p130/E2F4 Binds to and Represses the cdc2 Promoter in Response to p53*

p53 represses the transcription of cdc2 and cyclin B1 , causing loss of Cdc2 activity and G 2 arrest. Here we show that the region 2 22 to 2 2 of the cdc2 promoter called the R box is required for repression by p53 but not for basal promoter activity. The R box confers p53-de-pendent repression on heterologous promoters and binds to p130/E2F4 in response to overexpression of p53. R box-dependent repression requires p21/waf1, and overexpression of p21/waf1 also represses the cdc2 promoter. These observations suggest that p53 represses the cdc2 promoter by inducing p21/waf1, which inhibits cyclin-dependent kinase activity, enhancing the binding of p130 and E2F4, which together bind to and repress the cdc2 promoter. 10% fetal bovine serum (Life Technologies, Inc.) in a humidified atmosphere containing 10% CO 2 . The TR9-7 cell line, expressing a tetracycline-regulated p53 cDNA, was derived from MDAH041 cells (4). Cells were arrested in G 2 in response to p53 as described by Taylor et al. (28). Briefly, mimosine (Sigma), which arrests cells reversibly at the beginning of S-phase (49), was added at a final concentration of 200 m M to asynchronously growing TR9-7 cells for 48 h. The cells were released from the mimosine block and simultaneously

The p53 tumor suppressor protects mammals from neoplasia by eliminating cells containing damaged DNA through apoptosis or cell cycle arrest (1,2). p53-dependent arrest in response to DNA damage occurs in both G 1 and G 2 phases of the cell cycle (3)(4)(5)(6). p53 also controls progression through S phase when nucleotide pools are out of balance, thus avoiding DNA damage, and inhibits entry into mitosis when DNA synthesis is blocked (7)(8)(9). Damaged or unreplicated DNA, imbalances in nucleotide pools, or hypoxia cause the p53 protein to accumulate and activate its ability to bind to promoters, activating transcription of the genes that control several cellular responses to stress (1,2,9,10). For example, p53 causes G 1 arrest in part by stimulating transcription of the cyclin-dependent kinase (CDK) 1 inhibitor p21/waf1, thus reducing the activities of CDK2, 4 and 6, required for progression from G 1 into S phase (11)(12)(13)(14).
Repression of the transcription of several genes by p53 contributes to G 2 arrest. Topoisomerase II orchestrates the higher order compaction of chromatin that is required to form mitotic chromosomes (37). Inhibition of topoisomerase II blocks cells in G 2 because chromatin is not condensed and decatenated (38). p53 represses the topoisomerase II promoter, which may contribute to G 2 arrest (39,40). Repression of the cyclin B1 promoter by p53 is an important factor in the inhibition of Cdc2 (28,41). Furthermore, in some cell types overexpression of Cyclin B1 alone can abrogate p53-induced G 2 arrest (41). p53 also represses the transcription of cdc2 (15,28), and the CCAAT element is involved in this process (42). Also, downregulation of the level of Cdc2 protein in response to p53 requires p21/waf1 and can be abrogated by the E7 protein of human papilloma virus (43). E7 binds to several cellular proteins, including those of the Rb family (44 -48). In the present study, we demonstrate that p53 represses the cdc2 promoter through a mechanism involving induction of p21/waf1 and binding of the Rb family protein p130 and the transcription factor E2F4 to the CDE and CHR elements within the R box of the cdc2 promoter.

