BRCA1 Expression Restores Radiation Resistance in BRCA1-defective Cancer Cells through Enhancement of Transcription-coupled DNA Repair*

The breast cancer predisposition genes,BRCA1 and BRCA2, are responsible for the vast majority of hereditary breast cancer. Although BRCA2functions to help the cell repair double-stranded DNA breaks, the function of BRCA1 remains enigmatic. Here, we develop a human genetic system to study the role of BRCA1 in oxidative DNA damage. We show that human cancer cells containing mutated BRCA1 are hypersensitive to ionizing radiation. This hypersensitivity can be reversed by the expression of forms ofBRCA1 that are not growth suppressing. Reversal of hypersensitivity requires the ring finger of BRCA1, its transactivation domain, and its BRCT domain. Lastly, we show that unlike BRCA2, BRCA1 does not function in the repair of double-stranded DNA breaks. Instead, it functions in transcription-coupled DNA repair (TCR). TCR ability correlated with radioresistance as cells containing BRCA1 showed both increased TCR and radioresistance, whereas cells withoutBRCA1 showed decreased TCR and radiosensitivity. These findings give physiologic significance to the interaction ofBRCA1 with the basal transcription machinery.

The breast cancer predisposition genes, BRCA1 and BRCA2, are responsible for the vast majority of hereditary breast cancer. Although BRCA2 functions to help the cell repair double-stranded DNA breaks, the function of BRCA1 remains enigmatic. Here, we develop a human genetic system to study the role of BRCA1 in oxidative DNA damage. We show that human cancer cells containing mutated BRCA1 are hypersensitive to ionizing radiation. This hypersensitivity can be reversed by the expression of forms of BRCA1 that are not growth suppressing. Reversal of hypersensitivity requires the ring finger of BRCA1, its transactivation domain, and its BRCT domain. Lastly, we show that unlike BRCA2, BRCA1 does not function in the repair of doublestranded DNA breaks. Instead, it functions in transcription-coupled DNA repair (TCR). TCR ability correlated with radioresistance as cells containing BRCA1 showed both increased TCR and radioresistance, whereas cells without BRCA1 showed decreased TCR and radiosensitivity. These findings give physiologic significance to the interaction of BRCA1 with the basal transcription machinery.
BRCA1 and BRCA2, the breast cancer 1 and 2 genes, are responsible for over 90% of hereditary breast cancers (1)(2)(3)(4). Although BRCA2 has been shown to affect the repair of doublestranded DNA breaks (Refs. 5-8; reviewed in Ref. 9), a clear consensus has not been reached on the function of BRCA1. Unfortunately, BRCA1 Ϫ/Ϫ mice die at day 6.5 to day 8.5 of embryonic gestation because of lack of proliferation of the mouse blastocyst (10 -12). Despite this embryonic lethality, work in the mouse systems has resulted in two suggestive findings. First, when BRCA1 Ϫ/Ϫ mice are mated with p53 Ϫ/Ϫ mice to generate BRCA1 Ϫ/Ϫ p53 Ϫ/Ϫ mice, these double-knock out mice show reduced embryonic lethality (13,14), suggesting that BRCA1 and p53 may lie on a common functional pathway. The second finding is that cells from BRCA1 Ϫ/Ϫ mice have a defect in transcription-coupled DNA repair (TCR) 1 (15), implying that BRCA1 may be involved in DNA repair and/or the stress response of the cell.
Despite these suggestive findings in the mouse system, there are such large differences in mouse and human BRCA1 biology that it is unclear whether the DNA repair function of mouse BRCA1 is applicable to human BRCA1. Mouse BRCA1 is only 57% homologous to human BRCA1 (10 -12), and BRCA1 appears to function differently in the two systems. Although BRCA1 has been shown to be required for cellular proliferation during mouse development, BRCA1 has been shown to be a powerful growth suppressor in both yeast and human systems (16 -20). Although there are reports of living humans who are homozygous for BRCA1 mutations (21), mice carrying homozygous BRCA1 mutations die early in gestation. Lastly, DNA repair in a mouse cell is not necessarily indicative of repair in a human cell (22)(23)(24). Mouse and human cells show differences in the amount of damage sustained per given DNA-damaging dose, in the kinetics of DNA repair, and in cellular survival at a given dose of a DNA-damaging agent (22)(23)(24).
