p53 Mediates Repression of the BRCA2 Promoter and Down-regulation of BRCA2 mRNA and Protein Levels in Response to DNA Damage*

Adriamycin and other DNA-damaging agents have been shown to reduce BRCA2 mRNA levels in breast cancer cell lines, but the mechanism by which this oc-curs is unknown. In this study, we show that adriamycin and mitomycin C, but not other DNA-damaging agents, repress BRCA2 promoter activity in a dose- and time-de-pendent manner. We demonstrate that the effect is dependent on wild type p53 and that adriamycin and p53 mediate repression of the BRCA2 promoter by inhibiting binding of an upstream stimulatory factor protein complex to the promoter. In addition, we present evidence indicating that adriamycin and other DNA-damaging agents reduce BRCA2 mRNA and protein levels by alter-ing both BRCA2 mRNA stability and protein stability. Thus, BRCA2 levels in the cell are regulated by three independent mechanisms in a p53-dependent manner. The BRCA2 gene was identified in 1996 as a breast and ovarian cancer susceptibility gene 2). The BRCA2 gene encodes 3,418-amino acid, cell cycle-regulated, in the response to The came six-well g BRCA2 promoter luciferase reporter construct g of pRL-TK-Renilla M

The BRCA2 gene was identified in 1996 as a breast and ovarian cancer susceptibility gene (1,2). The BRCA2 gene encodes a 3,418-amino acid, cell cycle-regulated, nuclear phosphoprotein (3,4) that has been implicated in the response to DNA damage. The evidence for a role in DNA repair came initially from the observation that BRCA2 binds directly with RAD51 through the exon 11-encoded BRC repeats (5,6) and through an additional C-terminal binding site in the mouse (7). This association with a protein involved in meiotic and mitotic recombination and DNA double-stranded break repair suggests a similar role for BRCA2. Further support for a role in DNA repair comes from the observation that cells expressing a wild type BRCA2 BRC4 domain show hypersensitivity to ␥-irradiation, an inability to form RAD51 radiation-induced foci, and a failure of radiation-induced G 2 /M, but not G 1 /S, checkpoint control (8). Moreover, cells expressing mutant BRCA2 are more sensitive to methyl methanesulfonate-induced DNA damage than cells expressing wild type BRCA2 (9). Animal models have also been used to demonstrate an association between BRCA2 and the DNA damage response. Brca2-null mouse embryos that do not survive past day 8 of embryogenesis are highly sensitive to ␥-irradiation (7). Similarly, mouse embryo fibroblast cell lines derived from viable brca2 Ϫ/Ϫ animals are highly sensitive to DNA damage induced by a number of agents (10,11). BRCA2 has also been directly implicated in homologous recombination and gene conversion using CAPAN-1 BRCA2 mutant cell lines and homozygous mutant brca2 embryonic stem cells (12,13) and in transcription-coupled repair in response to 8-oxoguanine treatment (14). Most recently, the C terminus of BRCA2 has been shown to bind directly to singlestranded DNA and to promote strand transfer and RAD51 loading onto DNA during homologous recombination (15).
The finding that BRCA2 was involved in the response to DNA damage led to the hypothesis that DNA damage might result in the induction of BRCA2 expression. However, the opposite has proven true. Specifically, BRCA2 mRNA levels were significantly down-regulated in breast and ovarian cancer cell lines after exposure to various DNA-damaging agents including adriamycin (ADR) 1 and camptothecin (16,17). In an effort to characterize this response to DNA damage better we investigated whether ADR and other DNA-damaging agents affected BRCA2 promoter activity and BRCA2 mRNA and protein levels. Here we show that ADR down-regulates the BRCA2 promoter in a p53-dependent manner by inhibiting binding of a USF transcription factor complex to the minimal BRCA2 promoter. In addition, we demonstrate that ADR alters BRCA2 mRNA and protein stability in a p53-dependent manner, resulting in a significant reduction in BRCA2 mRNA and protein levels. This suggests that BRCA2 levels are regulated tightly in response to DNA damage.

EXPERIMENTAL PROCEDURES
Plasmids-The preparation of a series of pGL3 luciferase reporter constructs (Promega) containing partial fragments of the BRCA2 promoter has been described previously (18). A pcDNA3.1 plasmid containing the wild type p53 cDNA (wtp53) was provided by Wilma Lingle. CMV-USF1 and USF2-VP16 expression constructs were provided by Michele Sawadogo and Howard Towle, respectively. An R273L dominant negative p53 mutant construct (dnp53) was generated by sitedirected mutagenesis of the wtp53 construct using the QuikChange kit (Qiagen) according to the manufacturer's instructions. PCR primers for the site-directed mutagenesis were 5Ј-GGAACAGCTTTGAGGTGCTT-GTTTGTGCCTGTCCTGG-3Ј (forward) and 5Ј-CCAGGACAGGCACAA-ACAAGCACCTCAAAGCTGTTCC-3Ј (reverse). Plasmid DNA was isolated from colonies, and the presence of the mutation was confirmed by DNA sequencing.
