent manner, peaking at G 2 -M phases and decreasing in G 1 (27). It was thus possible that the increased expression of p53 triggered by doxycyclin resulted in downregulation of Chk2 expression as a result of the induction of G 1 arrest. To examine this possibility, we monitored cell cycle progression in the doxycyclin-treated cells by flow cytometry. The up-regulation of p53 expression did not arrest the cell cycle of Saos2 cells at G 1 phase (Fig. 1B); it did, however, increase the size of the cell population with a DNA content of less than 2 N. The p53-induced down-regulation of Chk2 expression was thus not likely to be due to G 1 arrest of the cell cycle by p53.
p53-dependent Down-regulation of Chk2 by Genotoxic Stress-Given that various types of genotoxic stress induce the accumulation of p53 (5), we next examined whether such stress also triggers down-regulation of Chk2. Both HCT116 (p53 ϩ/ϩ ) cells expressing wild-type p53 and its isogenic mutant HCT116 (p53 Ϫ/Ϫ ) cells lacking p53 were exposed to ionizing radiation, and the abundance of p53, Chk2, and ␣-tubulin was then determined by immunoblot analysis. As expected, the amount of p53 in HCT116 (p53 ϩ/ϩ ) cells was increased as early as 2 h after irradiation, presumably as a result of protein stabilization ( Fig. 2A). Also consistent with previous observations, Chk2 was phosphorylated (as revealed by a shift in electrophoretic mobility) in both cell types within 1 h of irradiation, confirming that ionizing radiation induced the activation of ATM in these cells. However, whereas the amount of Chk2 decreased in a time-dependent manner in irradiated HCT116 (p53 ϩ/ϩ ) cells, it remained unchanged in irradiated HCT113 (p53 Ϫ/Ϫ ) cells. The reduction in Chk2 abundance in HCT116 (p53 ϩ/ϩ ) cells was first detected at 2 h, with Ͻ30% of the protein remaining at 8 h. Similar reduction of Chk2 expression was also observed in HCT116 (p53 ϩ/ϩ ) but not in HCT116 (p53 Ϫ/Ϫ ) cells after UV irradiation, although the reduction of Chk2 and the accumulation of p53 in HCT116 (p53 ϩ/ϩ ) cells occurred somewhat later when compared with x-ray irradiation. Agents that affect cell cycle progression, such as hydroxyurea and aphidicolin, failed to induce p53 accumulation or to reduce the amount of Chk2 in HCT116 (p53 ϩ/ϩ ) cells (data not shown). HCT116 (p53 ϩ/ϩ ) cells expressing a dominant negative mutant of p53 (V143A) failed to down-regulate Chk2 protein level after x-ray irradiation. These results thus indicated that the accumulation of p53 induced by genotoxic stress triggers the down-regulation of Chk2 in HCT116 (p53 ϩ/ϩ ) cells.
Isolation and Characterization of the Human CHK2 Promoter Region-A full-length human Chk2 cDNA was used as a probe to screen a human genomic DNA library. 5Ј-rapid amplification of cDNA ends analysis of the positive genomic clones resulted in the identification of a 5Ј-noncoding exon and the consequent localization of the translational initiation codon (ATG) in exon 2 of CHK2. We isolated a 6.0-kb fragment of the CHK2 promoter region by PCR with genomic DNA from HeLa cells and primers based on available sequence information. To identify the transcriptional initiation site of CHK2, we performed an RNase protection assay with total RNA isolated from U937 human leukemia cells. An antisense RNA probe was synthesized from a PCR product that was obtained by amplification of the 6-kb fragment of the CHK2 promoter with primers thought to span the transcriptional start site. A protected RNA fragment of 103 bp was detected after treatment with RNase (Fig. 3A), indicating that the transcriptional initiation site (designated position ϩ1) is located 7151 bp upstream of the translation start site. To identify potential regulatory elements in the CHK2 promoter, we sequenced the region upstream of exon 1 (Fig. 3B). A typical TATA-related sequence was not apparent in the region immediately 5Ј to the transcription start site. Computer analysis did, however, reveal consensus binding sites for several transcription factors, including SP1, CCAAT box, C/EBP, AP1, and E2F sites.
