sign in sign up
<<
Home , Bloom syndrome protein, PALB2, RAD51

Scaffolding SPIDR/KIAA0146 connects the with repair

Li Wan1, Jinhua Han1, Ting Liu1, Shunli Dong, Feng Xie, Hongxia Chen, and Jun Huang2

Life Sciences Institute, Zhejiang University, Hangzhou, Zhejiang 310058, China

Edited by James E. Cleaver, University of California, San Francisco, CA, and approved February 26, 2013 (received for review December 1, 2012) The Bloom syndrome product, BLM, is a member of the highly of the SDSA pathway (6, 7). The ability of BLM to yield non- conserved RecQ family. An emerging concept is the BLM helicase crossover products is thought to play a critical role in the avoidance collaborates with the homologous recombination (HR) machinery to of chromosomal rearrangements during the homolog-directed re- help avoid undesirable HR events and to achieve a high degree of pair of chromosomal lesions. As a result, cells defective for BLM fidelity during the HR reaction. However, exactly how such coordina- exhibit elevated rates of sister chromatid exchange (SCE) (19–21). tion occurs in vivo is poorly understood. Here, we identified a protein Upon the occurrence of DNA damage, BLM is able to form termed SPIDR (scaffolding protein involved in DNA repair) as the link discrete foci, where it colocalizes with other DNA repair between BLM and the HR machinery. SPIDR independently interacts (22, 23). However, mechanistically how BLM is recruited to sites with BLM and RAD51 and promotes the formation of a BLM/RAD51- of DNA damage and how it collaborates with other proteins to containing complex of biological importance. Consistent with its role mediate recombination repair remain largely unexplored. In this as a scaffolding protein for the assembly of BLM and RAD51 foci, cells study, we used an affinity purification approach to isolate BLM- depleted of SPIDR show increased rate of sister chromatid exchange containing complex and identified a unique scaffolding protein and defects in HR. Moreover, SPIDR depletion leads to genome in- KIAA0146, which we refer to as SPIDR (scaffolding protein in- stability and causes hypersensitivity to DNA damaging agents. We volved in DNA repair). We demonstrate that SPIDR directly

propose that, through providing a scaffold for the cooperation of interacts with BLM and this interaction is required to target BLM CELL BIOLOGY BLM and RAD51 in a multifunctional DNA-processing complex, SPIDR to sites of DNA damage. Consistent with a role in BLM-mediated not only regulates the efficiency of HR, but also dictates the specific processes, cells depleted of SPIDR display an increased level HR pathway. of sister chromatid exchange. Moreover, we found that SPIDR promotes HR through a direct interaction with the recombinase DNA double-strand breaks | noncrossover | double RAD51. Finally, we show SPIDR promotes the formation of a BLM/RAD51-containing complex of biological importance. We omologous recombination (HR) repair allows precise repair therefore propose that, through licensing key cellular biochemical Hof DNA double-strand breaks and is critical for restarting properties of BLM and RAD51, SPIDR not only regulates the stalled or collapsed DNA replication forks (1–4). The process efficiency of HR but also dictates the specificHRpathway. of HR in eukaryotic cells requires several proteins, the central player of which is the RAD51 recombinase (1–4). HR repair is Results initiated by nuclease-mediated DNA end resection to generate Identification of SPIDR as a Unique BLM-Binding Partner. To search 3′-single-stranded DNA (ssDNA) tails that are initially coated by for previously undetected proteins present in BLM-containing the (RPA) complex (1–4). In a subsequent complex, we performed tandem affinity purification (TAP) using step, recombination mediator proteins catalyze the replacement 293T cells stably expressing streptavidin-flag-S protein (SFB)- of RPA with RAD51, resulting in the formation of ssDNA-RAD51 tagged wild-type BLM for the identification of BLM-interacting filament (1–4). The ssDNA-RAD51 filament then catalyzes strand proteins. Mass spectrometry analysis revealed several known invasion into homologous duplex DNA, leading to the formation of BLM-associated proteins, including TopoIIIα, RMI1/2, Rif1, a displacement loop (D-loop) (1–4). After DNA synthesis primed and FANCM (Fig. 1A). Interestingly, we also repeatedly iden- by the invading strand, the repair can bifurcate into two alternative tified a previously uncharacterized BLM-binding protein as subpathways referred to as synthesis-dependent strand annealing KIAA0146 (Fig. 1A). To ensure that KIAA0146 indeed associates (SDSA) and double-strand break (DSB) repair (5–9). In SDSA, the with BLM, we performed reverse TAP using a cell line stably extended D-loop can be dissolved by DNA and the newly expressing tagged KIAA0146, and identified BLM, TopoIIIα, synthesized strand is annealed to the ssDNA tail on the other break Rif1, RMI1/2, and FANCM as major KIAA0146-associated end, which is followed by gap-filling DNA synthesis and ligation proteins (Fig. 1B). These data strongly suggest that KIAA0146 (5–9). The repair products from SDSA are always noncrossover is a bone fide BLM-binding protein. The KIAA0146 gene enc- (5–9). In DSB repair, the second DSB end is captured to form an odes a deduced polypeptide of 915 amino acids with a predicted intermediate with two Holliday junctions, called double Holliday molecular mass of 105 kDa. Database searches did not identify junction (dHJ) (5–9). dHJ can be processed to yield crossover or noncrossover recombination products (5–9). There is mounting evidence that DNA repair mechanisms in mitotically proliferating Author contributions: L.W., J. Han, T.L., and J. Huang designed research; L.W., J. Han, T.L., cells avoid generating crossover recombination products, which can S.D., F.X., and H.C. performed research; L.W., J. Han, and J. Huang analyzed data; and contribute significantly to (5–9). J. Huang wrote the paper. In human cells, the Bloom syndrome gene product (BLM) The authors declare no conflict of interest. helicase, inactivated in individuals with Bloom syndrome, asso- This article is a PNAS Direct Submission. ciates with TopoIIIα and RMI1/2 (BLAP75/18) to form the BTR Freely available online through the PNAS open access option. (BLM-TopoIIIα-RMI1/2) complex (10–15). This complex pro- 1L.W., J. Han, and T.L. contributed equally to this work. motes the dissolution of dHJ to yield exclusively noncrossover 2To whom correspondence should be addressed. [email protected]. – recombination products (16 18). BLM can also drive HR toward This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. the formation of noncrossover products through the utilization 1073/pnas.1220921110/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1220921110 PNAS Early Edition | 1of6 Downloaded by guest on September 25, 2021 A B C To test whether there is a direct protein–protein interaction be- tween SPIDR and BLM, we performed GST pull-down assays with Myc-SPIDR SFB-BLMNegative control SFB-SPIDRNegative controlSFB-SPIDR fi kDa SFB-BLM kDa recombinant GST-BLM fusion proteins puri ed from baculovirus- 250 Protein Number of 250 Number of Protein SFB- Morc3BLM 150 peptides peptides infected insect cells and in vitro-translated full-length SPIDR. As 150 α BLM 208 -Myc 100 100 BLM 98 F TopoIIIα 66 KIAA0146 65 shown in Fig. 1 , BLM directly interacts with SPIDR in vitro. 75 75 RMI1 18 TopoIIIα 30 α-Flag

