Research ArticlesDNA damage

RNF4-Dependent Hybrid SUMO-Ubiquitin Chains Are Signals for RAP80 and Thereby Mediate the Recruitment of BRCA1 to Sites of DNA Damage

See allHide authors and affiliations

Sci. Signal.  04 Dec 2012:
Vol. 5, Issue 253, pp. ra88
DOI: 10.1126/scisignal.2003485

Abstract

The DNA repair function of the breast cancer susceptibility protein BRCA1 depends in part on its interaction with RAP80, which targets BRCA1 to DNA double-strand breaks (DSBs) through recognition of K63-linked polyubiquitin chains. The localization of BRCA1 to DSBs also requires sumoylation. We demonstrated that, in addition to having ubiquitin-interacting motifs, RAP80 also contains a SUMO-interacting motif (SIM) that is critical for recruitment to DSBs. In combination with the ubiquitin-binding activity of RAP80, this SIM enabled RAP80 to bind with nanomolar affinity to hybrid chains consisting of ubiquitin conjugated to SUMO. Furthermore, RNF4, a SUMO-targeted ubiquitin E3 ligase that synthesizes hybrid SUMO-ubiquitin chains, localized to DSBs and was critical for the recruitment of RAP80 and BRCA1 to sites of DNA damage. Our findings, therefore, connect ubiquitin- and SUMO-dependent DSB recognition, revealing that RNF4-synthesized hybrid SUMO-ubiquitin chains are recognized by RAP80 to promote BRCA1 recruitment and DNA repair.

Introduction

DNA double-strand breaks (DSBs) are highly cytotoxic lesions that, when not properly recognized and repaired, give rise to genome instability and can lead to cell death or to cancer. To maintain genome integrity, DSBs elicit a complex signaling cascade involving activation of cell cycle checkpoints and recruitment of chromatin-modifying and DNA repair factors to sites of DNA damage (1). DSBs are recognized by the MRE11-RAD50-NBS1 complex, which initiates damage signaling through recruitment and activation of the protein kinase ataxia telangiectasia mutated (ATM) (2, 3). Other posttranslational protein modifications, including ubiquitylation and sumoylation, act downstream of ATM-mediated phosphorylation to coordinate the assembly and regulation of repair factors at DSBs (4, 5).

Requirements for ubiquitylation in DSB repair are well established. Multiple ubiquitin E3 ligases are recruited to DSBs, including RNF8, RNF168, HERC2, and BRCA1 (6). RNF8 and RNF168 function at least in part to attach K63-linked polyubiquitin chains to histones H2A and H2AX (7). These polyubiquitin chains serve as signals that are recognized by ubiquitin-binding proteins, including the RAP80 subunit of the BRCA1-A complex (a complex containing the breast cancer–associated tumor suppressor BRCA1, RAP80, Abraxas, BRCC36, BRE, and NBA1). RAP80 contains tandem ubiquitin-interacting motifs (UIMs) that bind K63-linked polyubiquitin chains, a function critical for efficient recruitment of the BRCA1-A complex to DSBs (810). Specific roles for sumoylation in DSB repair are less well defined. SUMO-1, SUMO-2, and SUMO-3 are detected at sites of DSBs, but the modified substrates and the functional consequences of their sumoylation remain to be fully characterized. Depletion of the SUMO E3 ligases, PIAS1 and PIAS4, disrupts recruitment of BRCA1 to DSBs, at least in part through suppression of the accumulation of RNF168 and ubiquitin at sites of damage (1113). Thus, sumoylation is required at an early stage of DSB repair, upstream of ubiquitylation. The precise molecular mechanisms underlying the connections between sumoylation, ubiquitylation, and the recruitment of BRCA1 to DSBs, however, have remained unclear.

