BRCA1 Ubiquitinates RPB8 in Response to DNA Damage Free

Germ line mutation of the cancer susceptibility gene BRCA1 causes familial breast and ovarian cancer. BRCA1 acts as a hub protein that coordinates many cellular pathways to prevent tumor progression. Accordingly, down-regulation of this key protein by mechanisms other than BRCA1 gene mutation causes sporadic breast cancer (1). All cells defective in BRCA1 show genomic instability as evidenced by hypersensitivity to DNA damage, the presence of chromosomal abnormalities, and the loss of heterozygosity at multiple loci. These results are likely to stem from the failure of BRCA1 to function in DNA damage repair, transcriptional regulation, apoptosis induction, intra-S or G₂-M checkpoint function, and regulation of centrosome duplication (2–4).

Involvement of BRCA1 in multiple cellular processes is logical given its enzymatic function as a ubiquitin ligase (E3). In this capacity, it has the potential to interact with numerous protein substrates and subsequently influence the biological response of a cell at many points. BRCA1 contains an NH₂-terminal RING finger domain, a common motif found in ubiquitin ligases. It acquires significant ubiquitin ligase activity when bound to another conformationally similar RING finger protein, BARD1, as a RING heterodimer (5–8). The most common polyubiquitin chain is linked through Lys⁴⁸ of ubiquitin and serves as a signal for rapid degradation of substrates by the proteasome-dependent proteolysis pathway (9). However, BRCA1-BARD1 has the unique capacity to catalyze Lys⁶-linked polyubiquitin chains, and the ubiquitination mediated by BRCA1-BARD1 could signal a process other than degradation (10–13). Deleterious missense mutations in the RING finger domain of BRCA1 found in familial breast cancer abolish the E3 ligase activity of BRCA1-BARD1 (6, 7, 14), indicating that the E3 ligase activity is important for role of BRCA1 as a tumor suppressor.

One of the most significant functional features of BRCA1 is that it is a component of the RNA polymerase II holoenzyme (15, 16). BRCA1 specifically interacts with a large fraction of hyperphosphorylated, processive polymerase II (IIO), in preference to the hypophosphorylated polymerase II (IIA) found at promoters (17). It has been proposed that BRCA1 binds polymerase IIO complexes as part of a genome scanning function for DNA damage (18). However, how BRCA1 affects the polymerase II complexes after DNA damage remains partially understood. In this study, we identify RPB8 (also called hRPB17 or POLR2H), a common subunit of all three types of RNA polymerases, as a substrate of BRCA1 E3 ligase and show that BRCA1 ubiquitinates RPB8 immediately after DNA damage. HeLa cell lines stably expressing a ubiquitin-resistant form of RPB8 exhibited UV hypersensitivity, a known phenotype of BRCA1 deficiency. These results indicate a significant role of ubiquitin ligase activity of BRCA1 for cell survival after DNA damage and provide a new aspect of a common subunit of RNA polymerases in DNA damage responses.

Two-dimensional difference gel electrophoresis. Methods for fluorescence two-dimensional difference gel electrophoresis (DIGE) and mass spectrometric analysis are reported in the Supplementary Data.

Plasmids. cDNA for full-length human RPB8 was amplified by PCR from a MCF10A cell cDNA library using Pfx polymerase (Stratagene, La Jolla, CA). Mammalian expression plasmids for BRCA1, BARD1, ubiquitin, and their mutants were previously described (7, 11). The point mutations to substitute the Lys residue(s) of RPB8 with Arg were produced by site-directed mutagenesis (Stratagene). All plasmids used were verified by DNA sequencing.