Other attempts to determine the function of BRCA1 have centered on finding functional domains of BRCA1. BRCA1 contains a Zn 2ϩ finger in its N terminus (2), a transcriptional activation domain in its C terminus (25,26), and a BRCT domain at its far C terminus (27,28). BRCA1 has been shown to interact with the basal transcriptional machinery (RNA polymerase II, TFIIH, TFIIE, and RNA helicase A) (29,30), BRCA2 (31), and Rad51 (29). BRCA1 also colocalizes with Rad51 to discrete nuclear foci when exposed to UV radiation (32). Although BRCA1 and BRCA2 show no homology (2, 3), mutations of these genes both cause breast cancer, and these two genes show identical expression throughout development (33). Cells that lack BRCA2 have been shown to be hypersensitive to ionizing radiation and have a decreased capacity to repair double-stranded DNA breaks (5)(6)(7)(8). In addition, BRCA2 interacts directly with Rad51 (5,34). When coupled with the finding that BRCA1 Ϫ/Ϫ mouse cells are deficient in TCR, all these findings suggest a role for BRCA1 in the cellular response to DNA damage.
To determine the role of BRCA1 in the cellular response to DNA damage, we have developed a human genetic system. HCC1937 is a human breast cancer cell line originally established from a family carrying a known cancer-causing BRCA1 mutation, 5283insC. We show that: 1) HCC1937 cells are hy-persensitive to ionizing radiation; 2) replacement of critical regions of BRCA1 can reverse this radiation sensitivity; and 3) TCR correlates with radioresistance as cells containing BRCA1 show both increased TCR and radioresistance.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfections-HCC 1937 cells were grown in RPMI medium (Life Technologies, Inc.) supplemented with 5 mM glutamine, 1ϫ antibiotic/antimycotic (Sigma), 1ϫ ITSA (Sigma), and 5% fetal bovine serum (Summit). HCC1937 cells were transfected with 1 g of DNA using Cellfectin (Life Technologies, Inc.). Using ␤-galactosidase, we determined an average transfection efficiency of 20%. XRV15B cells were a gift of Gilbert Chu (Stanford University). These were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 5 mM glutamine, and 1ϫ antibiotic/antimycotic. V8 cells are a cell line generated by our lab from a breast cancer patient who carries an exon 11 BRCA1 mutation (R1203X). These cells are heterozygous for this mutation (data not shown). MCF-7 cells (hemizygous wild-type BRCA1), HBL100 cells and Capan-1 cells (hemizygous mutant BRCA2) were obtained from American Type Tissue Collection and grown as instructed.
Cell Survival Assays-Two types of cell survival assays were used: the colony forming assay and the transient transfection cell survival assay. The colony forming assay was performed as described previously (36). The transient transfection cell survival (34) assay was performed as described with the following changes. Approximately 1 ϫ 10 6 HCC1937 cells were transfected at 85% confluency with 1 g of indicated plasmid and Cellfectin (Life Technologies, Inc.). 18 h later, the plates of cells were exposed to the various doses of ionizing radiation. Cells were then refed every 2-3 days. 8 days following irradiation (10 days following transfection), the cells were trypsinized, and the total cell number was determined by Coulter Counter. Relative survival refers to the cell number from a given transfection at a given dose compared with the cell number from an unirradiated plate of the same transfected cells plated on the same day and grown for the same length of time.
