Report

Ubiquitin-Binding Protein RAP80 Mediates BRCA1-Dependent DNA Damage Response

See allHide authors and affiliations

Science  25 May 2007:
Vol. 316, Issue 5828, pp. 1202-1205
DOI: 10.1126/science.1139621

Abstract

Mutations in the breast cancer susceptibility gene 1 (BRCA1) are associated with an increased risk of breast and ovarian cancers. BRCA1 participates in the cellular DNA damage response. We report the identification of receptor-associated protein 80 (RAP80) as a BRCA1-interacting protein in humans. RAP80 contains a tandem ubiquitin-interacting motif domain, which is required for its binding with ubiquitin in vitro and its damage-induced foci formation in vivo. Moreover, RAP80 specifically recruits BRCA1 to DNA damage sites and functions with BRCA1 in G2/M checkpoint control. Together, these results suggest the existence of a ubiquitination-dependent signaling pathway involved in the DNA damage response.

Despite developing various DNA lesions generated during DNA replication or after exposure to environmental agents, cells normally maintain their genomic integrity and prevent neoplastic transformation because of the existence of several cell cycle checkpoints and DNA repair systems (13). Many proteins [including the protein kinase ataxia-telangiectasia mutated (ATM), γ-H2AX, mediator of DNA damage checkpoint protein 1 (MDC1), Nijmegen breakage syndrome 1 (NBS1), BRCA1, and checkpoint kinases 1 and 2 (Chk1 and Chk2)] are involved in the ionizing radiation (IR)–induced DNA damage response pathway (4). ATM is recruited to and activated at the sites of DNA breaks. Activated ATM transduces DNA damage signals to downstream proteins, including BRCA1. BRCA1 encodes a tumor suppressor gene that is mutated in ∼50% of hereditary breast and ovarian cancer patients (5, 6). The human BRCA1 protein contains an N-terminal RING finger domain that has intrinsic E3 ubiquitin ligase activity and tandem BRCA1 C-terminal (BRCT) domains at its C terminus, which are phosphoserine- or phosphothreonine-binding motifs (79). Many disease-causing mutations are detected within these two regions of BRCA1.

Although BRCA1 is recruited to the sites of DNA breaks and participates in cell cycle checkpoint control, it remains obscure how the recruitment of BRCA1 is controlled in the cell. We purified BRCA1-BRCT domains from human leukemia K562 cells stably expressing this protein with N-terminal S-tag, Flag epitope, and streptavidin-binding peptide (SFB) triple tags (SFB-BRCA1-BRCT). We detected three specific bands that eluted with the SFB-BRCA1-BRCT domain but not with the SFB-BARD1-BRCT domain (Fig. 1A), where BARD1 signifies the BRCA1-associated RING domain protein 1. Mass spectrometry analysis revealed that these three proteins (respectively) are BRCA1-associated C-terminal helicase (BACH1), C-terminal binding protein–interacting protein (CtIP), and RAP80. BACH1 and CtIP are two known BRCA1 BRCT domain-binding proteins (9, 10). RAP80 was originally identified as a retinoid-related testis-associated protein (11). The physiological function of RAP80 is unknown. We first confirmed the association between RAP80 and BRCA1 both in vitro and in vivo (Fig. 1B and fig. S1) (12). The interaction between BRCA1 and RAP80 remained the same before or after DNA damage (Fig. 1C).

Fig. 1.

Identification of RAP80 as a BRCA1-binding protein. (A) Silver staining of affinity-purified BRCA1-BRCT complexes. The cell extracts prepared from K562 cells stably expressing SFB-BRCA1-BRCT or SFB-BARD1-BRCT were subjected to two rounds of affinity purification. Final elutes were analyzed by SDS-PAGE and visualized by silver staining. The specific bands were excised from the silver-stained gel, and the peptides were identified by matrix-assisted laser desorption/ionization–time-of-flight mass spectrometry. Lines indicate protein bands corresponding to BACH1, CtIP, and RAP80. (B) The interaction between endogenous BRCA1 and RAP80. We performed immunoprecipitation (IP) reactions using preimmune serum or antibody to BRCA1. The immunoprecipitates were subjected to immunoblotting analyses with antibodies to BRCA1 or RAP80. (C) The interaction between BRCA1 and RAP80 before and after exposure of cells to IR. Lysates prepared from mock-treated or irradiated 293T cells were immunoprecipitated with antibody to BRCA1. The immunoprecipitates were separated by SDS-PAGE and immunoblotted with the indicated antibodies (top two lanes). The amount of endogenous RAP80 in cells before and after radiation was shown in the bottom lane.

BRCA1 relocalizes to sites of DNA breaks in cells exposed to IR. Immunostaining showed RAP80 to be evenly distributed in the nucleoplasm in normal cells (Fig. 2A). However, after exposure of cells to IR, RAP80 relocalized to foci that colocalized with γ-H2AX and BRCA1 (Fig. 2, A and B). RAP80 also associated with chromatin only in cells exposed to IR (Fig. 2C). Together, these data indicate that the localization of RAP80, like that of BRCA1, is regulated in response to DNA damage.