Cell cultures and transfections. T47D, HCC1937 breast carcinoma cells, HeLa cervical carcinoma cells, and 293T transformed human kidney cells were cultured in DMEM supplemented with 10% FCS and 1% antibiotic-antimycotic agent (Life Technologies, Inc., Grand Island, NY) in 5% CO₂ at 37°C. MCF10A normal human breast epithelial cells were grown in DMEM/Ham's F12 (1:1) medium supplemented with 2.5% FCS, 100 ng/mL cholera toxin, 20 ng/mL epidermal growth factor, 500 ng/mL hydrocortisone, 10 μg/mL insulin, and 1% antibiotic-antimycotic agent. For epirubicin treatment, cells were incubated in medium containing 0.2 μg/mL epirubicin (Pfizer, New York, NY). To examine the half-life of proteins in vivo, cells were incubated with 10 μg/mL cycloheximide (Wako, Osaka, Japan), a protein synthesis inhibitor, for the indicated time periods. 293T cells were transfected using the standard calcium phosphate precipitation method. To generate cell lines that stably expressed either wild-type (WT) or mutant FLAG-RPB8, HeLa cells were transfected using FuGENE6 (Roche, Indianapolis, IN) with pcDNA3 plasmids encoding each protein and selected with G418. For UV irradiation studies, cells were washed with PBS, irradiated with UV light (254 nm; UVP, Inc., Upland, CA) at the indicated doses, and grown in fresh medium for various times.

Antibodies. Mouse monoclonal antibodies to hemagglutinin (HA; Boehringer-Mannheim, Mannheim, Germany), Myc (BabCo, Richmond, CA), FLAG (Sigma, St. Louis, MO), polyubiquitin (Affiniti, Exeter, United Kingdom), conjugated ubiquitin (Affiniti; ref. 10), α- and β-tubulin (Neomarkers, Fremont, CA), and actin (Santa Cruz Biotechnology, Santa Cruz, CA) as well as rabbit polyclonal antibodies to BRCA1 (Santa Cruz Biotechnology), RPB1 (Covance), and cleaved caspase-3 (Cell Signaling Technology, Danvers, MA) were purchased commercially. Anti-FLAG cross-linked agarose beads (Sigma) were used for immunoprecipitation to detect in vivo ubiquitinated substrates. Rabbit polyclonal antibodies to BARD1 and RPC155 were generous gifts from Dr. Richard Baer (Columbia University, New York, NY) and Dr. Nouria Hernandez (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY), respectively. Rabbit polyclonal antibody to RPB8 was generated against full-length human glutathione S-transferase (GST)-RPB8 and purified by protein G agarose chromatography.

RNA interference. SMART pool BRCA1 small interfering RNA (siRNA) mix and control siRNA mix were purchased from Dharmacon Research, Inc. (Lafayette, CO). RNA duplexes (final concentration 50 nmol/L) were transfected into the cells with Oligofectamine (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Retrovirus expressing short hairpin RNA (shRNA) that targets BRCA1 mRNA sequence 5′-CUAGAAAUCUGUUGCUAUG-3′ was created by cotransfecting 293T cells with pGP vector, pE-ampho vector, and pSINsi-hU6 retroviral vector that has previously been subcloned with the oligonucleotide 5′-GATCCGCTAGAAATCTGTTGCTATGTTCAAGAGACATAGCAACAGATTTCTAGCTTTTTTAT-3′ according to the manufacturer's protocol (TaKaRa, Otsu, Japan). Oligonucleotide 5′-GATCCGTAAGGCTATGAAGAGATACTTCAAGAGAGTATCTCTTCATAGCCTTACTTTTTTAT-3′ was used for the retrovirus expressing control shRNA. For infection, HeLa cells were incubated with virus supernatants and fresh culture medium containing 8 μg/mL Polybrene (Sigma). Cells were analyzed 48 h after transfection or infection.

Immunoprecipitation and immunoblotting. Immunoprecipitation and immunoblotting methods were previously described (11). For the immunoblotting analysis after two-dimensional gel electrophoresis, cells were lysed with 7 mol/L urea/2 mol/L thiourea–containing buffer as described above. Soluble fractions were prepared with 0.5% NP40-based buffer as previously described (11). Denatured whole-cell lysates were prepared by boiling in Laemmli SDS-loading buffer with 0.1 mol/L DTT.