EXPERIMENTAL PROCEDURES
Cell Lines and Culture Conditions-Cells were grown in Dulbecco's minimal essential medium (Life Technologies, Inc.), supplemented with antibiotics and 10% fetal bovine serum (Life Technologies, Inc.) in a humidified atmosphere containing 10% CO 2 . The TR9-7 cell line, expressing a tetracycline-regulated p53 cDNA, was derived from MDAH041 cells (4). Cells were arrested in G 2 in response to p53 as described by Taylor et al. (28). Briefly, mimosine (Sigma), which arrests cells reversibly at the beginning of S-phase (49), was added at a final concentration of 200 M to asynchronously growing TR9-7 cells for 48 h. The cells were released from the mimosine block and simultaneously deprived of tetracycline to induce p53-mediated arrest, mainly in G 2 (4). Recombinant adenoviruses, provided by Joseph Nevins (Duke University, Durham, NC; Ref. 50), were amplified by infection of HEK293 cells (American Type Culture Collection, Manassas, VA). Cell supernatant fluids were collected 48 h later, and debris was removed with a 0.2-m filter. Viral titers were determined by infecting confluent monolayers of HEK293 cells for 2.5 h, removing the medium that contains the virus, and overlaying the cells with a solution of 1% low melting agarose (FMC Bioproducts, Rockland, ME) in Dulbecco's modified Eagle medium plus 10% fetal bovine serum. Infected cells were identified 72 h later by immunofluorescence with an antibody specific for the adenovirus hexon protein (Biodesign International, Kennebunk, ME).
Reporter Constructs-cdc2 promoter activities were determined by measuring luciferase activity in pools of TR9-7 cells stably transfected with constructs containing regions of the cdc2 promoter driving the expression of luciferase (51). Chimeric constructs were described by Sugarman et al. (51). Mutations were generated using the polymerase chain reaction and confirmed by DNA sequencing. Luciferase activity in extracts of stable pools was corrected for total protein concentrations in each lysate, determined by the Bradford method. Luciferase activity in transient transfections was corrected for transfection efficiency, determined by cotransfection of a plasmid constitutively expressing ␤-galactosidase.
To detect proteins binding to the cdc2 R box, gel mobility shift assays were carried out as described previously, with an oligonucleotide spanning this region (54). Briefly, cells were lysed for 20 min on ice in buffer I, containing 10 mM HEPES, pH 7.9, 1.5 mM MgCl 2 , 10 mM KCl, 0.3 M sucrose, 0.1 mM EGTA, 0.5% Nonidet P-40, 0.5 mM DTT, 0.5 mM PMSF, FIG. 1. Repression of cdc2/prolactin promoter chimeras by p53. A, luciferase activity after G 2 arrest induced by p53 overexpression. Stable pools of cells containing the promoter chimeras shown in B and C were released from a mimosine block in the presence or absence of tetracycline (TET). Removal of tetracycline induces the expression of high levels of p53, which causes cell cycle arrest mainly in G 2 . Luciferase activity was measured 48 h after removal of mimosine. Rlu: relative light units. Standard errors are shown by the bars. B, schematic diagrams of chimeras and average fold repression from two independent experiments (data shown in A). C, sequences of chimeric promoter constructs in the region of the R box.

Repression of the cdc2 Promoter by p53 Requires a Region
Proximal to the Start of Transcription-We have studied regulation of the cdc2 promoter by p53 using the TR9-7 cell line, which contains tetracycline-regulated p53 (4). Luciferase constructs linked to regions of the cdc2 promoter were transfected into TR9-7 cells, and pools of clones in which the reporter constructs had stably integrated into the genome were analyzed. A reporter construct containing cdc2 promoter sequences from Ϫ94 to ϩ75 was repressed ϳ3-fold by p53. 2 However, a construct extending from Ϫ74 to ϩ75 was not sufficient for basal promoter activity and was not repressed by p53 (28). A CCAAT element between Ϫ94 and Ϫ74 binds to the NF-Y transcriptional activator (42), suggesting that repression of the cdc2 promoter by p53 requires NF-Y. Because the CCAAT element is also required for basal activity of this promoter, it was possible that loss of repression by deletion of the CCAAT element might be due to a loss of basal promoter activity. For example, the construct extending to Ϫ74 might not be repressed by p53 because basal level activity is ϳ40-fold lower, with repression obvious only when promoter activity is higher (28,42). The Ϫ94 to ϩ75 region of the cdc2 promoter contains an additional regulatory region, called the R box (51), immediately upstream of the start of transcription. The R box is required for repression of the cdc2 promoter by phorbol ester (51).