The SUMO-targeted ubiquitin E3 ligase, RNF4, is a potentially important factor involved in integrating ubiquitin and SUMO signals at sites of DNA damage. RNF4 is critical for DSB repair, with functions in regulating the stability of MDC1 (mediator of DNA damage checkpoint) and the efficiency of DNA end resection at sites of DNA damage (1416). RNF4 contains N-terminal SUMO-interacting motifs (SIMs) that enable it to bind polysumoylated proteins and attach ubiquitin to the SUMO chains on those proteins, thus producing hybrid SUMO-ubiquitin chains (17). The best-characterized fate of sumoylated proteins recognized and ubiquitylated by RNF4 involves proteasome-mediated degradation, although other fates have been described, including changes in protein localization (1719). Here, we demonstrate that hybrid SUMO-ubiquitin chains synthesized by RNF4 are recognized as high-affinity signals by RAP80. Moreover, we demonstrate that RNF4, and the recognition of hybrid SUMO-ubiquitin chains by RAP80, is critical for the recruitment of BRCA1 to sites of DNA damage.

Results

RAP80 is a SUMO-binding protein

Multiple components of the BRCA1-A complex have ubiquitin-binding activity, including RAP80, Abraxas, BRE, and BRCC36 (20). However, interactions between these proteins and SUMO have not been reported. Using bioinformatic analysis, we identified conserved candidate SIMs within predicted β strands that are in close proximity to known or predicted UIMs in each of these four proteins, suggesting a potential to bind SUMO and possibly hybrid SUMO-ubiquitin chains (Fig. 1A). In vitro binding assays with an immobilized glutathione S-transferase (GST)–tagged SUMO-2 polymer (a linear fusion of three SUMO-2 molecules) and purified recombinant RAP80, Abraxas, BRE, or BRCC36 demonstrated that these proteins interacted with SUMO-2 at amounts above background interaction with GST alone (Fig. 1B and fig. S1). We focused on RAP80 because of its well-characterized role in BRCA1 recruitment to DSBs. We evaluated the binding of RAP80, expressed in vitro in rabbit reticulocyte lysate in the presence of [35S]methionine, to immobilized GST-tagged SUMO-1 or SUMO-2 monomers, as well as SUMO-1 or SUMO-2 polymers. Binding to all four forms of SUMO was observed, with appreciably greater interactions detected with SUMO-2 relative to SUMO-1, and with polymers relative to monomers (Fig. 1C). Binding was not observed between RAP80 and SUMO-2(QFI), a mutant form of SUMO-2 containing alanine substitutions at residues Gln35 (Q), Phe36 (F), and Ile38 (I), which are critical for binding proteins with consensus SIMs (Fig. 1C) (21).

Fig. 1

RAP80 contains a consensus SIM critical for recruitment to DSBs. (A) Diagram illustrating predicted ubiquitin-binding (blue) and SUMO-binding (orange) domains in components of the BRCA1-A complex. AIR, Abraxas-interacting region; ZnF, zinc fingers; MPN±, MPR1 and Pad1 N-terminal region ± active protease site; CC, coiled coil; UEV1, ubiquitin E2 variant. (B) GST or GST-tagged SUMO-2(x3) was immobilized on glutathione-coated plates and incubated with the indicated FLAG- or His-tagged proteins. Bound proteins were eluted and analyzed by immunoblotting. Data shown are representative of three experiments. (C) RAP80 was transcribed and translated in rabbit reticulocyte lysate in the presence of [35S]methionine and incubated with GST alone or GST-tagged SUMO-1 monomer, SUMO-1(x3), SUMO-2 monomer, SUMO-2(x3), or SUMO-2(QFI) mutant. Bound proteins were eluted and analyzed by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) and autoradiography. Data are representative of three experiments. (D) Diagrams of wild-type (WT) and mutant (mut) RAP80 N-terminal domains and RAP80 SIM. The UIM mutant is missing the second of the two UIMs. (E) GST or GST-tagged WT and mutant RAP80 N-terminal domains were immobilized on glutathione-coated plates and incubated with SUMO-2(x3) (left panel) or K63-Ub4 (right panel). Bound proteins were eluted and analyzed by immunoblotting with SUMO-, ubiquitin-, or GST-specific antibodies. Data shown are representative of three experiments. Quantification of binding results is shown in fig. S2. (F) U2OS cells were transfected with constructs encoding full-length, GFP-tagged WT, UIM, SIM, or UIM and SIM (UIM/SIM) mutants of RAP80. Cells were irradiated, and recruitment to DNA repair foci was analyzed by fluorescence microscopy after a 2-hour recovery. Scale bar, 5 μm. (G) Quantitative analysis of RAP80 recruitment to DNA repair foci after IR. Results are representative of three independent experiments. Error bars represent SDs. *P < 0.003; **P < 0.008, Student’s one-tailed t test.

