Transcriptional Regulation of the Human DNA Polymerase Catalytic Subunit Gene POLD1 by p53 Tumor Suppressor and Sp1*

The DNA polymerase catalytic subunit gene (POLD1) was studied as a transcriptional target of p53. Northern blotting showed that a significantly decreased steady-state level of POLD1 mRNA was associated with increased wild-type p53 expression in cells treated with methyl methanesulfonate. When ectopic wild-type p53 expression was induced to a physiologically relevant level in “tet-off” cultured cells in which p53 expression was tightly regulated by tetracycline, it was found that POLD1 steady-state mRNA was repressed by about 65%. Transient cotransfection experiments using a POLD1 promoter luciferase reporter construct showed that: (i) POLD1 promoter activity was inhibited by transfected wild-type p53 plasmid to a maximum of about 86%; (ii) p53 mediated a large part of the transcriptional repression through a sequence-specific interaction with a site identified as the P4 site of the POLD1 promoter; (iii) tumor-derived p53 mutations in the p53 DNA-binding domain completely abolished the p53 transrepression activity. Moreover, transfection assays demonstrated that p53 was able to repress Sp1-stimulated POLD1 promoter activity and that this repression was largely due to the loss of the sequence-specific interaction between Sp1 protein and the P4 Sp1-binding site, which overlaps the P4 p53-binding site. Finally, gel shift assays suggested that p53 competes with Sp1 protein for binding to the P4 sequence of the POLD1 promoter.

DNA replication is a highly organized process that involves many enzymes and proteins, including several DNA polymerases. DNA polymerase ␦ (pol ␦) 1 is one of the at least 12 DNA polymerases identified thus far in mammalian cells (1,2). Studies of in vitro reconstituted SV40 DNA replication established the requirement of two DNA polymerases, pol ␦ and pol ␣. pol ␦ is capable of highly processive DNA synthesis in association with the molecular sliding clamp, PCNA, and thus plays a role in the synthesis of the leading strand at the replication fork. pol ␣/primase functions to synthesize RNA-DNA primers for initiation on the leading strand, but it is thought that pol ␦ is required for the completion of Okazaki fragment synthesis (3,4). Several lines of evidences have shown that pol ␦ may also be involved in nucleotide excision repair (5,6) and in base excision repair of exogenous DNA methylation damage (7). Recently, pol ␦ has been found to be important for mismatch repair (8) and mitotic double-strand break-induced gene conversion (9).
The pol ␦ enzyme is now thought to consist of at least four subunits in human and Schizosaccharomyces pombe (10,11). The catalytic subunit is a 125-kDa protein that is encoded by the POLD1 gene in human cells (12). The transcriptional level of pol ␦ mRNA is regulated during the cell cycle (13) and is induced in serum-stimulated cells (14). The POLD1 gene promoter is TATA-less, GC-rich, and has been reported to be activated by Sp1 and Sp3 transcriptional factors that bind specifically to two repeat sequences in the promoter (15). Nevertheless, the effects of other transcriptional factors on POLD1 promoter remain unknown.
The p53 tumor suppressor gene product functions as a checkpoint in maintaining genome stability. More than half of a wide spectrum of human cancers contain mutations in the p53 tumor suppressor gene, and more than 90% of the p53 missense mutations are clustered within the sequence-specific DNAbinding domain, suggesting that functional inactivation of p53, especially the DNA binding activity of p53, is a crucial and often obligatory step in pathways to tumorigenesis. p53 protein is induced and/or activated in response to a variety of stimuli such as DNA damage or the expression of viral oncogenes. Activation of p53 leads to one of the two major cellular pathways, either apoptosis or cell cycle arrest, that prevents cells from progressing to S phase until the damaged DNA is fully repaired (16). The transcriptional functions of p53 are well established to be important in arresting the cell cycle, although this is less clear in p53-mediated apoptosis.
As a transcription activator, p53 recognizes a specific consensus DNA sequence consisting of two copies of a 10-bp motif, 5Ј-PuPuPuC(A/T)(T/A)GPyPyPy-3Ј, separated by a 0 -13-bp spacer. Wild-type p53 transactivates the expression of its target genes by specifically binding to p53-binding sites in the sequences of these genes (17). Among the p53-transactivated genes is p21/WAF1/Cip1, which encodes a cyclin-dependent kinase inhibitor (18). It is evident that the p53-induced transcription of p21/WAF1/Cip1 is responsible in large measure for p53-dependent G 1 arrest (19,20). Moreover, p21/WAF1/Cip1 has been shown to directly inhibit DNA replication by binding tightly to and masking the elements on PCNA that are required for its interaction with DNA polymerase ␦ (21). The product of another p53-transactivated gene, Gadd45, has been shown to interact with PCNA and inhibit DNA excision repair (22). Other evidence has also suggested that p53 may play one or more roles in regulating processes such as DNA replication, repair, and recombination (16). p53 not only acts as a transcriptional activator but also has been shown to repress the activity of a broad range of viral and cellular gene promoters that lack p53 binding sites, including but not limited to those of the ␣-fetoprotein gene (ARF), bcl-2, cdc2, c-fos, cyclooxygenase-2 gene (Cox-2), DNA topoisomerase II␣ gene, insulin-like growth factor I receptor gene (IGF-IR), microtubule-associated protein gene (Map4), retinoblastoma gene (Rb), and Werner helicase gene (WRN) (16,(23)(24)(25)(26)(27).
Because of its critical role in DNA synthesis, we investigated whether the DNA polymerase ␦ catalytic subunit gene, POLD1, is one of the transcriptional targets of the p53 tumor suppressor. The study described here demonstrates that the expression of the POLD1 gene mRNA is repressed by wild-type p53 in response to DNA damage. This p53-mediated repression is at the transcriptional level and is mediated largely through a specific interaction between p53 and a p53-binding site in the POLD1 promoter. Evidence was also obtained for a mechanism in which p53 exerts its repressive effect by exclusion of the binding of Sp1 transcriptional factor to a Sp1-binding site harbored within the p53 binding site.

EXPERIMENTAL PROCEDURES
Cell Culture and Cell Treatments-The cell lines Saos-2, H1299, and MDA MB231 and three lines of MCF7 were purchased from American Type Culture Collection (ATCC, Manassas, VA). Cells were cultured according to ATCC protocols. A human breast carcinoma MCF7 cell line that expresses relatively low levels of p53 under normal conditions was selected for this study. For DNA damaging agent treatment experiments, MCF7 or MDA MB231 cells were exposed to 100 g/ml methane methylsulfonate (MMS) (Aldrich). Human colon cancer cell lines RKO and RKO mp53-13 (28) were kindly provided by Dr. M. B. Kastan, St. Jude Children's Research Hospital, Memphis, TN. H24-p53-14, a H1299 human large-cell lung carcinoma cell line (29) that expresses tetracycline-induced wild-type p53, was provided by Dr. X. Chen, Medical College of Georgia, Augusta, GA. H24-p53-14 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 1.6 g/ml tetracycline, 0.3 mg/ml G418 (Life Technologies, Inc.), and 2 g/ml puromycin (Stratagene). For the tetracycline-dependent p53 induction experiments, H24-p53-14 cells were cultured in medium containing the indicated tetracycline concentrations (1.6 g/ml, 40 ng/ml, 20 ng/ml, 10 ng/ml, 5 ng/ml, 2 ng/ml, or without tetracycline) and harvested 24 h after tetracycline treatment. Drosophila Schneider line 2 (SL2) cells were purchased from ATCC and grown in Schneider medium (Life Technologies, Inc.) supplemented with 10% fetal bovine serum at 25°C.