Generation of Stably Transfected HCC1937 Cells-Stably transfected cells were generated by transfecting 1 g of pMT-BRCA1⌬702-834 into HCC1937 cells. Cells were then selected in 50 g/ml G418 (Sigma) for 4 weeks. Approximately 20 individual colonies were selected and analyzed for BRCA1⌬702-834 expression by Western blotting. Briefly, total cell lysates were prepared. 20 g of total protein was electrophoresed on a 7% SDS-polyacrylamide gel. The electrophoresed protein was then transferred to polyvinylidene difluoride membrane and blotted FIG. 1. Cells containing only mutant BRCA1 are hypersensitive to ionizing radiation. Cells were plated and then either exposed to 1, 2, 3, 4, or 6 Gy of ionizing radiation or were mock-exposed (0 Gy). After 2 weeks of colony growth, the cells were stained, and colonies were counted manually. Survival of the mock-exposed cells from each cell line was set equal to 1, and the irradiated cell survival was standardized to the survival of the mock-exposed cells from each line. The cell lines used in this experiment included the nontransformed HBL100 cells line, the breast cancer cell line V8 (containing one functional BRCA1 allele), and the breast cancer cell line HCC1937 (containing no nonmutant BRCA1).
Where error bars are not seen, the S.E. was too small to be visualized on the graph. The experiment was performed a total of six times with similar results each time.

FIG. 2. Expression of nongrowth suppressing forms of BRCA1
reverses radiation sensitivity in BRCA1-null HCC1937 cells. A, the deletion mutants generated and used to reverse the radiation sensitivity of HCC1937 cells are depicted. BRCA1 5283insC is the endogenous mutant BRCA1 allele present in the HCC1937 cells. This mutant deletes the BRCT region of BRCA1. BRCA1⌬3-29 and BRCA1 ⌬3-471 both make mutations in the N-terminal region reported to be important for transactivation by BRCA1. BRCA1⌬702-834 is the smallest deletion mutant able to destroy the ability of BRCA1 to suppress growth. BRCA1⌬512-1284 deletes a central region of BRCA1 but leaves the BRCT domain and the transactivation domains intact. BRCA1⌬922-1664 destroys a portion of the transactivation domain. BRCA1⌬1293-1864 deletes the C-terminal region of BRCA1, including the BRCT domain. The ability of these deletion mutants to suppress growth was measured by transfecting the mutants into HCC1937 cells and selecting for 2 weeks in G418. Each experiment was performed in triplicate. B, HCC1937 cells were transfected with the various deletion mutants (20% transfection rate as measured by ␤-galactosidase staining), acutely irradiated with the indicated dose of ionizing radiation 48 h later, and then allowed to grow for 8 days. The unirradiated cell count from each transfection mutant was set to equal 1, and relative survival of a given set of cells transfected with one BRCA1 mutant was standardized to that of the unirradiated cells transfected with the same mutant. Where error bars are not seen, S.E. was too small to be seen on the graph. Each experiment was performed four times with similar results each time.
Double-stranded DNA Break Repair Assays-This assay was performed as described previously (7). Briefly 1 ϫ 10 5 cells (XRV15B, HCC1937, or HCC1937 BRCA1D702-834; clone 1) were embedded into agarose plugs under isotonic conditions. The cell plugs were then irradiated at 4°C with 10 Gy ionizing radiation and placed into complete medium at 37°C to perform double-stranded break repair for the given amount of time. After the given amount of repair time, the cell plugs were digested overnight at 55°C in 10 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.5% SDS, 50 mM EDTA, and 2 g/ml proteinase K (Life Technologies, Inc.). The digested cell plugs were then embedded into a 0.7% agarose gel and subjected to pulsed field gel electrophoresis (3 V/cm, 45-s pulse time) for 72 h. The separated DNA was then transferred to a Gene-Screen filter, and Southern blotting was performed. The probe was generated by isolation of total genomic DNA from the given cell line followed by Random Prime labeling using [␣-32 P]ATP (Prime-It II, Stratagene).
