Molecular Cell Article

BRCA1 Is Required for after UV-Induced DNA Damage

Shailja Pathania,1,4 Jenna Nguyen,4 Sarah J. Hill,4 Ralph Scully,2,6 Guillaume O. Adelmant,3,4,5 Jarrod A. Marto,3,4,5 Jean Feunteun,7 and David M. Livingston1,4,* 1Department of Genetics 2Department of Medicine 3Department of Biological Chemistry and Molecular Pharmacology Harvard Medical School, Boston, MA 02115, USA 4Department of Biology 5Blais Proteomics Center Dana-Farber Cancer Institute, 450 Brookline Avenue, Boston, MA 02215, USA 6Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, MA 02215, USA 7Laboratoire Stabilite´ ge´ ne´ tique et Oncoge´ ne` se, UMR8200 Institut Gustave-Roussy, Villejuif 94805, France *Correspondence: [email protected] DOI 10.1016/j.molcel.2011.09.002

SUMMARY In NER, like XPC, XPA, and TFIIH are first attracted to sites of damage, and the damaged DNA is converted to a BRCA1 contributes to the response to UV irradiation. suitable repair substrate. Photoproduct excision is performed Utilizing its BRCT motifs, it is recruited during S/G2 to by the ERCC1/XPF and XPG endonucleases to create a UV-damaged sites in a DNA replication-dependent 30 nt, single-strandeded DNA (ssDNA)-gapped intermediate but nucleotide excision repair (NER)-independent (Huang et al., 1992), which is then filled by a DNA polymerase manner. More specifically, at UV-stalled replication active in postreplicative repair (e.g., DNA polymerase d)(Ogi forks, it promotes photoproduct excision, suppres- et al., 2010; Wood and Shivji, 1997). During DNA replication, other ssDNA gaps, some a kilobase sion of translesion synthesis, and the localization or longer, are generated by the stalling of replication forks at and activation of replication factor C complex (RFC) unrepaired damage sites (Byun et al., 2005; Cortez, 2005; subunits. The last function, in turn, triggers post-UV Sogo et al., 2002). Repair of these gaps is NER-independent, checkpoint activation and postreplicative repair. may involve postreplicative recombinational repair, and is re- These BRCA1 functions differ from those required quired to re-establish an intact replicon (Michel et al., 2004; for DSBR. Nagaraju and Scully, 2007; Petermann and Helleday, 2010). If not repaired, stalled fork gaps can evolve into genome-desta- bilizing, even lethal DSBs (Cox et al., 2000; Petermann and Hel- INTRODUCTION leday, 2010). Unrepaired bulky adducts also trigger translesional DNA BRCA1, a breast and ovarian cancer-suppressing , is synthesis (TLS) (Friedberg, 2005; Kannouche et al., 2004). In both an E3 ubiquitin ligase and a dynamic scaffold that functions TLS, an adduct can serve as a template for error-prone replica- in the maintenance of genome integrity. It is attracted to double- tion at an otherwise stalled fork. Members of the Y family of DNA stranded DNA breaks (DSBs) during the S and G2 cell-cycle polymerases catalyze TLS, activation of which involves a specific phases (Scully et al., 1997a) and supports error-free, homolo- E3 ubiquitin ligase, Rad18, that monoubiquitinates PCNA on gous recombination-type DSB repair (HR/DSBR), (Chen et al., lysine 164 (Moldovan et al., 2007). This modification triggers 1999; Moynahan et al., 1999; Narod and Foulkes, 2004; Scully the binding of a Y family polymerase that carries out bypass et al., 1997b). BRCA1 also supports DSB-driven intra-S and replication of an offending bulky adduct. -G2 checkpoint activation (Xu et al., 2001). Although there is speculation that BRCA1 is, primarily, dedi- Little evidence currently points to a BRCA1 role in the re- cated to DSBR, it and the BRCA1-associated protein, BACH1, sponse to DNA damage arising from bulky DNA adducts. also operate in the response to interstrand crosslinking (ICL) Such lesions develop frequently in most cells and are poten- agents and to reactive oxygen species (ROS) (Alli et al., 2009; tially mutagenic when unrepaired. Typically, they trigger nucle- Shen et al., 2009). PALB2, another known BRCA1-associated otide excision repair (NER), the primary pathway for eliminating protein, also associates with ICLs (Shen et al., 2009). We have DNA helix-distorting lesions like those that arise after UV expo- now identified a specific role for BRCA1 in key aspects of the sure (6–4 photoproducts [6-4PPs] and cyclobutane pyrimidine response to stalled replication forks generated by UV damage. dimers [CPDs]]. Analogous DNA lesions are generated by Its role in this setting differs fundamentally from its role in the a variety of chemical mutagens (Cleaver, 2005; de Laat et al., response to DSB, implying that its contribution to DNA damage 1999). control is more variegated than was previously suspected.

Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. 235 Molecular Cell BRCA1 Is Required for Postreplicative Repair

RESULTS localized at UV sites very soon after irradiation and considerably before BRCA1, we speculate that BRCA1 does not participate in Effect of BRCA1 Deficiency on the Response NER-driven photoproduct recognition. to UV-C damage To test whether BRCA1 concentration at UV sites is depen- Four BRCA1/ human breast cancer cell lines all proved to be dent on the prior localization of NER proteins therein, primary 2.5- to 3.5-fold more sensitive to UV-C irradiation in colony WT, XPA-deficient, and XPC-deficient fibroblasts were analyzed formation assays than two normal, primary, human breast for BRCA1 recruitment. This activity, assayed 1 hr after UV, was epithelial cell strains (BPE and IMEC, Figure 1A). When one of not dependent upon the presence of XPA and XPC; rather, it was these BRCA1/ lines (HCC1937) was reconstituted with WT modestly enhanced (Figure 2K). Thus, UV damage site recruit- BRCA1 cDNA, its UV hypersensitivity decreased (Figure 1B). In ment of BRCA1 is both replication dependent and NER indepen- addition, a BRCA1 shRNA led to significant BRCA1 depletion dent. This suggests that BRCA1 recruitment is associated (Figure 1C) and increased UV sensitivity (Figure 1D). Thus, with the development of stalled replication forks arising at sites endogenous BRCA1 function suppresses hypersensitivity to containing residual photoproducts. UV-C. Given this suggestion, the question of whether BRCA1 is at- tracted to collapsed forks (i.e., those containing DSBs) arose. BRCA1 Is Recruited to Sites of UV Damage To address it, neutral comet assays were performed 30 min U2OS cells were overlaid with filters that contain microscale after IR or UV. At 10 J/m2, a dose at which BRCA1 is efficiently pores through which UV energy can pass (Ge´ rard et al., 2006). recruited to damage sites, no increase was seen in the tail The overlaid cells were UV irradiated and fixed. Recruitment of moment compared to the undamaged cells. However, increases BRCA1 and other proteins to the resulting sites of UV damage were readily detected in IR-treated cells (10 Gy) (Figure S1B), was assayed by immunofluorescence (IF). Antibody to CPD implying that DSBs were not formed during the time of BRCA1 permitted the detection of individual damaged nuclear regions. recruitment to UV sites (i.e., 30 min). In asynchronous, micropore UV-irradiated U2OS cells, nuclear 53BP1 is recruited to DSB in IR-treated cells, implying that BRCA1 and CPD staining colocalized (Figure 2A). it might, at a minimum, be attracted to DSB at collapsed forks. Whether BRCA1 behaves similarly in G0 and G1 was also Unirradiated and UV-irradiated control and BRCA1-depleted tested. During these intervals, BRCA1 protein levels are low cells were each fixed at different times after UV and then analyzed and only increase when cells enter S phase (Scully et al., for 53BP1 foci (Figure 2L). There were no 53BP1 foci in control or 1997a). Indeed, BRCA1 was not detected at UV sites in G0 cells BRCA1-depleted cells 1 hr after irradiation (20 J/m2). Even 20 hr (Figures 2B and S1A), and the efficiency with which it concen- later, control cells lacked 53BP1 foci (Figure 2L). By contrast, trated there was much lower in G1 than in S phase cells (Fig- the frequency and intensity of foci at these times rose dramati- ure 2B). Moreover, when cells were synchronized in S, treated cally in UV-irradiated, BRCA1-depleted cells, implying an accu- with aphidicolin (a replication-blocking compound), and then mulation of collapsed forks (DSBs) in its absence. UV irradiated, BRCA1 recruitment was reduced relative to Thus, DSB developed at late times after UV irradiation and only controls (Figures 2C and 2D). Since extracts of aphidicolin- in cells depleted of BRCA1. Therefore, the data suggest that treated and control cells contained equivalent amounts of BRCA1 is not attracted to UV sites because of the concurrent BRCA1 protein (Figure 2E), its damage site recruitment is largely development of DSB. Furthermore, the increase in collapsed dependent upon ongoing DNA replication. forks in the absence of BRCA1 suggests that BRCA1 is required For BRCA1 recruitment to DSB, its BRCT motifs are required for the prevention of DSB at UV-stalled forks. (Au and Henderson, 2005). When U2OS cells were transfected with and equivalently expressed a WT copy of a C-terminal, Photoproduct Elimination myc-tagged BRCT motif-containing BRCA1 fragment The ability of asynchronous WT and BRCA1-depleted cells to (B1F6myc, aa 1314–1863) or a clinically relevant point mutant clear CPD and 6-4PP DNA adducts was assessed. There was derivative (B1F6_M1775Rmyc) (Miki et al., 1994)(Figure 2F), no significant change in the rates of 6-4PP or CPD elimination the former blocked the concentration of endogenous BRCA1 following BRCA1 depletion (Figure S2A), although the same cells at CPD-rich sites while the latter was much less effective in were, as expected, defective in HR post-IR, when assayed in this regard (Figures 2G and 2H). A fragment containing the parallel (data not shown) (Moynahan et al., 1999). Thus, BRCA1 paired BRCT motifs of 53BP1 also failed in this regard (data is not required for the majority of photoproduct elimination in not shown). Thus, the localization block associated with overex- an asynchronous culture. pression of the WT BRCA1 fragment is not a generic effect of Given the replication dependence of its recruitment to UV overexpressing any BRCT domain-containing fragment. Thus, damage sites, we asked whether BRCA1 contributes to photo- at least one of its BRCT motifs is needed for BRCA1 recruitment product excision in a subset of replicating cells. Here, a 20 J/m2 to UV damage sites. UV dose was employed to study 6-4PP excision (Figure 3A) Cells were also analyzed for the concentration of certain NER and 0.5 J/m2 to study CPD excision (Figure 3B). The assigned proteins at UV-damaged sites. TFIIH and XPA concentrated at UV dose was used for 6-4PP excision analysis to maximize these sites %10 min after irradiation (Figures 2I and 2J). stalled fork generation arising from unrepaired damage. CPD BRCA1 became similarly localized 30 min later and remained excision was analyzed at a relatively low UV dose to decrease there for R3 hr, a time when XPA staining had dissipated. Since the abundance of these lesions, increase the extent of their an NER protein (XPA) engaged in photoproduct recognition excision over time, and thereby facilitate excision analysis. CPD