To elucidate the mechanism by which p53 negatively regulates Chk2 expression, we generated several deletion and point mutant constructs of the 5Ј-flanking region of CHK2 and inserted them into a luciferase reporter plasmid. The resulting vectors were then introduced into HCT116 (p53 ϩ/ϩ ) and HCT116 (p53 Ϫ/Ϫ ) cells by transient transfection together with pCMV␤-gal (a control for normalization of transfection efficiency) and with an expression vector for wild-type human p53, a dominant negative mutant (V143A), or a phosphorylation site mutant (S20A) thereof. The relative luciferase activities of the various reporter constructs were then determined (Fig. 4). The activity of the largest construct, p[Ϫ2848,Ϫ24], containing nucleotides Ϫ2848 to Ϫ24 of CHK2, was dramatically suppressed by wild-type p53 but enhanced by the dominant negative mutant of p53 in HCT116 (p53 ϩ/ϩ ) cells, whereas such increased activity by the dominant negative mutant was not observed in HCT116 (p53 Ϫ/Ϫ ) cells. The phosphorylation mutant of p53 (S20A) suppressed the reporter activity to the same extent as did wild-type. This result was consistent with the previous report in which S20A still retains transcriptional activity but lacks the ability to be accumulated upon DNA damage (31,32). Similar effects of wild type and mutants of p53 on the reporter activity were obtained with the construct p[Ϫ160,Ϫ24], indicating that the regulatory element targeted by p53 is located between nucleotides Ϫ160 and Ϫ24 of CHK2. Further deletion of 53 bp from the 5Ј-end of this region (p[Ϫ107,Ϫ24]) resulted in an 60% decrease in basal transcriptional activity as well as a complete loss of p53 sensitivity. These results thus further localized both the p53-responsive element and basal promoter elements to the region between nucleotides Ϫ160 and Ϫ108 of CHK2.
This region of the upstream sequence of CHK2 contains a CCAAT box and a putative SP1 binding site. Mutation of the CCAAT box resulted in a complete loss of p53 sensitivity as well as a marked decrease in basal transcription (Fig. 4). In contrast, although mutation of the SP1 site resulted in a substantial decrease in basal transcriptional activity, it did not abolish p53 sensitivity. These results suggested that p53 inhibits the transcriptional activity of the CHK2 promoter through regulation of a transcription factor (or factors) that binds to a CCAAT box in the promoter region. Although SP1 also appears to contribute to the basal level of CHK2 expression, it probably does not mediate the effect of p53. Furthermore, although we did not detect a substantial effect of the AP1 or C/EBP sites on the transcriptional activity of CHK2 (data not shown), we cannot exclude the possibility that these sites also contribute to the regulation of CHK2 expression.
Negative Regulation of the CHK2 Promoter by p53 through Inhibition of NF-Y-Given that the transcription factor NF-Y interacts with typical CCAAT boxes (41), we next investigated the ability of the p53-sensitive CCAAT box in the CHK2 promoter to bind NF-Y. A 32 P-labeled oligonucleotide containing the wild-type CCAAT box sequence of the CHK2 promoter was synthesized for use as a probe in electrophoretic mobility shift assay analysis, and oligonucleotides containing either a mutant version of the CCAAT box or the C/EBP binding site located immediately upstream of the CCAAT box were prepared as competitors. Proteins present in nuclear extracts of HeLa cells formed complexes with the 32 P-labeled probe (Fig.  5). The formation of these complexes was inhibited by unlabeled probe in a dose-dependent manner, but it was not affected by either of the unlabeled competitor oligonucleotides containing the mutated CCAAT box or the C/EBP site.
We next examined which transcription factors contributed to the complexes formed with the CCAAT box probe by supershift analysis with antibodies specific for NF-YA, NF-YB, C/EBP␤, or C/EBP␣,␤,␦. The antibodies to NF-YA or to NF-YB reduced the mobility of the protein-probe complexes, although the effect of the antibodies to NF-YB was less pronounced than was that of those to NF-YA (Fig. 5). In contrast, the antibodies to C/EBP failed to affect the mobility of the protein-probe complexes. None of the antibodies affected the mobility of the probe in the absence of nuclear extract (data not shown), excluding the possibility that the antibodies to NF-Y interacted directly with the oligonucleotide.