Plasmid Constructs-The full-length POLD1 promoter/reporter pGLϪ1758 was generated as described (15). Mutated POLD1 promoter constructs are shown in Fig. 5A. The promoter constructs pGLϪ583 and pGLϪ1758/Ϫ272 were generated by PCR. PCR reactions were performed using the Expand High Fidelity PCR System (Roche Molecular Biochemicals). The pGL⌬Ϫ113/Ϫ58 has a 55-bp deletion from Ϫ113 to Ϫ58 in the full-length POLD1 promoter. The full-length promoter construct pGLϪ1758 was double-digested at the HindIII site at the 3Ј end of the promoter insertion and at the PstI site at Ϫ113. The larger of the two generated fragments contains the vector sequences and a 3Ј deleted promoter sequence (from Ϫ1758 to Ϫ113) and was isolated for subsequent ligation with a Ϫ58 to ϩ49 promoter PCR fragment containing a PstI site at Ϫ58 and a HindIII site at the 3Ј end. Site-directed mutagenesis was performed to generate the P4 p53-binding site-mutated promoter construct pGLϪ1758mP4 and the P4 Sp1-binding site-mutated promoter pGLϪ1758mSp1. Primers used for pGLϪ1758mP4 and pGLϪ1758mSp1 were 5ЈCCCTGCAGTCGATAAATAGGGGCGTG-GCATTTACCGCACTTGGGC3Ј and 5ЈGTCGAACAAGCGGTTATTG-GCCTTGCCCGC3Ј, respectively. The underlined letters indicate the P4 Sp1-binding site, and bold italic letters indicate changes of sequence from the original promoter sequence. Mutated promoter constructs were made using the QuikChange site-directed mutagenesis kit (Stratagene). The sequences of all promoter constructs were verified by automated sequencing by the DNA Core Laboratory, University of Miami. The control p53 reporter plasmid p53-Luc was purchased from Stratagene. It has an artificial p53 promoter containing 15 repeats of consensus p53-binding sites.
The p53 expression plasmid pCMV-p53 was renamed from the pC53-SN3-p53wt plasmid (30), which was provided by Dr. B. Vogelstein, Johns Hopkins University School of Medicine, Baltimore, MD. The control vector pCMV-0 was generated by religating the vector fragment generated by the BamHI digestion of pCMV-p53. p53 point mutant expression plasmids (pCMV-L22Q/W23S pCMV-V143A, pCMV-R175H, pCMV-R248W, pCMV-R273H, and pCMV-R282) were generated using the QuikChange site-directed mutagenesis kit (Stratagene). PCR was performed to prepare the ⌬Pro p53 mutant, which has a deletion from residues 63 to 91. The 1.8-kb BamHI fragment, which contains the coding sequence for p53 and other unknown sequences at both ends of the coding fragment, was isolated from the original pC53-SN3-p53wt plasmid. This fragment was ligated into the BamHI site of the pUC18 vector to generate pUC18-p53. The 5Ј p53 fragment containing residues 1-62 was generated by PCR. The forward primer, 5Ј-GGTCGACTCTA-GAGG-3Ј, was designed according to the multiple cloning site sequences on the pUC18 vector and included the GG sequence (underlined) of the GGATCC BamHI site. The reverse primer, 5Ј-CGGTACCGGACCT-GGGTCTTC-3Ј, was designed to include a GGTACC KpnI site (underlined). For the 3Ј p53 fragment containing residues 92-393, the forward primer was 5Ј-CCGGTACCCCTGTCATCTTCTG-3Ј and included a GG-TACC KpnI site (underlined). The reverse primer, 5Ј-GCTCGGTAC-CCGGGG-3Ј, was designed according to the multiple cloning site sequences on the pUC18 vector and included the GGATCC BamHI site (underlined). The PCR products from each reaction were subsequently ligated into BamHI-digested pCMV-0 vector fragment in a three-piece ligation. Because of a junctional GGTACC KpnI site, the ⌬Pro mutant has two extra residues (Gly and Thr) in the deleted region. Mutated sites in all of the p53 mutants generated were verified by sequencing from both directions using the T7 Sequenase method according to the manufacturer's protocol (Amersham Pharmacia Biotech).
The Sp1 expression plasmid pPacSp1 (31) was kindly provided by Dr. R. Tijian, University of California, Berkeley, CA. The control vector pPacU was constructed by religation after excision of the XhoI insert fragment. The p53 expression vector, pPac-p53, was generated such that the p53 cDNA was driven under the control of the Drosophila actin 5C promoter. The p53 cDNA was PCR-amplified using pCMV-p53 as the template. The forward primer was 5Ј-ggatccGAGGAGCCGCAGT-CAGATCC-3Ј, and the reverse primer was 5Ј-ggatccGTCTGAGTCAG-GCCC-3Ј, both of which contain BamHI sites (underlined). The PCR product was isolated by BamHI digestion and then ligated with BamHIlinearized pPacU vector fragment. Orientation of the resulting plasmid was positively identified by PvuII/XhoI double digestion.
Northern Blotting-Northern blotting was performed as described previously (14). Total RNA was extracted from cells using STAT60 (Tel-Test, Friendswood, TX). The POLD1 gene cDNA (12) was used as the template for PCR amplification of a probe fragment from nucleotide 2660 to 2920. The control GAPDH probe was purchased from CLON-TECH as a 1.1-kb cDNA fragment. The autoradiographic images were captured with a charge-coupled device (CCD) camera (Datavision, North Falmouth, MA), and the ratio of POLD1 mRNA to GAPDH mRNA was quantified by performing densitometric analysis within the linear range of each capture signal using the Image Pro Plus software (Media Cybernetics, Silver Spring, MD).
Western Blotting and Antibodies-Cells were lysed in Nonidet P-40 lysis buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml pepstatin, and 10 g/ml leupeptin). Samples of cell lysate supernatant (30 g protein) were resolved on SDS-PAGE and transferred to nitrocellulose membranes. The membranes were incubated with 0.5-1.0 g/ml primary antibodies. Immunocomplexes were detected by incubation with chemiluminescence-based SuperSignal substrate (Pierce) and subsequent exposure to x-ray film. The p53 monoclonal antibody PAb421 was provided by Dr. E. Harlow, Massachusetts General Hospital. The p53 monoclonal antibody DO-1 and Sp1 polyclonal antibody PEP2 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-actin polyclonal antibodies (Sigma) were incubated with the same membrane following membrane stripping.