Cell Culture-HCT116/p53 ϩ/ϩ and HCT116/p53 Ϫ/Ϫ cells were provided by Junjie Chen. HCT116/p21 ϩ/ϩ and HCT116/p21 Ϫ/Ϫ cells were provided by Wafik El-Deiry. MCF7/pCMV and MCF7/E6 cells were provided by Scott H. Kaufmann. All other cell lines were obtained from * This work was supported by Grant DAMD17-97-1-7048 from the United States Army Medical Research and Materiel Command. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Transient Transfection and Luciferase Reporter Assays-Transient transfections were performed in six-well plates using FuGENE 6 transfection reagent (Roche Molecular Biochemicals) with 0.5-2.0 g of BRCA2 promoter luciferase reporter construct and 0.1 g of pRL-TK-Renilla luciferase control vector (Promega). For ADR (Sigma) treatment experiments, cells were transfected, grown for 24 h, and exposed to 0, 2.5, 5.0, or 10 M ADR for 1 h in standard medium. The cells were washed with serum-free medium and incubated at 37°C in fresh culture medium for another 24 h or the indicated time. Because less than 10% of any cell type exhibited a toxic response to 5 M ADR, drug toxicity had no effect on the outcome of the study. In cotransfection experiments, cells were also transfected with 0.5 g of wtp53, dnp53, USF1, USF2-VP16, or pcDNA3.1 control. Protein lysates were prepared from the cells, and luciferase activities were measured as described previously (18). Renilla luciferase activity from a cotransfected pRL-TK control vector was used for normalization. In some lower dose experiments cells were exposed to 0.7 M ADR for 24 h, or the indicated time, in standard medium and processed immediately.
Electrophoretic Mobility Shift Assays (EMSAs)-Double-stranded oligonucleotides containing bp Ϫ10 to Ϫ35 of the BRCA2 promoter and either a wild type or mutated USF binding site (Ϫ20 to Ϫ13) were labeled with [␥-32 P]ATP and used in EMSAs (18). Double-stranded DNA probes were purified from the reaction mixture using a Bio-Gel P-100 column (Bio-Rad), incubated with whole cell extract from MCF7 cells, and separated on 5% polyacrylamide gels as described previously (18). Supershift assays using anti-USF1 (Santa Cruz), and anti-USF2 (Santa Cruz) antibodies were also performed as described previously (18).
Northern Blotting-Cells were transiently transfected with pcDNA3.1, wtp53, or dnp53 expression constructs or exposed to 5 M ADR for 1 h. After further incubation for 24 h, poly(A) ϩ RNA was isolated. RNA samples (1.5 g/lane) were used for Northern blotting as described previously (18).
Flow Cytometry-MCF7 cells treated with ADR were washed with phosphate-buffered saline, collected by centrifugation, and fixed in icecold 95% ethanol at Ϫ20°C for 12 h. The fixed cells were permeabilized and stained with 0.05% Triton X-100, 100 g/ml RNase A, and 40 g/ml propidium iodide at 37°C for 30 min and analyzed by flow cytometry. The data were processed with VERITY ModFit software, version 5.2, for DNA distribution analysis.
mRNA Stability-MCF7 cells were incubated in the presence or absence of 10 g/ml ␣-amanitin (Sigma) and 5 M ADR for 1 h. Medium was removed, and cells were incubated further in the presence or absence of ␣-amanitin. Total RNA was extracted from the cells by standard methods after different periods of incubation. Total cDNA was prepared by oligo(dT) priming of 1 g of RNA template from each time point using the cDNA preamplification kit (Invitrogen). The amount of BRCA2 mRNA present in the cells at each time point was measured by semiquantitative reverse transcription PCR using oligonucleotide primer pairs from the 5Ј-and 3Ј-ends of the BRCA2 cDNA. GAPDH was used as an internal control. The BRCA2 forward primer was 5Ј-CAA-GCAGATGATGTTTCCTGTCC; BRCA2 reverse, 5Ј-AGAACTAAGGGT-GGGTGGTGTAGC; BRCA2 forward, 3Ј-GCAGTGAAGAATGCAGC-AGA; BRCA2 reverse, 3Ј-CAATACGCAACTTCCACACG; GAPDH forward, CAACTACATGGTTTACATGTTC; GAPDH reverse, GCCA-GTGGACTCCACGAC. PCR products were separated on 10% polyacrylamide gels, stained with Sybr Green, and scanned on a PhosphorImager.
Protein Stability-MCF7 cells were incubated with 20 g/ml cycloheximide and 5 M ADR for 1 h. Medium was removed, and cells were incubated further in the presence of cycloheximide. Total protein was extracted from the cells by standard methods after different periods of incubation, and Western blotting with the BRCA2 Ab-2 antibody was performed. The intensities of Western blot signals were quantitated by densitometry using NIH Image software, version 1.62.