Fig. 2.

Localization and phosphorylation of RAP80 in cells exposed to IR. DNA damage–induced RAP80 foci formation and its colocalization with γ-H2AX (A) and BRCA1 (B) are shown. DAPI, 4′,6′-diamidino-2-phenylindole. Mock-treated or irradiated 293T cells were fixed and stained with monoclonal antibodies to γ-H2AX or BRCA1 and polyclonal antibody to RAP80. (C) Association of RAP80 with chromatin after DNA damage. The soluble and chromatin fractions were prepared from mock-treated or irradiated 293T cells and subjected to Western blot analysis with antibodies to RAP80 (top lanes), phosphorylated H2AX (p-H2AX) (middle lanes), and H2AX (bottom lanes). (D) Phosphorylation of RAP80 after DNA damage. Lysates prepared from control or irradiated HeLa cells were immunoprecipitated with antibody to RAP80 and then incubated with or without λ phosphatase for 1 hour at 30°C. λPPase, λ protein phosphatase. The samples were subjected to immunoblotting with antibody to RAP80. (E) Requirement of ATM for IR-induced phosphorylation of RAP80. ATM-deficient FT169A cells and cells reconstituted with WT ATM (YZ5) were exposed to IR. Immunoprecipitation and immunoblotting were performed as described in (D). (F and G) Dependence of DNA damage–induced RAP80 foci formation. MDC1+/+ and MDC1–/– mouse embryo fibroblasts (MEFs) (F) and H2AX+/+ and H2AX–/– MEFs (G) were exposed to IR. The immunostaining experiments were performed as described in (A). (H) Requirement of RAP80 for damage-induced BRCA1 foci formation. Control (con) or RAP80 siRNA-transfected 293T cells were exposed to IR. Immunostaining was conducted with monoclonal antibodies to BRCA1, MDC1, 53BP1, or γ-H2AX and polyclonal antibody to RAP80.

RAP80 isolated from irradiated cells migrated more slowly during SDS–polyacrylamide gel electrophoresis (SDS-PAGE) than did RAP80 isolated from unirradiated cells. Moreover, phosphatase treatment reversed the slow mobility of RAP80 prepared from irradiated cells (Fig. 2D), indicating that RAP80 may be phosphorylated in cells exposed to IR. We confirmed this using a phosphospecific antibody raised against a phosphorylation site that we identified (Ser101; fig. S2). The ATM protein kinase is activated in response to DNA damage and phosphorylates many proteins involved in the DNA damage response. Treatment of cells with two different ATM kinase inhibitors, wortmannin and caffeine, abolished the IR-induced mobility shift of RAP80 (fig. S3A). The mobility shift of RAP80 was only observed in cells expressing wild-type (WT) ATM but not in ATM-deficient cells (Fig. 2E). These data suggest that ATM is required for damage-induced phosphorylation of RAP80.

The accumulation of RAP80 to the sites of DNA breaks depended on MDC1 and H2AX (Fig. 2, F and G) but not on NBS1, p53 binding protein 1 (53BP1), or BRCA1 (fig. S3, B to D). When we reduced endogenous RAP80 expression using RAP80 small interfering RNAs (siRNAs), we still detected damage-induced foci formation of MDC1, γ-H2AX, and 53BP1. However, no BRCA1 foci were present in these RAP80-depleted cells (Fig. 2H), suggesting that RAP80 acts upstream of BRCA1 and is required for the accumulation of BRCA1 to sites of DNA breaks.

We also determined which regions of RAP80 are important for its focus localization. Full-length and several internal deletion mutants of RAP80 localized to nuclear foci in cells with DNA damage, whereas RAP80D1 and RAP80D2 did not (Fig. 3A and fig. S4A). Because RAP80D1 and RAP80D2 are the only two internal deletion mutants that lack the two putative ubiquitininteracting motifs (UIMs) (13), these results imply that the region containing UIMs may be required for RAP80 localization to DNA damage foci. The putative UIMs in RAP80 largely match with the UIM consensus sequence (fig. S4B). To determine whether the tandem RAP80 UIMs indeed bind to ubiquitin, we used a ubiquitin–glutathione S-transferase fusion protein (Ubi-GST). Ubi-GST specifically bound to the WT RAP80 but not to RAP80 lacking the two putative UIMs (RAP80D1; Fig. 3B). We also tested the binding of WT or mutant RAP80 UIMs [mutation of Ala88→Gly88 (A88G) (14) and S92A in the first UIM and A113G and S117A in the second UIM] with Ubi-GST. The Ubi-GST specifically interacted with RAP80 UIM but not with the UIMs containing point mutations (Fig. 3B). We further checked whether point mutants within RAP80 UIMs would disrupt RAP80 foci formation in vivo. WT RAP80 and the RAP80P4 mutant (mutation of the linker region between two UIMs) formed detectable damage-induced nuclear foci, whereas the RAP80P1, RAP80P2, and RAP80P3 point mutants did not (Fig. 3C and fig. S4A). RAP80P1, RAP80P2, and RAP80P3 contain mutations within the first UIM (A88G and S92A), the second UIM (A113G and S117A), or both UIMs (A88G, S92A, A113G, and S117A), respectively. Therefore, the ubiquitin-binding activity of RAP80 correlates with its ability to localize to damage-induced foci in vivo. Like RAP80, the Homo sapiens DnaJ1A (HSJ1A) protein localizes to nuclei and also contains two UIMs. However, full-length HSJ1A or a construct containing the HSJ1A UIM region did not form nuclear foci in cells with DNA damage (fig. S4C). Thus, the ability to form nuclear foci is specific for the RAP80 UIM region. Notably, RAP80 UIMs bind specifically to Lys63-linked but not to Lys48-linked polyubiquitin chains in vitro (fig. S5).