IP: α -Flag Upon the occurrence of DNA damage, BLM could form large RMI2 10 Rif1 21 50 50 Rif1 5 RMI1 15 37 37 α-Myc nuclear foci (22, 23). A physical interaction between SPIDR and FANCM 4 FIGNL1 15 KIAA0146 4 FANCM 12 Input α-Flag BLM, as demonstrated above, raises the possibility that SPIDR FANCA 4 RMI2 7 may colocalize with BLM at sites of DNA damage. Indeed, IP DEIP F Glutathione beads discrete foci of SPIDR were readily detected in cells following pulldown Input Input G IgG SPIDR SPIDR IgG BLM BLM GST + - hydroxyurea (HU) or camptothecin (CPT) treatment (Fig. 1 ). GST-BLM - + 5% IVT SPIDR Input SiRNA Con Con BLM Con Con BLM SiRNA Con Con SPIDR Con Con SPIDR Moreover, these foci colocalize with BLM foci, indicating that BLM SPIDR α-SPIDR SPIDR BLM kDa the localization of SPIDR, like that of BLM, is regulated in re- GST-BLM 250 150 G TopoIIIα TopoIIIα 100 sponse to DNA damage (Fig. 1 ). 75 RMI1 RMI1 50 To further define the binding between SPIDR and BLM, we 37 RMI2 RMI2 sought to identify the regions within SPIDR responsible for its in- GST 25 FAN1 FAN1 Coomassie staining teraction with BLM. Coimmunoprecipitation experiments revealed that SPIDR associated with BLM through its entire C terminus G SPIDR (Flag) BLM DAPI Merge Merged enlarged (residues 451–915), because deletion mutants lacking any part of this region failed to coprecipitate with BLM (Fig. S1 A and B).

HU Furthermore, pull-down experiments with recombinant proteins expressed in insect cells consolidated the above notion (Fig. S1C). Conversely, using a series of overlapping BLM truncations and deletion mutants spanning its entire coding sequence, we map-

CPT ped the SPIDR-binding region to residues 301–600 of BLM (Fig. S1 D–F).