To characterize elements in RAP80 required for SUMO interactions, we performed binding assays with a GST-tagged N-terminal fragment of RAP80 (amino acids 1 to 137) and mutant versions of this fragment containing either alanine substitutions at Ile41, Val42, and Ile43 within the predicted SIM or a previously characterized deletion of residues 103 to 108 that disrupted the UIM (Fig. 1D) (9). Wild-type and UIM mutant RAP80 bound equally well to polymeric SUMO-2, whereas SUMO-2 binding was undetectable with SIM mutant RAP80 (Fig. 1E and fig. S2). Similarly, wild-type and SIM mutant RAP80 bound equally well to K63-linked tetraubiquitin, whereas ubiquitin binding was undetectable with the UIM mutant (Fig. 1E and fig. S2). Thus, the predicted SIM and UIMs of RAP80 functioned independently to mediate SUMO and ubiquitin binding, respectively.

SUMO binding facilitates RAP80 recruitment to DSBs

To investigate the contributions of SUMO and ubiquitin binding to the recruitment of RAP80 to DSBs, we transfected U2OS cells with constructs encoding full-length, green fluorescent protein (GFP)–tagged wild-type RAP80 or mutants with individual or combined SIM and UIM mutations. Cells expressing equivalent amounts of each fusion protein (fig. S3) were treated with ionizing radiation (IR), and the recruitment of RAP80 to DSBs was evaluated by colocalization with phosphorylated histone H2AX (γH2AX), a marker for sites of DNA repair (Fig. 1F). Whereas wild-type RAP80 was efficiently recruited to repair foci (~90% of cells contained ≥10 foci colabeled with γH2AX and RAP80), this recruitment was significantly, but only partially, reduced in cells expressing individual UIM or SIM mutants (~55 to 65% of cells contained ≥10 colabeled foci) (Fig. 1, F and G). However, recruitment was reduced to background amounts in cells expressing the combined UIM and SIM mutant form of RAP80 (Fig. 1, F and G), indicating that ubiquitin binding and SUMO binding function additively to recruit RAP80 to DSBs.

RAP80 binds hybrid SUMO-ubiquitin chains

The close proximity of the SIM and UIMs in RAP80 raised the possibility that recognition of hybrid SUMO-ubiquitin chains underlies RAP80 recruitment to DSBs. To test this hypothesis, we used RNF4 together with Ube2W, followed by the complex of Ubc13 and Mms2, to synthesize hybrid SUMO-ubiquitin chains consisting of a single molecule of SUMO-2 and two ubiquitins linked to each other through K63 (Fig. 2A and fig. S4A). In vitro binding assays were performed to compare interactions between RAP80 and monomeric SUMO-2, K63-linked diubiquitin, and hybrid SUMO-2–diubiquitin chain. Under assay conditions in which interactions with SUMO-2 or K63-linked diubiquitin alone were below the level of detection, we detected interactions between the N-terminal fragment of RAP80 and hybrid SUMO-ubiquitin chains (Fig. 2B). About threefold reduction, compared to the wild-type RAP80 N-terminal fragment, in binding to hybrid chains was observed with N-terminal fragments of RAP80 bearing individual SIM or UIM mutations (Fig. 2, C and D). To quantitate the relative affinity of RAP80 for different binding partners, we used equilibrium analytical ultracentrifugation and modeled the associations of the N-terminal domain of RAP80 with SUMO-2, K63-linked diubiquitin, or hybrid SUMO-2–diubiquitin chains to calculate dissociation constants (Kd’s) (Fig. 2E and fig. S4B). The affinity with which RAP80 bound K63-linked diubiquitin was consistent with the previously reported values (22) and was similar to the value we obtained for the affinity of RAP80 for SUMO-2, whereas the affinity of RAP80 for hybrid SUMO-2–diubiquitin chains was greater than that for SUMO-2 or diubiquitin alone, with a Kd of <0.2 μM. Thus, RAP80 bound hybrid SUMO-ubiquitin chains with an affinity 80-fold greater than the affinity for SUMO-2 or diubiquitin alone, and this high-affinity binding was mediated by dual SUMO-ubiquitin recognition by the SIM and UIMs in RAP80.