236 Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. Molecular Cell BRCA1 Is Required for Postreplicative Repair

A UV IC50s 18 16 14 12 10 8

J/m2 6 4 2 0

BPE IMEC SUM149 HCC1937 SUM1315 L56BRC1

‘Normal’ mammary BRCA1 mutant lines controls

B 120

100

80

HCC1937 + WT BRCA1 60 HCC1937 + Vector

40

20 Cell survival (percentage)

0 0251015202530 UV dose (J/m2)

C D 120

100

80

shGFP shGFPshBRCA1 60 shBRCA1 BRCA1 40

20 GAPDH

Cell survival (percentage) 0 0J 4J 8J UV dose (J/m2)

Figure 1. BRCA1–/– Tumors and BRCA1-Depleted Cells Are Hypersensitive to UV-C (A) Normal primary human breast epithelial cells (BPE/IMEC) and BRCA1/ tumor cell lines were irradiated with increasing doses of UV-C and 7 days later assayed for colony formation. Error bars represent the standard deviation between two independent experiments. (B) HCC1937, a BRCA1 mutant breast cancer cell line that is hypersensitive to UV, was reconstituted with BRCA1 cDNA or backbone vector. Reconstituted cells were irradiated and assayed for colony formation. Error bars represent the standard deviation in two independent experiments. (C) Western blot analysis of BRCA1 protein abundance in control (shGFP) and BRCA1-depleted (shBRCA1) U2OS cells. (D) These cells were also exposed to increasing doses of UV-C and assayed for colony formation. lesions are excised slower than 6-4PP, because they do not U2OS cells were UV irradiated, as noted, and photo- distort the DNA helix enough to be an ideal NER trigger (Kim product excision was analyzed by FACS. No cell-cycle per- et al., 1995). turbation or change in replication timing was detected in the

Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. 237 Molecular Cell BRCA1 Is Required for Postreplicative Repair

238 Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. Molecular Cell BRCA1 Is Required for Postreplicative Repair

BRCA1-depleted culture (Figures S2B and S2C). Yet after damage and was NER dependent (Figure 3E) and BRCA1 inde- BRCA1 depletion, S phase and, to some extent, G2 cells re- pendent. Thus, BRCA1 appears not to be required for canonical vealed a decrease in their CPD and 6-4PP excision rate (30 nt), NER-dependent gap repair. compared to controls (Figures 3A and 3B). BRCA1-depleted G1 cells showed no such defect. Thus, in a fraction of S and Single-Stranded DNA Gaps at Stalled Replication Forks G2 cells, BRCA1 is required for optimal photoproduct elimi- NER-independent ssDNA gaps of various lengths are generated nation. Moreover, ERCC1, a key subunit of the ERCC1-XPF following fork stalling at photoproduct-containing sites. They complex and a participant in NER-driven photoproduct excision, develop when the polymerase at the leading strand halts at the concentrated weakly in UV-damaged chromatin following site of damage but the continues to unwind the DNA BRCA1 depletion (Figure 3C), although its intracellular abun- template (Figure 4A) (Byun et al., 2005; Lopes et al., 2006). dance did not change (data not shown). No such defect was Proper repair of these gaps contributes to emergence of an observed for other NER proteins (e.g., XPG and TFIIH) (Fig- intact replicon. ure S2D). The BRCA1 dependence for optimal ERCC1 recruit- To assess gap formation, we assayed ssDNA generated in ment was also replication dependent (cf. HU-treated cells, Fig- UV-irradiated, replicating cells using a BrdU immunoassay ure 3C), implying that ERCC1 recruitment, like that of BRCA1, performed under nondenaturing conditions (–HCl) (Rubbi and is triggered by stalled forks. This result predicts that BRCA1 Milner, 2001)(Figure 4B). The intensity of BrdU staining was depletion should render NER-deficient cells even more UV sensi- greatly diminished following aphidicolin addition (Figure S3A), tive, which was indeed the case (Figure 3D). and an intense signal appeared in a sizeable fraction of unper- Thus, a subpopulation of UV lesions, i.e., those that had not turbed, asynchronous, XPA/ and XPC/ fibroblasts (Fig- yet been excised by the NER pathway and had spawned ure S3B). Thus, the signal is largely replication dependent and stalled forks, are likely excised during S/G2 in a BRCA1- NER independent and thereby signals photoproduct-driven dependent manner in part through its effect on ERCC1/XPF ssDNA gaps that likely result from replisome stalling (Figure 4A). recruitment to stalled forks. In addition, much-reduced ssDNA (BrdU staining) was de- tected in BRCA1-depleted cells compared to controls (Fig- Post-UV Gap Repair in G1 Cells ure 4B). Yet when the DNA in the same cells was denatured NER-dependent gap repair is most efficient soon after UV expo- and stained for BrdU, equivalent incorporation was detected in sure (Luijsterburg et al., 2010; Mocquet et al., 2008) when XPF/ both control and BRCA1-depleted cells (Figure 4B, +HCl panel). ERCC1 and XPG endonucleases concentrate at UV damage Thus, the reduction in BrdU staining in BRCA1-depleted cells sites and generate canonical 30 nt gaps (Riedl et al., 2003). arose from the absence of ssDNA rather than a DNA synthesis DNA polymerase delta (Pol d), an enzyme likely involved in repair- block. ing these gaps, is also recruited to UV damage sites soon after After DNA damage, the RPA protein coats ssDNA and is UV exposure (Mocquet et al., 2008). Indeed, its recruitment phosphorylated (Carty et al., 1994). RPA immunostaining of was detected within 10 min after irradiation and gradually damaged DNA serves as a reporter of ssDNA gap formation after decreased over the next 2 hr (in cells arrested at G1/S; data UV irradiation (Zou and Elledge, 2003), and we used anti-Ser4/8 not shown). Similarly, in vivo uptake of biotinylated dUTP (bio- phospho-RPA32 IF as a gap assay. Together with the dramatic dUTP), an assay for gap repair, was detectable 30 min after reduction in UV-dependent, single-stranded BrdU staining in

Figure 2. Recruitment of BRCA1 to Sites of UV Damage (A) Asynchronously growing U2OS cells were UV irradiated (30 J/m2) through micropore filters. One hour postirradiation, cells were fixed and coimmunostained with CPD and BRCA1 Abs. (B) U2OS cells synchronized in S, G1, or in G0/G1 were UV irradiated and fixed 1 hr later for coimmunostaining with BRCA1 and CPD Abs. The error bars represent the standard deviation in three separate experiments. (C) S phase-synchronized U2OS cells were incubated in the presence of aphidicolin (2 mg/ml) for 1 hr to block replication before UV irradiation (30 J/m2). Two hours later, the cells were coimmunostained for BRCA1 and CPD. (D) Quantization of BRCA1 localization to CPD-positive sites in the experiment described in (C). Error bars represent the standard deviation in three independent experiments. (E) BRCA1 abundance in control (DMSO) and aphidicolin-treated U2OS cells. (F) A myc-tagged, BRCA1 BRCT domain containing fragment (B1F6-myc) and a point-mutated derivative (B1F6-M1775R-myc) were each transfected into U2OS cells, and extracts were analyzed for expression of endogenous BRCA1 and the myc-tagged fragments. (G) B1F6-myc- and B1F6-M1775R-myc-transfected U2OS cells were UV irradiated, and BRCA1 Ab (MS110—directed against an N-terminal epitope) was used to detect the recruitment of endogenous BRCA1 to the UV damage sites. Ab to myc was used to mark the cells expressing the myc-tagged fragments. (H) Myc-positive cells were analyzed for endogenous BRCA1 localization at UV damage sites. Error bars represent standard deviation in three independent experiments. (I and J) U2OS cells were irradiated and immunostained for BRCA1, XPA, and TFIIH. Cells were fixed at the indicated times and immunostained with the respective antibodies. (K) Localization of BRCA1 at CPD-rich UV damage sites was analyzed in two NER-deficient cell lines (XPA/ and XPC/). A primary fibroblast strain, IMR90, was used as a WT control. All three cultures were UV irradiated (30 J/m2) and coimmunostained for CPD and BRCA1. The fraction of cells containing BRCA1 and CPD colocalized at UV damage spots is plotted. Error bars represent standard deviation in three independent experiments. (L) U2OS cells infected with control (shLuc) and ShBRCA1-encoding lentiviruses were irradiated with 20 J/m2 and fixed at 1 hr and 20 hr after UV. Cells were immunostained with Ab to 53BP1.

Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. 239 Molecular Cell BRCA1 Is Required for Postreplicative Repair

6-4PP Excision CPD Excision AB110 100 100 90 80 80 G1 60 G1 70 60 40 50 ShLuc 40 20 ShBRCA1 % CPD left after UV (0.5J/m2)

% 6-4PP left after % 6-4PP UV (20J/m2) ShLuc 30 ShBRCA1 0 20 1hr 3hr 5hr 7hr 4.5hr 6.5hr 10hr Time after UV Time after UV 110 100 100 90 80 80 S 70 S 60 60 40 50

% CPD left after UV (0.5J/m2) 40 20 ShLuc 30 ShLuc % 6-4PP left after % 6-4PP UV (20J/m2) ShBRCA1 ShBRCA1 0 20 1hr 3hr 5hr 7hr 4.5hr 6.5hr 10hr Time after UV Time after UV 110 100 100 90 80 80 G2 G2 60 70 60 40 50

% CPD left after UV (0.5J/m2) 40 20 ShLuc ShLuc % 6-4PP left after % 6-4PP UV (20J/m2) ShBRCA1 30 ShBRCA1 0 20 1hr 3hr 5hr 7hr 4.5hr 6.5hr 10hr Time after UV Time after UV

C DEXPA-/-+ shLuc XPA-/-+shBRCA1 100 16 ShLuc ShBRCA1 ShLuc ShBRCA1 80 14 HU (16hrs) - --- + +++ 12 UV -+-+-+-+ 60 10 8 BRCA1 40 6 4

Percentage Survival 20 2

ERCC1 positive % of BiodUTP 0 0 0J 0.5J 1.0J 1.5J 2.0J shLuc XPC-/- XPA-/- UV Dose (J/m2) shBRCA1-

U2OS

Figure 3. Excision of UV-Induced Lesions Is BRCA1 Dependent in S and G2 Phase Cells (A and B) U2OS cells infected with control (shLuc) and shBRCA1-encoding lentiviruses were irradiated with 20 J/m2 and 0.5 J/m2 and the excision efficiencies of 6-4PP and CPD were analyzed, respectively, by FACS. Cells were gated for S, G2, and G1 phase to study damage excision during different cell-cycle phases. (C) Western blot analysis of ERCC1 and BRCA1 in chromatin extract prepared from unirradiated and UV-irradiated HeLa cells infected with lentiviruses encoding a Luc (control) or a BRCA1 hairpin. Where indicated, cells were pretreated with HU (1 mM) for 16 hr to block replication and arrest cells at the G1/S boundary. Cells were harvested 4 hr after UV exposure (30 J/m2).

240 Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. Molecular Cell BRCA1 Is Required for Postreplicative Repair

BRCA1-depleted cells, there was markedly lower phospho- Postreplicative Gap Repair at UV Damage Sites RPA32 staining (Figures 4C, 4D, and 4G). By contrast, the total One wonders whether BRCA1-dependent generation of ssDNA cellular abundance of RPA did not change (Figure S3C). at UV-stalled forks acts as a signal to attract gap binding and We then attempted to ‘‘rescue’’ this defect by restoring WT repair proteins to promote postreplicative repair of UV-stalled BRCA1 synthesis to the BRCA1 shRNA-expressing cells (Fig- forks. Such repair is predicted to occur later than the major ure 4E). Reconstitution was attempted with a WT BRCA1 waves of NER that occur in normal cells. cDNA vector, the RNA product of which is resistant to the afore- Indeed, unlike much of the NER repair, which can be detected mentioned shRNA, or with a vector control. When compared to within 30 min after damage (cf. above), the first detectable the control, WT BRCA1-reconstituted cells revealed a marked concentration of PCNA, Pol d, and bio-dUTP occurred at 3 hr. increase in post-UV RPA/chromatin loading. Thus, the defect The latter effect largely disappeared after BRCA1 depletion in RPA/chromatin loading is a specific consequence of BRCA1 (Figures 5A and 5B). Moreover, by comparison with control cells, depletion and not an off-target effect. Furthermore, three other, XPA depletion led to a significant increase in RPA and Pol d nonoverlapping BRCA1-directed shRNAs and one BRCA1- recruitment (Figures 5C and 5D), but the latter was completely specific siRNA also elicited a defect in RPA/chromatin recruit- abolished in cells codepleted of BRCA1 (Figure 5D). Thus, ment after UV exposure versus controls (Figure S3D). BRCA1 supports the accumulation of postreplicative repair How BRCA1-dependent gap formation relates to the post-UV machinery at stalled forks that arise at lesions not repaired by dissociation of MCM7 from the DNA polymerase complex was NER. Moreover, post-UV colony formation by BRCA1-depleted then probed. Following partial MCM7 depletion in asynchronous XPA/ cells was rendered even more defective than that by cells, we observed a significant decrease in post-UV ssDNA control XPA/ cells (Figure 3D). generation, as assayed by RPA staining (Figure S3E–S3G), Conceivably, there is a contribution of the homologous re- without incurring a significant effect on the overall rate of DNA combination pathway to repair of stalled forks that lack DSB replication (data not shown); yet BRCA1 recruitment to UV (cf. earlier discussion). Indeed, Rad51 accumulates at HU- damage sites was unperturbed (Figure S3F). This implies that, stalled replication forks (Hanada et al., 2007; Scully et al., for postreplicative gap formation, BRCA1 operates either up- 1997a) and is required for their restart (Petermann et al., 2010). stream of or parallel to MCM/replication complex dissociation. Rad51 did accumulate at UV sites but only in presence of RPA is required for the recruitment of ATRIP to ssDNA and BRCA1 (Figure 5E). However, depletion of Rad51 led to only for the subsequent activation of ATR (Zou and Elledge, 2003). a modest change in Pol d recruitment to these sites (Figure 5F). In keeping with the absence of RPA-coated ssDNA in BRCA1- As expected, there was a major HR defect in these Rad51- depleted cells, ATRIP recruitment was also compromised depleted cells, similar to that observed after BRCA1 depletion (Figures 4F and 4G). Thus, the post-UV accumulation of ssDNA, (Figure S4A). Thus, Rad51-dependent HR normally plays a coating of ssDNA by RPA, and recruitment of ATRIP to it are all relatively minor role in the postreplicative repair of UV-stalled BRCA1-dependent in replicating cells. forks.

BRCA1 Participation in Post-UV Checkpoint Activation Role of BRCT Binding Proteins in Post-UV BRCA1 Since ATRIP recruitment to RPA-coated DNA was compro- Function mised following BRCA1 depletion, we asked whether this Since the C-terminal BRCT motifs of BRCA1 are essential for defect also resulted in impaired Chk1 phosphorylation and the recruitment of BRCA1 to sites of UV damage, we asked G2/M checkpoint activation (Bulavin et al., 2001; Callegari whether any of the three known BRCA1-BRCT interacting and Kelly, 2007). BRCA1 depletion led to reduced UV-induced proteins—BACH1, CtIP, and Abraxas—contribute to ssDNA Chk1 phosphorylation and G2/M checkpoint activation versus gap formation at stalled forks. Using specific siRNAs, U2OS controls (Figures 4H and S3H). cells were depleted, in parallel, of CtIP, BACH1, and Abraxas; Whether BRCA1 contributes to the post-UV intra-S phase UV irradiated; and analyzed for gap formation. ssDNA/BrdU checkpoint was also tested. The Cdc45 association with staining revealed that CtIP depletion led to suppression of these chromatin is disrupted after bulky adduct-induced damage (Liu effects (Figure S4B). Parallel depletion of Abraxas and BACH1 et al., 2006), and when the abundance of chromatin-bound had no such effect, although all three depletion events, per- Cdc45 was compared before and after UV, there was a signifi- formed in parallel, led to a major perturbation of homologous cant reduction, except after BRCA1 depletion (Figure 4I). There recombination (HR) function (Figure S4C) (cf. legend for explana- was also a less dramatic, post-UV decrease in BrdU uptake in tion of Abraxas depletion effect on HR) (Hu et al., 2011). Thus, BRCA1-depleted versus control S phase cells (Figure S3I). BRCA1 and its known binding partner, CtIP, are both required Thus, BRCA1 function contributes to the post-UV intra-S and to generate ssDNA at UV-stalled forks. Furthermore, the G2/M checkpoints, much as it does after DSB formation. absence of a strong contribution by Abraxas and BACH1 is

(D) GM04312 (XPA/) cells were infected, in parallel, with control (ShLuc) and BRCA1 shRNA-encoding lentiviruses, plated in triplicate, and irradiated with the indicated doses of UV. Colony formation was assayed after 7 days. Error bars represent the standard deviation in two independent experiments. (E) XPA/, XPC/, and U2OS cells that were infected with shBRCA1 or shLuc lentiviruses were arrested at G1/S by incubation with 1 mM HU for 16 hr and then irradiated through micropore filters. Cells were fixed 30 min later, and gap repair was analyzed by measuring in situ DNA synthesis in the presence of bio-dUTP. The fraction of cells that revealed bio-dUTP staining in UV damage sites was plotted as a mean of three separate experiments. Error bars represent the standard deviation between these experiments.

Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. 241 Molecular Cell BRCA1 Is Required for Postreplicative Repair

A E ShLuc ShBRCA1 Vector BRCA1 Vector BRCA1 + UV - + - + - + - + UV Stalled lagging strand BRCA1

Myc CPD (BRCA1)

Stalled leading strand Phospho -RPA32 B No denaturation (- HCl) H3 Merge BrdU (ssDNA)

CPD ATRIP F - UV Denatured (+ HCl)

BrdU ShLuc

ShLuc (control) + UV ShBRCA1

shLuc G shBRCA1 + UV 40 35 30 ShBRCA1 25 20 C 15 shLuc (control) shBRCA1 10 - UV + UV (5J/m2 ) - UV + UV (5J/m2 ) 5 Percentage of Cells 0 Phospho-RPA ATRIP

H shGFP shBRCA1 - + + + - + + + UV Phospho RPA32 - 1 3 5 - 1 3 5 Time (hr)

BRCA1 DAPI Chk1S317

MCM7 D shLuc shBRCA1 - + + + + - + + + + UV - 1 3 5 7 - 1 3 5 7 Time (hr) I Chromatin extract BRCA1 shLuc shBRCA1 - + - + Phospho BRCA1 RPA32 MCM7 GAPDH Cdc45

Figure 4. Generation of RPA-Coated ssDNA after UV Damage Is Largely BRCA1 Dependent (A) Model: UV-induced damage in replicating cells generates stalled forks. (B) BrdU assay for detection of ssDNA after UV irradiation. U2OS cells were fixed 4 hr after irradiation and immunostained for BrdU with and without denaturation of DNA with HCl. (C) ShLuc- and shBRCA1-transduced U2OS cells were UV irradiated (5 J/m2), fixed 4 hr later, and immunostained with phospho-RPA32 Ab.

242 Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. Molecular Cell BRCA1 Is Required for Postreplicative Repair

consistent with BRCA1 function at UV-stalled forks not being Thus, like BRCA1, RFC complex subunits are conveyed to a simple outcome of its attraction to DSB in that setting. We UV-stalled forks at late times and support the generation of also found that BRCA1-depleted cells revealed inefficient sizable ssDNA gaps, Pol d recruitment, and predictably, postre- recruitment of CtIP-GFP to UV-damaged DNA, unlike control plicative repair at these sites. Unexpectedly, the RFC subunit cells (Figures S4D and S4E). However, CtIP-depleted cells re- behavior in this setting is BRCA1 dependent. vealed only a minimal defect in the recruitment of BRCA1 to Since an interaction between RFC-Rad17 and the 9-1-1 UV-damaged sites. complex is required for checkpoint activation after UV damage (Parrilla-Castellar et al., 2004; Sancar et al., 2004; Zou et al., BRCA1 Directs RFC Complex Function 2002), we assayed the loading of the 9-1-1 complex at UV- at UV-Stalled Forks damaged sites. Compared to controls, it was significantly per- How does BRCA1 execute a response to UV damage-driven turbed after BRCA1 depletion (Figure 6H), implying that at least replication stress? In this regard, the RFC complex is a compo- part of the BRCA1 effect on RFC function is manifest as support nent of a machine that contributes to checkpoint activation at for RFC-Rad17/9-1-1 complex formation at stalled forks. stalled replication forks (Parrilla-Castellar et al., 2004; Sancar We also asked whether Rad9 recruitment to sites of UV et al., 2004). Replication factor C (RFC1-5) is the clamp loader damage is compromised in the absence of BRCA1. Consistent for PCNA during DNA replication, and after UV damage, RFC1 with the aforementioned results, a marked reduction in Rad9 is substituted in this complex by Rad17 to form the RFC2-5/ localization at micropore sites was detected in Rad9-GFP-ex- Rad17 (RFC-Rad17) complex (Bermudez et al., 2003; Parrilla- pressing, BRCA1-depleted cells compared to controls (Figures Castellar et al., 2004). RFC-Rad17 interacts with and helps to 6I and 6J). These data further suggest that BRCA1 is required load the Rad9-Hus1-Rad1 (9-1-1) complex at stalled forks (Med- for the efficient development of an interaction between RFC- hurst et al., 2008; Zou et al., 2002), where 9-1-1 helps to trigger Rad17 and 9-1-1, which in turn helps to explain the role of the G2/M checkpoint. Since BRCA1 participates in eliciting this BRCA1 in post-UV, G2/M checkpoint activation. checkpoint, we asked whether BRCA1 also contributes to the Given that BRCA1 is required for efficient recruitment of Rad9 presence and activity of RFC at UV-stalled forks. to sites of UV damage, the question of whether BRCA1 joins the Antibodies directed at three RFC subunits (RFC1, RFC4, and Rad9-containing chromatin-bound complex was posed. Chro- RFC5) were effective in IF. Using them, we found that all three matin extract was prepared from cells expressing HA-tagged RFC subunits concentrated at UV-irradiated sites 3 hr post-UV Rad9 and was subjected to anti-HA IP. Like RFC4, BRCA1 (Figures 6A, 6B, and 6C), a time when BRCA1-driven postrepli- was present in the chromatin-bound Rad9 complex (Figure 6K). cative repair, an NER-independent process, also occurs. Thus, BRCA1 likely contributes to the formation and recruitment Surprisingly, they did so in a BRCA1-dependent manner (Fig- of the RFC-Rad17/9-1-1 complex at stalled forks and physically ure 6D). In return, RFC2 depletion led to significant codepletion interacts with at least some of its components during or after this of BRCA1 at UV-stalled forks (Figure 6E), suggesting that process. RFC2 promotes B1 concentration at UV-stalled forks. Individual At least one other RFC subunit-containing complex, CTF18- depletion of the other known RFC subunits did not elicit this RFC, an alternative to RFC1-5 as a clamp loader, exists in effect. mammalian cells where it operates, in part, in NER-associated The next question was whether BRCA1-dependent ssDNA repair synthesis (Ogi, et al., 2010). Currently, we have no indica- generation and postreplicative repair at stalled forks is affected tion whether or not it associates specifically with BRCA1 or is by RFC subunit depletion. Indeed, depletion of RFC1, RFC2, influenced by its function. RFC3, RFC4, or RFC5 (Figure S5A) in each case led to a marked reduction in phosphorylated RPA and Pol d recruitment to micro- BRCA1 Controls the Recruitment of Pol h pore sites (Figures 6F and 6G) and, one might predict, postrepli- to UV-Damaged Sites cative repair at these sites versus controls. To test whether this Bypass of UV-damaged DNA through TLS was also assayed was due to a reduction in DNA replication rate and, possibly, in BRCA1-depleted cells. Specifically, post-UV localization of the number of stalled forks resulting from depletion of any single Pol h (PolH) was assessed as a test of ongoing TLS activity RFC complex member, the rate of DNA replication was tested, (Kannouche et al., 2001). When studied in control and and it did not change (Figure S5B). BRCA1 shRNA-treated cells that had been transfected with

(D) ShLuc- and shBRCA1-infected U2OS cells were UV irradiated (30 J/m2) and collected at the indicated time points. The western blot was immunostained for phospho-RPA32, BRCA1, and GAPDH (loading control). (E) ShLuc- and shBRCA1-transduced U2OS cells were transfected with myc-tagged WT BRCA1 in parallel with a vector control. Forty-eight hours later, cells were UV irradiated (30 J/m2) and harvested 4 hr later to prepare chromatin extracts, which were then blotted. The western blot was immunostained for phospho- RPA32, BRCA1 (using MS110 and anti-myc), and histone H3 (loading control). (F) Localization of ATRIP in locally UV irradiated cells. ShLuc- and shBRCA1-transduced U2OS cells were UV irradiated (30 J/m2), fixed 2 hr later, and coim- munostained for CPD and ATRIP. (G) Data collected from experiments depicted in (C) and (F) are plotted. Error bars represent the standard deviations in three independent experiments. (H) Effect of BRCA1 depletion on Chk1S317 phosphorylation after UV. shRNA-transduced U2OS cells were harvested at the indicated times after UV exposure (30 J/m2) and analyzed for Chk1 phosphorylation. (I) ShLuc-and shBRCA1-transduced U2OS cells were UV irradiated (30 J/m2). Chromatin was prepared from cells harvested 4 hr post-UV. The western blot was immunostained for BRCA1, MCM7, and Cdc45.

Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. 243 Molecular Cell BRCA1 Is Required for Postreplicative Repair

A CPD PCNA CPD Pol δ BiodUTP Merge shBRCA1 shGFP

B CDshGFP+siGAPDH shGFP+siXPA shGFP shBRCA1+siGAPDH shBRCA1 shBRCA1+siXPA 30 120 45

25 100 40 35 20 80 30 15 60 25 20 10 40 15 5 20 10 5 0 0

PCNA Pol delta BiodUTP at UV damage sites % cells with RPA siGAPDH siXPA 0 -CPD colocalization % cells with indicated proteins/BiodUTP -CPD

% cells with pol delta at UV damage sites 3hr post UV

EF

CPD Rad51 Merge

25

20 ShLuc 15

10

5 ShBRCA1 0 % cells with pol delta at UV damage sites WT BRCA1 Rad51

Figure 5. BRCA1 Plays a Role in the Recruitment of PCNA and DNA Polymerase d to Sites of UV Damage (A) U2OS cells infected with shGFP (control)- and shBRCA1-encoding lentiviruses were UV irradiated at 30 J/m2. Cells were fixed 3 hr later and analyzed by IF for PCNA and Pol d recruitment to CPD-positive foci. Repair synthesis (via in situ bio-dUTP incorporation) was analyzed in parallel 3 hr after irradiation. (B) The percentage of cells with the indicated proteins or bio-dUTP localized at CPD-positive foci were plotted for shGFP- and shBRCA1-transduced U2OS cells. (C) Percentage of cells with RPA colocalized with CPD is plotted for siRNA-transfected U2OS cells. (D) ShGFP- and shBRCA1-transduced U2OS cells were transfected with siGAPDH and siXPA. Forty-eight hours after transfection, cells were UV irradiated (30 J/m2), fixed 3 hr later, and coimmunostained for Pol d and CPD. Percentage of CPD-stained cells that were also positive for Pol d was plotted. Error bars represent the standard deviation in three independent experiments. (E) ShLuc- and ShBRCA1-transduced U2OS cells were UV irradiated (30 J/m2), fixed 2 hr later, and immunostained for Rad51 and CPD. (F) Percentage of cells with Pol d at sites of UV damage in siRNA-treated cells.