Transcription-coupled Repair Assays-1 ϫ 10 7 HCC1937 cells or 1 ϫ 10 7 BRCA1⌬702-834 (clone 1) cells for each time point were incubated in complete medium supplemented with 10 M bromodeoxyuridine (Sigma) and 25 Ci of [ 3 H]thymidine immediately prior to irradiation with 10 Gy. The zero time point was iced immediately. Remaining samples were incubated for the indicated times at 37°C. Cells were then washed three times with phosphate-buffered saline, and DNA purification was performed by standard methods (proteinase K, phenol/chloroform/ethanol precipitation). The DNA was then digested with KpnI, repurified by phenol/chloroform and ethanol precipitation, and incubated in phosphate-buffered saline containing 0.1% bovine serum albumin with an antibody directed against bromodeoxyuridine (Sigma) overnight at 4°C. Following this incubation, 10 l of goat anti-mouse antibody (Life Technologies, Inc.) was added incubated by 4 h and then collected by microcentrifugation for 15 min. Unbound DNA diluted 1:5 with TE and used directly for Southern blotting experiments. Aliquots of both the bound and unbound DNA fractions were scintillation counted to deter-mine the total amount of repaired DNA. Both bound and unbound DNA samples from each time point were then electrophoresed on a 1% agarose gel, blotted to nylon membrane, and probed with strand-specific probes for the human DHFR gene.
Strand-specific DHFR Probes-To generate strand-specific probes for transcribed regions of the human DHFR genes, Bluescript II (Stratagene) was digested with HindIII and BamHI. The following oligonucleotide linker from the DHFR gene was then inserted into Bluescript II (F, 5Ј-AGCTTCAGAGAACTCAAGTAAGTACCTTAACATAAATTCAC-CACAAG-3Ј, and R, 5Ј-GATCCTTGTGGTGAATTTATGTTAAGGTAC-TTACTTGAGTTCTCTGA-3Ј (43). After construction, this plasmid containing DHFR plus the linker was digested with either HindIII or BamHI and transcribed with T3 polymerase or T7 polymerase in the presence of [␣-32 P]GTP to generate strand-specific probes.

RESULTS
To study the role of BRCA1 in the cellular response to DNA damage, we have utilized a human breast cancer cell line (HCC1937) that contains only a mutated BRCA1 (35). The functional allele in these cells contains a frameshift mutation that deletes the BRCT domain of BRCA1. To determine whether cells lacking functional BRCA1 are sensitive to ionizing radiation, we performed colony forming assays (36). In comparison with MCF-7 cells, V8 cells and HBL100 cells, HCC1937 cells were significantly more radiation-sensitive. HCC1937 cells were approximately 2 logs more sensitive to ionizing radiation than HBL100 cells at 6 Gy and were approximately 30 to 40 times more sensitive than either V8 or MCF-7 cells (Fig. 1). Radiation sensitivity seemed to correlate with the genetic status of BRCA1. Cells with two copies of wt BRCA1 (HBL100) were more resistant than cells with one copy of wt BRCA1 (MCF-7), which were more resistant than cells without wt BRCA1 (HCC1937). Survival curves were generated on the XRV15B Ku86-deficient cell line (37) and the BRCA2-deficient Lower blot: I-20. B, representative colony forming assay after the exposure of the two stable clones to ionizing radiation as described in the legend to Fig. 1. C, quantitation of the colony forming assay. Where error bars are not seen, the S.E. was too small to be seen on the graph. Large error seen at low doses is reflective of the log scale used in the curve. Each experiment was performed three times with similar results each time. D, double-stand DNA repair assays were performed as described previously (7). Equal cell numbers were exposed to 10 Gy ionizing radiation and allowed to repair DNA for the indicated times. Under these pulsed field gel electrophoretic conditions, damaged DNA runs as a discrete band below that of undamaged or repaired DNA. Capan-1 cell line, and these survival curves were nearly identical to that of the HCC1937 cell line (Fig. 1). Because only a handful of cell lines have been shown to be as radiation-sensitive as the XRV15B cell line (38), this finding indicates that the sensitivity of the HCC1937 cell line is likely to be a real effect.