Fig. 2

RAP80 binds hybrid SUMO-ubiquitin chains with high affinity. (A) Schematic outline of the synthesis of hybrid SUMO-ubiquitin chains. RNF4, Ube2W, and the complex Ubc13-Mms2 represent E2 and E3 enzymes used in each step of synthesis. (B) GST-tagged RAP80 N-terminal domain was immobilized on glutathione-coated plates and incubated with SUMO-2, K63-linked diubiquitin, or hybrid SUMO-2–diubiquitin chains. Bound proteins were eluted and analyzed by immunoblotting with SUMO-2–, ubiquitin-, or GST-specific antibodies. Data shown are representative of three experiments. (C) Binding assays that are similar to those described in (B) were performed with the WT RAP80 N terminus and UIM or SIM mutants (mut). (D) Quantitative analysis of in vitro binding to hybrid SUMO-2–diubiquitin chains shown in (C). Results are representative of three independent experiments. Error bars represent SDs. AFU is defined as arbitrary fluorescence units. (E) Quantitative binding results based on analytical ultracentrifugation analysis of RAP80 N-terminal domain in complex with SUMO-2, K63-linked diubiquitin, or hybrid SUMO-2–diubiquitin chains.

RNF4 is required for RAP80 and BRCA1 recruitment to DSBs

Because RNF4 is involved in DSB repair (2325) and is a SUMO-targeted ubiquitin E3 ligase that synthesizes hybrid SUMO-ubiquitin chains, we investigated its potential role in the recruitment of RAP80 and BRCA1 to DNA repair foci in vivo. To demonstrate that RNF4 is recruited to IR-induced DSBs, we first treated U2OS cells with IR and analyzed them 2 hours later by immunofluorescence microscopy with antibodies specific for RNF4 and γH2AX. In the absence of DNA damage, RNF4 localized diffusely and in small foci throughout the nucleoplasm (Fig. 3A). In response to IR treatment, a population of RNF4 accumulated at DNA repair foci colabeled with γH2AX (Fig. 3A), consistent with its role in DSB repair.

Fig. 3

RNF4 is recruited to DSBs and required for recruitment of RAP80 to DSBs. (A) Control and IR-treated U2OS cells were analyzed by immunofluorescence microscopy with antibodies specific for RNF4 or γH2AX. Scale bar, 5 μm. (B) U2OS cells were transfected with control or RNF4-specific siRNAs and subsequently treated with IR. Cells were allowed to recover for 2 hours, and recruitment of RAP80 to γH2AX-labeled repair foci was analyzed by immunofluorescence microscopy. Scale bar, 5 μm. (C) Quantitative analysis of RAP80 recruitment to γH2AX-labeled repair foci in cells transfected with control or RNF4-specific siRNAs. Data are presented as the percentage of cells with more than 10 foci that were positive for both proteins. Results are representative of three independent experiments. Error bars equal SDs. (D) Quantitative analysis of RAP80 recruitment to γH2AX-labeled repair foci in cells cotransfected with RNF4-specific siRNA and RNA interference (RNAi)–resistant RNF4-GFP cDNA. Analysis was performed as described in (C).

Because we found that RNF4 was recruited to IR-induced DSBs, we next evaluated the role of RNF4 in recruiting RAP80 to DNA repair foci. U2OS cells were depleted of RNF4 by transfection with two independent RNF4-specific small interfering RNAs (siRNAs) (fig. S5A), treated with IR, and subsequently colabeled with antibodies recognizing RAP80 or γH2AX. We observed ~4.5-fold reduction in the amount of RAP80 and γH2AX that colocalized in RNF4-depleted cells compared to the amount of colocalization observed in control cells (Fig. 3, B and C, and fig. S5B). Cotransfecting cells with an siRNA-resistant complementary DNA (cDNA) coding for wild-type RNF4, but not a RING domain mutant, rescued this effect on RAP80 recruitment to DNA repair foci (Fig. 3D). Both wild-type and mutant proteins were expressed at equal amounts (fig. S6). These findings support a role for RNF4 ligase activity in generating signals required for efficient RAP80 recruitment to DSBs.