RESULTS
The BRCA2 Promoter Is Repressed by ADR-ADR has recently been shown to down-regulate BRCA2 mRNA levels in human breast cancer cells such as MCF7 (16). To address whether the reduction in BRCA2 mRNA levels by ADR is dependent on BRCA2 promoter regulation, MCF7 cells were transiently transfected with a BRCA2 promoter reporter gene construct (pGL3Prom) (18) and exposed to various concentrations of ADR for 1 h. After removal of the drug and further propagation for 24 h, cell lysates were harvested for luciferase reporter assays. As shown in Fig. 1A, ADR treatment reduced BRCA2 promoter activity in a dose-dependent manner, with 5 M ADR down-regulating the promoter activity by 85%. Similar effects of ADR treatment on the BRCA2 promoter were detected in both MCF10A normal breast epithelial cells (Fig. 1B) and U2OS osteosarcoma cells (data not shown). Continuous exposure of MCF7 cells to lower doses of ADR (0.7 M) for 24 h FIG. 1. Dose-and time-dependent repression of BRCA2 promoter activity by ADR. A, ADR inhibition of BRCA2 promoter activity in MCF7 cancer cells is dose-dependent. Luciferase activity was measured in MCF7 cells that were transfected with a BRCA2 promoter reporter construct (pGL3Prom), exposed to different concentrations of ADR for 1 h, and incubated for a further 24 h. B, ADR inhibition of BRCA2 promoter activity in normal MCF10A cells is dose-dependent. Luciferase activity in MCF10A cells was measured as described in A. C, low dose ADR inhibits the BRCA2 promoter. Luciferase activity was measured in pGL3Prom-transfected MCF7 cells after exposure to 0.7 M ADR for 24 h. D, ADR-associated repression of the BRCA2 promoter is time-dependent. Luciferase activity was measured in MCF7 cells that were transfected with pGL3Prom, exposed to 5.0 M ADR for 1 h, and incubated further for different times. also reduced promoter activity by 60% (Fig. 1C). Furthermore, treatment of MCF7 cells with 5 M ADR for 1 h resulted in a gradual reduction in promoter activity over time (Fig. 1D).
Thus, ADR appears to repress the BRCA2 promoter in a doseand time-dependent fashion in both tumor and normal cell lines. To confirm that the ADR-dependent repression was specific to the BRCA2 promoter and was not a result of a generalized effect on gene transcription, several other reporter constructs were evaluated using luciferase assays. MCF7 cells were transiently transfected with reporter constructs containing the Rous sarcoma virus, CMV, nuclear factor-〉, and ␤galactosidase promoters and treated with 5 M ADR. No repression of these promoters was observed (data not shown), suggesting that ADR specifically represses the BRCA2 promoter.
Repression of the BRCA2 Promoter by ADR Is Dependent on p53-ADR is a potent DNA-damaging agent that induces p53 accumulation and p53-dependent cell death in wtp53-expressing cell lines (19). Given that ADR regulates the BRCA2 promoter and induces p53, we investigated whether ADR downregulates the BRCA2 promoter in a p53-dependent manner. Initially, BRCA2 reporter assays were performed in p53-positive U2OS and p53-null Saos2 osteosarcoma cells in an effort to address the role of p53 in regulation of the promoter while minimizing tissue-specific differences. Treatment with either 5 M ADR for 1 h or 0.7 M ADR for 24 h did not affect BRCA2 promoter activity in Saos2 cells, whereas inhibition of the promoter was observed in U2OS cells ( Fig. 2A). Similarly, ADR treatment of matched p53-null (HCT116/p53 Ϫ/Ϫ ) and p53 wild type (HCT116/p53 ϩ/ϩ ) HCT116 cells resulted in repression of the BRCA2 promoter in p53 wild type cells but not in p53-null or mutant cells (Fig. 2B). In addition, reporter assays were performed in MCF7 cells stably expressing either the human papilloma virus type 16 (HPV-16) E6 gene (MCF7/E6) or a CMV vector control (MCF7/CMV). The E6 protein stimulates degradation of p53 through a ubiquitin pathway (20,21), resulting in very low levels of p53 expression in these cells. Reduced levels of p53 in MCF7/E6 cells was verified by Western blotting (data not shown). A decrease in BRCA2 promoter activity in response to ADR treatment was only detected in the MCF7/CMV cells (Fig. 2C). Taken together, these data suggest that repression of BRCA2 promoter activity by ADR is dependent on the presence of wtp53.
Given that ADR-associated repression of the BRCA2 promoter is dependent on p53, it seemed likely that p53 could regulate the BRCA2 promoter independently of ADR treatment. To address this hypothesis a series of BRCA2 promoter reporter assays was performed in MCF7 and Saos2 cells transiently transfected with wtp53, dnp53, or a pcDNA 3.1 vector control. Wtp53 expression in MCF7 (Fig. 2D) and Saos2 cells (Fig. 2E) reduced BRCA2 promoter activity 10fold relative to a vector control, and transfection of increasing concentrations of wtp53 into these cells resulted in dose-dependent repression of the promoter (data not shown). In contrast, ectopic expression of the dnp53 mutant markedly up-regulated the BRCA2 promoter (Fig. 2D). This apparent induction of promoter activity is likely caused by dnp53associated inactivation of endogenous p53 and consequent FIG. 2. p53 represses the BRCA2 promoter. A, repression of the BRCA2 promoter in osteosarcoma cells is p53-dependent. Luciferase activity from pGL3Prom was measured 24 h after treatment with 5 M ADR or Me 2 SO (DMSO) for 1 h in matched U2OS (p53 ϩ/ϩ ) and Saos2 (p53 Ϫ/Ϫ ) osteosarcoma cell lines. B, ADR does not repress BRCA2 promoter activity in p53-deficient HCT116 cells. Luciferase activity from pGL3Prom was measured 24 h after treatment with 5 M ADR or Me 2 SO for 1 h in matched HCT116/p53 ϩ/ϩ and HCT116/p53 Ϫ/Ϫ cells. C, ADR does not repress the BRCA2 promoter after E6-associated degradation of p53. Luciferase activity from pGL3Prom was measured 24 h after treatment with 5 M ADR or Me 2 SO for 1 h in matched MCF7/CMV and MCF7/E6 cells. D, ectopic expression of p53 in MCF7 (p53 ϩ/ϩ ) cells inhibits BRCA2 promoter activity. Luciferase activity from pGL3Prom was measured 24 h after treatment with 5 M ADR or Me 2 SO for 1 h in MCF7 cells transiently transfected with wtp53, dnp53 (R273L), or control vector. E, ectopic expression of p53 inhibits BRCA2 promoter activity in Saos2 (p53 Ϫ/Ϫ ) cells. Luciferase activity from pGL3Prom was measured 24 h after treatment with 5 M ADR or Me 2 SO for 1 h in Saos2 cells transfected with wtp53, dnp53 (R273L), or control vector. F, MMC and actinomycin D (Act-D) inhibit BRCA2 promoter activity in a p53-dependent manner. Luciferase activity from pGL3Prom was measured after 24 h of treatment with 10 ng/ml actinomycin D or 30 g/ml MMC in MCF7 (wtp53), Saos2 (p53-null), and T47D (p53 mutant) cells.