We used a recently described method to measure cell survival after transient transfection and irradiation (34) to determine whether reintroduction of BRCA1 into the HCC1937 cells could reverse the radiation sensitivity of these cells. Surprisingly, we were unable to demonstrate that reintroduction of full-length BRCA1 into HCC1937 reverses radiosensitivity because this assay depends on the continued growth of the cells after transfection. Thus, because BRCA1 is such a potent suppressor of cell growth (16 -19, 39), transfection of BRCA1 stops cell growth and does not allow us to accurately measure cell survival following irradiation. To overcome this problem, we constructed a variety of BRCA1 deletion mutants and tested them for the ability to suppress growth in the HCC 1937 cells. Each of these mutants contains an intact nuclear localization signal (40,41); however, these mutants lack important functional regions of BRCA1 (Fig. 2A). Because both BRCA1 and BRCA1⌬3-29 suppress cell growth, they cannot be scored in an assay requiring cell growth. All other mutants tested, however, destroyed the growth suppressive effect of BRCA1 and could therefore be used in the cell growth and survival assay (Fig. 2). The HCC1937 cells contain an endogenous BRCA1 mutation that eliminates the BRCT domain but leaves the transactivation domain intact. Because the HCC1937 cells are radiationsensitive, this indicates that the BRCT domain is necessary for radiation resistance. In this assay, a deletion mutant that destroyed the N-terminal Zn 2ϩ finger domain plus additional N-terminal sequences (BRCA1⌬3-471) could not rescue radiation sensitivity (Fig. 2B). Mutants that deleted both the BRCT domain and the transactivation domain (BRCA1⌬1293-1864) or the transactivation domain alone (BRCA1⌬922-1664) could not rescue radiation sensitivity (Fig. 2B). These findings demonstrate that the Zn 2ϩ finger, the transactivation domain and the BRCT domain are all necessary for radiation resistance. In contrast to these findings, two deletion mutants (BRCA1⌬702-834 and BRCA1⌬512-1284) were able to restore radiation resistance (Fig. 2B). These two mutants can eliminate the role of BRCA1 as a growth suppressor but can reverse radiation sensitivity, showing that these functions can be separated.
Because BRCA1 interacts with both Rad51 and BRCA2 (29,31), two double-stranded DNA repair proteins, and because mouse cells lacking BRCA1 are deficient in TCR (15), we were interested in determining both the double-stranded break repair capacity and the TCR capacity of the HCC1937 cells. To study the repair capacity of HCC1937 cells lacking BRCA1 and those containing a form of BRCA1 that reverses radiation sensitivity (BRCA1⌬702-834), stably transfected cell lines were generated. HCC1937 cells were transfected with BRCA1⌬702-834 on a vector also containing the neomycin resistance gene. After 4 weeks of selection in G418, approximately 20 colonies were selected and propagated. Of these 20 colonies, two colonies expressed high levels of BRCA1⌬702-834 as shown by Western blotting (clones 1 and 2; Fig. 3, A and B). The two antibodies used for Western blotting were Ab-D (42) and I-20. These antibodies recognize the far C terminus of BRCA1. HCC1937 cells contain a mutant form of BRCA1 that lacks this epitope. BRCA1⌬702-834 restores this epitope; therefore, the two clones generated show expression of BRCA1⌬702-834 by Western blotting (Fig. 3A, upper blot shows Ab-D, and bottom blot shows I-20). Because zinc is required in HCC cell medium, addition of Zn 2ϩ gave little induction (Fig. 3A). These two clones were then tested for radiation sensitivity in the colony forming assay. HCC1937 cells that contain only empty vector are hypersensitive to ionizing radiation, whereas those expressing BRCA1⌬702-834 are approximately 2 logs more radiation-resistant at the 6 Gy dose (Fig. 3, B and C). Thus by both transient and stable transfection BRCA1⌬702-834 reverses the radiation sensitivity of the HCC1937 cells.