To test whether the R box is important for repression by p53, several chimeric constructs were stably transfected into TR9-7 cells (Fig. 1). Cells were arrested in G 2 before analysis of luciferase activity (28). The region Ϫ36 to ϩ25 containing the R box was replaced with a region from the prolactin promoter from the same position relative to the start of transcription (51). Whereas the wild-type cdc2 promoter construct CC was repressed 4.4-fold, replacement of the R box region with prolactin sequences (the PP construct) completely eliminated repression by p53 (Fig. 1, A and B). Importantly, this replacement did not substantially affect the basal activity of the promoter, suggesting that the swapped region is required for repression specifically (Fig. 1A). Construct PCP generated by inserting 21 bp of the cdc2 promoter into the prolactin region, was repressed 3.8-fold by p53 (Fig. 1, A and B). However, construct CPC in which 10 bp from the prolactin promoter were inserted into the cdc2 promoter, showed a smaller repression of 2.0-fold (Fig. 1, A and B). These experiments localized the minimal region required for repression to the R box, located at nucleotides Ϫ22 to Ϫ2.
The R box contains several DNA sequence elements that regulate the transcription of the cdc2 and other promoters. The cell cycle-dependent element (CDE) is 4 bp upstream from the cell cycle gene homology region (CHR) (55). Mutation of these elements in the cdc25c, cyclin A, and cdc2 promoters relieves the repression that normally occurs upon arrest of the cells in G 0 (55). The CDE in the cdc2 promoter overlaps a putative E2F-binding site. Binding of p130/E2F4 to this element was correlated with repression of the promoter in G 0 -arrested cells (54). The factor CDF-1 can also bind to the CDE/CHR elements of the cdc2 promoter (56). We tested the importance of these elements for repressing the cdc2 promoter by p53. Constructs containing mutations in either the CDE or CHR were transfected into TR9-7 cells, and stable pools were incubated in the presence or absence of tetracycline to induce expression of p53. The wild-type promoter was repressed 5.6-fold by p53 ( Fig. 2A). Mutation of either the CDE or CHR elements substantially inhibited repression by p53. The promoter with mutations in the CDE was repressed 2.3-fold by p53 ( Fig. 2A). Similarly, mutations in the CHR reduced repression to 2.2-fold ( Fig. 2A). A comparison of the sequences in the chimeric and mutant promoters revealed several important points. Replacement of the first 5 bp of the putative E2F site, without altering the CDE, did not substantially affect repression (Figs. 1, A and B, and 2B, construct PCP), whereas mutations in the CHR that maintain the putative E2F site caused a loss of repression (Figs. 1A and 2, constructs CPC and mCHR). Mutations in the CDE that disrupt the putative E2F site also caused a loss of repression (Fig. 2, construct mCDE). These analyses localized repression of the cdc2 promoter by p53 to two distinct elements.
The R Box of the cdc2 Promoter Can Mediate p53-dependent Repression of Heterologous Promoters-We placed the region Ϫ74 to ϩ75 of the cdc2 promoter, which contains the R box downstream of two different elements capable of stimulating transcription. Protein(s) bound to the R box might interact specifically with NF-Y bound to the CCAAT box between Ϫ74 and Ϫ94, thereby blocking the ability of NF-Y to activate transcription. Alternatively, the R box might bind to a repressor capable of blocking several different types of upstream activators. TR9-7 cells were stably transfected with constructs containing the R box downstream of binding sites for either AP1 or SP1, two efficient activators of transcription. AP1-and SP1driven transcription was reduced by 2.5-and 3.0-fold by p53, respectively (Fig. 3). Wild-type cdc2 promoter sequences extending to Ϫ94 were used as a positive control. Luciferase activity driven by the cdc2 promoter was reduced by 3.3-fold by p53 (Fig. 3). These observations suggest that NF-Y is not specifically targeted by the protein(s) bound to the R box. Instead, the repressor recruited to the R box in response to p53 blocks transcription induced by three different transcriptional activators. Also, the ability of the R box to inhibit transcription driven by upstream activators may explain why deletion of the CCAAT box eliminates repression (28,42).