Fig. 3.

Focus localization of RAP80 depends on its UIMs. (A and C) HeLa cells were transfected with SFB-tagged wild type and internal serial deletion mutants (A) or several point mutants (C) of RAP80. Cells were exposed to IR 24 hours later. Immunostaining experiments were conducted with monoclonal antibody to Flag and polyclonal antibody to γ-H2AX. (B) Direct binding of RAP80 UIMs to ubiquitin in vitro. GST or Ubi-GST protein was incubated with cell lysates containing exogenously expressed Flag-tagged WT RAP80, RAP80D1, RAP80 UIMs, or RAP80 UIMDMs (double mutations in the UIMs). After extensive washing, the bound RAP80 proteins were analyzed by immunoblotting with antibody to Flag.

Cells carrying BRCA1 mutants display increased sensitivity to IR and defective G2/M checkpoint control (15). We examined whether the loss of the RAP80 would result in similar defects in the DNA damage response. Both RAP80 siRNAs that we synthesized efficiently decreased RAP80 expression in cells (Fig. 4A). Using a previously established G2/M checkpoint assay (16), we showed defective G2/M checkpoint control in RAP80-depleted cells (Fig. 4B). Similar G2/M checkpoint defects were also observed in BRCA1- or CtIP-depleted cells (fig. S6). The protein kinase Chk1 is required for the G2/M checkpoint control (17, 18) and acts downstream of BRCA1 in response to IR (19, 20). If RAP80 functions upstream of BRCA1, we would expect a defective Chk1 activation in RAP80-depleted cells. This is indeed the case (Fig. 4C). RAP80-depleted cells were also more sensitive to radiation than control cells (Fig. 4D). These data together indicate that RAP80 acts upstream of BRCA1 and specifically regulates BRCA1 functions after DNA damage.

Fig. 4.

Requirement of RAP80 for the IR-induced G2/M checkpoint. (A) Western blot analysis for RAP80 expression level. RAP80 protein levels were analyzed by immunoblotting with antibodies to RAP80 with the use of control or RAP80 siRNA-transfected cell lysates. (B) G2/M checkpoint in RAP80-depleted cells. HeLa cells transfected with control or RAP80 siRNAs were exposed to 0 or 4 grays (Gy) of IR. Cells were incubated for 1 hour before fixation and subjected to staining with antibody to phosphorylated histone H3 (pH3) and propidium iodide. The percentages of mitotic cells were determined by fluorescence-activated cell sorting analysis. The boxed area in the top left panel indicates mitotic cells. (C) Requirement of RAP80 for Chk1 phosphorylation after DNA damage. Control or RAP80 siRNA-transfected HeLa cells were exposed to IR. Cells were harvested 2 hours later, and lysates were immunoblotted with indicated antibodies. (D) Radiation sensitivity of cells lacking RAP80. HeLa cells were transfected with control or RAP80 siRNAs. Cells were counted and irradiated with various doses of IR. Percentages of surviving colonies were determined 11 days later. These experiments were performed in triplicate, and the results represent the average of two or three independent experiments. Error bars indicate SD for different doses of irradiation.

Exactly how RAP80 is recruited to DNA damage sites is still unknown. Because RAP80 UIMs bind directly to ubiquitin in vitro, we reason that one or several ubiquitinated proteins might bind RAP80 and recruit RAP80 to the DNA damage sites. There are several proteins known to be ubiquitinated and localized to the sites of DNA damage (2123). One of them is Fanconi anemia complementation group D2 (FANCD2). However, RAP80 foci still form normally after irradiation in FANCD2-deficient cells (fig. S7), suggesting that there may be other as-yet-unidentified ubiquitinated proteins that act early in DNA damage response and regulate RAP80 localization.

Many cell cycle checkpoint proteins, including ATM, Chk2, BRCA1, and p53, play critical roles in the maintenance of genomic stability. Their mutation often results in increased tumor incidence, highlighting the importance of the integrity of DNA damage pathways in tumor suppression. As a BRCA1-associated protein involved in DNA damage checkpoint control, RAP80 may also function as a tumor suppressor and be dysregulated or mutated in human patients. Future genetic studies will allow us to test this possibility.

Supporting Online Material

www.sciencemag.org/cgi/content/full/316/5828/1202/DC1

Materials and Methods

Figs. S1 to S9

References

References and Notes

View Abstract

Navigate This Article