relief of repression. Thus, wtp53 appears to have a specific inhibitory effect on the BRCA2 promoter.
To evaluate further the relevance of p53 to the ADR response, we repeated these reporter assays in the presence of ADR. As before, ADR treatment caused significant down-regulation of BRCA2 promoter activity in vector-transfected MCF7 cells, but had no effect in Saos2 cells (Fig. 2, D and E). However, when wtp53-transfected MCF7 cells were treated with ADR a 40-fold down-regulation of the promoter, or 2-fold greater repression than with wtp53 expression alone, was observed (Fig. 2D). The lack of a synergistic effect between ADR and wtp53 expression is further support for the concept that wtp53 and ADR influence the BRCA2 promoter through a common signaling pathway.
Repression of the BRCA2 Promoter by Other DNA-damaging Agents-We also evaluated whether other DNA-damaging agents associated with induction or stabilization of wtp53 were capable of repressing the BRCA2 promoter. MCF7 cells were treated with 30 g/ml mitomycin C (MMC), 10 ng/ml actinomycin D, 5-10 M camptothecin, 10 J/m 2 UV irradiation, and 2-10 Grays of ␥-irradiation, and luciferase assays were performed. MMC and actinomycin D repressed the BRCA2 promoter similarly to ADR, whereas the other agents had no apparent effect. In contrast, none of the agents had an effect on the BRCA2 promoter in the p53-null Saos2 and p53 mutant T47D cell lines (Fig. 2F). These data suggest that only certain DNA-damaging agents can repress the BRCA2 promoter and that the effect is uniformly dependent on the presence of wtp53.
In Vivo Repression of BRCA2 Expression by ADR Requires Wtp53-To confirm an in vivo effect of ADR and wtp53 on endogenous BRCA2 expression, BRCA2 mRNA levels were measured in MCF7 and Saos2 cells treated with 5 M ADR. BRCA2 expression was decreased substantially in MCF7 cells but was not affected in Saos2 cells after ADR treatment (Fig.  3A). Similarly, ectopic expression of wtp53 resulted in decreased BRCA2 expression in MCF7 and Saos2 cells, and the combination of ADR and wtp53 expression did not appear to have a synergistic effect. In addition, BRCA2 mRNA levels were not substantially affected by ADR treatment of MCF7 cells ectopically expressing dnp53 (Fig. 3B). The combination of these data and the results from the reporter assays described above strongly suggest that ADR and wtp53 can regulate BRCA2 expression in vivo through inhibition of BRCA2 promoter activity.
In Vivo Reduction in BRCA2 Protein Levels by ADR and Wtp53-Although ADR appears to down-regulate BRCA2 promoter activity and BRCA2 expression levels, down-regulation of BRCA2 protein levels must also be evident to speculate that ADR-dependent repression of BRCA2 transcription can affect BRCA2 function. To evaluate the effect of ADR on in vivo BRCA2 protein levels, MCF7, U2OS, and MCF10A cell lines were treated with 5 M ADR. As shown in Fig. 4A, BRCA2 levels were decreased significantly in ADR-treated cells. In addition, the amount of BRCA2 in MCF7 cells decreased in a dose-responsive manner after treatment with 0.5, 2.0, and 5.0 M ADR (Fig. 4B). The expected dose-dependent increase in p53 levels was also observed. To evaluate whether ADR treatment had a time-dependent effect on BRCA2 protein expression, MCF7 cells were treated with 5 M ADR for 1 h, and BRCA2 protein levels were measured over a 24-h period. A gradual decrease in BRCA2 levels over time was observed with the most significant changes occurring between 5 and 11 h post-treatment (Fig. 4C). Interestingly, the decrease in BRCA2 correlated inversely with changes in p53 levels. Similar results were obtained in response to continuous treatment with 0.7 M ADR over a 24-h period, although the most significant changes in BRCA2 and p53 were delayed and occurred between 9 and 13 h after initiation of treatment (Fig. 4D).