Transient Cotransfection Assays and Luciferase Reporter Gene Assays-Saos-2 cells (1.5 ϫ 10 5 ) were plated onto a 6-well plate and grown overnight in Opti-MEM I medium (Life Technologies, Inc.). pCMV-p53 expression plasmid or pCMV-0 control vector DNA (0.1 g) was cotransfected with pGLϪ1758 reporter (0.2 g) into Saos-2 cells using Lipo-fectAMINE/PLUS (Life Technologies, Inc.). The medium was changed 3 h after transfection. The cells were lysed in 70 l of reporter lysis buffer (Promega) 24 h after transfection. Transient cotransfection assays in Drosophila SL2 cells were performed at room temperature using CellFectin reagent according to the manufacturer's protocol (Life Tech-nologies, Inc.). pPac-p53 expression plasmid DNA (0.4 g) and/or pPac-Sp1 expression plasmid (0.2 g) were used together with each POLD1 promoter/reporter construct (0.5 g). pPacU plasmid was used as the control vector. The total amount of DNA for cotransfection was adjusted to 3 g by adding pUC18 DNA. The medium was changed 5 h after transfection. Forty-eight hours after transfection, cells were collected in 1ϫ reporter lysis buffer (Promega).
Cells transfected in 6-well plates were gently washed with cold phosphate-buffered saline (80 mM Na 2 HPO 4 , 20 mM NaH 2 PO 4 , 100 mM NaCl, pH 7.6), lysed in 70 l of reporter lysis buffer in a luciferase assay system (Promega) for 10 min at room temperature with gentle vortexing. Cell lysates were collected and centrifuged briefly. Aliquots (10 l) of cell extract were used for luciferase assays using 50 l of substrate solution (Promega). Luciferase activity was measured by a luminometer (Turner Designs) with settings of 3-s delay time and 10-s integration time. Relative luciferase units were evaluated after normalization against protein concentration of each sample. Fold change was expressed as (relative luciferase unit with pCMV-p53/relative luciferase unit with pCMV-0). Values are the means Ϯ standard deviation (error bars) of relative luciferase activity from triplicate samples in three separate experiments.
Human wild-type p53 was over-expressed in Sf9 insect cells using p53 recombinant baculovirus h-p53wt (provided by Dr. C. Prives, Columbia University). 2 ϫ 10 8 cells were collected 48 h after infection. Nuclei extraction and the Sepharose Q step of p53 purification were performed as described previously (32). Eluted fractions were collected, resolved on 10% SDS-PAGE, and stained by Coomassie blue or detected by Western blot using p53 monoclonal antibody DO-1. Purified PAb421 antibody was coupled to ProA beads to make a p53 immunoaffinity column. The p53 peak fractions from Sepharose Q were combined and dialyzed against buffer containing 50 mM Tris-HCl, pH 7.6, 50 mM NaCl, 1 mM DTT, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml pepstatin, and 50 g/ml leupeptin. Clarified proteins were loaded onto the PAb421 immunoaffinity column. p53 was then eluted with 30% ethylene glycol and dialyzed into a buffer containing 10 mM Tris-HCl, pH 7.6, 25% glycerol, 5 mM NaCl, 0.1 mM EDTA, 10 mM DTT, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml pepstatin, and 50 g/ml leupeptin. All purification steps were performed at 4°C. p53 protein was aliquoted and stored at Ϫ70°C.
Affinity-purified recombinant p53 (50 ng) or 30 ng of recombinant Sp1 protein (Promega) was incubated with 50 ng of ds-poly(dI-dC)⅐poly(dI-dC) (Pharmacia) and ϳ0.5 ng (14,000 cpm) of each oligonucleotide probe in 50 l of EMSA buffer (20 mM Hepes, pH 7.9, 25 mM KCl, 2 mM MgCl 2 , 0.1 mM EDTA, 0.5 mM DTT, 0.025% Nonidet P-40, 2 mM spermidine, 0.1 mg/ml bovine serum albumin, 10% glycerol) for 30 min at room temperature. For supershift assays, 0.1 g of p53 monoclonal antibody DO-1 or Sp1 polyclonal antibody PEP2 was incubated individually with p53 or Sp1 protein in EMSA buffer at room temperature for 10 min before the addition of the ds-poly(dI-dC)⅐poly(dI-dC) and probe. For competition assays, an excess amount (5ϫ, 30ϫ, and 200ϫ molar ratio) of cold P4 -40-mer or other 40-mers (Gadd45-wt or Gadd45-mt) that contained either the consensus p53-binding site or mutated p53-binding sites were used. Cold oligonucleotides were added to the reaction mixtures and incubated for 10 min before the addition of the probe. For EMSA using both Sp1 and p53, 30 ng of Sp1 was used in combination with 15, 50, and 250 ng of p53. Other steps were the same as described previously. EMSA reaction mixtures were loaded onto a native 4% polyacrylamide gels containing 0.5ϫ Tris borate-EDTA buffer (TBE; 44.5 mM Tris-HCl, 44.5 mM boric acid, 1 mM EDTA, pH 8.3), 0.05% Nonidet P-40, and 2.5% glycerol. Gels were run in 0.5ϫ TBE running buffer at 200 -230 V for about 2 h at 4°C, dried, and exposed to x-ray film.
Southwestern Blotting-Southwestern blotting was performed as described by Zhao and Chang (15). Samples were resolved on 5-15% gradient SDS-PAGE and then transferred to a polyvinylidene difluoride membrane (PVDF-P SQ , Millipore, Bedford, MA) for 4 h at 4°C. The membrane was prehybridized in HB buffer (10 mM Tris-HCl, pH 7.5, 50 mM NaCl, 1 mM MgCl 2 , 0.5 mM EDTA, and 10 mM DTT) with 5% nonfat dry milk at room temperature for 30 min. Following two rinses with HB buffer containing 0.25% milk, the membrane was hybridized with the 32 P-labeled 40-mer P4 oligonucleotide probe (5 ϫ 10 6 cpm/ml) in HB buffer containing 0.25% milk and 100 ng/ml of ds-poly(dI-dC)⅐poly(dI-dC) at room temperature for 2 h. The membrane was then washed four times with HB buffer containing 0.25% milk, for 10 min each time, and exposed overnight to a x-ray film. MCF7 and RKO cell extract (30 g) protein were used for each lane. 20 ng of Sp1 protein (Promega) was loaded as a control.