Fig. 1. Identification of SPIDR as a BLM-binding partner. (A and B) 293T cells SPIDR Is Required for BLM Foci Formation and Suppresses SCE. Be- stably expressing SFB-tagged BLM or SPIDR were used for TAP of protein com- cause SPIDR exists in a complex with BLM, we sought to de- plexes. Tables are summaries of proteins identified by mass spectrometry anal- termine whether SPIDR has a role in the assembly of BLM foci. ysis. Letters in bold indicate the bait proteins. (C) 293T cells were transiently Thus, three different SPIDR-specific siRNA oligonucleotides were transfected with plasmids encoding SFB-tagged BLM or Morc3 together with transfected into U2OS cells. Whereas siRNA#1 and siRNA#3 plasmids encoding Myc-tagged SPIDR. Coprecipitation and immunoblotting significantly reduced SPIDR expression, siRNA#2 resulted in were carried out as indicated. (D and E) SiRNA-treated HeLa cells were lysed in being ineffective (Fig. 2 A and B). Moreover, the protein level of the presence of benzonase, cell lysates were then incubated with protein A agarose beads conjugated with indicated antibodies and Western blot analysis SPIDR remained the same in cells with BLM depletion and vice was carried out as indicated. (F) Direct binding between recombinant GST-BLM versa, suggesting that although SPIDR and BLM interact, they purified from baculovirus-infected insect cells and in vitro translated SPIDR (IVT do not influence each other’sstability(Fig.2B). Remarkably, SPIDR). (Upper) SPIDR was detected by immunoblotting. (Lower) Input GST- HU- or CPT-induced BLM focus formation was severely impaired proteins visualized by Coomassie staining. (G) SPIDR colocalizes with BLM. SFB- in SPIDR-depleted U2OS cells (Fig. 2 C and D). Similar results tagged SPIDR was expressed in HeLa cells. (Magnification: 100×.) Foci assembled were obtained using HeLa cells (Fig. S2 A and B). In contrast, by this fusion protein and by BLM following exposure to HU (2 mM) for 16 h or SPIDR depletion had no effect on γ-H2AX focus formation, in- μ fl CPT (1 M) for 3 h were detected by immuno uorescence using anti-Flag and dicating that SPIDR depletion does not generally interfere with anti-BLM antibodies, respectively. A merged image shows colocalization. DNA damage detection and signaling (Fig. 2C). To further ensure the specificity of the SPIDR-siRNA phenotype on BLM focus formation, we performed recovery experiments. similarity to known proteins in any species or to any known We engineered a U2OS cell line to express siRNA#1-resistant functional motifs. On the basis of its functions described below, form of SPIDR under the control of a tetracycline-inducible we designated this protein as SPIDR. promoter and examined BLM focus formation as a readout of SPIDR function. Induced expression of siRNA#1-resistant SPIDR SPIDR Interacts with BLM in Vivo and in Vitro. To validate our TAP fi was no more decreased by siRNA#1, whereas it was reduced as results, we rst performed coimmunoprecipitation experiments expected upon transfection of SPIDR-siRNA#3 (Fig. 2E). No- and found that Myc-tagged SPIDR interacts strongly with SFB- tably, the induced siRNA#1-resistant SPIDR was able to restore C tagged BLM but not with the Morc3 control protein (Fig. 1 ). BLM focus formation conferred by siRNA#1, but not siRNA#3 To examine the interaction between endogenous BLM and (Fig. 2 F and G), showing that the effect on recovery is truly SPIDR, cell extracts from HeLa cells were immunoprecipitated dependent on SPIDR. with the anti-BLM antibody or with the control IgG. As expec- The hallmark feature of BLM-deficient cells is their elevated ted, SPIDR was detected in the immunoprecipitations obtained frequency of SCEs. The observation that SPIDR is required for with the anti-BLM antiserum but not with the control IgG (Fig. efficient recruitment of BLM to sites of DNA damage led us to 1D). To prove the specificity of the BLM antibody, we performed propose that SPIDR may also be involved in SCE suppression. coimmunoprecipitation in BLM-depleted HeLa cells treated with As expected, the SCE level in SPIDR-depleted HeLa cells was BLM-specific siRNA. Indeed, we failed to detect any SPIDR from significantly elevated (Fig. 2H, and Fig. S2 C and D). The ob- the anti-BLM immunoprecipitations with these BLM-depleted cells served two- to threefold elevated frequency in SCE formation is (Fig. 1D). We also performed a reciprocal coimmunoprecipitation equivalent to that seen following depletion of BLM or RMI1/ assay. As shown in Fig. 1E, the endogenous BLM complex was BLAP75 in HeLa cells (11). Moreover, that depleting SPIDR readily immunoprecipitated with the SPIDR-specific antibody, but and BLM together did not cause a further increase in SCE fre- not with the control IgG. In these experiments, benzonase was in- quency compared with SPIDR or BLM depletion alone supports cluded in the lysis buffer to exclude the possibility that the in- the idea that BLM and SPIDR may suppress SCE via a common teraction occurs indirectly via DNA/RNA bridging (Fig. 1 D and E). pathway (Fig. S2 C and D). Flow cytometry analyses showed that

2of6 | www.pnas.org/cgi/doi/10.1073/pnas.1220921110 Wan et al. Downloaded by guest on September 25, 2021 comparable to that of control cells, SPIDR deletion mutants A 120 B C BLM γ-H2AX DAPI Merge A–C 100 defective in BLM binding failed to do so (Fig. S3 ). Con- 80 SiConSiSPIDR#1SiSPIDR#2SiSPIDR#3SiBLM sistently, although distinct nuclear foci of full-length BLM were

60 SPIDR SiCon Δ – 40 readily detected in HU-treated cells, the 301 600 mutant, 20 BLM which does not bind to SPIDR, failed to form foci after HU relative mRNA abundance 0 GAPDH treatment (Fig. S3D). Taken together, these results suggest that SiCon fi SiSPIDR#1 SPIDR and its interaction with BLM are required for the ef - D SiSPIDR#1SiSPIDR#2SiSPIDR#3 cient accumulation of BLM at sites of DNA damage. 60 Strikingly, the Δ1–300 mutant of BLM also exhibited a defect 50 D 40 SiSPIDR#2 in HU-induced foci formation (Fig. S3 ). Early studies have 30 shown the entire RMI (RecQ-mediated genome instability) 20 complex is essential for BLM recruitment or retention at DNA 10 SiSPIDR#3 0 damage sites primarily via a role in maintaining the protein stability of the BTR complex (11–13). Thus, it is possible that the SiCon SiCon % of Cells with >5 BLM foci SiBLM SiBLM