Because RAP80 is required for the efficient recruitment of BRCA1 to DSB, we also evaluated the effect of RNF4 depletion on BRCA1 localization in response to IR. We observed about threefold reduction in BRCA1 and γH2AX colocalization in RNF4-depleted cells compared to that in control cells at 30 min after IR (Fig. 4, A and B). Reduction in the recruitment of BRCA1 to DSBs persisted at later time points (Fig. 4B and fig. S5B), although to a lesser extent (~1.7-fold reduction). These findings may reflect temporal differences in mechanisms regulating BRCA1 recruitment to DNA repair foci after IR, or effects of incomplete RNF4 depletion. As observed for RAP80, cotransfecting cells with an siRNA-resistant cDNA coding for wild-type RNF4, but not a RING domain mutant, rescued BRCA1 recruitment to DSBs (Fig. 4C and fig. S6).

Fig. 4

RNF4 is required for the recruitment of BRCA1 to DSBs. (A) U2OS cells were transfected with control or RNF4-specific siRNAs and subsequently treated with IR. Cells were allowed to recover for 2 hours, and recruitment of BRCA1 to γH2AX-labeled repair foci was analyzed by immunofluorescence microscopy. Scale bar, 5 μm. (B) Quantitative analysis of BRCA1 recruitment to γH2AX-labeled repair foci in cells transfected with control or RNF4-specific siRNAs. Analyses were performed 0.5, 1.0, and 2.0 hours after treatment with IR. Results are representative of three independent experiments. Error bars equal SDs. (C) Quantitative analysis of BRCA1 recruitment to γH2AX-labeled repair foci in cells cotransfected with RNF4-specific siRNA and RNAi-resistant RNF4-GFP cDNA. Analysis was performed 0.5 hour after treatment with IR. (D) Model for hybrid SUMO-ubiquitin chain-dependent recruitment of RAP80 and the BRCA1-A complex to DSBs.

Discussion

Our findings provide a molecular explanation for the previously documented observation that sumoylation is required upstream of ubiquitylation and BRCA1 localization at DSBs (11, 12). Our model invokes the synthesis of hybrid SUMO-ubiquitin chains by RNF4 at sites of DNA damage and suggests that the use of signaling through ubiquitylation extends beyond polyubiquitin chains with unique linkages (26) and includes hybrid chains containing SUMO (Fig. 4D). Although a model in which distinct SUMO and ubiquitin chains attached to closely juxtaposed lysines is also possible, we favor hybrid chains for several reasons.

First, hybrid SUMO-ubiquitin chains appear to be ubiquitous in the cell and are particularly prominent during cellular responses to stress including arsenic exposure and proteasome inhibition (17, 2729). Hybrid chains containing K48-linked polyubiquitin are believed to function as signals for proteasome-mediated degradation, as illustrated by arsenic-induced degradation of the promyelocytic leukemia protein PML (17). Hybrid chains containing K63-linked polyubiquitin accumulate in cells after prolonged proteasome inhibition, although their functions in this context remain uncharacterized (28). We also favor hybrid SUMO-ubiquitin chains on the basis of our demonstration that they are recognized as distinct entities by RAP80. Recognition by RAP80 was achieved through closely juxtaposed UIMs and a SIM that together define a previously unreported, high-affinity binding element. The affinity of RAP80 for hybrid SUMO-ubiquitin chains was nearly two orders of magnitude greater than the observed affinity for K63-linked polyubiquitin chains alone and comparable to only the highest affinities measured for other known ubiquitin-binding proteins (22). Although future structural studies will be required to define hybrid chain recognition in detail, we propose that the specific spacing and orientation of SUMO and ubiquitin within hybrid chains determine recognition and high-affinity binding by RAP80. Note that tandem recognition motifs that bind unique protein sequence elements are an emerging theme among DNA repair proteins. Srs2, for example, binds sumoylated proliferating cell nuclear antigen (PCNA) through tandem motifs that independently bind SUMO and PCNA (30). Similarly, RNF168, RAD18, and RAP80 contain leucine-arginine motifs that, through proximity to UIMs, are thought to facilitate interactions with specific ubiquitylated proteins at sites of DNA damage (31).