In vitro ubiquitin ligation assay. Full-length His-FLAG-RPB8 was obtained from BL21/DE3 bacteria cells with isopropyl-l-thio-β-d-galactopyranoside induction by two-step purification using nickel agarose beads followed by anti-FLAG cross-linked agarose beads (Supplementary Fig. S3). Complexes of WT or I26A mutant of FLAG-BRCA1^1-772 with BARD1 were purified from transfected 293T cells by anti-FLAG affinity chromatography and FLAG peptide elution. Both WT and I26A mutant complexes contained an ∼1:1 ratio of BRCA1 and BARD1 proteins (Supplementary Fig. S3). Rabbit E1 (BIOMOL, Plymouth Meeting, PA) and mammalian ubiquitin (Boston Biochem, Cambridge, MA) were purchased commercially. The in vitro reaction was done as previously described (11) with a reaction mixture (30 μL) that contained 0.5 μg His-FLAG-RPB8, 40 ng E1, 0.3 μg UbcH5c, and 0.3 μg each of FLAG-BRCA1^1-772 and BARD1.

Runoff transcription assay. The runoff transcription assay used was described elsewhere (17). Briefly, the runoff template was created by annealing 50 pmol each of a 65-mer oligonucleotide 5′-ATTGGGTAAAGGAGAGTATTTGAGCGGAGGACAGTACTCCGGGTCCCCCCCCCCCCCCCCCCCCC-3′and a complementary 45-mer oligonucleotide 5′-GACCCGGAGTACTGTCCTCCGCTCTTTTACTCTCCTTTACCCAAT-3′ in a 200 μL annealing mixture containing 20 mmol/L Tris (pH 7.4), 1 mmol/L EDTA, and 0.2 mol/L NaCl. Runoff transcription reactions (20 μL) contained 8.25 mmol/L MgCl₂, 5 μg of bovine serum albumin, 250 nmol/L nucleotide triphosphates, 5 units of RNase inhibitor, 50 ng of poly(deoxyinosinic-deoxycytidylic acid), 0.05% NP40, 1 pmol of annealed oligonucleotides, and 0.5 μCi of [α-³²P]CTP. Equilibrated FLAG-RPB8 immunocomplexes bound to M2 beads (10 μL) were added to the reactions (20 μL) and incubated for 40 min at 30°C and stopped with 50 μL of PK buffer (300 mmol/L sodium acetate, 0.2% SDS, 10 mmol/L EDTA, 100 ng tRNA, and 10 μg proteinase K). Reactions were then incubated at 55°C for 20 min, extracted with phenol/chloroform, and precipitated with ethanol. Single-stranded RNA transcripts were resolved under denaturing conditions on 12% polyacrylamide/urea gels and scanned with the Typhoon 9400 image analyzer (Amersham, Piscataway, NJ).

Identification of RPB8 as a protein modified in BRCA1-positive cells after epirubicin treatment. To search for candidate substrates for the BRCA1-BARD1 E3 ligase in response to DNA damage, we used two-dimensional DIGE technology. Breast cancer–derived, BRCA1–positive T47D cells and BRCA1-defective HCC1937 cells were incubated for 3 h with epirubicin, a topoisomerase II inhibitor that induces DNA double strand breaks. Cells were lysed with 7 mol/L urea/2 mol/L thiourea–containing buffer, and the proteomes were compared with untreated cells using two-dimensional DIGE. Interestingly, whereas the expression levels of only a few proteins were affected by the epirubicin treatment in T47D cells, that of ∼100 proteins were altered in HCC1937 cells (Fig. 1A). Conversely and even more interesting, two proteins whose expression levels were dramatically reduced in T47D cells were not changed in HCC1937 cells (Fig. 1A , red arrows). Therefore, we speculated that the reduction could depend on the presence of BRCA1. The protein spots were in-gel digested and subjected to nanoscale capillary liquid chromatography-tandem mass spectrometry (LC/MS/MS) analysis. LC/MS/MS analysis revealed that the samples were RPB8, a common subunit of three types of RNA polymerases, and myosin light chain. RPB8 is a very acidic, small protein with a calculated molecular mass of 17.1 kDa and an isoelectric point of 4.34 (19). One of the most significant functional features of BRCA1 is that it is a component of the RNA polymerase II holoenzyme (15, 16). Therefore, we focused on RPB8 for further analyses.