Exposure of Cells to DNA-damaging Agents Induces Accumulation of Wild-type p53 and Down-regulates Steady-state Levels of POLD1
Gene mRNA-Exponentially growing human breast carcinoma cells MCF7 with wild-type p53 (33) were exposed to MMS, an alkylating agent that induces DNA base modifications. Northern blot analysis showed a reduction of POLD1 mRNA levels at 6 and 12 h (Fig. 1, top panel). p53 levels were elevated soon after MMS treatment as shown by Western blotting ( Fig. 1, bottom panel). In contrast, POLD1 mRNA levels showed no significant change in MDA MB-231 cells (Fig. 1, top  panel), a human breast adenocarcinoma cell line that constitutively expresses a transcriptionally inactive mutation of p53 (R280K) (34). Similar results were observed (data not shown) when human colon cancer cell lines RKO and RKOmp53-13 (28) that express wild-type p53 and a dominant negative p53, respectively, were treated with MMS, in that POLD1 mRNA levels were reduced only in the cell line expressing a wild-type p53. The inverse correlation between wild-type p53 expression and POLD1 mRNA steady-state levels suggested the involvement of wild-type p53 in the down-regulation of POLD1 transcription in the cellular response to DNA damage.
Steady-state Levels of POLD1 mRNA Are Reduced by Induction of Physiologically Relevant Levels of Wild-type p53 in "tetoff " Cells-The transcriptional response of the POLD1 gene to induced ectopic p53 expression was investigated in a tet-off cell FIG. 1. Northern blot analysis of POLD1 gene mRNA after exposure to a DNA-damaging agent. MCF 7 and MDA MB-231 cells were treated with 100 g/ml MMS for the indicated periods. Cells were harvested for total RNA isolation or protein extraction. Total RNA (20 g) was subjected to Northern blotting using probes for POLD1 and GAPDH. Thirty g of protein from each sample lysate was loaded on a gel and immunoblotted for p53 using p53 monoclonal antibody DO-1 (see "Experimental Procedures").
line, H24-p53-14, in which p53 expression was tightly suppressed by tetracycline concentration in the culture medium (29). p53 expression under the conditions used for these experiments was monitored by Western blotting at different levels of tetracycline ( Fig. 2A). As expected, p53 expression was induced to varying levels in accordance with decreasing tetracycline concentration. p53 was detectable at 40 ng/ml tetracycline or less. When tetracycline was completely removed from the medium, p53 expression in these cells was maximally induced to a level that was comparable with the levels elicited by MMS treatment of MCF7 cells. The transcriptional control of POLD1 expression by p53 under these conditions, which likely reflects physiologically relevant levels of p53, was then investigated by Northern blotting (Fig. 2B). POLD1 mRNA levels were reduced by about 65% at maximal p53 expression (either with 2 ng/ml tetracycline or without tetracycline). The levels of POLD1 mRNA decreased only when tetracycline was reduced below 10 -5 ng/ml. Overall, these results show that the POLD1 gene mRNA is suppressed by p53 levels that are physiologically relevant in cultured cells. In addition, it is seen that POLD1 mRNA could be seen to be reduced only after a certain threshold of expression of p53. When the parental p53 null cell line H1299 was treated with tetracycline, no changes in POLD1 mRNA levels were observed (data not shown), ruling out a nonspecific inhibition of POLD1 mRNA by tetracycline.
Repressive Effect of Exogenous Wild-type p53 Expression on POLD1 Promoter Activity in Saos-2 Cells-Transient cotransfection experiments were performed in the p53 null cell line, Saos-2, using the POLD1 promoter luciferase reporter construct pGLϪ1758 and various amounts of a wild-type p53 expression plasmid, pCMV-p53, to establish that the repression of POLD1 steady-state mRNA levels by p53 is at the level of the POLD1 gene promoter. The pGLϪ1758 reporter contains the 1.8-kb 5Ј flanking region of the POLD1 gene fused with the firefly luciferase reporter gene (15). p53 levels were monitored by Western blotting (Fig. 3A) and were shown to increase in correspondence with the amount of transfected plasmid DNA. Results from the luciferase assays (Fig. 3B) showed no repression when 3 or 10 ng of wt-p53 expression plasmid was transfected even though low levels of p53 protein were detected by Western blot. Thirty ng of p53 plasmid DNA caused a 2-fold increase of p53 expression level compared with that at 10 ng and about a 5-fold increase if compared with 3 ng of p53 plasmid. At 30 ng/ml p53 plasmid, inhibition of POLD1 promoter activity was significant (73%). This result is consistent with the repression of POLD1 mRNA observed in the previous experiments with tet-off cells. Increasing the amount of transfected p53 plasmid resulted in a maximal repression of promoter activity by 83% at 0.1 g of p53 plasmid DNA. Promoter activity was not repressed further by greater amounts of p53 expression plasmid at 0.3 and 1 g, indicating a saturable system for POLD1 repression. These experiments confirm that exogenous wild-type p53 inhibits POLD1 promoter activity.
The repression of the POLD1 promoter activity by p53 was examined at different time points after transfection (Fig. 3C). Promoter activity was inhibited by about 65% at the 12-h time point after transfection when exogenous p53 expression was increased more than 5-fold compared with the p53 expression level at the 6-h time point. The activity of the POLD1 promoter was further repressed by about 85% at 24 and 36 h, whereas the corresponding levels of exogenous p53 protein were maintained at the peak. At 48 h, the p53 expression level decreased by about half, possibly because of its instability. However, the inhibited promoter activity at 48 h was not affected, indicating the lack of an immediate effect on POLD1 promoter upon the decline of p53 expression. The time course results confirmed that there was a time-dependent repression of the POLD1 promoter activity by wild-type p53 protein. The results also justified the promoter repression results of Fig. 3B, which were performed at 24 h after transfection when p53 was maximally expressed.