Role of p21/waf1 in Repressing the cdc2 Promoter in Response to p53-Ionizing radiation causes down-regulation of cdc2 mRNA in a p53-and p21/waf1-dependent manner (57,58). Our results show that overexpression of p53 without exogenous activation can repress the cdc2 promoter, suggesting that the down-regulation of cdc2 mRNA caused by DNA damage is due to repression of its promoter by p53. We tested whether promoter repression was dependent on p21/waf1 by using a p21/ waf1-null HCT116 colorectal tumor cell line, which contains wild-type p53 (59). Parental and p21/waf1-null cells were transfected transiently with a construct containing cdc2 promoter sequences up to Ϫ245 linked to luciferase (Fig. 1B, CC) and a plasmid expressing p53 under the control of the constitutive cytomegalovirus promoter. As a negative control, we used a reporter construct in which prolactin sequences replaced the R box (Fig. 1B, PP). Overexpression of p53 repressed the cdc2 promoter by 60 -70-fold in parental HCT116 cells (Fig.  4A). Deletion of p21/waf1 substantially impaired the ability of p53 to repress the cdc2 promoter; however, repression by 10fold was still observed (Fig. 4A). The prolactin promoter chimera was repressed by 10-fold in parental cells but not in p21/waf1-null cells when transfected with 0.5 g of the plasmid expressing p53 (Fig. 4A). The fold repression of cdc2 by p53 in transient assays was much higher than in pools of TR9-7 cells stably transfected with cdc2 reporter constructs, probably because of differences in the levels of expression of p53 and differences in chromatin structure of the reporter construct.
Constructs containing either the wild-type cdc2 promoter (CC) or the prolactin promoter chimera (PP) were transfected stably into parental and p21/waf1-null HCT116 cells. cdc2 promoter activity was reduced 2.6-fold upon treatment of the parental cells with adriamycin, a DNA damaging agent known to induce functional p53 in these cells (Fig. 4B and Ref. 59). cdc2 promoter activity was reduced only 1.5-fold upon treatment of p21/waf1-null cells with adriamycin (Fig. 4B). Transcription driven by the prolactin chimera was not affected substantially after treatment of either cell type with adriamycin (Fig. 4B). Therefore, repression of the cdc2 promoter in response to DNA damage requires p21/waf1. We tested whether p21/waf1 was sufficient for repression of the cdc2 promoter by using a recombinant adenovirus in which the expression of p21/waf1 is driven by a constitutive cytomegalovirus promoter (50). 2 A pool of TR9-7 cells transfected stably with a construct containing the region up to Ϫ94 of the wild-type cdc2 promoter was incubated in the presence of tetracycline to maintain low levels of p53. Infection with adenovirus expressing p21/waf1 repressed the cdc2 promoter by 2.1-and 7.1-fold at multiplicities of infection of 20 or 100, respectively (Fig. 4C). These results suggest that the elevated levels of p21/waf1 resulting from overexpression of p53 are important for repressing the cdc2 promoter.
E2F4/p130 Binds to the R Box of the cdc2 Promoter to Mediate Repression by p53-We next investigated which downstream targets of p53 and p21/waf1 might mediate repression of the cdc2 promoter. p130 and E2F4 bind to the putative E2F site that overlaps the CDE of this promoter and have been implicated in the repression observed in G 0 and G 1 (Ref. 54 and Fig. 2B). p21/waf1 can inhibit the ability of Cyclin/CDK complexes (12) to phosphorylate the Rb family members pRb, p107 and p130 (60,61). Complexes of Rb family members with E2F/DP heterodimers repress the transcription of genes containing E2F-binding sites (62). Phosphorylation of Rb family members by CDKs blocks their interaction with E2Fs (60). Therefore, p53 might repress the cdc2 promoter by inducing p21/waf1 which, by inhibiting CDKs, causes Rb/E2F complexes FIG. 3. Effects of the R box from the cdc2 promoter on different transcriptional activators. The R box was placed downstream of either three AP1 sites or one SP1 site, and TR9-7 cells were stably transfected with these constructs. cdc2 promoter sequences extending to Ϫ94 were used as a positive control. Luciferase activity was measured after incubation of asynchronously growing pools of cells in the presence or absence of tetracycline for 48 h. Standard errors are shown by the bars.