The ability to generate HCC1937 cells containing stably transfected BRCA1⌬702-834 allowed us to compare the DNA repair capabilities of parental HCC1937 cells and HCC1937 cells containing BRCA1⌬702-834. Pulsed field gel electrophoresis was then performed with Southern blotting using total genomic DNA as a probe. Under the electrophoresis conditions chosen, damaged DNA runs as a discrete band below the undamaged or repaired DNA (7). In this experiment, HCC1937 cells repaired most damaged DNA within 2 h (Fig.  3D, upper panel). HCC1937 BRCA1⌬702-834 cells also repaired most of the damaged DNA within 2 h (Fig. 3D, bottom  panel). These cells showed slightly faster repair than parental HCC1937 cells, but the difference is too small to account for the 2 log difference in radiation sensitivity between these two cell lines.
We next analyze TCR in HCC cells, with and without BRCA1 FIG. 4. HCC1937 cells are defective in transcription-coupled DNA repair, and this defect can be reversed by expression of BRCA1⌬702-834. A, a representative TCR assay. The experiment was performed as described in Ref. 15 and under "Experimental Procedures"). Cells were exposed to 10 Gy ionizing radiation and allowed to repair damaged DNA for the indicated times. Note that virtually no signal is present in the bound sample at time zero because insufficient time has passed for any repair to have occurred. Upper panels, HCC cells show little difference in the transcribed DNA either 2 or 4 h after repair (compare bound and free DNA). The same result is seen in the nontranscribed strand. Lower panels, HCC1937 cells stably transfected with BRCA1⌬702-834 (clone 1) show that transcribed DNA is preferentially repaired (compare bound and free DNA). As expected, the nontranscribed strand shows little repair. The zero time point for the free untranscribed sample for clone 1 was diluted 5-fold because of a standardization error. B, quantitation of TCR assays. The expression of BRCA1⌬702-834 greatly increases TCR repair in the HCC1937 cells. Each experiment was performed three times with similar results each time. Where error bars are not seen, error was too small to be seen on the graph. Similar results were seen using HCC1937 BRCA1⌬702-834 (clone 2). gene replacement. Mouse cells containing defective BRCA1 have been shown to be deficient in TCR (15). HCC1937 cells showed a marked inability to undergo TCR (Fig. 4A, upper  panels). Although the nontranscribed strand shows the expected result (little repair of the nontranscribed strand), the transcribed strand also shows little repair (compare bound and free fractions at both 2 and 4 h of repair). In contrast, cells containing BRCA1⌬702-834 were able to undergo TCR. Within 2 h after exposure, a much greater proportion of DHFR is present in the bound fraction (Fig. 4A, lower panels). Cells containing BRCA1⌬702-834 undergo TCR approximately 15 times faster than cells lacking functional BRCA1 (Fig. 4B). DISCUSSION We have shown that cells containing a mutant BRCA1 that is known to cause cancer in human patients are hypersensitive to ionizing radiation. This radiation sensitivity can be reversed by forms of BRCA1 that do not suppress growth but maintain the Zn 2ϩ finger, the transcription activating domain, and the BRCT domain intact. Although BRCA1 does not affect doublestranded break repair, cells lacking BRCA1 are deficient in their ability to perform TCR. These findings suggest that the interaction of BRCA1 with the transcription machinery (RNA pol II, TFIIE, TFIIF, TFIIH, and RNA helicase A) are physiologically important. This work also suggests functional differences between BRCA1 and BRCA2. Whereas BRCA1 predominantly affects TCR, cells lacking BRCA2 have been shown to be deficient in double-stranded break repair (5)(6)(7)(8). Further work can help to determine whether these differing DNA repair defects lead to the higher cancer penetrance seen in BRCA1 carriers, the higher incidence of male breast cancer seen in BRCA2 carriers, and the different histology seen in BRCA1 or BRCA2-defective tumors.