In Vitro EMSA of the Interaction of p53 with the Sequence in the P4 Site of the POLD1 Promoter-A direct binding of wt-p53 to the consensus p53-binding site (RRCWWGYYY-RRCWW-GYYY) has been shown to be essential for the transcriptional activity of p53 (35, 36). Inspection of the POLD1 promoter shows five potential p53-binding sites, designated P1-P5. Sequence matches for the five sites are shown in Fig. 4A. The interaction between p53 and these putative p53-binding sites, which might provide a possible explanation for the p53 repression of POLD1 promoter, was studied by EMSA using purified recombinant p53 protein. Human recombinant p53 was overexpressed in insect cells and was purified to about 90% homogeneity (Fig. 4B). Gel shift assay results (Fig. 4C) showed that only a 40-mer ds-oligonucleotide spanning the P4 site (see "Experimental Procedures") was shifted by purified p53 protein (lane 11) and was further supershifted in the presence of p53 monoclonal antibody DO-1 (lane 12). The P4 site matched the canonical p53-binding site in 17 of 20 nucleotides and harbors a Sp1-binding site between the two halves of the p53 binding FIG. 4. EMSA analysis of the in vitro interaction of p53 protein with the POLD1 promoter P4 site sequence. A, putative p53-binding sites P1 to P5. Sites are shown with the consensus sequence below each site. There are four sites (P1-P4) as shown and one half-site (P5). A Sp1 site located in the P4 site is also shown. Matches with the p53 consensus site are underlined. R ϭ A or G; W ϭ A or T; Y ϭ C or T. B, p53 was over-expressed in insect cells and was purified as described under "Experimental Procedures." The purified p53 was loaded onto a 5-15% gradient SDS-PAGE and silver-stained. C, EMSA analysis of p53 binding to putative sites P1-P5. EMSA was performed using labeled oligonucleotides containing each putative site as indicated. A Gadd45 probe was used as a control. Purified p53 (50 ng) was incubated with a labeled 40-mer probe corresponding to each putative site. Lanes 1, 4, 7, 10, 13, and 16: probes only. Lanes 2,5,8,11,14, and 17: p53 ϩ probes. Lanes 3,6,9,12,15,and 18: p53 monoclonal antibody DO-1 ϩ p53 ϩ probes. Bands for the free probe, protein-DNA complex, or antibody-protein-DNA complex are indicated. D, EMSA analysis of specific p53 binding to the P4 site on POLD1 promoter by competition and supershift assays. Indicated amounts of different unlabeled oligonucleotides were added to test for competition with the labeled probes (see "Experimental Procedures"). Lane 1: p53 ϩ probe; lanes 2-4: p53 ϩ probe ϩ cold wild-type Gadd45-oligonucleotide; lanes 5-7: p53 ϩ probe ϩ cold mutated Gadd45-oligonucleotide; lanes 8 -10: p53 ϩ probe ϩ cold P4-oligonucleotide; lanes 11-14: p53 ϩ probe ϩ antibodies as indicated. E, EMSA analysis of sequence specific p53 binding to the P4 p53-binding site. Purified p53 (50 ng) was incubated with each labeled 40-mer probe. Lanes 1, 4, 7, and 10: probes only. Lanes 2, 5, 8, and 11: p53 protein ϩ probes. Lanes 3, 6, 9, and 12: p53 ϩ p53 antibody DO-1 ϩ probes. The Gadd45 probe was used as a control. Each oligonucleotide containing mutated sequences was listed. The P4 p53-binding site is underlined, and the P4 Sp1-binding site is double underlined. Letters in bold italics indicate mutated sequences. sequence (Fig. 4A). A 40-mer Gadd45 ds-oligonucleotide containing a known p53-binding site (see "Experimental Procedures") was used as a positive control (lanes 16 to 18). P4 oligonucleotide/p53 bands were weaker than the Gadd45 shifted bands, suggesting that p53 has a lower affinity for the P4 oligonucleotide than the Gadd45 sequence. A 40-mer oligonucleotide containing the P5 site was shifted and supershifted but with a very weak signal (lanes 14 and 15), whereas oligonucleotides containing P1, P2, and P3 sites were not shifted by p53.
If the putative P4 p53-binding site is a sequence-specific p53-binding site, mutations in this site should result in the elimination of its binding to p53. A 40-mer oligonucleotide (mP4) containing a mutated P4 p53-binding site was tested by EMSA (Fig. 4E). Four matched nucleotides in the first half of the p53 10-bp motif and three matched nucleotides in the second half motif of the P4 sequence were mutated (sequence described in Fig. 4E, lower panel). The interaction of mP4 with p53 was greatly weakened compared with the P4 oligonucleotide, as only very weak signals for the shifted and supershifted bands were observed (Fig. 4E, lanes 5 and 6). In contrast, the mSp1 oligonucleotide (sequence described in lower panel), in which the intervening sequence between the two halves of the p53 consensus was mutated and which has an unchanged P4 p53-binding site, was able to bind and retain an interaction with p53 (Fig. 4E, lanes 8 and 9). These results provide additional evidence for a sequence-specific interaction between p53 and the P4 site. The results from the above experiments showed that only the P4 site possesses a sequence that exhibits significant interaction with p53 and provided the initial identification of the P4 site as the likely locus for the actions of p53 on the transcriptional activity of the intact POLD1 promoter.
Interaction between p53 and the P4 p53-binding Site Plays a Major Role in the Regulation of POLD1 Promoter by p53-The function of the P4 site in the intact promoter as the site of p53 action was analyzed by the use of deletion and point mutations in the pGLϪ1758 POLD1 reporter construct. The P4 p53-binding site in the promoter was either deleted (pGL⌬Ϫ113/Ϫ58) or mutated (pGLϪ1758mP4). Promoter/reporter constructs and site-directed mutation sequences are shown in Fig. 5A. The effect of p53 on these promoter constructs was examined by co-transfection into Saos-2 cells. p53 repressed the promoter activity of pGL⌬Ϫ113/Ϫ58 by about 55%, whereas it repressed that of pGLϪ1758mP4 by about 35% (Fig. 5B). The transrepression activity of p53 on both promoters was significantly diminished but not fully lost compared with that of the full-length pGLϪ1758 promoter (about 86% inhibition). Nevertheless, these results indicate that p53 mediated its major inhibitory effects on POLD1 promoter through the P4 site. The diminished p53 inhibitory effect on the P4-mutated POLD1 promoter is likely due to the loss of p53 binding to the mutated P4 site (mP4), as shown in Fig. 4E. The results also suggest a possibility that there may be other p53 negative response elements in addition to the P4 site present in the POLD1 promoter, because deletion of the P4 site did not completely ablate p53 repression.
In addition, when the P4 site was deleted (pGL⌬-113/-58), the basal promoter activity in the absence of p53 was reduced by about 39% compared with that for the pGLϪ1758 full-length promoter. This was attributed to the fact that the deleted region in pGL⌬Ϫ113/Ϫ58 also includes a Sp1-binding site (see below). The deletion of the Sp1-binding site could contribute to the significantly reduced basal promoter activity for pGL⌬Ϫ113/Ϫ58. This possibility was confirmed by the results with the pGLϪ1758mSp1 promoter, where the P4 Sp1-binding site was mutated and the P4 p53 binding site was unchanged (Fig. 5A). In this promoter, basal promoter activity was similarly reduced by about 34% in the absence of p53 (Fig. 5B). When a 5Ј deletion promoter construct (pGLϪ583) in the 5Јmost 1.2-kb region was truncated, resulting in the deletion of the P1, P2, and P3 sites, p53 inhibited the promoter activity by about 83%; this indicated that the three putative p53-binding sites (P1, P2, and P3) are unlikely to be involved in p53 transcriptional repression.