to accumulate. These might bind to the R box and repress transcription. We tested the requirement of various Rb family members in down-regulating Cdc2 in response to DNA damage. Mouse embryo fibroblasts (MEFs), derived from mice in which Rb, p107, p130, or both p107 and p130 were inactivated by gene targeting (63), were treated with adriamycin for 24, 48, or 72 h to induce functional p53. Western analysis of Cdc2 protein levels indicated that Cdc2 was down-regulated efficiently in MEFs from wild-type, heterozygous, and p107-or p130-null mice (Fig. 5). However, the down-regulation of Cdc2 was less efficient in Rb-null cells and was highly defective in cells lack-ing both p107 and p130 (Fig. 5). Therefore, either Rb or p107 or p130 are required for efficient Cdc2 down-regulation in response to DNA damage in MEFs.
To test the importance of Rb family members and E2F proteins in repressing the human cdc2 promoter, we analyzed TR9-7 cells for expression of the relevant proteins. Rb, p107, and p130 were detected in TR9-7 cells by Western blotting and the electrophoretic mobility of all three proteins was increased following induction of p53 (Fig. 6A), consistent with inhibition of CDKs in response to p53 overexpression, because phosphorylation of Rb family proteins by CDKs decreases their mobility (for example see Ref. 64). Similarly, removal of serum, which reduces CDK activity, also increased the mobility of Rb, p107, and p130 (Fig. 6A). Removal of serum caused a reduction in the level of Rb and p107 but an increase in the amount of p130 compared with the levels in asynchronously growing cells (Fig.  6A). TR9-7 cells also contain E2F4, which has been implicated in suppressing cdc2 promoter activity (Fig. 6A and Ref. 54). Removal of serum caused an increase in the intensity of more slowly migrating bands, detected with the E2F4 antiserum. These bands may represent phosphorylated species (65). Gel mobility shift analysis using a probe containing a consensus E2F-binding site was used to characterize E2F-binding complexes in TR9-7 cells. We focused on E2F4 and p130, given their role in regulating the human cdc2 promoter (54). Several complexes containing p130 and E2F4 in TR9-7 cells were detected before and after induction of p53 (Fig. 6B). p53 induction changed the complement of E2F complexes, consistent with CDK inhibition and changes in the phosphorylation of Rb family members and in their affinities for E2Fs (Fig. 6B). Thus, TR9-7 cells contain E2F4 and p130 complexes capable of binding to a well defined E2F site.
Gel mobility shift analysis with a probe derived from the R box was used to test whether Rb/E2F complexes were capable of binding to the putative E2F site in the cdc2 promoter. p53 induced the binding of a complex to the R box that could be supershifted with antibodies to either E2F4 or p130 (Fig. 7, A  and B). The antibody to p130 cross-reacts with p107. However, an antibody to p107 did not supershift the p53-induced complex, suggesting that the complex contains p130 and not p107 (Fig. 7B). An antibody to pRb did not supershift the p53induced complex (Fig. 7B). The p53-induced complex was reduced by a 100-fold molar excess of wild-type unlabeled probe. However, unlabeled probes containing mutations in either the CDE or CHR that impaired repression by p53 could not efficiently compete with the labeled probe ( Figs. 2A and 7C). These results suggest that p53 induces a complex on the cdc2 promoter containing p130 and E2F4 and that formation of this complex requires the CDE and CHR elements. DISCUSSION The arrest of mammalian cells in G 2 in response to damaged DNA is ensured by multiple p53-dependent and -independent pathways (15,28). p53 represses the transcription of cyclin B1 and cdc2 and induces p21/waf1, Gadd45, and 14-3-3 to inhibit residual Cdc2/Cyclin B1 (15,28,34,35). p53 affects proteins other than Cdc2 to contribute to G 2 arrest. For example, it causes the down-regulation of topoisomerase II (39,40), required for entry into mitosis, and induces the expression of B99 (66), which can induce G 2 arrest, possibly because of its interaction with the microtubule cytoskeleton. Several cell types lacking p53 still arrest in G 2 in response to DNA damage, suggesting