SiBLM RMI complex regulates the accumulation of BLM at sites of SiSPIDR#1SiSPIDR#2SiSPIDR#3 SiSPIDR#1SiSPIDR#2SiSPIDR#3 HU CPT DNA damage through its interaction with the N-terminal region missing in the Δ1–300 mutant of BLM. We then mapped the E SFB-SPIDR-SiR F SiCon SiSPIDR#1SiSPIDR#1 SiSPIDR#3 SiBLM Dox: + - + + + RMI binding domain on BLM. Pull-down experiments revealed that BLM interacted with RMI1 via its N terminus, because

SiCon SiSPIDR#1SiSPIDR#1SiSPIDR#3SiBLM BLM Δ – Dox: + - + + + deletion mutant lacking the 300 N-terminal amino acids ( 1 SPIDR 300) failed to pull-down with RMI1 (Fig. S3E). Moreover, con- (α-Flag) BLM sistent with previous studies (24), we also found that BLM DAPI α F GAPDH interacted with TopoIII via its N terminus (Fig. S3 ). Taken G together, these results suggest that both SPIDR and the RMI 80 H SiCon SiSPIDR#1 SiSPIDR#3 complex are required for optimal BLM foci formation. 60 40 Ability of BLM to Function in SCE Suppression Correlates with Its

20 Association with SPIDR. To further decipher the biological signif- CELL BIOLOGY 0 Dox: + - + + + + - + + + icance of the association between SPIDR and BLM, we used the SCE assay to investigate whether this interaction might be in- SiCon SiBLMSiCon SiBLM volved in suppressing SCE. Consistent with the above notion that % of Cells with >5 BLM foci SiSPIDR#1SiSPIDR#1SiSPIDR#3 SiSPIDR#1SiSPIDR#1SiSPIDR#3 HU CPT the Δ301–600 mutant of BLM lost its focus forming ability, the Δ301–600 mutant failed to restore the elevated SCE phenotype Fig. 2. SPIDR is required for BLM foci formation and suppresses SCE. (A) in BLM-depleted HeLa cells (Fig. S3 G and H). Quantitative RT-PCR showing SPIDR mRNA levels are down-regulated by To rule out the possibility that the phenotypes observed in siRNAs. Bars represent the average of three experiments, and error bars are SDs. Δ – (B) Representative immunoblotting of cells transfected with indicated siRNAs. (C BLM- 301 600 reconstituted BLM-depleted cells may be and D) SPIDR is required for BLM foci formation. U2OS cells were transfected twice caused by a failure in forming the BTR complex, we performed with control siRNA, or siRNAs specific for SPIDR or BLM. Forty-eight hours after coimmunoprecipitation experiments. As shown in Fig. S3I, the transfection, cells were treated with HU (2 mM) for 16 h or CPT (1 μM) for 3 h interaction between TopoIIIα or RMI1 with Δ301–600 mutant is before fixing and processed for BLM and γH2AX immunofluorescence. Repre- similar to that with wild-type BLM. sentative BLM foci (HU treatment) are shown (C). (Magnification: 100×.) Quan- tification results were the average of three independent experiments and were SPIDR Is Required for RAD51 Foci Formation and Promotes Homologous presented as mean ±SEM (D). (E–G) Rescue of BLM foci formation by expression of Recombination. Given that multiple roles have been proposed for a siRNA#1-resistant SPIDR cDNA. A U2OS cell line to express siRNA#1-resistant BLMinHRrepair(25–27), it will be of interest to determine SPIDR (SFB-SPIDR-SiR) under the control of a tetracycline-inducible promoter whether SPIDR plays a possible role in the early steps of HR. We was generated. The resulting cell line was transfected with indicated siRNAs and was either left uninduced (−) or was induced (+) by doxycycline (Dox) then examined whether depletion of SPIDR in U2OS cells affected addition for 24 h before HU (2 mM) or CPT (1 μM) treatment. After 16 h of RAD51 focus formation in response to irradiation (IR) or CPT treatment with HU or 3 h of treatment with CPT, cells were fixed and pro- and found that in the absence of SPIDR, RAD51 focus formation cessed for BLM immunofluorescence. The exogenous SPIDR expression was was severely impaired (Fig. 3 A and B). Similar results were confirmed by immunoblotting using anti-Flag antibody (E). Representative obtained using HeLa cells (Fig. S4 A and B). We did not observe BLM foci (HU treatment) were shown (F). (Magnification: 100×.) Quantifica- an obvious decrease in end resection following SPIDR depletion, tion results were the average of three independent experiments and were as measured by RPA2 focus formation, suggesting that SPIDR ± presented as mean SEM (G). (H) SPIDR suppresses SCE. Representative may promote RAD51 nucleofilament formation downstream of metaphase spreads showing SCEs from HeLa cells transfected with control or C–E fi SPIDR siRNAs (red arrows). (Magnification: 100×.) ssDNA generation (Fig. S4 ). To further con rm the role of SPIDR in HR, we performed a gene-conversion assay using the DR-GFP reporter system (28) (Fig. 3C). Consistent with a critical cell-cycle distribution was not affected by SPIDR down-regula- role in RAD51 recruitment, SPIDR depletion resulted in a sig- nificant reduction in HR (Fig. 3 D and E). Moreover, using the tion (Fig. S2E), ruling out the possibility that the phenotypes same method as described in Fig. 2F, we ruled out the possibility observed in SPIDR-depleted cells may be a result of any change that the loss of RAD51 focus formation after SPIDR depletion in cell-cycle distribution. was because of off-target effects of siRNAs (Fig. 3 F–H).