Transrepression of POLD1 by p53 Mutants-Tumor-derived mutations in p53 significantly reduce p53 transrepression activity on many promoters such as the mouse DP1 promoter (37), the DNA topoisomerase II␣ promoter (25), and the insulin-like growth factor I receptor gene promoter (26). To determine whether the inhibition of POLD1 promoter was affected by tumor-derived mutations in p53, wt-p53 or individual tumor-derived p53 point mutants (pCMV-V143A, pCMV-R175H, pCMV-R248W, pCMV-R273H, and pCMV-R282W) were transfected into Saos-2 cells. All of these mutants are ones that have lost p53 consensus binding and transactivation activities. Only wt-p53 was able to inhibit POLD1 promoter activity (by about 85%) compared with the control vector (Fig. 6B). All of the tumor-derived p53 point mutants exhibited a disruption of transrepression activity under identical assay conditions. The ⌬Pro deletion mutant (Dlt-Pro), which has been reported to have normal p53-mediated transcriptional activation but is defective in p53-mediated apoptosis, did not exhibit a significantly reduced repression activity. Results from these p53 mutants indicated that the specific DNA binding activity of p53 was indispensable for the p53 transcriptional regulation of POLD1 promoter, confirming the results from the above experiments.
Extracts from appropriately transfected Saos-2 cells were examined to show that wild-type and p53 mutants were produced at equivalent levels (Fig. 6A). All of the p53 constructs transfected in Saos-2 cells were able to express their corresponding proteins to approximately similar levels. In addition, wild-type and mutant p53 protein levels were stable between 18 and 36 h after transfection (data not shown). Therefore, these results eliminated the possibility that the lack of significant repression activity on POLD1 promoter by p53 mutants is due to deficient protein levels of p53 mutants in transfected Saos-2 cells or that the results are due to overexpression of the mutants compared with the wild type. To confirm the transcriptional activation functions of all the ectopic p53 mutants, reporter gene assays were performed (Fig. 6C). Experiments were performed under the same conditions as described in Fig.  6B, except that an artificial p53 reporter promoter containing 15 repeats of consensus p53-binding site was used instead of the full-length POLD1 promoter/reporter pGLϪ1758. The results showed that wild-type p53 and the ⌬Pro p53 deletion mutant were able to activate promoter activity substantially whereas the other mutants were transactivation defective, con-sistent with the reported behavior of p53 in promoter activation. These results indicated that the inhibition of POLD1 promoter activity by wild-type p53 was not an artificial effect because of general inactivation of the transcriptional machinery or squelching of general transcription factors. This conclusion is based on the result that the wild-type p53 and the ⌬Pro p53 deletion mutant were able to substantially activate the p53-Luc reporter, whereas an inhibition would be expected in the general inactivation or transcriptional squelching mechanism.
Sp1 Transcription Factor Interacts with the P4 Site Sequence in Vitro-Southwestern blotting of cell extracts was performed to study other transcriptional factors that may participate in the regulation of POLD1 promoter by p53 repressor at the basal transcriptional level. MCF7 and RKO cells were treated with and without MMS. The radiolabeled 40-mer P4 oligonucleotide was used as a probe. Two signals were detected when the labeled P4 oligonucleotide was hybridized with the membrane, indicating that the P4 oligonucleotide interacted with two proteins of about 90 and 30 kDa in size (Fig. 7A, left panel,  lanes 1-4). The 90-kDa signal was of the same molecular size as that in lane 5, where purified recombinant Sp1 protein was used. The same membrane was Western-blotted with Sp1 polyclonal antibody, which detected a band of 90 kDa (Fig. 7A, right  panel). The 90-kDa Sp1 band superimposed well with the 90-kDa signal in the Southwestern blot. These data suggested that the P4 oligonucleotide complexed with the Sp1 protein from nuclear extracts in vitro. However, no signal was observed at the position of p53 in the Southwestern blot (Fig. 7A, left panel,  lanes 1-4). The lack of detectable interaction between the P4 oligonucleotide with p53 is not likely because of a lack of p53 protein. Western blotting of the same membrane showed that cellular p53 was induced by MMS treatment (Fig. 7A, right  panel, lanes 1-4). Because Southwestern blotting employs a denaturation step, it might result in the loss of native conformation and binding activity in some sensitive DNA-binding proteins such as p53. A stepwise denaturation-renaturation procedure using 7 M guanidine HCl under the same conditions did not show positive results either (data not shown). The 30-kDa signal in the Southwestern blot showed that the P4 oligonucleotide also interacted with an unknown cellular protein.
To confirm a sequence-specific interaction between Sp1 factor and the P4 Sp1-binding site, EMSA was performed using purified recombinant Sp1 protein and 40-mer oligonucleotides. The P4 oligonucleotide was shifted by Sp1 protein (Fig. 7B, lane  1) and supershifted by the addition of Sp1 polyclonal antibody (lane 2). Sp1 was able to bind to the mP4 oligonucleotide, in which the p53-binding site was mutated (lanes 3 and 4). However, when the mSp1 oligonucleotide, which contains a mutated Sp1-binding site, was used as a probe, Sp1/DNA binding was eliminated (lanes 5 and 6), indicating a specific binding of Sp1 to the Sp1-binding site in the P4 sequence.
p53 Partially Suppresses Sp1-stimulated POLD1 Promoter Activity via the P4 Site-The binding of the Sp1 factor to the P4 site adds another layer of complexity to the regulation of the POLD1 promoter, because Sp1-family proteins have been found to stimulate the basal transcriptional level of POLD1 promoter (15). To investigate the effect of p53 on the Sp1stimulated promoter activity, Sp1 and p53 expression plasmids were co-transfected into Drosophila SL2 cells, which lack endogenous Sp1 factors and have been useful in studying the function of Sp1 proteins (31). As shown in Fig. 8A (the second  bar in the pGLϪ1758 group), when the full-length POLD1 promoter/reporter construct pGLϪ1758 was co-transfected with Sp1 expression plasmid pPacSp1, the promoter was stim-ulated about 7-fold compared with the results for the control vector (first bar in pGLϪ1758 group). Similar results were obtained with pGLϪ1758mP4 in which the P4 p53-binding site  4), and 20 ng of purified recombinant Sp1 protein (lane 5) were resolved on 5-15% gradient SDS-PAGE and then transferred to a polyvinylidene difluoride membrane (PVDF-P SQ , Millipore). The membrane was hybridized with the 32 P-labeled P4 site 40-mer, the same probe used in EMSA. After washing, the membrane was exposed to x-ray film. Molecular size markers are labeled on the left. Right panel: Western blot of the same membrane probed with antibodies against Sp1 (upper section) or p53 (lower section). Positions of the p53 or Sp1 protein band are indicated. B, EMSA analysis of Sp1-P4 oligonucleotide interaction. Electrophoretic mobility shift assays were performed using Sp1 and oligonucleotides containing each mutated P4 site as indicated. A Gadd45 probe was used as a control. Purified Sp1 (30 ng) was incubated with each labeled 40-mer probe. Probe preparation and EMSAs were performed as described under "Experimental Procedures." Lanes 1, 3, 5, and 7: Sp1 protein ϩ probes. Lanes 2, 4, 6, and 8: Sp1 ϩ Sp1 polyclonal antibody PEP2 ϩ probes. Bands for free probe, protein-DNA complex, or antibody-protein-DNA complex are indicated. Each oligonucleotide contains mutated sequences as listed. The P4 p53-binding site is underlined, and the P4 Sp1-binding site is double underlined. Letters in bold italics indicate mutated sequences. was mutated. Elimination of the P4 Sp1-binding site by mutation (pGLϪ1758mSp1) resulted in an approximately 50% reduction of promoter activity, indicating that the P4 Sp1-binding site is responsible for about half of the Sp1-dependent stimulation. Because the POLD1 promoter contains multiple Sp1 sites, this result is not surprising, but nevertheless shows that the Sp1 site in P4 is a major site for Sp1 activation of the POLD1 promoter.