Next we evaluated whether the effect of ADR on BRCA2 protein levels was dependent on p53. Matched p53-null (HCT116/p53 Ϫ/Ϫ ) and p53 wild type (HCT116/p53 ϩ/ϩ ) HCT116 cells, matched MCF7/E6 and MCF7/CMV cells, SW480 p53 mutant cells, and Saos2 p53-null cells were treated with 5 M ADR, and cell lysates were Western blotted for BRCA2 with the Ab2 antibody. As shown in Fig. 4E, an ADR-dependent reduction in BRCA2 protein levels was observed in cells expressing wtp53 but not in p53-null or p53 mutant cells. Further analysis showed that ectopic expression of wtp53 in Saos2 cells (Fig. 4F) and MCF7 cells (data not shown) also caused a significant reduction in BRCA2 levels. When combining all of these data, it is evident that there is a strong correlation between the effects of ADR and p53 on BRCA2 protein levels and on BRCA2 promoter activity.
The ADR-and p53-responsive Element Is Adjacent to the Transcription Initiation Site-Having determined that ADR and p53 repress BRCA2 promoter activity, we sought to understand the mechanism of repression. We began by identifying the ADR-responsive region within the BRCA2 promoter using a series of luciferase reporter constructs containing deleted forms of an 8-kb form of the BRCA2 promoter in reporter assays (18). MCF7 cells were transfected with the various constructs, and luciferase activities were measured before and after treatment with 5 M ADR. All promoter constructs that contained the minimal promoter, located between nucleotides Ϫ58 and Ϫ1, were repressed equivalently by ADR treatment (data not shown), suggesting that the ADR-responsive cis-element is located in the minimal BRCA2 promoter. However, as deletion of the minimal promoter between Ϫ58 and Ϫ19 eliminated basal promoter activity (18), we could not directly evaluate the role of this region in the ADR response. To overcome this problem a series of promoter reporter constructs containing substitution mutations in the putative ATF and MLTF binding sites, in the known USF-binding element, and in the two repeats of the GCGTCACG tandem repeat sequence within the Ϫ58 to Ϫ19 region of the BRCA2 promoter was used (18). The reporter constructs containing the mutated Ϫ13 to Ϫ20 USF transcription factor binding site and the overlapping Ϫ17 to Ϫ24 repeat sequence were 3-fold less repressed in the pres- ence of ADR than any other reporter construct (data not shown). Similar results were obtained after ectopic expression of wtp53. These data suggest that this USF binding site that has been implicated previously in regulation of basal activity of the BRCA2 promoter (18) also regulates the response to ADR and wtp53.
Wtp53 Inhibits Binding of USF to the BRCA2 Promoter-To determine whether the USF transcription factor is directly involved in p53-and ADR-dependent regulation of BRCA2 promoter activity, the ability of USF to bind to the BRCA2 promoter was evaluated. Two 26-bp (Ϫ10 to Ϫ35) oligonucleotide probes containing either a wild type or mutated USF binding site (18) were used in gel shift assays with lysates from MCF7 cells that had been treated with Me 2 SO or 5 M ADR. As shown in Fig. 5A, a specific protein complex bound to the wild type oligonucleotide but not to the mutant oligonucleotide and the ability of this complex to bind DNA decreased over time when cells were treated with ADR (Fig. 5A).
To determine whether wtp53 was needed for ADR-dependent inhibition of complex formation, we repeated the gel shift assays using lysates from matched Saos2 and U2OS cells, matched HCT116/p53 Ϫ/Ϫ and HCT116/p53 ϩ/ϩ cells, and matched MCF7/CMV and MCF7/E6 cells (Fig. 5, B-D). A similar protein-oligonucleotide complex was evident in all cells, but the relative intensity of the complex was significantly greater in p53 mutant or p53-null cells than in matched cells with endogenous wtp53. Moreover, exposure to ADR further reduced the intensity of the protein⅐DNA complex in wtp53-expressing cells but had no effect in p53-null or mutant cells (Fig. 5, B-D).
In addition, complex formation was reduced substantially by ectopic expression of wtp53, similarly to the ADR treatment, but was enhanced by expression of dnp53. This suggests that dnp53 inactivates endogenous or induced wtp53 and prevents p53-dependent inhibition of complex formation (Fig. 5E).
To verify that the protein complex binding to the wild type oligonucleotide contained USF, it was demonstrated that anti-USF2 antibodies could supershift the protein⅐DNA complex (Fig. 5F). Furthermore, Western blotting with anti-USF1 and anti-USF2 antibodies demonstrated that the levels of these proteins were not altered in response to ADR or p53 expression (data not shown). Together these data suggest that the repressive effect of ADR and p53 on the BRCA2 promoter does not involve reduction in USF1 and USF2 levels but is dependent on altered binding of USF2 and/or USF1 to the promoter.
p53 Inhibits USF-dependent Induction of the BRCA2 Promoter-To establish further the relevance of USF to p53-and ADR-dependent repression of the BRCA2 promoter, we tested whether p53 could block induction of the BRCA2 promoter by USF. Initially, a USF2-VP16 fusion protein that binds to the USF site was used to transactivate the BRCA2 promoter in MCF7 cells (18), and then the effect of wtp53 and ADR on the activated promoter was evaluated. As shown in Fig. 6A, wtp53 partially inhibited USF2-VP16-associated activation of the BRCA2 promoter, whereas dnp53 further activated the promoter. Subsequently, the ability of wtp53 to inhibit activation of the promoter by USF1 alone was evaluated. Ectopic expression of USF1 in SW480 (p53 mutant) cells up-regulated the BRCA2 promoter 4.5-fold. However, ectopic expression of wtp53 completely blocked this effect (Fig. 6B). Similar effects were observed in the Saos2 p53-null and T47D p53 mutant cell lines (data not shown). Thus, p53 inhibits USF-dependent promoter activation.