that p53 provides a redundant level of control (3,15). Under some circumstances, the loss of p53 can affect the duration of G 2 arrest, clearly implicating p53 as a significant factor in the control of the G 2 /M transition (6). p53-dependent mechanisms of G 2 control are probably important in maintain- FIG. 4. Role of p21/waf1 in repression of the cdc2 promoter by p53. A, HCT116 cells in which p21/waf1 was inactivated by gene targeting were transfected transiently with various amounts of a p53 expression plasmid and either the cdc2 promoter (CC) or prolactin chimeric (PP) reporter constructs described in Fig. 1. Luciferase activity measured 48 h after transfection was corrected for transfection efficiency by using a cotransfected ␤-galactosidase construct. A representative experiment is shown. B, role of p21/waf1 in repression of the cdc2 promoter in response to DNA damage. HCT116 cells with or without p21/waf1 were transfected stably with the CC or PP reporter constructs. Stable pools of cells were treated with adriamycin to induce DNA damage, followed by measurement of luciferase activity 48 h later. Standard errors were less than 15% of the mean in all experiments. C, effects of overexpression of p21/waf1 on the cdc2 promoter. A pool of TR9-7 cells stably transfected with a reporter construct extending up to Ϫ94 of the cdc2 promoter was infected at the indicated multiplicities with an adenovirus to overproduce p21/waf1. Luciferase activity was measured 72 h after infection. Standard errors are shown by the bars.
ing genomic stability and in suppressing neoplasia.
p53 Represses Transcription by Several Distinct Mechanisms-Repression of the transcription of a large number of genes by p53 occurs through several distinct mechanisms. p53 can bind to TATA-binding protein-associated factors to activate the transcription of genes containing p53-binding sites (67). At the same time, high levels of p53 bound to TATA-binding protein-associated factors block their interaction with the TATA regions of genes that do not contain p53-binding elements, resulting in transcriptional repression (68 -70). This squelching mechanism can also affect genes that do not contain TATA regions, presumably because TATA-binding protein-associated factors are still required for the transcription of those genes (71). Similarly, p53 can bind to TFIIIB, a TATA-binding protein-containing complex that is specifically required by RNA polymerase III, resulting in repression of the transcription of the tRNA, 5 S rRNA, and U6 snRNA genes (72). Also, repression of AP1-driven transcription by p53 requires binding of p53 to the p300 histone acetyltransferase. This event blocks the binding of p300 to c-Jun at the AP1 site, which is normally required for transcriptional activation (73).
Repression can also involve interference by p53 with sequence-specific transcriptional activators. The ␣-fetoprotein gene contains a p53-binding site in its promoter (74) and p53 represses expression of the ␣-fetoprotein gene in hepatocytes that contain the HNF-3 transcriptional activator, which also binds to the ␣-fetoprotein promoter (74). p53 bound to the ␣-fetoprotein promoter may interfere with the ability of HNF-3 to activate transcription. p53 also binds to the bcl2 promoter and interferes with the ability of the simultaneously bound Brn-3a transcription factor to drive the transcription of this gene (75). Repression of transcription of the map4 gene also requires p53, which binds to the corepressor mSin3a, recruiting histone deacetylases that cause repression (76). Consistent with this mechanism, the repression of stathmin and map4 expression by p53 can be reversed by inhibiting histone deacetylases with trichostatin A (76). p53 can also repress transcription by inhibiting the binding of sequence-specific transcriptional activators to DNA. The binding of p53 to the SP1 transcriptional activator blocks its ability to bind to promoters (77). This mechanism is responsible for the repression of the Werner helicase gene by p53 (78). Repression of the cyclin B1 promoter by p53 maps to a region between Ϫ287 and Ϫ123, suggesting that the effects of p53 on transcriptional activators bound to this region are involved (28).