To confirm the role of NF-Y in regulation of the CHK2 promoter, we examined the effect of a dominant negative NF-Y FIG. 5. Identification of NF-Y as a transcription factor that binds to the p53-sensitive CCAAT box of the CHK2 promoter. Electrophoretic mobility shift assay analysis was performed with nuclear extract of HeLa cells and a 32 P-labeled oligonucleotide probe containing the p53-sensitive CCAAT box of the CHK2 promoter (nucleotides Ϫ158 to Ϫ134). For competition assays, 5-or 50-fold molar excesses of the unlabeled probe (wt) or of oligonucleotides containing either a mutant CCAAT box (mut) or the adjacent C/EBP binding site, as indicated, were included in the reaction mixture together with the labeled probe. For supershift assays, antibodies specific for NF-YA, NF-YB, C/EBP␤, or C/EBP␣,␤,␦ were included in the reaction mixture.
The arrows indicate specific complexes formed by the labeled probe and proteins present in the nuclear extract as well as supershifted complexes formed in the additional presence of antibodies to NF-Y. The first lane corresponds to a control reaction performed in the absence of nuclear extract. mutant (NF-YA13m29) on p53-mediated repression of CHK2 expression. Three essential amino acids in the DNA binding domain of NF-YA were mutated to yield NF-YA13m29; transcriptional protein complexes containing this mutant have been shown to be functionally inactive both in vitro and in vivo (42). A172 cells (p53 status: wild type) were transfected with an expression vector for the NF-YA mutant together with the p[Ϫ2848,Ϫ24] reporter construct and vectors for either wildtype or dominant negative forms of p53. Measurement of the luciferase activity of lysates of the transfected cells revealed that the NF-YA mutant inhibited the basal transcriptional activity of the CHK2 promoter in the absence of wild-type or mutant p53 (Fig. 6); it did not affect SV40 promoter activity measured as a control (data not shown). The NF-YA mutant also blocked the stimulatory effect of the dominant negative mutant of p53 as well as the inhibitory effect of wild-type p53 on CHK2 promoter activity.
To examine whether functional p53 might affect the expression of NF-YA, we analyzed the expression of this protein in HCT116 (p53 ϩ/ϩ ) and HCT116 (p53 Ϫ/Ϫ ) cells after x-ray irradiation or UV treatment. The expression level of NF-YA was similar in both cells (Fig. 7). Taken together, these results suggested that p53 represses CHK2 transcription by targeting the interaction of NF-Y with the cognate CCAAT box of the CHK2 promoter. DISCUSSION Several lines of evidence have suggested the existence of an interaction between Chk2 and p53. For example, p53-dependent transcription is markedly inhibited in Chk2-deficient cells (34 -36), and heterozygous germ line mutations in CHK2 have been detected in a subset of individuals with Li-Fraumeni syndrome whose p53 alleles are intact (43). We have now shown that Chk2 expression is negatively regulated by p53 and that this effect is mediated at the level of the transcription factor NF-Y, indicating the existence of a negative feedback loop between p53 and Chk2.
We have previously shown that the abundance of Chk2 mRNA was inversely correlated with p53 status, being low in cells expressing wild-type p53 and high in cells lacking functional p53 (27). A similar inverse relation between CHK2 expression and p53 status was recently demonstrated in tumor samples from individuals with gastric cancer (37). However, it has remained unclear whether the Chk2 gene is a direct target for transcriptional regulation by p53 or whether other effects of p53, such as regulation of cell cycle progression, might be indirectly responsible for repression of Chk2 expression. To clarify this issue, we examined Tet-on Saos2 cells in which the expression of p53 is inducible by doxycyclin and found that the induction of p53 resulted in down-regulation of the abundance of both Chk2 mRNA and protein. Our previous data showed that the expression levels of both Chk2 and Chk1 are reduced in G 1 and increased in S to M phases of the cell cycle (27,44). However, the down-regulation of Chk2 mRNA and protein in response to p53 induction in Tet-on Saos2 cells was not an indirect consequence of cell cycle arrest in G 1 ; p53 induction thus did not affect the proportion of cells in G 1 , but rather induced apoptosis in these cells. In addition, whereas exposure to ionizing radiation induced rapid phosphorylation of Chk2 in HCT116 cells that express or lack functional p53, the subsequent down-regulation of Chk2 abundance was apparent only in the former cells. These results thus indicate that the negative regulation of Chk2 expression by p53 occurs under physiological conditions.