ADR Enhances BRCA2 mRNA and Protein Degradation-The data described above show that ADR, MMC, and actinomycin D repress the BRCA2 promoter in a p53-dependent manner. However, although the effect is significant it does not entirely account for the rapid decrease in BRCA2 mRNA and protein levels. In fact, a comparison among promoter activity, mRNA levels, and protein levels at various time points after ADR exposure (Figs. 1D and 4C) indicates that BRCA2 mRNA and protein levels decreased more rapidly than BRCA2 promoter activity. This suggests that ADR can influence BRCA2 mRNA and protein levels through additional mechanisms and provides support for earlier observations that DNA-damaging agents can decrease BRCA2 mRNA levels (16,17)  certain damaging agents such as UV irradiation can enhance BRCA2 protein degradation (22).
To determine whether ADR altered the stability of BRCA2 mRNA, MCF7 cells were treated with ␣-amanitin to block transcription, and the rate of decrease in BRCA2 mRNA in cells treated with 5 M ADR and Me 2 SO was compared by semiquantitative reverse transcription PCR. As shown in Fig.  7, A and B, ADR substantially enhanced the rate of BRCA2 mRNA degradation. Quantitation by densitometry showed that the half-life of BRCA2 mRNA in MCF7 cells is normally 9 h, whereas in the presence of ADR it is reduced to ϳ3 h. In addition, by comparing mRNA levels from MCF7 cells treated with ADR alone or with ADR and ␣-amanitin at various time points we determined that promoter inhibition accounted for 50% of the reduction in BRCA2 mRNA levels.
We also attempted to measure the rate of BRCA2 protein degradation in the presence and absence of ADR. MCF7 cells were treated with cycloheximide, and BRCA2 protein levels were measured by Western blotting and densitometry. The half-life of the BRCA2 protein was estimated to be 5 h (data not FIG. 5. ADR and p53 inhibit binding of USF to the BRCA2 promoter. A, ADR-associated disruption of a BRCA2 promoter USF binding site protein⅐DNA complex is time-dependent. EMSAs were performed using oligonucleotide probes containing a wild type (W) or mutated (M) USF binding site from the BRCA2 promoter, and whole cell protein extract from MCF7 cells was incubated for 0, 5, 13, and 24 h after treatment with 5 M ADR for 1 h. B, disruption of the USF protein⅐DNA complex by ADR in osteosarcoma cells is p53-dependent. EMSAs of protein extracts from Saos2 (p53-null) and U2OS (p53 wild type) cells incubated for 24 h after treatment with 5 M ADR (ϩ) or Me 2 SO (Ϫ) for 1 h are shown. C, ADR does not disrupt protein⅐DNA complex formation in HCT116/p53 Ϫ/Ϫ cells. EMSAs of protein extracts from HCT116/p53 Ϫ/Ϫ and HCT116/p53 ϩ/ϩ cells incubated for 24 h after treatment with 5 M ADR or Me 2 SO for 1 h are shown. D, ADR does not disrupt protein⅐DNA complex formation after E6-associated degradation of p53. EMSA were performed of protein extracts from MCF7/CMV and MCF7/E6 cells incubated for 24 h after treatment with 5 M ADR or Me 2 SO for 1 h. E, ectopic expression of p53 disrupts protein⅐DNA complex formation. EMSAs were performed of protein extracts from MCF7 cells that had been transfected with vector, wtp53, and dnp53 and incubated for 24 h after treatment with 5 M ADR or Me 2 SO for 1 h. F, p53 disrupts binding of USF to the BRCA2 promoter. Supershift assays of protein⅐DNA complexes formed between wild type USF binding site oligonucleotides and protein extracts from MCF7 cells transfected with vector or wtp53 expression constructs were performed using anti-USF2 antibody.
shown). MCF7 cells were subsequently treated with cycloheximide and 5 M ADR, and the half-life was estimated to be 4 h (data not shown). Although the rate of protein degradation in response to ADR treatment was enhanced, it was not sufficient to account for a 50% reduction in the BRCA2 protein level 5 h after ADR treatment as shown in Fig. 4C. Interestingly, a previous study of UV-associated BRCA2 protein degradation suggested that an inducible factor required for active degradation of BRCA2 in response to UV treatment could not be synthesized in the presence of cycloheximide and that the rate of UV-dependent BRCA2 protein degradation could not be measured accurately (22). It is likely that a similar effect was encountered in our experiments and that the actual rate of BRCA2 protein degradation in response to ADR is more rapid than the rate detected in our cycloheximide experiments. Our finding that BRCA2 protein levels are reduced by ϳ50% and that BRCA2 mRNA levels are also reduced by ϳ50% in MCF7 cells treated for 5 h with ADR supports this observation because it is unlikely that the reduction in mRNA levels can account for the entire reduction in protein. As a result it is not possible to determine the relative contribution of promoter inhibition, mRNA degradation, and protein degradation to the reduction in BRCA2 protein in response to ADR, but it is clear that each mechanism makes a substantial contribution.

