Jour

nal o

f Cel

l Sci

ence

RESEARCH ARTICLE

RRP6/EXOSC10 is required for the repair of DNA double-strandbreaks by homologous recombination

Consuelo Marin-Vicente`, Judit Domingo-Prim`, Andrea B. Eberle* and Neus Visa§

ABSTRACT

The exosome acts on different RNA substrates and plays important

roles in RNA metabolism. The fact that short non-coding RNAs are

involved in the DNA damage response led us to investigate whether

the exosome factor RRP6 of Drosophila melanogaster and its human

ortholog EXOSC10 play a role in DNA repair. Here, we show that

RRP6 and EXOSC10 are recruited to DNA double-strand breaks

(DSBs) in S2 cells and HeLa cells, respectively. Depletion of RRP6/

EXOSC10 does not interfere with the phosphorylation of the histone

variant H2Av (Drosophila) or H2AX (humans), but impairs the

recruitment of the homologous recombination factor RAD51 to the

damaged sites, without affecting RAD51 levels. The recruitment of

RAD51 to DSBs in S2 cells is also inhibited by overexpression of

RRP6-Y361A–V5, a catalytically inactive RRP6 mutant. Furthermore,

cells depleted of RRP6 or EXOSC10 are more sensitive to radiation,

which is consistent with RRP6/EXOSC10 playing a role in DNA repair.

RRP6/EXOSC10 can be co-immunoprecipitated with RAD51, which

links RRP6/EXOSC10 to the homologous recombination pathway.

Taken together, our results suggest that the ribonucleolytic activity of

RRP6/EXOSC10 is required for the recruitment of RAD51 to DSBs.

KEY WORDS: RRP6, EXOSC10, DNA repair, Exosome, Non-coding

RNA, RAD51

INTRODUCTIONThe exosome is a multiprotein complex with ribonucleolytic

activity that plays important roles in RNA metabolism, including

processing of RNA precursors and surveillance of mRNA

biogenesis (reviewed in Houseley et al., 2006; Chlebowski

et al., 2013; Eberle and Visa, 2014). The eukaryotic exosome is

composed of a core that is catalytically inactive (reviewed in

Lorentzen et al., 2008; Januszyk and Lima, 2014). The RNases

RRP6 (also known as EXOSC10 in mammals) and DIS3 (also

known as RRP44) associate with this core and provide catalytic

activity (reviewed by Lykke-Andersen et al., 2011). RRP6 and

DIS3, however, have other interaction partners than the exosome

core and can act independently of the exosome (Synowsky et al.,

2009; Graham et al., 2009; Kiss and Andrulis, 2011).

The activity of the mRNA surveillance machinery of

Saccharomyces cerevisiae has been linked to the DNA damage

response (Hieronymus et al., 2004; Manfrini et al., 2014), and the

human exosome, when complexed with the RNA and DNA

helicase senataxin, is targeted to sites of transcription-induced

DNA damage in human cells (Richard et al., 2013). However, the

contribution of the exosome to the repair of DNA lesions has not

been investigated.

The DNA-damage response (DDR) integrates signaling

cascades that lead to the arrest of cell proliferation and the

activation of the cellular DNA repair machineries (reviewed by

Polo and Jackson, 2011). Double-strand breaks in the DNA

(DSBs) trigger the recruitment and activation of the sensor

protein kinases ATM or ATR to the damaged sites. These kinases

phosphorylate a number of specialized substrates, including

the histone variant H2AX in humans or its functional homolog

H2Av in D. melanogaster (Madigan et al., 2002). The local

phosphorylation of H2AX/H2Av in the vicinity of DSBs

generates binding sites for adaptor proteins and DDR effectors.

Two major DNA repair mechanisms, non-homologous end-

joining (NHEJ) and homologous recombination, are responsible

for the repair of DSBs. NHEJ is the predominant mechanism in

mammalian cells, whereas homologous recombination depends

on the availability of homologous sequences and is more frequent

in Drosophila (Preston et al., 2006; Lieber, 2008). Regardless of

the repair mechanisms used, the sequential recruitment of adaptor

proteins and effector repair factors to the DSBs leads to the

formation of cytologically detectable foci at the damaged DNA

sites. These foci are known as ‘DDR foci’.