A similar down-regulation of Chk1 expression by p53 was recently described (45). This regulation required functional p21 (a cyclin-dependent kinase inhibitor) and pRb (the retinoblastoma protein), suggesting that it might be mediated at the level of E2F-dependent transcription of the Chk1 gene. Overexpression of p21 failed to affect the abundance of Chk2 mRNA or protein in Saos2 cells (data not shown), however, suggesting that the inhibitory effects of p53 on Chk1 and Chk2 expression are mediated by distinct mechanisms. Down-regulation of Chk2 protein as a result of genotoxic stress-induced p53 expression was not obvious in this previous study (45). Although the reason for this apparent discrepancy with our data remains unclear, it might be due to differences in the extent of accumulation of p53 in response to stress signals. Indeed, the abundance of p53 has been shown to determine the specificity of target gene activation (46). Alternatively, the genotoxic stresses imposed in the previous study (45) may not have been sufficient to induce full activation of Chk2, given that rapid phosphorylation of Chk2 was not obvious. Although Chk2 is dispensable for both p53 stabilization and p53-mediated G 1 arrest, phosphorylation of p53 by activated Chk2 might determine target gene specificity and be indispensable for p53-mediated down-regulation of Chk2 expression.
Although a typical p53 binding sequence was not apparent in a 2848-bp region immediately upstream of the transcription start site of human CHK2, a transient transfection assay with a luciferase reporter gene in HCT116 (p53 ϩ/ϩ ) and HCT116 (p53 Ϫ/Ϫ ) cells revealed that CHK2 transcription was repressed by wild-type p53 and was enhanced by a dominant negative mutant of p53, the latter effect presumably being due to interference with the endogenous wild-type p53 present in HCT116 (p53 ϩ/ϩ ) cells. Consistent with this notion, the enhancement of CHK2 transcription was not obvious in HCT116 (p53 Ϫ/Ϫ ) cells. Subsequent deletion and point mutation analyses resulted in the identification of a CCAAT box as a cis-acting element in CHK2 necessary for both basal promoter activity and negative regulation by p53. Transcriptional repression by p53 has been suggested to be mediated by a direct interaction with the TATA-binding protein (12,13) or with TATA-binding proteinassociated factors (47,48), resulting in inhibition of the basal transcription machinery. However, p53-induced repression of the transcription of certain genes has been shown to be dependent on the existence of a cis-acting element in the promoter. For example, p53 represses C/EBP-mediated transactivation of the albumin gene promoter (14). Furthermore, repression of Hsp70 gene transcription by p53 is mediated through an interaction between p53 and CCAAT-binding protein (CBF, NF-Y), a transcriptional activator of the Hsp70 gene promoter (49,50).
The eukaryotic transcription factor NF-Y specifically recognizes the CCAAT regulatory element in either orientation in the proximal or distal enhancer regions of many genes (41). NF-Y is a heterotrimeric complex of NF-YA, NF-YB, and NF-YC subunits, all of which are required for binding to the CCAAT box (51). We have now identified NF-Y as a protein that binds to the p53-sensitive CCAAT box located at nucleotides Ϫ152 to Ϫ138 in the human CHK2 promoter. In addition, p53-dependent repression of CHK2 transcription was abolished (and basal CHK2 promoter activity was reduced) by expression of a dominant negative mutant of NF-YA in A172 cells. These results thus indicate that NF-Y plays an essential role both in p53-mediated repression of CHK2 transcription as well as in the basal activity of this gene. Transcription of the cdc2 gene has also been shown to be negatively regulated by p53 at the level of NF-Y (15). Furthermore, the crystal structure of the NF-YB-NF-YC dimer implicated the ␣-helical structure of NF-YC in binding to regulatory proteins such as MYC and p53 (52). In this regard, we could not detect the direct interaction between p53 and NF-Y complex by the immunoprecipitation-Western method under physiological conditions in HCT116 (p53 ϩ/ϩ ) cells (data not shown). This result, however, cannot exclude the possibility that the interaction between p53 and NF-Y complex might be transient but still regulates the transcriptional activity of NF-Y. NF-Y may thus be a common target of p53 in its role as a transcriptional repressor, although a detailed mechanism by which p53 regulates NF-Y complex is still elusive.
Several feedback loops between p53 and factors that regulate p53 function have been identified. The best characterized of these regulatory circuits is that involving p53 and Mdm2, in which p53 induces the expression of Mdm2 and Mdm2 then targets p53 for degradation (53). Our present results indicate the existence of a similar regulatory circuit involving p53 and Chk2; in this instance, Chk2 activated in response to DNA damage phosphorylates and activates p53, which, in turn, down-regulates the expression of Chk2. This down-regulation may allow the shut down of Chk2 pathways once DNA damages are repaired, thereby ensuring reentry into the cell cycle.