Repression of the cdc2 Promoter by p53 Requires the CDE

FIG. 6. Detection of E2F4 and Rb family members in TR9-7 cells. A,
Western analysis of TR9-7 cells with antibodies to p53, Rb, p107, p130, and E2F4 before and after induction of p53. Asynchronously growing TR9-7 cells were incubated for 48 h in the presence or absence of tetracycline (TET) to induce p53 expression. Cells incubated for 48 h in 0.25% serum and then stimulated for 24 h with 10% serum are also shown. B, gel mobility shift analysis using a consensus binding site for E2F-containing complexes. TR9-7 cells were released from a mimosine block in the presence or absence of tetracycline to induce p53dependent G 2 arrest. Nuclear lysates were incubated with a radioactively labeled probe in the presence or absence of antibodies to p130 or E2F4. and CHR Elements-p53 represses the cdc2 promoter by inducing p21/waf1, which inhibits CDKs, resulting in the binding of a complex containing p130/E2F4 to the cdc2 promoter. Repression of cdc2 requires the CDE and CHR elements. The cdc25c, cyclin A, and cdc2 genes contain CDE and CHR elements and are repressed during G 0 and G 1 (55). Repression during G 0 and G 1 correlates with the binding of the factor CDF-1 to the CDE and CHR elements (55). Mutations in CDE or CHR cause loss of CDF-1 binding and abolish repression (56). Thus, it has been postulated that CDF-1 is a repressor of transcription. Supershift analyses suggest that CDF-1 does not contain E2F or Rb proteins (56). Purification and cloning of CDF-1 will allow assessment of the mechanism used to repress transcription. The factor CHF can bind to the CHR element of the cyclin A but not cdc2 nor cdc25c promoters. The relationship of CHF to CDF-1 is not known (79). We have found that repression of the cdc2 promoter requires both the CDE and CHR elements, which mediate the binding of a p53-inducible complex containing E2F4 and p130. Consistent with our results, DNA damage leads to repression of the cdc2 promoter in a manner dependent on both the CDE and CHR elements (80). Repression was observed in WI38, HT1080, and HCT116 cells, all of which contain wild-type p53, but not in HeLa cells in which p53 is inactivated by the human papilloma virus E6 protein. Our results suggest that the repression of cdc2 by DNA damage depends on p53 and not on other factors that are defective in HeLa cells. We observed that the cdc2 promoter was partially repressed even after mutation of the CDE or CHR elements. These mutations may not completely abolish E2F4/ p130 binding in vivo, or p53 may target other elements in the cdc2 promoter to cause complete repression.
p21/waf1 and Rb Family Members Are Essential for Repression of cdc2 by p53-Mouse fibroblasts lacking p130 show nor-mal down-regulation of Cdc2 protein after DNA damage. However, Cdc2 was not down-regulated in cells after deletion of both p130 and p107, and down-regulation was less efficient in cells lacking Rb. These results suggest that all three Rb family members play overlapping roles in the repression of cdc2 by p53 in mouse cells. Using gel mobility shift analysis with extracts of human cells, we did not detect p107 or Rb bound to the R box, suggesting that the p53-dependent repression of the human promoter may be more specific in its requirement for p130. Alternatively, the interaction of Rb and p107 with the human cdc2 promoter may not be detectable by gel mobility shift analysis under the conditions employed. It is also possible that Rb and p107 contribute to Cdc2 down-regulation by a mechanism that does not involve their binding to the cdc2 promoter.
When CDKs are inactive, Rb family proteins are hypophosphorylated. In this state they can bind to E2F/DP complexes, which target them to E2F-binding sites. Binding to Rb family members converts E2F/DP from a sequence-specific transcriptional activator to a sequence-specific repressor (62). Repression is at least partly dependent on the ability of Rb proteins to recruit histone deacetylases, which inhibit transcription through their effects on chromatin condensation (81,82). Trichostatin A, an inhibitor of histone deacetylases, only partially relieved repression of the cdc2 promoter by p53, 2 suggesting that other mechanisms may also be important and consistent with the ability of p130 to interact with the CtIP/CtBP corepressor that has been postulated to function independently of histone deacetylases (83). E2F/Rb complexes have been implicated in the cell cycle-regulated expression of cdc25c. The E2Fbinding site mediating this effect is distinct from the CDE/CHR element (84).