Last, we favor a role for hybrid chains in DNA repair on the basis of the demonstration that RNF4 localized to DNA repair foci and was required for RAP80 and BRCA1 recruitment to IR-induced DSBs. RNF4 is an E3 ligase that specifically recognizes polysumoylated proteins and facilitates the conjugation of ubiquitin onto SUMO (17). We propose that hybrid SUMO-ubiquitin chains, synthesized at least in part by RNF4 at DSBs, offer advantages of both affinity and specificity. Specificity would be achieved through localized synthesis of hybrid chains at DSBs by an E2 and E3 cascade involving up to three independent E2 and E3 pairs. This cascade likely involves initial polysumoylation through the Ubc9/PIAS1 E2 and E3 pair because PIAS1 is particularly critical for the accumulation of proteins modified by SUMO-2 and SUMO-3 and the accumulation of RAP80 at DSBs (11, 12). Subsequent polyubiquitylation could involve RNF4 monoubiquitylation of SUMO chains followed by extension of K63-linked chains through the previously characterized Ubc13-Uev1A (also known as RNF168) E2 and E3 pair (7, 32). Alternatively, RNF4 could interact directly with Ubc13-Uev1A to stimulate the synthesis of K63-linked polyubiquitin chains on SUMO. Consistent with this latter model, in vitro studies indicate that RNF4 and the yeast homolog, the Sxl5/Slx8 heterodimer, can function with multiple E2 enzymes, including Ubc13 (33). Depletion of RNF4 from cells also reduces the accumulation of K63-linked polyubiquitin chains at sites of DNA damage (25). Substrates modified by hybrid SUMO-ubiquitin chains at DSBs remain to be fully characterized, but likely include MDC1 (2325), as well as histones or other DNA repair factors that are modified by polymeric SUMO-2/3 in response to cell stress (34). Functions of RNF4 extend beyond DNA repair and include regulation of gene expression and DNA methylation (35, 36). It is thus anticipated that hybrid SUMO-ubiquitin chains, and their recognition by proteins containing tandem UIMs and SIMs, will have important roles in multiple cellular contexts in addition to DNA repair.

Materials and Methods

Plasmids

Human RAP80 cDNA was provided by S. J. Elledge (Harvard Medical School, Boston, MA). Constructs for expression of RNF4-GFP wild type and RNF4-GFP RING mutant in mammalian cells were provided by O. J. Semmes (Eastern Virginia Medical School, Norfolk, VA). Full-length RAP80 cDNA was cloned into pcDNA6/myc–His A (Invitrogen) for in vitro transcription and translation in rabbit reticulocyte lysate. For expression in mammalian cells, full-length RAP80 cDNA was cloned into pEGFP-C1 (Clontech) and pFLAG-CMV-1 (Sigma). GST-tagged RAP80 proteins (amino acids 1 to 137) were produced in bacteria from pGEX-6P-1 (GE Healthcare). RAP80 mutants were generated by site-directed mutagenesis. RAP80 UIM mutant contained deletion of amino acid residues 103 to 108; RAP80 SIM mutant contained alanine substitution of amino acid residues Ile41, Val42, and Ile43; and RAP80 UIM/SIM mutant contained both deletion of residues 103 to 108 and substitution of residues I41A, V42A, and I43A.

Cell culture and treatment with IR

U2OS cells were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum and 10 mM Hepes. Cells were cultured at 37°C in a humidified incubator with 5% CO2. Exposure to 10 Gy of IR was carried out with a Gammator M-38 (Isomedix).

siRNA interference

Transfections were performed in 35 × 10–mm tissue culture dishes with 30 nM siRNA duplex and 3.75 μl of Lipofectamine RNAiMAX (Invitrogen) per transfection. Transfections were repeated 24 hours later. Experiments were performed 24 hours after the second transfection. RNF4 siRNA oligo 1 (GACTCACAATGACTCTGTTGTGATT) was obtained from Invitrogen, and RNF4 siRNA oligo 2 (GAAUGGACGUCUCAUCGUUUU) was synthesized by Dharmacon. Cotransfections were performed with 30 nM siRNA duplex, 3.75 μl of Lipofectamine 2000 (Invitrogen), and 1.0 μg of plasmids encoding for either RNAi-resistant RNF4-GFP wild type or RNF4-GFP RING mutant.