To confirm our mass spectrometry data, we generated a rabbit polyclonal antibody to GST-RPB8 for immunoblot analysis. Cells were treated as in Fig. 1A, and immunoblot analysis of the proteins resolved by two-dimensional gels verified that the protein spot was indeed RPB8. It was again severely reduced by epirubicin treatment only in T47D cells (Fig. 1B). The difference in RPB8 expression in response to epirubicin treatment could be due to the different genetic backgrounds of these two cell lines, not just the absence or presence of BRCA1. Therefore, we next compared RPB8 expression between isogenic cells with and without knockdown of BRCA1 expression. T47D cells were transfected with either control siRNA or BRCA1 siRNA and then treated as in Fig. 1A. The siRNA-transfected cells were successfully silenced for BRCA1 expression (Supplementary Fig. S1). Immunoblot analysis of the proteins resolved by two-dimensional gels showed that RPB8 was reduced by epirubicin treatment only in control cells, not in cells with BRCA1 knockdown, supporting the idea that this modification depends on BRCA1 expression (Fig. 1C). The reduction of RPB8 at its normal migrating position could be due to protein degradation or to covalent modification.

BRCA1-BARD1 interacts with and ubiquitinates RPB8. The polymerase II holoenzyme interacts with BRCA1 and BARD1 (15, 16). Consistent with the previous reports, a significant amount of endogenous BARD1 coimmunoprecipitated with RPB8 isolated from HeLa cells or MCF10A cells compared with controls (Fig. 2A). The same results were observed with MCF7, T47D, and 293T cells (data not shown). Exogenously expressed RPB8 also interacted with BRCA1 and BARD1 (Supplementary Fig. S2). Then, we tested whether RPB8 is ubiquitinated by BRCA1-BARD1 in vivo. FLAG-RPB8 was coexpressed in 293T cells with HA-ubiquitin, Myc-BRCA1^1-772, and BARD1. Cells were collected 36 h after transfection and boiled in 1% SDS–containing buffer, and FLAG-RPB8 was immunoprecipitated. Immunoblotting of the RPB8 precipitates resolved by SDS-PAGE using anti-HA antibody showed a ladder characteristic of polyubiquitinated RPB8 (Fig. 2B). Omission of FLAG-RPB8, HA-ubiquitin, Myc-BRCA1^1-772, or BARD1 all abolished the RPB8 ladders, supporting the idea of BRCA1-BARD1–dependent RPB8 ubiquitination.