We found that the CDE of the cdc2 promoter is required for FIG. 7. Gel mobility shift analysis of the cdc2 promoter. The experiments were carried out with a radioactively labeled probe derived from the R box of the human cdc2 promoter. TR9-7 cells released from a mimosine block and incubated in the presence or absence of tetracycline (TET) for 72 h were lysed and analyzed. A, effect of p53 expression on complexes bound to the R box region of the human cdc2 promoter. The analysis was carried out in the presence or absence of an antibody to p130. B, supershift analysis of a p53-induced complex with antibodies to E2F1, E2F4, and Rb family members. Lysates from cells arrested in G 2 by overexpression of p53 were used for analyses in the presence or absence of the indicated antibodies. C, competition analysis of a p53-induced complex. Binding reactions were carried out in the absence or presence of the indicated competitor probes, added at a 100-fold molar excess over the labeled probe. Competition was carried out with probes with a wild-type sequence or with mutations in either the CDE or CHR, as shown. Mutated bases are shown in lowercase letters. The mutations are the same as used in Fig. 2. repression by p53 and for binding of E2F4/p130. E2F sites contain the consensus TTTGCGCGC sequence, whereas in the cdc2 CDE region, the T triplet is immediately followed by A (85). van der Waal's contacts between tyrosine 124 of E2F4 and the cytosine after the T triplet stabilize binding of E2F4/DP2 to DNA (85). Replacement of cytosine with adenine is expected to disrupt these interactions and may impair DNA binding. Furthermore, we found that replacement of the T triplet with the sequence AAA reduced repression by p53 only slightly (Fig.  2B). The T triplet interacts with arginine 17 of E2F4, suggesting that the replacement we have made might also impair the binding of an E2F/DP heterodimer to this site (85). These results suggest that E2F4/p130 might bind relatively weakly to the CDE. Furthermore, mutations in the CHR 6 bp downstream of the putative E2F site caused a loss of E2F4/p130 binding and greatly reduced repression by p53. E2F4/DP2 appears not to make specific base contacts at this location downstream from the core binding site (85). These results suggest that other proteins perhaps bound to the CHR might stabilize binding of E2F4/p130 to the noncanonical E2F site in the CDE region of the cdc2 promoter.
The R box mediates repression by p53 when placed downstream of SP1 or AP1 sites, suggesting that the E2F4/p130containing complex induced by p53 does not specifically target NF-Y. However, interference of p53 with SP1 or AP1 could also contribute to the repression of these constructs. We also observed that repression mediated by the R box depended on p21/waf1, consistent with the ability of p21/waf1 to inhibit CDKs, which is expected to stimulate the formation of E2F4/ p130 complexes. cdc2 mRNA is suppressed upon overexpression of p21/waf1 (86), consistent with our result that p21/waf1 down-regulates the cdc2 promoter. p53 was much more efficient in repressing the cdc2 promoter in experiments using transient transfection compared with experiments using pools of cells with stably integrated cdc2 reporter constructs and tetracycline-regulated p53. This effect may be due to higher levels of p53 expression and the failure of the reporter construct to integrate into the chromosomes in the transient assay, which may make it more susceptible to repression. Using transient transfections we also observed that a cdc2 chimeric construct lacking the R box was still repressed, although at a lower level. This effect was dependent on p21/waf1, suggesting that high levels of p21/waf1 may affect the cdc2 promoter independently of the R box.
Physiological Relevance of Repression of cdc2 by p53-The down-regulation of Cdc2 protein by p53 is delayed compared with the down-regulation of Cyclin B1, suggesting that the repression of cdc2 may contribute only partially to the initial arrest (28). However, the down-regulation of Cdc2 may be more important for the stability of the arrest. For example, cell cycle checkpoints in yeast that block entry into mitosis in response to DNA damage (87) eventually become attenuated, and the cells eventually do enter mitosis, presumably by reactivating Cdc2, without having repaired the damaged DNA (88). Continued proliferation of unicellular organisms may be more important than eliminating all cells with damaged DNA (88). In mammals, a single cell with damaged DNA may develop into a neoplasm. One way to ensure that damaged mammalian cells do not continue to proliferate even if checkpoint pathways become attenuated is to eliminate the proteins required for entry into mitosis. Mammals may utilize p53 to repress the cdc2 promoter for this reason.