Short non-coding RNAs are required for the assembly of DDR

foci and for DNA repair by homologous recombination (Francia

et al., 2012; Wei et al., 2012; Gao et al., 2014). The fact that the

exosome acts on virtually all types of RNA in eukaryotic cells led

us to investigate whether the catalytic subunit of the nuclear

exosome, RRP6/EXOSC10, plays a role in DSB repair. We show

that RRP6/EXOSC10 interacts, directly or indirectly, with the

homologous recombination factor RAD51 and is required for the

proper assembly of DDR foci in fly and human cells. Moreover,

experiments with a catalytically inactive RRP6 mutant suggested

that the ribonucleolytic activity of RRP6 is required for this

process. In summary, our results identify RRP6/EXOSC10 as a

new player in the DDR and link RRP6/EXOSC10 to the

homologous recombination pathway.

RESULTSRRP6 is recruited to sites of DNA damageWe used a Drosophila S2 cell line (S2-RRP6–V5) that expresses

a V5-tagged version of RRP6 (Hessle et al., 2009) to analyze

whether the interaction network of RRP6 was affected in response

to DNA damage. The RRP6–V5 protein is imported into the

nucleus, is able to engage in functional interactions and can be

immunoprecipitated by an anti-V5 antibody (Hessle et al., 2009).

Department of Molecular Biosciences, The Wenner-Gren Institute, StockholmUniversity, SE-106 91 Stockholm, Sweden.*Present address: Department of Chemistry and Biochemistry, University of Bern,CH-3012 Bern, Switzerland.`These authors contributed equally to this work

§Author for correspondence (neus.visa@su.se)

Received 26 June 2014; Accepted 21 January 2015

� 2015. Published by The Company of Biologists Ltd | Journal of Cell Science (2015) 128, 1097–1107 doi:10.1242/jcs.158733

1097

Jour

nal o

f Cel

l Sci

ence

We exposed S2-RRP6–V5 cells to gamma irradiation (cIR) andimmunoprecipitated RRP6–V5 30 min after the irradiation. We

identified 49 proteins that co-immunoprecipitated with RRP6–V5in non-irradiated control cells (supplementary material Tables S1,S2). Only 23 of these 49 interactors were detected in theirradiated cells, which also revealed two unique interactions with

the linker histone H1 and the histone variant H2Av (Fig. 1A).These results suggest that cIR rapidly remodels the interactionnetwork of RRP6.

We then analyzed the distribution of RRP6–V5 byimmunofluorescence using antibodies against the V5 tag andagainst the phosphorylated form of H2Av (cH2Av) as a marker for

DDR foci (Fig. 1B,C). Exposure to cIR increased the percentage ofcells with cH2Av-rich foci to almost 100% 10 min after theirradiation. At the same time, the distribution of RRP6–V5 became

more granular, and 66% of the cells displayed distinct RRP6–V5-rich foci (Fig. 1B,C). The same relocalization was observed withan antibody directed against the endogenous RRP6 protein. Many

cH2Av-rich foci overlapped (58%) or were juxtaposed (27%) withthe RRP6–V5-rich foci (Fig. 1D,E). The distributions of cH2Av

and RRP6–V5 in cIR-treated cells did not overlap completely,which is not surprising considering the many roles of RRP6 inRNA metabolism. Similar immunofluorescence experiments wereperformed with S2 cells that expressed RRP4–V5, a subunit of the

exosome core (Hessle et al., 2009). The distribution of RRP4–V5was not affected by the cIR treatment (Fig. 1C).

In another series of experiments, we asked whether the cIR-

induced relocalization of RRP6 was dependent on thephosphorylation of H2Av. S2 cells were treated with CGK733,an inhibitor of ATM and ATR, and stained with antibodies

against cH2Av or endogenous RRP6. Phosphorylation of H2Avwas impaired in CGK733-treated cells, as expected. Furthermore,the formation of RRP6-rich foci was drastically reduced by the

CGK733 treatment (Fig. 1F; supplementary material Fig. S1).In summary, cIR induces a rapid and transient relocalization of

RRP6, but not of RRP4, to sites of radiation-induced DNA

Fig. 1. RRP6 is relocated to DSBs in response to ionizing radiation. (A) S2-RRP6–V5 cells were irradiated, allowed to recover for 30 min, and analyzed byimmunoprecipitation and high-performance liquid chromatography followed by tandem mass spectrometry. The Venn diagram shows the number of proteinsassociated with RRP6–V5 in each condition. (B) Immunofluorescence imaging of RRP6–V5 and cH2Av in S2 cells 10 min after cIR. Overlapping pixels are shownin white in the colocalization image. Scale bar: 5 mm. (C) The percentages of cells showing RRP6–V5-rich foci, RRP4–V5-rich foci or cH2Av-rich foci werequantified in samples prepared as in B. The histogram also shows the percentages of cells showing RRP6-rich foci in non-transfected, parental S2 cells stained withan antibody against the endogenous RRP6 protein. (D) Example of colocalization between RRP6–V5 and cH2Av. The graph shows the relative intensities foreach channel, in arbitrary units, along the line defined by the white arrow. The small arrows in the graph show overlapped foci. (E) Quantification of colocalizationpatterns. The cH2Av foci were classified as being overlapped, juxtaposed or non-overlapped with RRP6–V5 foci. (F) S2 cells were incubated with or withoutCGK733 before irradiation and staining with antibodies against either cH2Av or RRP6. The histogram shows the percentage of cells showing RRP6-rich or cH2Av-rich foci. The histograms in C, E and F show the mean6s.d. obtained from three independent experiments. *P,0.05 (Student’s t-test).