BRCA1-BARD1 is the only known E3 ligase to catalyze Lys⁶-linked polyubiquitin chains (10, 11, 13). To show that the in vivo RPB8 ubiquitin ladders were directly due to BRCA1-BARD1 ligase activity, we verified that RPB8 was modified by ubiquitin through Lys⁶ linkages. HA-tagged ubiquitins that have a single lysine residue available for conjugation were used for in vivo ubiquitination assays. As expected, BRCA1-BARD1–dependent RPB8 polyubiquitination was predominantly detected when HA-ubiquitin with only Lys⁶ available, but not Lys⁴⁸ or Lys⁶³, was coexpressed (Fig. 2C). However, it has been suggested that ubiquitin mutants could fold incorrectly and may cause artifacts (20). Recent quantitative analysis of in vitro ubiquitination revealed that even for cyclin B1 ubiquitination catalyzed by the anaphase-promoting complex, heterogeneous ubiquitin chains, including Lys⁶³, Lys¹¹, and Lys⁴⁸, or monoubiquitin attached to multiple lysine residues on the substrate. Further, some types of linkages are dependent on the combination of E2 and E3 enzymes (21). Thus, it is possible that ubiquitination mediated by BRCA1-BARD1 also resulted in multiple polyubiquitin chains, including Lys⁶. The preference for Lys⁶ ubiquitination observed in the in vivo experiment was not enough evidence to support the direct role of BRCA1-BARD1 for RPB8 ubiquitination. Therefore, we further tested whether BRCA1-BARD1 directly catalyzes RPB8 polyubiquitination by in vitro ubiquitination using recombinant RPB8 protein (Supplementary Fig. S3). His-FLAG-RPB8 incubated with ubiquitin, E1, E2/His-UbcH5c, and FLAG-BRCA1^1-772/BARD1 complex (Supplementary Fig. S3) resulted in a ladder and smear detected by anti-RPB8 immunoblot (Fig. 2D). Omission of substrate RPB8, ubiquitin/E1/E2, or FLAG-BRCA1^1-772/BARD1 complex, as well as substitution of BRCA1^1-772 with the E2-nonbinding mutant I26A, all abolished RPB8 ubiquitination. Hence, the results suggest that the RPB8 polyubiquitination is directly catalyzed by BRCA1-BARD1.

BRCA1-BARD1 does not destabilize RPB8 in vivo. Our previous results suggested that BRCA1-BARD1 catalyzed untraditional polyubiquitin chains that served as a signal for a process other than degradation (7, 11, 12). However, the reduced expression of RPB8 after epirubicin treatment detected by two-dimensional DIGE or two-dimensional immunoblot (Fig. 1) suggested the possibility of BRCA1-mediated RPB8 degradation. Therefore, we tested if BRCA1-BARD1 destabilized RPB8 in vivo under several different conditions, including BRCA1-BARD1 overexpression and BRCA1 knockdown by siRNA. FLAG-RPB8 was coexpressed in 293T cells with Myc-BRCA1^1-772 and HA-BARD1 (Fig. 3A). The steady-state level of FLAG-RPB8 increased upon coexpression of BRCA1-BARD1 in a dose-dependent manner in the soluble fraction (lanes 1–4) but not in whole-cell lysates (lanes 5–8). We then examined protein half-life of FLAG-RPB8 in the soluble fraction using cycloheximide, a protein synthesis inhibitor. The FLAG-RPB8 protein half-life was prolonged by BRCA1-BARD1 overexpression (Fig. 3B). We also tested the effect of BRCA1-BARD1 on endogenous RPB8 (Fig. 3C and D). The steady-state level of RPB8 only slightly increased upon coexpression of BRCA1-BARD1 in the soluble fraction (Fig. 3C,, lane 4) and no effect was observed when whole-cell lysates were evaluated (lanes 5–8). However, RPB8 protein half-life was detectably shortened by BRCA1 knockdown (Fig. 3D). This observation was not detected when whole-cell lysates were analyzed (data not shown). Together, analyses of steady-state levels and protein half-lives indicated that only soluble RPB8 was stabilized, whereas that in whole-cell lysate was unchanged (Fig. 3). Alternatively, it was also possible that BRCA1-BARD1 shifted RPB8 from the insoluble fraction, such as the chromatin fraction, to the soluble fraction. However, we could not detect such a shift by fractionation analyses (data not shown). In either case, these findings at least suggest that BRCA1-BARD1–mediated RPB8 ubiquitination is not a signal for its degradation.