244 Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. Molecular Cell BRCA1 Is Required for Postreplicative Repair

eGFP-PolH (Figure 7A), the fraction of cells containing eGFP- that it promotes the generation of ssDNA gaps at HU-stalled PolH localized at UV damage sites increased 4-fold after forks through its helicase activity (Gong et al., 2010). Since BRCA1 depletion (Figures 7B and 7C). Most cells in the irradi- a reproducible defect in gap generation was not apparent in ated control culture revealed little or no PolH-GFP at micropore BACH1-depleted cells after UV exposure, the difference in the sites. responses observed in the two DNA damage settings might A prerequisite for the recruitment of Pol h is PCNA monoubi- reflect a genuine difference in the response to stalled forks quitination (Moldovan et al., 2007). Thus, we asked whether arising from HU and UV. They might also reflect inadequate Pol h recruitment in BRCA1-depleted cells is, at least in part, BACH1 protein depletion during our assays. associated with increased PCNA ubiquitination. Compared to Not only do ssDNA stretches serve as checkpoint initiators, controls, BRCA1 depletion led to a super-normal accumulation some can also serve as intermediates in postreplicative repair of monoubiquitinated PCNA (Figure 7D). To test whether this (Lopes et al., 2006) performed, at least in part, by DNA polymer- particular PCNA species is the ubiquitinated form of interest, ases like d. Indeed, in UV-irradiated cells depleted of BRCA1, these cells were transfected with HA-tagged ubiquitin and there was markedly less Pol d staining at UV damage sites and immunoprecipitated with HA Ab. A western blot of the extracts clear suppression of NER-independent postreplicative repair probed with PCNA Ab revealed a parallel increase in both versus controls. One outcome at unresolved stalled replication HA-tagged (which migrates more slowly than the naive ubiquiti- forks is DSB formation. Given the marked increase in the number nated protein) and endogenous monoubiquitinated PCNA of 53BP1 foci in BRCA1-depleted cells at late times after UV (Figure 7D). (Figure 2L), BRCA1-dependent repair of stalled replication forks To study the role of BRCA1 in UV-induced TLS, a SupF-based is likely essential for their resolution. By contrast, BRCA1 is not mutagenesis assay was used (Choi and Pfeifer, 2005)(Fig- required for NER-dependent repair of canonical (30 nt) ssDNA ure S6A). Consistent with the increase in PCNA ubiquitination gaps. These results reinforce the view that BRCA1 operates in BRCA1-depleted cells, there was a 2- to 3-fold increase in in post-UV gap repair primarily, if not exclusively, at photo- frequency in these cells versus controls (Figures 7E product-stalled replication forks. Data presented here also and S6B). These results imply that UV-triggered TLS is a show that these newly detected BRCA1 functions are, at least BRCA1-regulated event. in part, different from those it exerts after the development of DSB. DISCUSSION Unexpectedly, BRCA1 also participates in photoproduct exci- sion at stalled forks and for ERCC1 recruitment there (Figure 3). In addition to its other DNA repair roles, BRCA1 operates in the This suggests (albeit does not prove) that ERCC1 recruitment physiological response to bulky adduct DNA damage. In prin- is an important step during BRCA1-dependent photoproduct ciple, a proposal that BRCA1 induces expression of XPC (Hart- elimination at these sites. Recruitment and function of ERCC1/ man and Ford, 2002) might explain this super-sensitivity. XPF at stalled forks is also required for adduct removal during However, after BRCA1 depletion, there was little or no defect ICL repair (Al-Minawi et al., 2009; Kuraoka et al., 2000). Since in photoproduct excision during G1 when BRCA1 is not attracted BRCA1/ cells are supersensitive to cis-platin (Tassone to UV damage sites (Figures 3A and 3B). et al., 2003), an ICL agent, perhaps, BRCA1 also contributes to BRCA1 is efficiently recruited to sites of UV damage, but only this repair event, at least in part, through ERCC1 recruitment. in cells that are actively replicating their DNA, implying that the From the results reported here, it is not clear whether, in recruitment process represents more than a simple response the absence of NER, BRCA1 can promote ERCC1 localization to the presence of photoproducts. Rather, the protein is at- at photoproducts associated with stalled forks. While the data tracted to sites containing photoproduct-stalled replication reported suggest such a possibility, they do not directly address forks, and its BRCT motif(s) are needed for stable recruitment it. Notably, even in replicating cells, significant albeit incomplete (Figures 2G and 2H). photoproduct excision was observed after BRCA1 depletion. At UV-stalled forks, relatively long stretches of RPA-coated This might be a result of incomplete BRCA1 depletion and/or ssDNA are normally generated (Cortez, 2005). Canonical, NER- the BRCA1-independent excision that occurs at UV sites not dependent 30 nt gaps generated during G1 could not be associated with stalled replication forks. imaged by RPA IF, although they could be imaged by Pol d IF. How is BRCA1 attracted to UV damage sites? Given the repli- One RPA complex requires at least a 30 nt ssDNA stretch for cation dependence of the process, it appears that BRCA1 is binding (Cai et al., 2007); thus, only one RPA complex can bind attracted to UV-stalled forks and not simply to photoproducts. to a traditional NER gap. In our experiments, the sensitivity of One potential explanation for this particular behavior was pro- RPA staining in this setting was insufficient to detect this event. vided by Paull et al. (Paull et al., 2001), who showed that purified Thus, the BRCA1-dependent, post-UV RPA IF signals detected BRCA1 exhibits an intrinsic binding affinity for ds/ssDNA joints in our experiments likely reflect the attraction of RPA to >30 nt or branched structures, such as those that might be expected long ssDNA gaps that form at stalled forks as a result of MCM- at a stalled fork. Conceivably, this contributes to BRCA1 locali- replisome uncoupling. zation at these structures. Among the three known BRCT binding proteins, only CtIP Surprisingly, RFC2 depletion led to a significant defect in appeared to be involved in BRCA1-dependent ssDNA gap BRCA1 binding at UV damage sites (Figure 6E). In return, formation. Others have reported that BACH1 is required for BRCA1 is required for the stable association of at least three RPA loading on chromatin after HU treatment and proposed RFC complex subunits (RFC1, 4, and 5) (cf. Figure 6D) at stalled

Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. 245 Molecular Cell BRCA1 Is Required for Postreplicative Repair

AB C BRCA1 RFC1 BRCA1 RFC4 RFC1 RFC5

DE F RFC1 RFC4 BRCA1 CPD + BRCA1 RFC5 20 18 25 16 14 20

12 ShLuc 10 15 8 6 10 4

Percentage of cells with RFC1, RFC4 or RFC5 2 5 0 ShLuc shBRCA1

ShRFC2 0 % of cells with Pol delta in UV damaged spots

ShLuc ShRFC1 ShRFC2 ShRFC3 ShRFC4 ShRFC5 G H Input 25

20

15 ShLuc ShBRCA1 BRCA1

10 IP: Rad9 IP: Rad1

5 Rad9 ShLuc ShBRCA1 ShLuc ShBRCA1 % of cells with RPA in UV % of cells with RPA damaged spots 0 - + - + - + - + UV

ShLuc RFC4 RFC4 ShRFC1 ShRFC2 ShRFC3 ShRFC4 ShRFC5

I JK CPD Rad9-GFP Merge Input IP: HA

30

25 Rad9-HA Vector-HA Rad9-HA Vector-HA ShLuc 20 BRCA1 15

10 Rad9-HA

5 % of cells with Rad9-GFP in the micropores

ShBRCA1 RFC4 0 shLuc shBRCA1

246 Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. Molecular Cell BRCA1 Is Required for Postreplicative Repair