Direct Interaction Between SPIDR and BLM Is Required for BLM Foci Functional Interaction Between SPIDR and RAD51. We next tested Formation. As shown above, SPIDR is involved in the recruitment the possibility that the SPIDR might recruit RAD51 to sites of of BLM to site of DNA damage. Therefore, it would be in- DNA damage through a direct SPIDR–RAD51 physical inter- teresting to further test whether the binding to SPIDR might be action. Coimmunoprecipitation experiments revealed that Myc- required for BLM foci formation. Interestingly, whereas wild- tagged SPIDR interacted strongly with SFB-tagged RAD51 (Fig. type SPIDR successfully restored BLM focus formation to levels 4A). We also examined this interaction by using in vitro GST

Wan et al. PNAS Early Edition | 3of6 Downloaded by guest on September 25, 2021 RAD51 focus formation, SPIDR deletion mutants defective in ABγ RAD51 -H2AX DAPI Merge 70 RAD51bindingfailedtodoso(Fig. S5 B–D), indicating that 60 50 SPIDR may help to recruit RAD51 to DNA damage sites through

SiCon 40 adirectSPIDR–RAD51 interaction. In agreement with this 30 20 hypothesis, the localization of BRCA2 or Palb2 was unaffected 10 following SPIDR knockdown (Fig. S6). 0 SiSPIDR#1 SiCon SiCon SPIDR Provides a Link Between the BLM Helicase and RAD51. Previ- SiRAD51 SiRAD51 % of Cells with >5 RAD51 foci SiSPIDR#1SiSPIDR#2SiSPIDR#3 SiSPIDR#1SiSPIDR#2SiSPIDR#3 ous studies have shown that BLM can form a complex with RAD51 IR CPT

SiSPIDR#2 6 (29). The ability of SPIDR to interact independently with BLM C DR-GFP D I-SceI BcgI 5 SceGFP iGFP 4 and RAD51 indicates that SPIDR, BLM, and RAD51 may form a +I-SceI 3 ternary complex. To determine whether this is the case, HeLa 2 SiSPIDR#3 HR 1 extracts were resolved by gel filtration and the individual fractions BcgI BcgI 0 GFP+ iGFP % GFP-positive cells SiCon SiRAD51 SiRAD51 SiSPIDR#1SiSPIDR#2SiSPIDR#3 E F SiCon SiSPIDR#1 SiSPIDR#1 SiSPIDR#3 SiRAD51 Dox: + - + + + A Myc-SPIDR B

SiConSiSPIDR#1SiSPIDR#2SiSPIDR#3SiRAD51 Glutathione beads SPIDR pulldown RAD51 RAD51 SFB- RAD51 Morc3 GST + - 5% IVT SPIDR Input GAPDH α-Myc GST-RAD51 - + G DAPI α 80 -SPIDR

60 H SFB-SPIDR-SiR α -Flag α kDa 40 -Flag 20 IP: 250 SiCon SiSPIDR#1SiSPIDR#1SiSPIDR#3SiRAD51 0 Dox: + - + + + 150 Dox: + - + + + + - + + + SPIDR 100 (α -Flag) α 75 SiCon SiCon RAD51 -Myc SiRAD51 SiRAD51 GST-RAD51 % of Cells with >5 RAD51 foci SiSPIDR#1SiSPIDR#1SiSPIDR#3 SiSPIDR#1SiSPIDR#1SiSPIDR#3 GAPDH IR CPT 50

Fig. 3. SPIDR promotes homologous recombination. (A and B) SPIDR is re- Input 37 quired for RAD51 foci formation. U2OS cells were treated with IR (10 Gy) or α-Flag CPT (1 μM) for 3 h before fixing and processed for RAD51 and γH2AX im- munofluorescence. Representative RAD51 foci (IR treatment) are shown (A). GST 25 (Magnification: 100×.) Quantification results were the average of three in- Coomassie staining dependent experiments and were presented as mean ± SEM (B). (C) Sche- matic representation of HR assay. (D) U2OS DR-GFP cells were transfected C IP with the indicated siRNA and 24 h later were electroporated with an I-SceI expression plasmid. Forty-eight hours after electroporation, cells were har- Input vested and assayed for GFP expression by FACS analysis. Results were the IgG SPIDR SPIDR average of three independent experiments and were presented as mean± SiRNA Con Con SPIDR Con Con SPIDR SEM (E). Knockdown efficiency was confirmed by immunoblotting. (F and G) Rescue of RAD51 foci formation by expression of a siRNA#1-resistant SPIDR SPIDR cDNA. Immunostaining experiments were performed as described in Fig. 2F. Representative RAD51 foci (IR treatment) are shown (F). (Magnification: 100×.) Quantification results were the average of three independent RAD51 experiments and are presented as mean ± SEM (G). (H) The exogenous SPIDR expression was confirmed by immunoblotting. D IP