Repression of the BRCA2 Promoter by ADR and p53
Is Independent of Cell Cycle Arrest-It is not known whether p53-dependent cell cycle arrest is required for repression of the BRCA2 promoter or whether these events are independent of each other and are simply concurrent responses to p53 induction by ADR. To address this question, cell cycle profiles of cells were measured at various time points after treatment with ADR, and the results were compared with time course studies of BRCA2 promoter activity and BRCA2 mRNA and protein levels from cells that had been treated similarly. It was noted that ADR-dependent repression of the BRCA2 promoter preceded ADR-associated induction of an S and G 2 /M cell cycle checkpoint arrest (Fig. 8, A and B). This was noted after exposure to a short pulse of 5 M ADR and in response to continuous exposure to 0.7 M ADR. Down-regulation of BRCA2 mRNA and protein also appeared to precede the onset of cell cycle arrest.
Given that induction of cell cycle arrest by p53 is associated with induction of p21 Waf1/Cip1 , we also investigated whether p21 Waf1/Cip1 was required for the ADR and p53 effect on the BRCA2 promoter. The BRCA2 luciferase reporter construct was transiently transfected into HCT116/p21 ϩ/ϩ and matched p21 Ϫ/Ϫ cells, and promoter activity was measured after treatment with 5 M ADR and ectopic expression of wtp53. BRCA2 promoter activity was repressed equivalently in the matched cell lines by ADR (Fig. 8C) and p53 (Fig. 8D), indicating that the ADR and p53 effect is independent of p21 Waf1/Cip1 and suggesting that the BRCA2 promoter is repressed independently of the induction of G 1 /S checkpoint arrest. DISCUSSION The effect of cellular stress and DNA-damaging agents on BRCA2 has been a topic of much interest. Initially, BRCA2 mRNA levels were shown to be down-regulated in cells in response to treatment with DNA-damaging agents such as ADR, MMC, and UV irradiation (16). In this study, we show that as part of the cellular response to ADR-associated DNA damage the BRCA2 promoter is significantly repressed in a p53-dependent manner. We have established that in response to ADR, MMC, and actinomycin D, p53 inhibits binding of the USF transcription factor to the BRCA2 minimal promoter, resulting in repression of basal promoter activity and substantial decreases in BRCA2 mRNA and protein levels. The inability of ADR treatment to repress the promoter and decrease mRNA and protein levels in p53 mutant or null cells or in cells expressing the dominant negative R273Lp53 indicates that this is a p53-dependent process. However, it is important to note that only ADR, MMC, and actinomycin D repressed the promoter by inhibiting USF binding in a p53-dependent manner, whereas other DNA-damaging agents such as UV irradiation, ␥-irradiation, and camptothecin, which are also associated with induction of p53, had no effect. This raises the possibility that only certain forms of DNA damage can induce other factors, in addition to p53, which are required for inhibition of USF binding to the promoter.
Although UV irradiation, ␥-irradiation, and camptothecin fail to inhibit USF binding, these agents all reduce BRCA2 mRNA and protein levels. This suggests that these agents either regulate transcription of the BRCA2 gene through ciselements and enhancer/repressor binding sites that are not present in our 8-kb promoter construct or that these agents influence the stability of BRCA2 mRNA and protein. In this study we determined that ADR and other DNA-damaging agents enhance BRCA2 mRNA and protein turnover, resulting in decreased BRCA2 mRNA and protein levels. Given that BRCA2 mRNA and protein are only reduced by ADR in cells expressing wtp53, the suggestion is that p53 mediates BRCA2 promoter activity and BRCA2 mRNA and protein turnover in response to ADR, MMC, and actinomycin D but only regulates mRNA and protein turnover in response to other DNA-damaging agents. Thus, BRCA2 protein levels may be regulated by various DNA-damaging agents at the level of the promoter, mRNA stability, and protein stability.
In this work we determined that ADR and p53 repress the BRCA2 promoter by inhibiting the ability of USF to bind to the promoter. To our knowledge, this is the first report showing that p53-mediated transcriptional repression is functionally associated with USF binding. We have shown previously that binding of USF to the BRCA2 minimal promoter is required for basal transcription of BRCA2 and for induction of the promoter and increased expression of BRCA2 during the S and G 2 phases of the cell cycle (18). Thus, by inhibiting binding of USF to the promoter it appears that ADR treatment and p53 actually prevent activation of the BRCA2 promoter during S and G 2 phases of the cell cycle. Although p53 appears to regulate USF binding in response to ADR treatment, p53 does not seem to bind directly to USF, as evidenced by an inability of anti-p53 antibodies to supershift the USF complex and by failure of p53 to coimmunoprecipitate with USF1 or USF2 (data not shown). Because we have also shown that p53 does not regulate USF expression levels, it seems likely that ADR or p53 may regulate and p53 may interact with other protein(s) that in turn modulate the ability of USF to bind to the promoter. However, it is also possible that p53 may be a part of the USF protein complex, but its presence in the complex may not be detected easily by certain antibodies. USF1 has recently been identified as a phosphoprotein (23, 24) whose DNA binding activity is dependent on cyclin-dependent phosphorylation and can be inhibited by the p53-inducible cyclin-dependent kinase inhibitor, p21 Waf1/Cip1 , which blocks phosphorylation of USF1 (24). This suggests that ADR and p53 regulate BRCA2 expression through p21 Waf1/Cip1 . However, we found that p53 and ADR repressed BRCA2 promoter activity in cells lacking p21 Waf1/Cip1 , suggesting that p21 Waf1/Cip1 plays no role in USF-dependent regulation of the BRCA2 promoter.