forks. Hence, once RFC2 concentrates at a stalled fork as part of and RFC-Rad17. The exact nature of the operative RFC complex a process that is BRCA1 dependent for the three RFC subunits in the latter case is not clear, although it is conceivable that that could be so tested, BRCA1 can in return concentrate there RFC-Rad17, RFC1-5, or both operate there, as well. Such maximally. This might reflect a positive feedback loop, whereby considerations aside, these are unexpected findings. the presence of RFC2 further stabilizes the BRCA1 that has RPA is required for the concentration of RFC-Rad17 at artifi- already concentrated at a stalled fork. cially generated DNA gap structures (Zou et al., 2003). The It is also possible that, among the various RFC subunits, only finding that BRCA1 participates in the concentration of RFC RFC2 could be sufficiently depleted, experimentally, to elicit subunits at stalled fork-associated gaps in UV-treated cells is an effect on BRCA1 localization. Other RFC subunits might not inconsistent with this finding. It is possible that RFC-Rad17 also participate in this event, but these effects were not attraction to stalled forks depends upon both BRCA1 and RPA detected because of insufficient knockdown. Finally, we cannot that normally accumulates in relatively low but sufficient quantity rule out the possibility that the effect of RFC2 on BRCA1/stalled at newly stalled forks to attract this complex. fork recruitment occurs independently of its role as a member Furthermore, an MCM7-Rad17 interaction is also believed to of the RFC1-5 or Rad17-RFC2-5 complexes. Clearly, a more contribute to post-UV checkpoint activation (Tsao et al., 2004). detailed understanding of the attraction process is needed. This fosters speculation that RFC-Rad17, in conjunction with At least one RFC complex (RFC-Rad17) operates in check- MCM7, contributes to the generation of long ssDNA gaps at point activation triggered by events at stalled forks. After locali- stalled forks. zation at a stalled fork, the RFC-Rad17 complex interacts with Where CtIP, a known participant in 50 end excision during the 9-1-1 complex, and this results in loading of the latter onto DSBR/HR and an apparently required element in long gap chromatin. Like RFC subunit concentration at stalled forks, this formation at UV stalled forks, fits in the latter process is intercomplex interaction and the subsequent concentration of unknown. Nonetheless, its involvement is BRCA1 driven, Rad9 at UV-stalled forks also proved to be BRCA1 dependent. because it did not gain access to UV sites after BRCA1 deple- Given that an interaction between Rad9 and BRCA1 was de- tion. One might speculate that, following BRCA1-driven recruit- tected (Figure 6K), it is possible that BRCA1 plays a more direct ment to a stalled fork, it directs a ssDNA excision process that role in recruiting the 9-1-1 complex to a stalled fork. The RFC- in turn creates initial RPA binding sites, which, following RPA Rad17/9-1-1 interaction and the 9-1-1 stalled fork loading events binding, attract RFC-Rad17. G2/M checkpoint activation might are, together with ATRIP binding at such sites, required for G2/M then follow. checkpoint function. These events also proved to be BRCA1 Replication fork stalling at UV-damaged sites can trigger TLS dependent. Therefore, the aforementioned collection of B1- (Lehmann, 2005). This activity was normally suppressed by and RFC-directed events can explain a significant part of the BRCA1. Given that BRCA1 is required for excision of stalled process by which BRCA1 triggers G2/M post-UV checkpoint fork-associated photoproducts in S phase cells, perhaps the activation. increase in PCNA ubiquitination observed in the absence of Like BRCA1, each of the known RFC subunits appeared BRCA1 is simply a manifestation of an increase in unexcised necessary for long ssDNA gap formation at these sites and their lesions/unresolved forks (i.e., TLS templates) that arise in this repair thereafter. Taken together, these results strongly suggest setting. that, at UV-stalled forks, BRCA1 promotes the localization of The finding that TLS was enhanced in the absence of BRCA1 is RFC subunits, which we infer become components of an consistent with the finding that a BRCA binding-defective mutant RFC-based complex and thereby support the process leading of BACH1 actually stimulates TLS in BRCA1 WT cells (Xie et al., to long RPA-coated gap formation and its postreplicative repair 2010). This suggests that BRCA1 suppression of TLS is, at least (cf. Figures 6F and 6G). Thus, BRCA1-to-RFC communication in part, linked to its ability to interact with BACH1. is essential for both G2/M checkpoint activation and a key set of Taken together, these findings suggest hypothetical bio- steps in stalled fork resolution/repair. In the former instance, logical outcomes that might occur in BRCA1-deficient repli- the likely communication is predicted to occur between BRCA1 cating cells. Should BRCA1 heterozygous cells prove to be

Figure 6. BRCA1 Is Required to Direct RFC Complex Localization and Function at UV-Stalled Forks (A–C) U2OS cells were irradiated (30 J/m2), fixed 3 hr later, and coimmunostained for BRCA1 and RFC1 (A), BRCA1 and RFC4 (B), or RFC1 and RFC5 (C). (D) ShLuc- and ShBRCA1-infected U2OS cells were irradiated (30 J/m2), fixed 3 hr later, and coimmunostained for CPD and RFC1 or RFC4 or RFC5. CPD-positive cells were analyzed for the recruitment of RFC1, 4, or 5. (E) ShLuc- and ShRFC2-infected cells were irradiated (30 J/m2), fixed 1 hr later, and immunostained for BRCA1. (F and G) ShLuc-, shRFC1-, shRFC2-, shRFC3-, shRFC4-, and shRFC5-infected U2OS cells were analyzed for recruitment of Pol d and RPA, respectively, to sites of UV damage (30 J/m2) at 3 hr after irradiation. (H) Chromatin extracts from unirradiated and UV-irradiated (30 J/m2) shLuc- and shBRCA1-transduced U2OS cells were immunoprecipitated with Rad9- and Rad1-specific antibody. The western blots were immunoblotted for BRCA1, Rad9, and RFC4. (I) Rad9-GFP-transfected shLuc- and shBRCA1-infected cells were irradiated (30 J/m2), fixed 3 hr later, immunostained for CPD, and examined directly for Rad9-GFP. (J) The fraction of CPD-positive cells that also were positive for Rad9-GFP from (I) were plotted. Error bars represent standard deviation between three separate experiments. (K) Chromatin extracts from cells transfected with an HA-tagged Rad9-encoding vector or the vector backbone plasmid were immunoprecipitated with HA antibody. The western blot was immunoblotted for BRCA1, Rad9, and RFC4.

Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. 247 Molecular Cell BRCA1 Is Required for Postreplicative Repair

A B C + PolH-eGFP DAPI eGFP-PolH 45 40 shGFP shBRCA1 35

BRCA1 ShLuc 30 25 20 GFP 15 (PolH) 10

% of cells with PolH at UV damage sites 5 0 GAPDH shLuc shBRCA1 ShBRCA1

D E IP: HA WCE 200 )

shGFP shBRCA1 shGFP shBRCA1 -4 180 UV - + + - + + - + + - + + 160 140 Time (hr) - 3 5 - 3 5 - 3 5 - 3 5 120 BRCA1 100 ShLuc ShBRCA1 80 60 HA-Ub-PCNA 40 20

Ub-PCNA Mutation Frequency (x10 0 PCNA -Damage +Damage

F

In Replicating Cells UV induced DNA lesion

PCNA

Stalled Replication Fork

RPA MCM2-7

BRCA1 + RPA BRCA1 BRCA1

RFC based complex ERCC1/XPF

9 + CtIP 1 1 ? Translesion Ub polymerase PCNA ATRIP PCNA PCNA

RPA

MCM2-7 MCM2-7 MCM2-7

ssDNA/RPA/ATRIP accumulation Damage Excision Suppression of translesion synthesis Activation of G2/M and S phase checkpoints Post-replicative gap repair

248 Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. Molecular Cell BRCA1 Is Required for Postreplicative Repair

haploinsufficient for the response to bulky adduct damage, previously (Scully et al., 1999). The primary NER-deficient fibroblast cell which is not yet known, certain clinical possibilities would lines GM04312 (XPA) and GM16093 (XPC) were purchased from NIGMS come to mind. BRCA1 breast cancer development is believed Human Genetic Cell Repository (Coriell Institute for Medical Research, Camden, NJ). They were grown in medium recommended by the purveyor. to be an estrogen-dependent process, even though most 2 All cells were grown at 37 C in a humidified incubator containing 5% CO . BRCA1 tumors are estrogen receptor (ER)-negative (Eisen Where indicated, cells were cultured in presence of 1 mM HU (Sigma), 2 mM et al., 2005; Narod and Offit, 2005). If ER function is not a 4NQO (Sigma), BrdU (BD PharMingen, 50 mM for BrdU ssDNA assay and key factor in the mechanism underlying the estrogen depen- 10 mM for intra-S phase checkpoint assay), 2 mg/ml aphidicolin (Calbiochem), dency of BRCA1 breast carcinogenesis, one wonders whether and 200 mM mimosine (Sigma). abnormal manifestations of estrogen metabolism, an ER-inde- pendent process, play a role instead. Most estradiol is normally SUPPLEMENTAL INFORMATION converted by catechol-o-methyltransferase to biologically inac- Supplemental Information includes six figures and Supplemental Experimental tive metabolites (Yager and Davidson, 2006). However, there is Procedures and can be found with this article online at doi:10.1016/j.molcel. another pathway that directs estradiol conversion to certain 2011.09.002. quinone derivatives capable of generating bulky adduct DNA damage (Yager and Davidson, 2006). Some of the most power- ACKNOWLEDGMENTS ful adducting estradiol metabolites are also mammary carcino- gens in rodents (Cavalieri et al., 2000). These observations raise We would like to thank Steve Elledge for helpful discussions and for ATRIP anti- body, Lucian Moldovan and Alan D’Andrea for help and advice in setting up the the question of whether BRCA1 heterozygous cells are haploin- supF assay, Alan Lehmann for the eGFP-PolH plasmid, Veronique Smits for the sufficient for BRCA1 participation in the response to bulky Rad9-GFP-HA plasmid, Zhongsheng You for the eGFP-CtIP plasmid, Richard adduct damage. If so, one might wonder whether they respond Baer for CtIP antibody, Margaret Shipp for the HA-tagged ubiquitin plasmid, to bulky adduct-generating compounds, such as the aforemen- and the RNAi Consortium at the Broad Institute for shRNAs directed at tioned estradiol derivatives, by acquiring permanent forms of BRCA1 mRNA. We also thank Daniel Silver, Hua Wang, and multiple members DNA damage arising from an inadequate response to replica- of the Livingston laboratory for invaluable discussions. This work was sup- ported in part by grants to one of us (D.M.L.) from the National Cancer Institute, tion stress. Were this the case, perhaps, replicating cells in by a SPORE grant in breast cancer research to the Dana-Farber/Harvard BRCA1 heterozygous females that accumulate estradiol and Cancer Center, and by grants from the Breast Cancer Research Foundation certain of its DNA damage-generating metabolites (e.g., mam- and the Susan B. Komen Foundation for the Cure. Generous support for this mary epithelial cells during pubertal development, pregnancy, work was also provided by the Dana-Farber Cancer Institute and the National and during intervals between menstruation in premenopausal Institutes of Health, NHGRI (P50HG004233), and NINDS (P01NS047572). women) are at special risk of accumulating the kinds of chronic bulky adduct damage that significantly elevates the risk of Received: December 23, 2010 Revised: May 6, 2011 carcinogenesis. This question is also driven by the knowledge Accepted: July 25, 2011 that chronic replication stress is a clinically significant driver Published online: September 29, 2011 of carcinogenesis (Bartkova et al., 2005; Gorgoulis et al., 2005; Halazonetis et al., 2008; Negrini et al., 2010). REFERENCES