BRCA1-dependent RPB8 ubiquitination after UV irradiation. BRCA1-mediated RPB8 ubiquitination prompted us to investigate the biological implications of this activity. We examined if RPB8 is ubiquitinated in response to DNA damage. Rather than exposing cells continuously to epirubicin, and because RPB1 is ubiquitinated after UV irradiation, we used UV irradiation to accurately determine the timing of RPB8 ubiquitination after DNA damage (22–25). We established HeLa cell lines that stably express FLAG-RPB8 at a low level (approximately one third of endogenous RPB8; Fig. 4A) to avoid artifacts caused by overexpression and analyzed ubiquitination of anti-FLAG immunoprecipitates with anti-ubiquitin (FK2) antibody. Because it has been reported that RPB1 ubiquitination occurs 1 to 2 h after UV irradiation (22–25), we first analyzed these time points. However, we did not detect any ubiquitination of FLAG-RPB8 (Fig. 4B and data not shown). Instead, ubiquitinated FLAG-RPB8 readily, and only transiently, appeared 10 min after UV irradiation (Fig. 4B , top). Reprobing the membrane with anti-RPB8 antibody verified that the detected ladder was ubiquitinated RPB8 (bottom).

To verify that UV irradiation–induced RPB8 ubiquitination requires endogenous BRCA1, RNA interference was used to knock down BRCA1 expression. HeLa cells stably expressing FLAG-RPB8 were transfected with BRCA1-specific siRNA. As a second alternative, we constructed a retrovirus engineered to express shRNA for BRCA1. Forty-eight hours after transfection or infection, cells were irradiated with UV (35 J/m²) and then harvested 10 min later. Both the siRNA-transfected and the shRNA retrovirus–infected cells were successfully silenced for BRCA1 expression (>90% and >75% reduction, respectively) compared with their controls (Fig. 4C , top). As expected, RPB8 ubiquitination after UV irradiation was dramatically reduced by BRCA1 knockdown in both cases (lower middle). Reprobing the membrane with anti-RPB8 antibody again verified the ubiquitinated RPB8 that became completely undetectable upon BRCA1 knockdown (bottom). These results support the idea that RPB8 is polyubiquitinated by BRCA1-BARD1 in an early phase after UV irradiation.

A ubiquitin-resistant form of RPB8 retains its polymerase activity. For the purpose of studying the physiologic consequences induced by the BRCA1-mediated RPB8 ubiquitination after UV irradiation, we generated a mutant of RPB8 that is incapable of being ubiquitinated by BRCA1-BARD1. RPB8 possesses eight Lys residues in the whole protein (Fig. 5A). We first mutated single Lys residues of RPB8 and tested its capacity to be ubiquitinated. However, RPB8 ubiquitination was not dramatically reduced by each single mutation (Fig. 5B,, lanes 2 and 7; data not shown). Instead, the ubiquitination of RPB8 was reduced as the number of Lys to Arg substitutions increased. This result recapitulates what we observed during studies of BRCA1 auto-ubiquitination and of BRCA1-mediated NPM1/B23 ubiquitination. When five of the eight Lys residues were substituted with Arg (5KR), RPB8 ubiquitination became undetectable (Fig. 5B , lane 5), although its binding capacity to BRCA1-BARD1 was not reduced (data not shown).

To confirm that the many mutations required to make RPB8 resistant to ubiquitination did not impair its fundamental function as a subunit of RNA polymerases, we verified that the 5KR mutant is capable of binding to RPB1 or RPC155 (the largest subunit of polymerase III) in vivo. WT FLAG-RPB8 or 5KR was transfected into 293T cells, and anti-FLAG immunocomplexes were isolated. Bound proteins were resolved by SDS-PAGE and analyzed by immunoblotting using anti-RPB1 or anti-RPC155 antibodies. Both RPB1 and RPC155 were detected in the FLAG-5KR immunocomplexes as well as the WT immunocomplexes (Fig. 5C). We measured catalytic activity of the anti-FLAG immunoprecipitates using a runoff transcription assay. The 5KR mutant immunocomplexes contained the ability to generate in vitro transcripts equal to that of WT immunocomplexes (Fig. 5D). Thus, the 5KR mutation of RPB8 constitutes a viable RNA polymerase complex in vivo that sustains its polymerase activity. This indicates that RPB8 ubiquitination by BRCA1-BARD1 is not required for RNA polymerase activity.