Repression of transcription of several genes by p53 is thought to be the consequence of p53-dependent inhibition of other transcriptional activators (25)(26)(27) or components of the basal transcription machinery (28 -30). One mechanism of p53associated repression utilizes histone deacetylases, mediated by interaction with Sin3a, to regulate target genes such as map4 and stathmin negatively (19) and to repress the CHK1 gene through the p21 Waf1/Cip1 protein (31). Another mechanism of p53-dependent repression involves binding of p53 to p300/ CBP and subsequent interference in coactivation of p300/CBPdependent factors, such as AP-1 (26), hypoxia-inducible factor 1 (32), and nuclear factor-B (33,34). Interestingly, it has been reported that p300 interacts functionally with USF to potentiate the activation of USF target genes (35). Whether the USFdependent regulation of the BRCA2 promoter by p53 is actually mediated by p300/CBP or histone deacetylases or as yet unidentified factors remains to be determined.
There is substantial evidence that BRCA2 plays a role in DNA repair. Therefore, it is surprising that BRCA2 is downregulated by p53 in response to ADR-associated DNA damage. One possible explanation is that p53 must down-regulate BRCA2 to induce cell cycle arrest and DNA damage repair because BRCA2 appears to interact with p53 in a RAD51⅐p53⅐BRCA2 complex and may partially repress p53dependent transactivation of target promoters such as p21 Waf1/CIP1 (36). Alternatively, BRCA2 may interfere with induction of p53-dependent apoptosis in response to DNA damage in a similar manner. However, because BRCA2 is itself involved in DNA repair this seems unlikely. It is also possible that down-regulation of BRCA2 by p53 allows cells to initiate apoptosis in response to DNA damage more efficiently, while the presence of BRCA2 provides sufficient DNA repair to pro-FIG. 8. ADR and p53 repression of the BRCA2 promoter is independent of p21 Waf1/CIP1 and cell cycle arrest. A, continuous exposure to low dose ADR induces a G 2 /M cell cycle arrest. Fluorescence-activated cell sorter analysis of the percentage of PI-stained MCF7 cells in each phase of the cell cycle was performed after various periods of exposure to 0.7 M ADR. B, high dose ADR induces cell cycle arrest at S and G 2 /M checkpoints. Fluorescence-activated cell sorter analysis of the percentage of PI-stained MCF7 cells in each phase of the cell cycle was performed after different periods of incubation after treatment with 5 M ADR for 1 h. C, p21 Waf1/CIP1 is not required for ADR repression of the BRCA2 promoter. Luciferase activity from the BRCA2 promoter reporter construct was measured 24 h after exposure of HCT116/p21 ϩ/ϩ and HCT116/p21 Ϫ/Ϫ cells to 5 M ADR or Me 2 SO for 1 h. D, p21 Waf1/CIP1 is not required for repression of the BRCA2 promoter by p53. Luciferase activity from the BRCA2 promoter reporter construct was measured 24 h after transfection of HCT116/p21 ϩ/ϩ and HCT116/p21 Ϫ/Ϫ cells with vector, wtp53, and dnp53 expression constructs. long cell viability. An alternative possibility is that p53 may down-regulate BRCA2 after DNA repair is complete to inactivate the DNA repair machinery and to release the cell from checkpoint arrest. This could also be a mechanism by which p53 regulates the extent and timing of DNA damage repair. The observation that p53 only significantly affects the BRCA2 promoter and BRCA2 protein levels 6 h after induction of DNA damage supports this possibility.
Previous reports have indicated that BRCA2 mRNA and protein levels change in a cell cycle-dependent manner, with low levels in G 1 and significantly elevated levels in S and G 2 phases of the cell cycle (3,4). In the work described above we have shown that repression of the BRCA2 promoter and associated down-regulation of BRCA2 mRNA and protein levels preceded p53-dependent S and G 2 /M phase cell cycle arrest in response to ADR treatment. This indicates that BRCA2 downregulation is dependent on p53 but is independent of the effect of p53 induction on the cell cycle. Future experiments will determine whether BRCA2 down-regulation is required or contributes to p53-dependent cell cycle arrest in response to DNA damage.
In conclusion, we have demonstrated that the BRCA2 promoter is down-regulated by ADR in a p53-dependent manner and that repression of the promoter is mediated by altered binding of USF to the minimal promoter. We have also shown that inhibition of promoter activity results in decreased BRCA2 mRNA and protein levels over time, suggesting a direct effect on BRCA2 function. Furthermore, we found that ADR and other DNA-damaging agents increase the rate of BRCA2 mRNA and protein turnover in a p53-dependent manner.
Given that BRCA2 appears to regulate the transactivation activity of p53 and that p53 inhibits BRCA2 expression, it appears that BRCA2 and p53 share a complex regulatory loop that may be directly associated with the cellular response to DNA damage. Further studies are needed to identify the specific mechanism by which p53 inhibits USF binding to the BRCA2 promoter and to }better understand why BRCA2 is down-regulated in response to specific DNA-damaging agents.