The regulatory effects of p53 on the Sp1-stimulated POLD1 promoter were examined by co-transfection of the p53 expression plasmid pPac-p53 with the Sp1 plasmid pPas-Sp1, as shown in the third bar in each group (Fig. 8A). The Sp1stimulated activity of the full-length promoter was reduced by about 62% by exogenous p53 but was not completely eliminated. This indicated that p53 was able to inhibit the Sp1stimulated POLD1 promoter activity and also showed that part of the Sp1-stimulated promoter activity was independent of p53 repression. However, bearing in mind that mutation of the P4 Sp1 site (pGLϪ1758mSp1) led to a reduction of about 50% in Sp1-stimulated promoter activity, the results are consistent with those expected if p53 completely inhibited the ability of Sp1 to act at the P4 site. When the p53 plasmid was co-transfected with the pGLϪ1758mP4 reporter in which the P4 p53binding site is mutated, only a slight inhibition was observed. Similar results were obtained with the pGLϪ1758mSp1 reporter where the P4 Sp1-binding site is mutated. These data provide evidence that the action of p53 on the POLD1 indicated is largely mediated through the P4 Sp1 site. However, the results also indicate that in addition to the P4 p53-binding site, there are some other sites through which p53 was able to down-regulate POLD1 promoter in a minor way. In Fig. 8B, extracts from appropriately transfected SL2 cells were examined to show equivalent levels of exogenous Sp1 or p53 expression.
p53 Competes with Sp1 Protein for Binding to the P4 Sequence in EMSA-The fact that p53 and Sp1 individually bind to corresponding sites in close proximity suggested a mechanism whereby p53 may interact with Sp1 protein on the P4 Sp1 cis-element to suppress Sp1-stimulated POLD1 promoter activity. To investigate this possibility, EMSA was performed using purified recombinant Sp1 and p53 protein for binding to the P4 oligonucleotide. In this experiment, different molar ratios of Sp1 and p53 were tested (these were calculated on the assumption that both protein are monomers, although it should be noted that both p53 and Sp1 bind to DNA as oligomers). At a nominal 1:1 molar ratio, p53 was unable to affect Sp1-P4 oligonucleotide interaction (Fig. 9, lane 6). However, as the ratio of p53 increased, there was a corresponding appearance of the p53-P4 oligonucleotide complex and an attenuated binding of Sp1 to the P4 oligonucleotide (Fig. 9, lane 7). A 15-fold molar excess of p53 completely abrogated Sp1-DNA complex formation and resulted in the appearance of a strong p53-DNA complex and an additional band due to nonspecific binding (Fig. 9,  lane 8). DISCUSSION p53 is a powerful tumor suppressor that acts as a checkpoint in maintaining genomic stability by inducing cell cycle arrest or apoptosis after DNA damage, largely through its function as a transcriptional regulator. Following enhanced expression and activation of p53 as a transcription factor, it also participates in a process by which DNA replication pauses, presumably until DNA repair is complete. Only a few p53 target genes have been FIG. 8. partial suppression of Sp1-stimulated POLD1 promoter activity by p53 via the P4 site. A, luciferase assays of POLD1 promoter activity for co-transfection in Drosophila SL2 cells, Promoter/ reporter constructs together with Sp1 expression plasmid and p53 expression plasmid were used for co-transfection assays (see "Experimental Procedures"). For Sp1 ϩ p53, the pPac-Sp1 expression plasmid and pPac-p53 expression plasmid were used. For Sp1, pPac-Sp1 expression plasmid and pPacU control plasmid were used. For the control, pPacU control plasmid was used. pGL2 is the pGL2 basic control reporter vector; pGLϪ1758 is the full-length POLD1 promoter/reporter construct; pGLϪ1758mSp1 and pGLϪ1758mp53 are promoter/reporters with mutations in the Sp1-binding site and the p53-binding site, respectively. Aliquots of cell extracts were assayed for luciferase activ- identified as involved in the inhibition of DNA replication. Among these are p21/Cip1/WAF1 (18) and Gadd45 (22). Nevertheless, as evidenced by the example of p21/Cip1/WAF1, regulation of a single p53 transcriptional target gene is generally not sufficient to account for the biological functions of p53 (19,20). In this regard, the identification of p53 target genes that participate in the regulation of DNA replication is an important objective, because the control of DNA synthesis is clearly one of the key cellular processes that is impacted by p53. In this study, the POLD1 gene was identified as a novel target of p53 transcriptional regulation, and evidence was obtained for a competition by p53 with Sp1 at a novel p53 binding sequence (P4), which harbors an internal Sp1 site.
Care was taken in this study to use a tet-off cell system in which ectopic wt-p53 expression was induced to a level similar to that observed in MMS-treated MCF7 cells. The H24-p53-14 cell line used in this study is one of the tet-off cell lines that has been utilized to investigate the role of p53 in G 1 cell cycle arrest or apoptosis. Transcripts of genes such as p21/Waf1/Cip1, mdm2, Gadd45, and bax were also shown to be activated in these cell lines (29). In this study, POLD1 mRNA was found to be repressed in response to both MMS-and tetracycline-induced p53 expression, suggesting that POLD1 repression by p53 is physiologically relevant.