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 1097–1107 doi:10.1242/jcs.158733

1098

Jour

nal o

f Cel

l Sci

ence

damage. This relocalization is dependent on the activity of theDDR kinases ATM and ATR.

Depletion of RRP6 impairs the efficient repair of DSBs inS2 cellsWe next analyzed whether the kinetics of assembly and clearance

of DDR foci were affected by depletion of RRP6. S2 cells weretreated with doubled-stranded RNA (dsRNA) to deplete RRP6 byRNA interference (RNAi), and control cells were treated in

parallel with dsRNA complementary to GFP (Fig. 2A). The cIRtreatment caused a significant increase in the percentage of cellswith DDR foci, and there were no differences between RRP6-

depleted and control cells 1 h after irradiation (Fig. 2B,C).However, 20 h after the irradiation the fraction of cells thatdisplayed prominent cH2Av-rich foci was significantly higher in

the cells depleted of RRP6 (Fig. 2B,C), which suggests thatRRP6 is required for efficient DNA repair.

We also analyzed whether the depletion of RRP6 affected thesensitivity of S2 cells to cIR, which would be expected if RRP6

were necessary for DNA repair. The growth rate of the RRP6-depleted cells irradiated with 10 Gy was almost 50% lower thanthat of GFP control cells (Fig. 2D), and RRP6-depleted cellsirradiated with doses as high as 50 Gy did not recover (Fig. 2E).

These experiments demonstrate that physiological levels of RRP6are required for the efficient repair of DSBs, and that depletion ofRRP6 renders S2 cells more sensitive to cIR.

RRP6 interacts with RAD51 and is required for the assemblyof DDR fociThe recruitment of the ATM and ATR kinases to the DSBs andthe local phosphorylation of H2Av trigger the assembly of eitherthe NHEJ or homologous recombination machineries, depending

on the nature of the damage and the stage in the cell cycle. Wehypothesized that RRP6 is involved in the homologous

Fig. 2. Depletion of RRP6 affects the DDR. S2 cells were depleted of RRP6 or RRP4 by RNAi (KD). Control cells were processed in parallel (GFP-KD).(A) The depletion was monitored by quantitative RT-PCR. The RNA levels were normalized to the levels of Actin 5C mRNA. (B) The cells were irradiated andstained with anti-cH2Av antibody. The image shows examples of the most common staining patterns in each condition. (C) Quantitative analysis of theexperiment described in B. The histogram shows the mean6s.d. from four independent experiments. (D) S2 cells were treated as above and the growth rateswere measured 3 days after irradiation. The histogram shows data from four independent experiments. (E) S2 cells were exposed to 50 Gy, and cell growth wasanalyzed 20 days after the irradiation. RRP6-depleted cells did not grow, but S2 colonies grew in the GFP-KD control cultures. *P,0.05 (Student’s t-test).Scale bars: 5 mm (B); 50 mm (E).

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 1097–1107 doi:10.1242/jcs.158733

1099

Jour

nal o

f Cel

l Sci

ence

recombination pathway because homologous recombination is thepredominant repair mechanism in D. melanogaster (Preston et al.,

2002), because homologous recombination requires short RNAs(Gao et al., 2014) and because the homologous recombinationreplication factor RPA1 (also known as RFA1 or RpA-70 inDrosophila) is a putative component of the exosome interaction

network in both S2 cells (supplementary material Table S2) andhuman cells (Lubas et al., 2011). We thus analyzed therecruitment of the homologous recombination machinery to

DDR foci using an antibody against RAD51 (reviewed by Krejciet al., 2012) as a marker for homologous recombination. At 1 hafter irradiation, 81% of the GFP control cells showed RAD51-

rich foci whereas the frequency of RAD51-rich foci did notincrease in RRP6-depleted cells (Fig. 3A,B). This effect wasspecific for RRP6, because the formation of RAD51 foci in cells

depleted of RRP4 was the same as in the GFP controls. It isimportant to point out that the depletion of RRP6 did not reduce

the levels of RAD51 protein (Fig. 3C). These observations led usto conclude that RRP6 is required for the efficient assembly of thehomologous recombination machinery at the DSBs.