Ubiquitin-resistant mutant of RPB8 causes UV hypersensitivity. BRCA1 deficiency causes hypersensitivity to DNA damage (14, 26, 27). Because RPB8 is ubiquitinated by BRCA1 after UV irradiation (Fig. 4), it was possible that failure to perform this function could cause the same phenotype. To test this possibility, we established HeLa cell lines that stably express the 5KR mutant of FLAG-RPB8. Two clones each of the WT (WT-1 and WT-2) and of the 5KR (5KR-1 and 5KR-2) cell lines were obtained (Fig. 6A). Polyubiquitination of FLAG-RPB8 after UV irradiation was detected in WT cells, but not in mutant cells (Fig. 6B). Using these cells, we examined if the expression of the mutant RPB8 affected cell survival after UV irradiation. The cell viabilities of the 5KR clones 48 h after 20 or 35 J/m² of UV irradiation were ∼38% and 23% of untreated cells at 0 h, respectively, whereas WT clones were ∼72% and 53%, respectively (Fig. 6C). Parental HeLa cells exhibited viabilities similar to that of WT clones (Fig. 6C). Representative data for cells observed by phase contrast microscopy 48 h after UV irradiation (35 J/m²) and for culture plates stained with Lillie's crystal violet stain are shown (Supplementary Fig. S4). Thus, expression of a ubiquitin-resistant RPB8 form in cells causes UV hypersensitivity.

Because UV-induced cell death is largely ascribable to caspase-induced apoptosis, we next tested whether activation of the caspase pathway by UV irradiation was enhanced in 5KR cells. HeLa cell lines expressing WT or 5KR mutant of FLAG-RPB8 were UV irradiated, and caspase activity was measured by immunoblotting with an antibody to cleaved caspase-3. As shown in Fig. 6D, 5KR cells expressed larger amount of cleaved caspase-3 than WT cells did at each time point after UV irradiation. This result suggests that failure to ubiquitinate RPB8 after UV irradiation activates the caspase pathway, resulting in apoptotic cell death.

BRCA1 exists in several different supercomplexes to execute diverse cellular processes. In most of these complexes, BRCA1 exists as a RING heterodimer with BARD1 (28), the form that acquires significant ubiquitin ligase activity (6–8). Revealing the substrates specific for each BRCA1 protein complex is crucial to understand the mechanisms underlying its tumor-suppressor functions.

BRCA1-BARD1 complexes bind to BRCA2 and Rad51 and localize to discrete nuclear foci during S phase. After DNA damage, BRCA1 is phosphorylated by ATM/ATR family kinases (29, 30), and the BRCA1 foci disperse within 30 min (31). The BRCA2-Rad51–containing complex, as well as the BRCA1 complex with Mre11-Rad50-Nbs1, gradually reassemble into different foci (sites of DNA damage) and play important roles in homologous recombination repair. The BRCA1-containing foci begin to appear ∼1 h after DNA damage has occurred, reach their peak after 6 to 8 h, and remain until 12 h after damage (31, 32). BRCA1-BARD1 also associates with the RNA polymerase II holoenzyme (15, 16). In contrast to the cases of other complexes described above, BRCA1 dissociates from hyperphosphorylated, processive polymerase II 1 h after DNA damage (17). However, how BRCA1 affects the polymerase II complexes, if at all, during the early stages after DNA damage and before the translocation of BRCA1 to the repair machinery remains to be elucidated. Our results suggest that BRCA1 polyubiquitinates a component of the polymerase II complex, RPB8, at this early stage after DNA damage.