pull-down assays and found that GST-tagged human RAD51, obtained Input from , interacts with in vitro-translated SPIDR (Fig. 4B). IgG RAD51 RAD51 To further determine whether endogenous SPIDR and RAD51 SiRNA Con Con RAD51 Con Con RAD51 interact, benzonase-treated HeLa cell lysates were prepared and RAD51 subjected to immunoprecipitation with either a control or anti- SPIDR antibody. As expected, endogenous RAD51 was immuno- precipitated by anti-SPIDR antibody, but not by control antibody, SPIDR and the reverse experiment confirmed this result (Fig. 4 C and D). We next attempted to identify the regions within SPIDR re- Fig. 4. Functional interaction between SPIDR and RAD51. (A) 293T cells sponsible for its interaction with RAD51. Pull-down assays dem- were transfected with plasmids encoding SFB-tagged RAD51 or Morc3 to- onstrated that a domain spanning amino acids 151–450 of SPIDR gether with plasmids encoding Myc-tagged SPIDR. Coprecipitation and im- is responsible for RAD51 binding (Fig. S5A). munoblotting were carried out as indicated. (B) Direct binding between fi To further investigate the biological significance of the SPIDR– recombinant GST-RAD51 puri ed from bacteria and in vitro-translated SPIDR. (Upper) SPIDR was detected by immunoblotting. (Lower) Input GST- RAD51 interaction, we performed rescue experiments similar to proteins visualized by Coomassie staining. (C and D) SiRNA-treated HeLa cells those described above to examine whether the RAD51-binding were lysed in the presence of benzonase, cell lysates were then incubated region on SPIDR is required for efficient RAD51 foci forma- with protein A agarose beads conjugated with indicated antibodies, and tion. Interestingly, whereas wild-type SPIDR successfully restored Western blot analysis was carried out as indicated.

4of6 | www.pnas.org/cgi/doi/10.1073/pnas.1220921110 Wan et al. Downloaded by guest on September 25, 2021 A B Discussion BLM BLM Ferritin (440 kDa) BSA (67 kDa) IgG fi SiRNA Con Con SPIDR In this study, we report the identi cation of a unique protein

6 7 8 9 10 12 14 16 18 20 22 24 26 28 IgG α -RAD51 α -BLM α -SPIDR RAD51 SPIDR SPIDR designated SPIDR, which may serve as a scaffolding protein that IP RMI1 helps to promote the formation of BLM/RAD51-containing BLM BLM BLM RAD51 complexes of biological importance. RMI1 RAD51 RAD51 It is well-established that accumulation of repair proteins on

Input BLM SPIDR damaged is required to restore genome stability. C D E Like many DNA damage/repair proteins, the BLM helicase is able SiCon SiSPIDR 0.3 100 0.25 to accumulate into foci structures in response to DNA damage 0.2 0.15 (22, 23). However, mechanistically how BLM is recruited to sites 0.1 10 of DNA damage remains unclear. Previous studies have shown Aberrations/cell 0.05 SiCon SiSPIDR#1 0 % cell viability that FANCM can target the BLM complex to nuclear foci after SiSPIDR#3 SiCon SiRAD51 treatment of cells with CPT or mitomycin-C (19). Targeting by 1 SiSPIDR#1 SiSPIDR#3 0246 IR (Gy) FANCM is dependent on a direct interaction between the con- F G H served MM2 domain of FANCM and the subunits RMI1 and 100 100 SPIDR α RAD51 BLM TopoIII of the BLM complex (19). In addition, RMI1/2 has been shown to regulate BLM retention at sites of DNA damage pri- 10 10 SiCon SiCon RAD51- double marily via a role in maintaining the protein stability of the BLM D-loops