Antibodies

Primary antibodies were obtained from the following sources: γH2AX mouse monoclonal antibody (Millipore, clone JBW301, catalog no. 05-636), BRCA1 rabbit polyclonal antibody (Bethyl Laboratories Inc., catalog no. A300-000A), RAP80 rabbit polyclonal antibody (Bethyl Laboratories Inc., catalog no. A300-763A), ubiquitin mouse monoclonal antibody (Santa Cruz Biotechnology, catalog no. sc-8017), and SUMO-2/3 mouse monoclonal antibody 8A2 (37). RNF4 antibody was provided by J. Palvimo (University of Eastern Finland).

Immunofluorescence microscopy

Cells grown on coverslips were washed two times with phosphate-buffered saline (PBS) and then fixed and permeabilized with PBS containing 2% formaldehyde/0.2% Triton X-100 for 10 min at room temperature. Cells were washed with PBS and fixed with 2% formaldehyde in PBS for 20 min at room temperature. After washing with PBS, cells were incubated with the indicated primary antibodies for 1 hour. Cells were washed and incubated with Alexa Fluor–conjugated goat anti-rabbit immunoglobulin G (IgG) or Alexa Fluor–conjugated goat anti-mouse IgG (Invitrogen) for 1 hour. Coverslips were washed with PBS and mounted with mounting solution containing 100 mM tris (pH 8.8), 50% glycerol, 2.5% DABCO (Sigma), and 4′,6′-diamidino-2-phenylindole (0.2 μg/ml). Images were obtained on a Zeiss Observer.Z1 microscope.

GST pull-down assays

GST-tagged SUMOs and GST-tagged RAP80 (1 to 137) were expressed in Escherichia coli and purified with Glutathione Sepharose 4B (GE Healthcare) according to the manufacturer’s procedure. Recombinant GST or GST-tagged SUMOs (8 μg of protein) were diluted into assay buffer (1× PBS, 0.05% Tween 20) and incubated in glutathione-coated 96-well plates (Pierce Biotechnology Inc.). After overnight incubation at 4°C, wells were blocked for 1 hour at room temperature with 2% bovine serum albumin in assay buffer. RAP80 wild-type and mutant proteins were produced by in vitro transcription and translation in rabbit reticulocyte lysate in the presence of [35S]methionine according to the manufacturer’s instructions (Promega). In vitro translated proteins (10 μl) were diluted in 100 μl of assay buffer and incubated with the immobilized proteins for 1 hour at room temperature. Unbound proteins were removed by washing, and bound proteins were eluted with SDS sample buffer and resolved by SDS-PAGE and autoradiography. Equal loading of immobilized proteins was verified by anti-GST immunoblot. All binding assays were repeated in triplicate. Where indicated, GST-tagged RAP80 wild-type or GST-tagged RAP80 mutant proteins were immobilized on glutathione-coated plates and incubated with linear fusions of polymeric SUMO-2, K63-Ub4, or hybrid chains consisting of SUMO-2 linked to K63-Ub2. Eluted proteins were analyzed by immunoblotting with antibodies recognizing SUMO, ubiquitin, or GST.

SUMO-ubiquitin chain synthesis

Human ubiquitin and K63-Ub4 were expressed and purified following the protocol of Pickart and Raasi (38). SUMO was ubiquitylated with Ube2W and RNF4 in reaction buffer containing 10 mM Hepes (pH 7.6), 50 mM NaCl, 1 mM dithiothreitol, 1 mM adenosine 5′-triphosphate (ATP), 10 mM MgCl2, 250 nM human E1, 10 μM Ube2W, 15 μM RNF4, 300 μM ubiquitin, and 250 μM SUMO-2 (containing substitution of Lys11 to cysteine). The reaction was incubated overnight at 37°C. The reaction was then diluted 1:10 in 50 mM ammonium acetate (pH 4.5) and loaded onto a Mono S 10/100 column (GE Healthcare). Protein was eluted from the column in a linear gradient of 50 mM ammonium acetate (pH 4.5) and 500 mM NaCl. Fractions containing ubiquitylated SUMO were pooled and concentrated to <500 μl by centrifugation. The second ubiquitin was added in a reaction containing 30 mM Hepes (pH 7.6), 200 mM NaCl, 2 mM ATP, 10 mM MgCl2, 500 nM human E1, 5 μM yeast Ubc13-Mms2, 75 μM human ubiquitin (containing substitution of Lys63 to arginine), and 50 μM Ub-SUMO. The reaction was incubated overnight at 37°C. The reaction was diluted and purified as described for Ub-SUMO.