Recently, ubiquitination of phosphorylated RPB1 by BRCA1-BARD1 has been reported (23, 25). Because double knockdown of BRCA1 and BARD1 restored the expression level of the phosphorylated polymerase II that had been repressed by UV irradiation, it was proposed that BRCA1-BARD1 could initiate the degradation of stalled RPB1 (23). However, the BRCA1-BARD1 double knockdown did not detectably affect RPB1 ubiquitination after UV irradiation. In addition, BRCA1-BARD1–mediated polyubiquitination of other substrates, including NPM1/B23 and phosphorylated CtIP, is not a signal for degradation (12, 33). Therefore, the restored expression level of the phosphorylated polymerase II by BRCA1-BARD1 double knockdown could be due to an indirect effect (23), for example, through the failure to ubiquitinate RPB8. Nonetheless, the clearly shown in vitro ubiquitination of phosphorylated RPB1 by BRCA1-BARD1 (23) strongly supports its direct role. The key to solving this discrepancy may be to analyze the timing of RPB1 ubiquitination in vivo. RPB1 ubiquitination shown in the previous report occurred 2 h after UV irradiation, when BRCA1 should already be dissociated from polymerase II and relocalized to the Rad50 or Rad51 DNA repair machineries. It is possible that early after DNA damage, RPB1 and RPB8 are transiently ubiquitinated by BRCA1 at the same time, and it may result in dissociation of the polymerase II holoenzyme from the damaged DNA site. RPB1 ubiquitination and degradation occurring in late phases could be mediated by other E3 ligases, such as the CSA-DDB1-CUL4A-ROC1 complex (34, 35).

It is well known that cells with impaired BRCA1 function display hypersensitivity to a range of DNA-damaging agents, including IR and UV irradiation (3, 26). However, the mechanism underlying this phenomenon is not fully understood. Although the failure of checkpoint function is a possible mechanism responsible for the hypersensitivity, it has been reported that neither selective abrogation of the S-phase checkpoint nor the G₂ checkpoint itself results in decreased cell survival after DNA damage (36, 37). Therefore, it has been proposed that some function of BRCA1 other than S-phase or G₂ cell cycle control may affect cell survival after DNA damage (37). The UV hypersensitivity of the cells stably expressing a ubiquitin-resistant mutant of RPB8 shown in this report provides a possible new role for BRCA1 that may compensate for this theoretical defect. Because hyperphosphorylated stalled polymerase II at damaged sites is an extremely cytotoxic ramification of DNA damage (38), the observed UV hypersensitivity could be caused by trapped polymerase II or prolonged polymerase II hyperphosphorylation. In this process, the ubiquitination of RPB8 could be an important step either for polymerase II disassembly, polymerase II dissociation from DNA, or polymerase II dephosphorylation by FCP1. It is interesting that there is considerable expression of endogenous WT RPB8 in the ubiquitin-resistant RPB8 mutant cells (Fig. 4A). This indicates that only partial interference of the RNA polymerase recovery is enough to induce cell death, probably by silencing a gene critical for cell survival. Alternatively, polymerase II complexes containing mutant RPB8 could stall at the damaged sites, subsequently causing a gridlock of all polymerase II complexes, including WT complexes. Supporting this idea, induction of local damage by microbeam UV irradiation in the nucleus led to transcription inhibition throughout the nucleus (39).

Lastly, it is noteworthy that RPB8 is shared by all three classes of RNA polymerases (19, 40). Whereas polymerase II synthesizes mRNA, which is only ∼5% of all RNAs, polymerase I and polymerase III synthesize the remaining 95% of all RNAs. Therefore, modification of those complexes, rather than polymerase II, might enormously influence cellular conditions. Whereas RPB1 has been intensively studied, the role of RPB8 in the DNA damage response has been poorly understood. The ubiquitination of RPB8 by BRCA1 reported here provides additional evidence for the role of RNA polymerases in the DNA damage response as well as in carcinogenesis.

Grant support: Japan Society for the Promotion of Science and the Japanese Ministry of Education, Culture, Sports, Science and Technology.

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.

We thank Drs. Yanping Zhang, Minoru Takata, and Masamichi Ishiai for helpful discussions and critical reading of the manuscript; and Drs. Richard Baer and Nouria Hernandez for their generous contribution of materials.