The results from this study provide evidence for a mechanism involving competition between p53 and Sp1 for binding to the P4 site of the POLD1 promoter, which contains both a p53 and a Sp1 site. This type of exclusion mechanism for p53 transcriptional repression may be more general, as evidence has been found for competition between p53 and other transcriptional factors in other systems. The repression of ␣-fetoprotein expression by p53 was found to involve competition between binding of p53 and hepatic nuclear factor-3 for overlapping sites in the ␣-fetoprotein promoter (23), a situation similar to that observed for the POLD1 promoter. Studies of the inhibition of the cyclooxygenase-2 gene by p53 have revealed that p53 competes with TATA-binding protein in a 40-bp region in the promoter close to the transcription start site; however, the specific sequence elements involved have not been identified (24). In vitro competition between p53 and Sp1 for binding to overlapping DNA sequences has been reported for the SV40 and HIV-LTR (human immunodeficiency virus) viral promoters (38,39). The requirement for multiple Sp1 binding sites within the p53-responsive region for the transcriptional inhibitory effect of wild-type p53 has recently been shown on cellular promoters for vascular endothelial growth factor gene (VEGF) (40) and human telomerase catalytic subunit gene (hTERT) (41). Therefore, the findings that p53 is able to exclude the binding of Sp1 by overlapping binding sites and exert a repressive effect on POLD1 gene transcription provides an important example of Sp1-p53 exclusion on a cellular gene promoter, considering a wealth of recent evidence that p53 is able to inhibit Sp1-mediated promoter activity of many genes such as insulin-like growth factor-I receptor gene (26), Werner helicase gene (27), and the multiple drug resistance-associated protein gene (42). In addition, Sp1-binding sites are present in the promoters of many growth-regulated genes (43). It may be important to consider that such inhibitory effects of p53 on Sp1-mediated promoter activity could be extended to the regulation of other cellular promoters that contain overlapping p53-and Sp1-binding sites. There is also evidence that there are structural similarities in the binding motifs for Sp1 and p53 (44), providing a structural rationale for the existence of overlapping Sp1-p53 binding sites.
Biochemical evidence for physical interactions between p53 and Sp1 has been lacking. Attempts to detect a stable Sp1-p53 heterocomplex using cell extracts from MMS-treated MCF7 cells in the presence or absence of P4 oligonucleotide were negative, although positive signals were obtained from experiments with extracts in which p53 was expressed ectopically (data not shown). In addition, we were unable to detect a Sp1-p53-DNA complex in our EMSA experiment (Fig. 9). The negative results for Sp1-p53 heterocomplex formation are consistent with the co-immunoprecipitation and EMSA results obtained by Bargonetti et al. (39). Nevertheless, in some studies, Sp1 has been found to form a heterocomplex with mutant or proliferative forms of p53 in regulating promoter activities (45,46). Furthermore, we cannot exclude the operation of a second level of transrepression involving p53 interaction with other transcription factors, as the co-transfection experiments involving Sp1 and p53 indicated that a competition at the P4 site would account for only about half of the observed p53 effects on Sp1-driven promoter activity.
p53 has been shown to transrepress several other DNA replication-related genes. In a "DEX-on" cell line, dexamethasoneinduced wild-type p53 protein expression led to cell growth inhibition and a significant decrease in steady-state mRNA level of DNA polymerase ␣ catalytic subunit gene. The mechanism of such inhibition has not been investigated (47). Wildtype p53 also confers cell line-dependent repression of the promoter of PCNA gene, the product of which is well established to be involved in DNA replication/repair and cell cycle regulation (48,49). Wild-type but not mutant p53 induced by DNA damaging agents was able to cause a decrease in DNA topoisomerase II␣ mRNA and promoter activity. However, such transrepression is not by p53 sequence-specific binding (25,50). Recent studies of genes regulated by p53 have revealed additional DNA replication genes that are repressed by p53, including DNA primase polypeptide I, chromosomal protein HMG-17, and RFC 37, which encodes one of the small subunits of replication factor C (RFC) that loads and unloads PCNA clamps (51)(52)(53).
The exact biological functions/consequences of the transrepression of DNA replication-related genes are still unclear. These genes are generally expressed in late G 1 /S-phase when cells have already passed the cell cycle restriction point and are committed to enter S phase and initiate DNA synthesis. It is conceivable that p53-mediated POLD1 repression serves to halt DNA replication as the consequence of p53-induced cell cycle arrest, which involves immediate-early genes such as c-jun, c-fos, c-myc, and other cell cycle regulation genes such as p21/WAF1/Cip1, Cyclin/Cdk, and Rb. It is also possible that the transcriptional repression of POLD1 and other DNA replication genes may be the result of p53-induced prolonged, probably permanent, G 1 delay based on the fact that the expression of DNA polymerase ␣ (54) and DNA polymerase ␦ 2 are significantly inhibited during myeloid and erythroleukemia cell terminal differentiation. Growing lines of evidence have argued that loss of p53 results in decreased genomic stability, not because of loss of transient checkpoint controls but because of a failure to prevent severely damaged cells from surviving and propagating, underscoring the important role of prolonged G 1 arrest in maintaining DNA integrity (55). Further studies are needed for insightful understanding of the biological effects of the p53 transrepression of DNA polymerase ␦ catalytic subunit gene. However, it is possible that loss of wild-type p53 function in cancer cells might rescue the POLD1 promoter activity normally repressed by wild-type p53, contributing to unregulated cancer cell growth and proliferation.
It seems paradoxical that after the genome is harmed by 2 B. Li and M. Y. W. Lee, unpublished data.
DNA-damaging agents, p53, the "genome guardian," represses the transcription of POLD1 gene, the product of which is also suggested to be involved in DNA repair. A possible scenario is that p53 arrests cell cycle, halts DNA replication, and leads to the repression of POLD1 transcription. Simultaneously, p53 stimulates initiation of DNA repair, which involves DNA polymerase ⑀ and DNA polymerase ␤. There are some reports that provide support for this possibility. 1) pol ⑀ functions as a sensor for DNA replication coordinating the transcriptional and cell cycle response to replication block (56). 2) pol ⑀ or pol ␦ components perform equally well in reconstituted nucleotideexcision repair assays (6). 3) In yeast genetic studies, pol ⑀ and pol ␦ are able to efficiently substitute for each other in nucleotide excision repair after uv damage (57). (4) Transcription of pol ␤, the major DNA polymerase in base-excision repair, is not repressed by overexpressed p53 (49). In summary, the present studies show that POLD1 gene expression is repressed by wild-type p53. The evidence indicates that the majority of transcriptional repression was mediated by sequence-specific binding of p53 to the P4 p53-binding site in the POLD1 promoter. Examination of the mechanism by which p53 represses the POLD1 promoter led to the identification of a competition between p53 and Sp1 for overlapping binding sites in the P4 sequence of the POLD1 promoter. Thus, the present finding that POLD1 is subject to p53 transrepression adds the DNA polymerase ␦ catalytic subunit gene, a central enzyme of DNA replication, to the growing group of DNA replication/repair genes inhibited by p53. These findings also add to the currently emerging body of evidence that sequence-specific binding of p53 to promoter sequences can result in transcriptional repression in addition to the known effects of p53 as a transcriptional activator. However, the biological functions/consequences of the transrepression of POLD1 by p53 in response to DNA damage is more complicated and remain to be determined. Future investigation of POLD1 transcriptional regulation by other cis-elements and trans-factors and the interaction between p53 and trans-factors may be helpful in understanding the biological significance of POLD1 gene regulation by p53 in response to DNA damage in cellular processes such as DNA replication/repair, cell cycle arrest, and apoptosis.