Determination of Kd values

Kd values for K63-Ub2, SUMO-2 (containing substitution of Lys11 to cysteine), and hybrid SUMO-Ub chains were determined by sedimentation equilibrium analytical ultracentrifugation. RAP80 and ligands were dialyzed overnight into 20 mM tris (pH 8.0), 100 mM NaCl, and 1 mM tris(2-carboxylethyl)phosphine. After dialysis, the concentration of each protein was determined with absorbance at 280 nm, except for RAP80, which was measured at 260 nm. The extinction coefficients were calculated with SEDNTERP (39) and were 1490 M−1 cm−1 for SUMO-2, 2980 M−1 cm−1 for K63-Ub2, 4470 M−1 cm−1 for hybrid SUMO-Ub chains, and 294 M−1 cm−1 at 260 nm for RAP80. RAP80 was mixed with ligand at three ratios (1:1, 2:1, and 1:4, RAP80/ligand), and 100 μl of each mixture was inserted into one sector of a six-channel epon centerpiece. The reference sectors were filled with 120 μl of dialysis buffer. Cells were then loaded into an An-60 Ti Rotor (Beckman Coulter Inc.) and equilibrated to 10°C and <50-μm pressure. Data were collected over a period of 72 hours with 24 hours of data collection at rotor speeds of 15,000, 20,000, and 25,000 rpm. Data were collected every 1 to 2 hours at 240 and 280 nm with Rayleigh interference optics. Data were sorted, and approach to equilibrium was confirmed in SEDFIT (40). Data were globally fit to an A + B → AB model in SEDPHAT (29).

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/5/253/ra88/DC1

Fig. S1. Multiple components of the BRCA1-A complex bind SUMO.

Fig. S2. Quantitative analysis of in vitro binding of GST-tagged RAP80 proteins to SUMO and ubiquitin.

Fig. S3. Analysis of the abundance of wild-type, UIM, and SIM mutant GFP-RAP80 fusion proteins expressed in U2OS cells.

Fig. S4. RAP80 binds with high affinity to hybrid SUMO-diubiquitin chains.

Fig. S5. RNF4 is required for RAP80 and BRCA1 recruitment to DSBs.

Fig. S6. Transfected U2OS cells express equivalent amounts of wild-type RNF4-GFP and RING mutant RNF4-GFP.

References and Notes

Acknowledgments: We would like to acknowledge members of the Matunis and Wolberger laboratories for helpful advice and comments during the course of this work. We would also like to acknowledge the following individuals for providing reagents: S. Elledge for RAP80 cDNAs, O. J. Semmes for RNF4-GFP constructs, and J. Palvimo for RNF4-specific antibodies. Funding: This work was supported by grants from the NIH to M.J.M. and from the American Association for Cancer Research to C.M.G. C.E.B. was supported in part by a Ruth Kirchstein National Research Service Award. C.W. is a Howard Hughes Medical Institute investigator. Author contributions: C.M.G. and M.J.M. conceived the study; C.M.G. performed all cloning, pull-down assays, and immunofluorescence microscopy experiments; V.G. prepared recombinant BRCA1-A complex proteins; C.E.B., A.D., and C.M.G. synthesized hybrid SUMO-Ub chains; C.E.B. performed analytical ultracentrifugation analysis and data interpretation; J.Z. optimized RNF4 knockdown conditions; M.J.M. wrote the manuscript; and R.A.G. and C.W. edited the paper. Competing interests: The authors declare that they have no competing interests. Data and materials availability: Materials and reagents are available upon request.
View Abstract

Navigate This Article