We hypothesized that RRP6 interacts physically with

components of the DDR machinery based on the relocalizationof RRP6 to DDR foci and on the fact that RRP6 is required for therecruitment of homologous recombination factors. To test this

possibility, we carried out co-immunoprecipitation experimentsin RRP6–V5-expressing S2 cells, and we searched forinteractions of RRP6 with proteins detected in the mass

spectrometry analysis, including histone H1, histone H2Av andRPA1/RFA1 (supplementary material Tables S1, S2). We couldnot demonstrate interactions of RRP6 with any of these proteins

Fig. 3. RRP6 interacts with RAD51 and is required for the assembly of RAD51-rich foci in S2 cells. S2 cells were depleted of RRP6 or RRP4 by RNAi(KD), irradiated and allowed to recover for 1 h or 20 h before fixation and immunofluorescence with anti-RAD51 antibody. (A) Examples of the most commonstaining pattern in each condition. (B) Quantitative analysis of the number of cells with RAD51 foci in samples depleted of RRP6 or RRP4, and in controlsamples. The histogram shows the mean6s.d. from four independent experiments. *P,0.05 (Student’s t-test). (C) The levels of RRP6 and RAD51 wereanalyzed by western blotting. An antibody against the endogenous RRP6 protein was used in this experiment. RRP6 is indicated with an arrow, and theadditional band observed in the blot is due to crossreactivity of the antibody with an unidentified protein. (D) The expression of S2-RRP6–V5 was induced with100 mM CuSO4. Nuclear protein extracts were prepared from irradiated S2-RRP6–V5 cells and used for immunoprecipitation using the anti-RAD51 antibody.The immunoprecipitated proteins were analyzed by western blotting using the anti-V5 antibody. An anti-NONA antibody was used as a negative control.NONA is an RNA-binding protein that is not co-immunoprecipitated by the RAD51 antibody. (E) S2-RRP6–V5 cells were fixed before or 10 min after irradiation(10 Gy) and a PLA was carried out with antibodies against RAD51 and V5. The graph shows the number of PLA dots per cell in each condition. The mean isindicated. *P,0.001 (Student’s t-test with correction for multiple comparisons; n5200 cells analyzed in each condition). Scale bars: 5 mm.

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 1097–1107 doi:10.1242/jcs.158733

1100

Jour

nal o

f Cel

l Sci

ence

(data not shown). We found instead that RRP6 co-immunoprecipitates specifically with RAD51 (Fig. 3D). The

association of RRP6 with RAD51 was further confirmed using aproximity ligation assay (PLA) to reveal protein–proteininteractions in situ (Fig. 3E). The RAD51–RRP6 interactionwas detected in both control and irradiated cells, but

quantification of PLA data revealed a significant increase ofRRP6–RAD51 association after exposure to cIR (right panel inFig. 3E).

In summary, the experiments reported above support theconclusion that RRP6 interacts with RAD51 and is required forthe recruitment of RAD51 to DDR foci.

Overexpression of a catalytically inactive RRP6 mutantinhibits the recruitment of RAD51 to DDR fociWe used a catalytically inactive RRP6 mutant to analyze whetherthe ribonucleolytic activity of RRP6 is necessary for homologousrecombination. This mutant, RRP6-Y361A–V5, carries a singleamino acid substitution in the active site (Phillips and Butler,

2003; Fig. 4A). In a first series of experiments, we overexpressedeither RRP6–V5 or RRP6-Y361A–V5 in S2 cells (Fig. 4B), andanalyzed the assembly and clearance of cH2Av foci (Fig. 4C).

The expression of RRP6-Y361A–V5 did not affect the cIR-induced phosphorylation of cH2Av compared to the expression ofthe wild-type RRP6 protein, but 20 h after the irradiation the

fraction of cells that displayed prominent cH2Av-rich foci wassignificantly higher in cells expressing the catalytically inactiveRRP6-Y361A–V5 than in cells that expressed the wild-type

RRP6–V5 (Fig. 4C). This result suggests that the activity ofRRP6 is required for efficient DNA repair.