% cell viability SiSPIDR#1 % cell viability SiSPIDR#1 ssDNA Holliday SiSPIDR#3 SiSPIDR#3 filaments junctions complex (11–13). In contrast, SPIDR recruits the BLM complex to SDSA SiRAD51 SiRAD51 DSBR 1 1 sites of DNA damage through a direct interaction with BLM. It is 0 5 10 20 0 0.0006 0.0008 0.001 CPT (nM) MMS (%) Non-crossover products also interesting to note that the MRN (MRE11, RAD50, and NBS1) complex can increase the affinity of BLM for DNA ends Fig. 5. SPIDR is required for maintaining genome integrity. (A) Endogenous (30). The existence of multiple distinct regulatory mechanisms for SPIDR, BLM, and RAD51 are present in a trimeric complex. HeLa extracts were resolved by gel filtration and the fractions were analyzed by immunoblotting the BLM complex recruitment in mammalian cells underscores (Left). Anti-SPIDR, anti-BLM, and anti-RAD51 immunoprecipitaton of fractions the importance of this process in DNA repair. containing SPIDR, BLM, and RAD51 were analyzed by immunoblotting with RAD51 and BLM have been implicated in the regulation of HR antibodies specific for SPIDR, BLM, and RAD51 (Right). (B)HeLacellstrans- at the early and late steps, respectively, according to the timing of fected with control or SPIDR siRNA were subjected to coimmunoprecipitation respective involvement in the recombination process. To achieve CELL BIOLOGY using anti-BLM antibody. Immunoblotting was performed using antibodies as maximal efficiency for HR repair, these steps must be tightly indicated. (C) Representative images of metaphase spreads prepared from regulated and coordinated at the molecular level. However, the HeLa cells treated with the indicated siRNAs. Representative aberrations are precise mechanisms for this coordination are poorly understood. fi × fi marked by arrows. (Magni cation: 100 .) (D)Quanti cation of chromosomal Here, we have described one example of such a mechanism in aberrations in control and SPIDR-depleted HeLa cells. The average of two which SPIDR coordinates both BLM and RAD51 functions fol- experiments is shown; at least 50 cells were counted in each experiment. Error – bars represent the SD. (E–G) Clonogenic survival assays in SPIDR-depleted HeLa lowing DNA damage. The SPIDR RAD51 interaction might fa- cells following IR, CPT, or MMS treatment. Experiments were done in tripli- cilitate the loading of RAD51 on resected DNA to initiate HR, cates. Results shown are averages of three independent experiments. (H)A such that via its direct interaction with BLM, SPIDR may serve as proposed model of SPIDR functions in DNA repair. a signaling platform to dictate the specificHRpathway(Fig.5H). It is noteworthy to point out that although cells lacking SPIDR or the breast susceptibility (BRCA) proteins show were analyzed by immunoblotting (Fig. 5A). The fractions con- similar defects in RAD51 focus formation, they have opposite taining SPIDR, BLM, and RAD51 were then immunoprecipitated effects on the frequency of SCE. Down-regulation of SPIDR with anti-SPIDR, anti-BLM, or anti-RAD51 antibody. As shown in exhibits an elevated SCE rate, whereas lack of BRCA1/2 leads to Fig. 5A, SPIDR, BLM, and RAD51 were present in all of the three a reduction in the SCE frequency (31, 32). The simplest explana- immunoprecipitations, thereby demonstrating the formation of a tion for this difference is that SPIDR not only promotes assembly trimeric complex. To further examine whether SPIDR may serve as of subnuclear RAD51 foci, but also regulates the ratio of crossover- a scaffolding protein to promote the BLM–RAD51 complex for- to-noncrossover recombinants via its ability to recruit the BLM mation,weusedsiRNAtoknockdownSPIDRandexaminedthe helicase. As a result, although the overall HR efficiency is reduced association between BLM and RAD51. As shown in Fig. 5B,when in SPIDR-depleted cells, most of the residual HR intermediates cells were treated with SPIDR siRNA, but not a control siRNA, the are specifically resolved into crossover products, which in turn interaction between BLM and RAD51 was significantly reduced. contribute, at least partially, to the elevated SCE frequency. These results suggested that SPIDR might mediate, at least in part, In summary, our study reveals a previously undescribed link the interaction between the BLM helicase and RAD51. between the BLM helicase and the HR machinery. Our results support the idea that SPIDR contributes to genomic integrity by SPIDR Is Critical for Maintaining Genome Stability. Given that SPIDR licensing the assembly of the key DNA-processing BLM is required for proper DNA repair, we tested the effect of SPIDR and RAD51 at DNA lesions. Defective HR repair is associated on genome integrity. As shown in Fig. 5 C and D, SPIDR-depleted with genomic instability and cancer prone syndromes, and it will be HeLa cells displayed a significant increase in chromosomal aber- particularly interesting to determine whether defects in SPIDR rations visualized in metaphase spreads. Furthermore, knockdown are also relevant to human diseases. of SPIDR in HeLa cells conferred cellular hypersensitivity to IR, CPT, and methyl-methane sulfonate (MMS) (Fig. 5 E–G and Fig. ACKNOWLEDGMENTS. We thank all our colleagues in the J. Huang laboratory S7A). More importantly, these cellular phenotypes can be rescued for insightful discussions and Dr. X. H. Feng for comments on the manuscript. This work was supported by National Basic Research Program of China Grants by reconstitution of SPIDR-depleted cells with wild-type SPIDR, 2012CB944402 and 2013CB911003; National Natural Science Foundation of but not with the deletion mutants defective in BLM or RAD51 China Grant 31071243; Natural Science Foundation of Zhejiang Province binding, indicating that SPIDR plays a functionally important role Grant R2110569; and Open Research Fund of State Key Laboratory of Cellular in allowing cells to repair genotoxic damage and maintain chro- Stress Biology, Xiamen University Grant SKLCSB2013KF002. T.L. is a member of the X. H. Feng laboratory and supported by National Science Foundation mosomal integrity, possibly through its direct interaction with BLM of China Grants 31171347 and 31090360 and the Ministry of Science and and RAD51 (Fig. S7 B–D). Technology Grant 2012CB966600.