Al-Minawi, A.Z., Lee, Y.-F., Ha˚ kansson, D., Johansson, F., Lundin, C., Saleh- EXPERIMENTAL PROCEDURES Gohari, N., Schultz, N., Jenssen, D., Bryant, H.E., Meuth, M., et al. (2009). The ERCC1/XPF endonuclease is required for completion of homologous recom- UV Irradiation bination at DNA replication forks stalled by inter-strand cross-links. Nucleic Cells were irradiated with a 254 nm UV-C lamp (UVP Inc., Upland, CA), and Acids Res. 37, 6400–6413. doses were measured with a UVX radiometer (Upland). UV irradiation through 3 mm isopore/micropore polycarbonate filters (Millipore) was undertaken as Alli, E., Sharma, V., Sunderesakumar, P., and Ford, J.M. (2009). Defective described (Polo et al., 2006) and at 30 J/m2. Filter-free UV irradiation was Repair of Oxidative DNA Damage in Triple-Negative Breast Cancer Confers 69 performed at 5 J/m2 to analyze phospho-RPA32 and g-H2AX staining. For Sensitivity to Inhibition of Poly(ADP-Ribose) Polymerase. Cancer Res. , ssDNA/BrdU assay, cells were irradiated with 20 J/m2 UV. 3589–3596. Au, W.W.Y., and Henderson, B.R. (2005). The BRCA1 RING and BRCT Cell Lines and Culture Conditions domains cooperate in targeting BRCA1 to -induced nuclear 293T, HeLa, and U2OS cells were cultivated in Dulbecco’s modified Eagle’s foci. J. Biol. Chem. 280, 6993–7001. medium (DMEM) supplemented with 10% FBS and pen/strep. HCC1937 cells Bartkova, J., Horejsı´, Z., Koed, K., Kra¨ mer, A., Tort, F., Zieger, K., Guldberg, P., reconstituted with WT BRCA1 or a vector control were cultivated as described Sehested, M., Nesland, J.M., Lukas, C., et al. (2005). DNA damage response

Figure 7. BRCA1 Plays a Role in Suppressing Translesional Synthesis (A) Western blot analysis of the expression of GFP-tagged PolH in shLuc- and shBRCA1-infected U2OS cells. (B) Localization of eGFP-PolH in virus-transduced U2OS cells irradiated with UV (30 J/m2), fixed 4 hr later, and immunostained with GFP Ab. (C) GFP-positive cells with eGFP-PolH in the local UV-irradiated spots were counted. Percentages plotted are the mean of three independent experiments. (D) HA-tagged ubiquitin was expressed in U2OS cells transduced with shGFP- and shBRCA1-encoding lentiviruses. Cells were irradiated (30 J/m2) and harvested at the indicated time points. The whole-cell extract (WCE) and the anti-HA immunoprecipitates (IP:HA) were electrophoresed and blotted. The blot was reacted with BRCA1 and PCNA antibodies. (E) SupF assay. Unirradiated and UV-irradiated (500 J/m2) SupF plasmid was transfected into shLuc- and shBRCA1-transduced 293T cells. Details of this assay are described in Supplemental Information. Data analyzed by counting blue (wild-type) and white (mutant) colonies from four separate experiments are plotted. (F) Model: BRCA1 function after UV-induced DNA damage.

Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. 249 Molecular Cell BRCA1 Is Required for Postreplicative Repair

as a candidate anti-cancer barrier in early human tumorigenesis. Nature 434, by generating double-strand DNA breaks. Nat. Struct. Mol. Biol. 14, 1096– 864–870. 1104. Bermudez, V.P., Lindsey-Boltz, L.A., Cesare, A.J., Maniwa, Y., Griffith, J.D., Hartman, A.-R., and Ford, J.M. (2002). BRCA1 induces DNA damage recogni- Hurwitz, J., and Sancar, A. (2003). Loading of the human 9-1-1 checkpoint tion factors and enhances nucleotide excision repair. Nat. Genet. 32, 180–184. complex onto DNA by the checkpoint clamp loader hRad17-replication factor Hu, Y., Scully, R., Sobhian, B., Xie, A., Shestakova, E., and Livingston, D.M. 100 C complex in vitro. Proc. Natl. Acad. Sci. USA , 1633–1638. (2011). RAP80-directed tuning of BRCA1 function Bulavin, D.V., Higashimoto, Y., Popoff, I.J., Gaarde, W.A., Basrur, V., at ionizing radiation-induced nuclear foci. Dev. 25, 685–700. Potapova, O., Appella, E., and Fornace, A.J., Jr. (2001). Initiation of a G2/M Huang, J.C., Svoboda, D.L., Reardon, J.T., and Sancar, A. (1992). Human checkpoint after radiation requires p38 kinase. Nature 411, nucleotide excision nuclease removes thymine dimers from DNA by incising 102–107. the 22nd phosphodiester bond 50 and the 6th phosphodiester bond 30 to the Byun, T.S., Pacek, M., Yee, M.-C., Walter, J.C., and Cimprich, K.A. (2005). photodimer. Proc. Natl. Acad. Sci. USA 89, 3664–3668. Functional uncoupling of MCM helicase and DNA polymerase activities Kannouche, P.L., Broughton, B.C., Volker, M., Hanaoka, F., Mullenders, activates the ATR-dependent checkpoint. Genes Dev. 19, 1040–1052. L.H.F., and Lehmann, A.R. (2001). Domain structure, localization, and function Cai, L., Roginskaya, M., Qu, Y., Yang, Z., Xu, Y., and Zou, Y. (2007). Structural of DNA polymerase eta, defective in variant cells. characterization of human RPA sequential binding to single-stranded DNA Genes Dev. 15, 158–172. using ssDNA as a molecular ruler. Biochemistry 46, 8226–8233. Kannouche, P.L., Wing, J.F., and Lehmann, A.R. (2004). Interaction of human Callegari, A.J., and Kelly, T.J. (2007). Shedding light on the DNA damage DNA polymerase eta with monoubiquitinated PCNA: a possible mechanism for checkpoint. Cell Cycle 6, 660–666. the polymerase switch in response to DNA damage. Mol. Cell 14, 491–500. Carty, M.P., Zernik-Kobak, M., McGrath, S., and Dixon, K. (1994). UV light- Kim, J.K., Patel, D., and Choi, B.S. (1995). Contrasting structural impacts induced DNA synthesis arrest in HeLa cells is associated with changes in induced by cis-syn cyclobutane dimer and (6-4) adduct in DNA duplex deca- phosphorylation of human single-stranded DNA-binding protein. EMBO J. mers: implication in mutagenesis and repair activity. Photochem. Photobiol. 62 13, 2114–2123. , 44–50. Cavalieri, E., Frenkel, K., Liehr, J.G., Rogan, E., and Roy, D. (2000). Estrogens Kuraoka, I., Kobertz, W.R., Ariza, R.R., Biggerstaff, M., Essigmann, J.M., and as endogenous genotoxic agents—DNA adducts and . J. Natl. Wood, R.D. (2000). Repair of an interstrand DNA cross-link initiated by 275 Cancer Inst. Monogr. Chapter 4, 75–93. ERCC1-XPF repair/recombination nuclease. J. Biol. Chem. , 26632– 26636. Chen, J.J., Silver, D., Cantor, S.B., Livingston, D.M., and Scully, R. (1999). BRCA1, BRCA2, and Rad51 operate in a common DNA damage response Lehmann, A.R. (2005). Replication of damaged DNA by translesion synthesis in 579 pathway. Cancer Res. 59 (7, Suppl), 1752s–1756s. human cells. FEBS Lett. , 873–876. Choi, J.-H., and Pfeifer, G.P. (2005). The role of DNA polymerase eta in UV Liu, P., Barkley, L.R., Day, T., Bi, X., Slater, D.M., Alexandrow, M.G., Nasheuer, mutational spectra. DNA Repair (Amst.) 4, 211–220. H.-P., and Vaziri, C. (2006). The Chk1-mediated S-phase checkpoint targets initiation factor Cdc45 via a Cdc25A/Cdk2-independent mechanism. J. Biol. Cleaver, J.E. (2005). Cancer in xeroderma pigmentosum and related disorders Chem. 281, 30631–30644. of DNA repair. Nat. Rev. Cancer 5, 564–573. Lopes, M., Foiani, M., and Sogo, J.M. (2006). Multiple mechanisms control Cortez, D. (2005). Unwind and slow down: checkpoint activation by helicase integrity after replication fork uncoupling and restart at irrepa- 19 and polymerase uncoupling. Genes Dev. , 1007–1012. rable UV lesions. Mol. Cell 21, 15–27. Cox, M.M., Goodman, M.F., Kreuzer, K.N., Sherratt, D.J., Sandler, S.J., and Luijsterburg, M.S., von Bornstaedt, G., Gourdin, A.M., Politi, A.Z., Mone´ , M.J., Marians, K.J. (2000). The importance of repairing stalled replication forks. Warmerdam, D.O., Goedhart, J., Vermeulen, W., van Driel, R., and Ho¨ fer, T. 404 Nature , 37–41. (2010). Stochastic and reversible assembly of a multiprotein DNA repair de Laat, W.L., Jaspers, N.G.J., and Hoeijmakers, J.H.J. (1999). Molecular complex ensures accurate target site recognition and efficient repair. J. Cell mechanism of nucleotide excision repair. Genes Dev. 13, 768–785. Biol. 189, 445–463. Eisen, A., Lubinski, J., Klijn, J., Moller, P., Lynch, H.T., Offit, K., Weber, B., Medhurst, A.L., Warmerdam, D.O., Akerman, I., Verwayen, E.H., Kanaar, R., Rebbeck, T., Neuhausen, S.L., Ghadirian, P., et al. (2005). Breast cancer risk Smits, V.A.J., and Lakin, N.D. (2008). ATR and Rad17 collaborate in modu- following bilateral oophorectomy in BRCA1 and BRCA2 mutation carriers: an lating Rad9 localisation at sites of DNA damage. J. Cell Sci. 121, 3933–3940. international case-control study. J. Clin. Oncol. 23, 7491–7496. Michel, B., Grompone, G., Flore` s, M.-J., and Bidnenko, V. (2004). Multiple Friedberg, E.C. (2005). Suffering in silence: the tolerance of DNA damage. Nat. pathways process stalled replication forks. Proc. Natl. Acad. Sci. USA 101, Rev. Mol. Cell Biol. 6, 943–953. 12783–12788. Ge´ rard, A., Polo, S.E., Roche, D.M.J., and Almouzni, G. (2006). Methods for Miki, Y., Swensen, J., Shattuck-Eidens, D., Futreal, P.A., Harshman, K., studying chromatin assembly coupled to DNA repair. Methods Enzymol. Tavtigian, S.V., Liu, Q., Cochran, C., Bennett, L.M., Ding, W., et al. (1994). A 409, 358–374. strong candidate for the breast and ovarian cancer susceptibility BRCA1. Science 266, 66–71. Gong, Z., Kim, J.-E., Leung, C.C.Y., Glover, J.N.M., and Chen, J. (2010). BACH1/FANCJ acts with TopBP1 and participates early in DNA replication Mocquet, V., Laine´ , J.P., Riedl, T., Yajin, Z., Lee, M.Y., and Egly, J.M. (2008). checkpoint control. Mol. Cell 37, 438–446. Sequential recruitment of the repair factors during NER: the role of XPG in initi- ating the resynthesis step. EMBO J. 27, 155–167. Gorgoulis, V.G., Vassiliou, L.-V.F., Karakaidos, P., Zacharatos, P., Kotsinas, A., Liloglou, T., Venere, M., Ditullio, R.A., Jr., Kastrinakis, N.G., Levy, B., Moldovan, G.-L., Pfander, B., and Jentsch, S. (2007). PCNA, the maestro of the et al. (2005). Activation of the DNA damage checkpoint and genomic instability replication fork. Cell 129, 665–679. in human precancerous lesions. Nature 434, 907–913. Moynahan, M.E., Chiu, J.W., Koller, B.H., and Jasin, M. (1999). Brca1 controls 4 Halazonetis, T.D., Gorgoulis, V.G., and Bartek, J. (2008). An oncogene- homology-directed DNA repair. Mol. Cell , 511–518. induced DNA damage model for cancer development. Science 319, 1352– Nagaraju, G., and Scully, R. (2007). Minding the gap: the underground func- 1355. tions of BRCA1 and BRCA2 at stalled replication forks. DNA Repair (Amst.) 6 Hanada, K., Budzowska, M., Davies, S.L., van Drunen, E., Onizawa, H., , 1018–1031. Beverloo, H.B., Maas, A., Essers, J., Hickson, I.D., and Kanaar, R. (2007). Narod, S.A., and Foulkes, W.D. (2004). BRCA1 and BRCA2: 1994 and beyond. The structure-specific endonuclease Mus81 contributes to replication restart Nat. Rev. Cancer 4, 665–676.