Wan et al. PNAS Early Edition | 5of6 Downloaded by guest on September 25, 2021 1. Moynahan ME, Jasin M (2010) Mitotic homologous recombination maintains genomic 18. Bussen W, Raynard S, Busygina V, Singh AK, Sung P (2007) Holliday junction processing stability and suppresses tumorigenesis. Nat Rev Mol Cell Biol 11(3):196–207. activity of the BLM-Topo IIIalpha-BLAP75 complex. JBiolChem282(43):31484–31492. 2. Petermann E, Helleday T (2010) Pathways of mammalian replication fork restart. Nat 19. Deans AJ, West SC (2009) FANCM connects the genome instability disorders Bloom’s Rev Mol Cell Biol 11(10):683–687. Syndrome and . Mol Cell 36(6):943–953. 3. West SC (2003) Molecular views of recombination proteins and their control. Nat Rev 20. Goss KH, et al. (2002) Enhanced tumor formation in mice heterozygous for Blm Mol Cell Biol 4(6):435–445. . Science 297(5589):2051–2053. 4. San Filippo J, Sung P, Klein H (2008) Mechanism of eukaryotic homologous 21. Gruber SB, et al. (2002) BLM heterozygosity and the risk of colorectal cancer. Science recombination. Annu Rev Biochem 77:229–257. 297(5589):2013. 5. Colavito S, Prakash R, Sung P (2010) Promotion and regulation of homologous 22. Wang Y, et al. (2000) BASC, a super complex of BRCA1-associated proteins involved in – recombination by DNA helicases. Methods 51(3):329 335. the recognition and repair of aberrant DNA structures. Dev 14(8):927–939. 6. Adams MD, McVey M, Sekelsky JJ (2003) Drosophila BLM in double-strand break 23. Bischof O, et al. (2001) Regulation and localization of the in – repair by synthesis-dependent strand annealing. Science 299(5604):265 267. response to DNA damage. J Cell Biol 153(2):367–380. 7. Bachrati CZ, Borts RH, Hickson ID (2006) Mobile D-loops are a preferred substrate for 24. Hu P, et al. (2001) Evidence for BLM and Topoisomerase IIIalpha interaction in ge- ’ – the Bloom s syndrome helicase. Nucleic Acids Res 34(8):2269 2279. nomic stability. Hum Mol Genet 10(12):1287–1298. 8. Barber LJ, et al. (2008) RTEL1 maintains genomic stability by suppressing homologous 25. Chu WK, Hanada K, Kanaar R, Hickson ID (2010) BLM has early and late functions in recombination. Cell 135(2):261–271. homologous recombination repair in mouse embryonic stem cells. Oncogene 29(33): 9. Bzymek M, Thayer NH, Oh SD, Kleckner N, Hunter N (2010) Double Holliday junctions 4705–4714. are intermediates of DNA break repair. Nature 464(7290):937–941. 26. Bugreev DV, Yu X, Egelman EH, Mazin AV (2007) Novel pro- and anti-recombination 10. Karow JK, Chakraverty RK, Hickson ID (1997) The Bloom’s syndrome gene product is activities of the Bloom’s syndrome helicase. Genes Dev 21(23):3085–3094. a3′-5′ DNA helicase. J Biol Chem 272(49):30611–30614. 27. Gravel S, Chapman JR, Magill C, Jackson SP (2008) DNA helicases and BLM pro- 11. Yin J, et al. (2005) BLAP75, an essential component of Bloom’s syndrome protein mote DNA double-strand break resection. Genes Dev 22(20):2767–2772. complexes that maintain genome integrity. EMBO J 24(7):1465–1476. 28. Weinstock DM, Nakanishi K, Helgadottir HR, Jasin M (2006) Assaying double-strand 12. Xu D, et al. (2008) RMI, a new OB-fold complex essential for Bloom syndrome protein break repair pathway choice in mammalian cells using a targeted endonuclease or to maintain genome stability. Genes Dev 22(20):2843–2855. – 13. Singh TR, et al. (2008) BLAP18/RMI2, a novel OB-fold-containing protein, is an essential the RAG recombinase. Methods Enzymol 409:524 540. component of the Bloom helicase-double Holliday junction dissolvasome. Genes Dev 29. Wu L, Davies SL, Levitt NC, Hickson ID (2001) Potential role for the BLM helicase in 22(20):2856–2868. recombinational repair via a conserved interaction with RAD51. J Biol Chem 276(22): – 14. Seki M, et al. (2006) Bloom helicase and DNA topoisomerase IIIalpha are involved in 19375 19381. the dissolution of sister chromatids. Mol Cell Biol 26(16):6299–6307. 30. Nimonkar AV, et al. (2011) BLM-DNA2-RPA-MRN and EXO1-BLM-RPA-MRN constitute two 15. Ellis NA, et al. (1995) The Bloom’s syndrome gene product is homologous to RecQ DNA end resection machineries for human DNA break repair. Genes Dev 25(4):350–362. helicases. Cell 83(4):655–666. 31. Tutt A, et al. (2001) Mutation in Brca2 stimulates error-prone -directed 16. Wu L, et al. (2006) BLAP75/RMI1 promotes the BLM-dependent dissolution of ho- repair of DNA double-strand breaks occurring between repeated sequences. EMBO J mologous recombination intermediates. Proc Natl Acad Sci USA 103(11):4068–4073. 20(17):4704–4716. 17. Raynard S, et al. (2008) Functional role of BLAP75 in BLM-topoisomerase IIIalpha- 32. Feng Z, Zhang J (2012) A dual role of BRCA1 in two distinct homologous recombination dependent holliday junction processing. J Biol Chem 283(23):15701–15708. mediated repair in response to replication arrest. Nucleic Acids Res 40(2):726–738.

6of6 | www.pnas.org/cgi/doi/10.1073/pnas.1220921110 Wan et al. Downloaded by guest on September 25, 2021