250 Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. Molecular Cell BRCA1 Is Required for Postreplicative Repair

Narod, S.A., and Offit, K. (2005). Prevention and management of hereditary Scully, R., Ganesan, S., Vlasakova, K., Chen, J., Socolovsky, M., and breast cancer. J. Clin. Oncol. 23, 1656–1663. Livingston, D.M. (1999). Genetic analysis of BRCA1 function in a defined tumor 4 Negrini, S., Gorgoulis, V.G., and Halazonetis, T.D. (2010). Genomic insta- cell line. Mol. Cell , 1093–1099. bility—an evolving hallmark of cancer. Nat. Rev. Mol. Cell Biol. 11, 220–228. Shen, X., Do, H., Li, Y., Chung, W.-H., Tomasz, M., de Winter, J.P., Xia, B., Elledge, S.J., Wang, W., and Li, L. (2009). Recruitment of fanconi anemia Ogi, T., Limsirichaikul, S., Overmeer, R.M., Volker, M., Takenaka, K., Cloney, and breast cancer proteins to DNA damage sites is differentially governed R., Nakazawa, Y., Niimi, A., Miki, Y., Jaspers, N.G., et al. (2010). Three DNA by replication. Mol. Cell 35, 716–723. polymerases, recruited by different mechanisms, carry out NER repair synthesis in human cells. Mol. Cell 37, 714–727. Sogo, J.M., Lopes, M., and Foiani, M. (2002). Fork reversal and ssDNA accu- mulation at stalled replication forks owing to checkpoint defects. Science 297, Parrilla-Castellar, E.R., Arlander, S.J.H., and Karnitz, L. (2004). Dial 9-1-1 for 599–602. DNA damage: the Rad9-Hus1-Rad1 (9-1-1) clamp complex. DNA Repair (Amst.) 3, 1009–1014. Tassone, P., Tagliaferri, P., Perricelli, A., Blotta, S., Quaresima, B., Martelli, M.L., Goel, A., Barbieri, V., Costanzo, F., Boland, C.R., and Venuta, S. Paull, T.T., Cortez, D., Bowers, B., Elledge, S.J., and Gellert, M. (2001). Direct (2003). BRCA1 expression modulates chemosensitivity of BRCA1-defective DNA binding by Brca1. Proc. Natl. Acad. Sci. USA 98, 6086–6091. HCC1937 human breast cancer cells. Br. J. Cancer 88, 1285–1291. Petermann, E., and Helleday, T. (2010). Pathways of mammalian replication Tsao, C.-C., Geisen, C., and Abraham, R.T. (2004). Interaction between human fork restart. Nat. Rev. Mol. Cell Biol. 11, 683–687. MCM7 and Rad17 proteins is required for replication checkpoint signaling. Petermann, E., Orta, M.L., Issaeva, N., Schultz, N., and Helleday, T. (2010). EMBO J. 23, 4660–4669. Hydroxyurea-stalled replication forks become progressively inactivated and Wood, R.D., and Shivji, M.K. (1997). Which DNA polymerases are used for require two different RAD51-mediated pathways for restart and repair. Mol. DNA-repair in ? Carcinogenesis 18, 605–610. Cell 37, 492–502. Xie, J., Litman, R., Wang, S., Peng, M., Guillemette, S., Rooney, T., and Cantor, Polo, S.E., Roche, D.M.J., and Almouzni, G. (2006). New histone incorporation S.B. (2010). Targeting the FANCJ-BRCA1 interaction promotes a switch from 127 marks sites of UV repair in human cells. Cell , 481–493. recombination to poleta-dependent bypass. Oncogene 29, 2499–2508. Riedl, T., Hanaoka, F., and Egly, J.M. (2003). The comings and goings of nucle- Xu, B., Kim, S.-t., and Kastan, M.B. (2001). Involvement of Brca1 in S-phase 22 otide excision repair factors on damaged DNA. EMBO J. , 5293–5303. and G(2)-phase checkpoints after ionizing irradiation. Mol. Cell. Biol. 21, Rubbi, C.P., and Milner, J. (2001). Analysis of nucleotide excision repair by 3445–3450. detection of single-stranded DNA transients. Carcinogenesis 22, 1789–1796. Yager, J.D., and Davidson, N.E. (2006). Estrogen carcinogenesis in breast Sancar, A., Lindsey-Boltz, L.A., Unsal-Kac¸maz, K., and Linn, S. (2004). cancer. N. Engl. J. Med. 354, 270–282. Molecular mechanisms of mammalian DNA repair and the DNA damage Zou, L., and Elledge, S.J. (2003). Sensing DNA damage through ATRIP recog- checkpoints. Annu. Rev. Biochem. 73, 39–85. nition of RPA-ssDNA complexes. Science 300, 1542–1548. Scully, R., Chen, J., Ochs, R.L., Keegan, K.S., Hoekstra, M.F., Feunteun, J., Zou, L., Cortez, D., and Elledge, S.J. (2002). Regulation of ATR substrate and Livingston, D.M. (1997a). Dynamic changes of BRCA1 subnuclear location selection by Rad17-dependent loading of Rad9 complexes onto chromatin. and phosphorylation state are initiated by DNA damage. Cell 90, 425–435. Genes Dev. 16, 198–208. Scully, R., Chen, J., Plug, A., Xiao, Y., Weaver, D., Feunteun, J., Ashley, T., and Zou, L., Liu, D., and Elledge, S.J. (2003). -mediated Livingston, D.M. (1997b). Association of BRCA1 with Rad51 in mitotic and recruitment and activation of Rad17 complexes. Proc. Natl. Acad. Sci. USA meiotic cells. Cell 88, 265–275. 100, 13827–13832.

Molecular Cell 44, 235–251, October 21, 2011 ª2011 Elsevier Inc. 251