We also analyzed the effect of overexpressing RRP6-Y361A–V5 on the recruitment of RAD51 to DSBs. The formation of

RAD51 foci was inhibited (from 64% to 27%) in cells thatoverexpressed the inactive RRP6-Y361A–V5 mutant (Fig. 4D).Moreover, overexpression of RRP6-Y361A–V5 significantly

reduced the growth rate of the cells after irradiation with 10 Gy(Fig. 4E). Importantly, the overexpression of RRP6–V5 or RRP6-Y361A–V5 did not affect the abundance of the RAD51 protein

(Fig. 4B), and the Y361A mutation did not inhibit the interactionof RRP6 with RAD51 (Fig. 4F). These results link the assemblyof RAD51 foci to the catalytic activity of RRP6, and suggest thata 39R59 RNA degradation event is required for homologous

recombination. The fact that RRP6 interacts with RAD51 and isrecruited to DSBs supports the conclusion that RRP6 contributesto local RNA degradation at or near the DSBs.

Depletion of EXOSC10 inhibits recruitment of RAD51 toradiation-induced DSBs in HeLa cellsGiven the structural and functional conservation of the exosome,we hypothesized that the role of RRP6 in DNA repair isconserved in human cells. In a first series of experiments, we

asked whether EXOSC10, the human RRP6 ortholog, is recruitedto sites of radiation-induced DNA damage in HeLa cells. The useof homologous recombination is limited to the S and G2 phases ofthe cell cycle in mammalian cells. For this reason, and in order to

be able to study homologous-recombination-related events, westudied the localization of EXOSC10 in HeLa cells synchronizedin G2 using an antibody against the endogenous protein.

EXOSC10 was predominantly concentrated in the nucleoli ofnon-irradiated cells (Fig. 5A), and cIR induced a progressiverelocation of EXOSC10 from the nucleoli to the nucleoplasm.

This relocation could be detected 15 min after irradiation and was

very prominent after 45 min (Fig. 5A,B). EXOSC10 did notaccumulate in prominent foci, nor was EXOSC10 concentrated to

areas irradiated by ultraviolet laser irradiation (data not shown).However, PLA experiments with antibodies against theendogenous EXOSC10 and RAD51 proteins revealed that asignificant increase in the interaction was induced by cIR

(Fig. 5C). These results demonstrate that EXOSC10 isrelocalized to the nucleoplasm in response to cIR and suggestthat EXOSC10 interacts with RAD51.

We then asked whether EXOSC10 depletion by smallinterfering RNA (siRNA) in human cells (Fig. 6A) had anyeffect on the assembly and disassembly of DDR foci. The

percentage of cells with cH2AX-positive foci was almost 100% at1 h after cIR, regardless of the siRNA treatment (Fig. 6B,C). ThecH2AX frequency recovered to background levels within 24 h in

control cells treated with scrambled (Scr) siRNA, but was stillhigh in EXOSC10-depleted cells, which implies a delay in theclearance of the DDR foci in the depleted cells. Furthermore, theRAD51-rich foci that formed in EXOSC10-depleted cells were

fewer and did not colocalize with the cH2AX-rich foci(Fig. 6B,D). Importantly, depletion of EXOSC10 did not affectthe levels of RAD51 (Fig. 6E) and did not cause arrest in G1

(data not shown), which would in itself suppress homologousrecombination. These results support the conclusion thatEXOSC10 is required for the recruitment of RAD51 to DSBs

in human cells. In agreement with a defect in DNA repair,depletion of EXOSC10 increased the sensitivity of the cells tocIR (Fig. 6F). Similar immunofluorescence experiments were

carried out with antibodies against 53BP1, a NHEJ factor.Depletion of EXOSC10 did not affect the recruitment of 53BP1to sites of radiation-induced DNA damage (Fig. 6G), whichsuggests that the role of EXOSC10 in DNA repair is restricted to

the homologous recombination pathway.Cells depleted of EXOSC10 presumably contain large amounts

of aberrant RNA that could engage in DNA–RNA hybrid

formation and hijack RAD51 to DNA–RNA hybrids (Wahbaet al., 2013). However, the depletion of EXOSC10 per se did notinduce the appearance of RAD51-rich foci (Fig. 6D), which

argues against indirect effects due to competition between DNA–RNA hybrids and DSBs.

Given the recently reported role of AGO2 in homologousrecombination (Gao et al., 2014), we also considered the possibility

that AGO2 is mislocalized in EXOSC10-depleted cells owing toincreased levels of dsRNA, for example pre-rRNA processingintermediates, which could indirectly cause homologous recombination

defects. However, immunofluorescence experiments did not revealany effect of EXOSC10 depletion on the distribution of AGO2(data not shown).

EXOSC10 is recruited to I-PpoI-induced DSBs and isnecessary for the assembly of RAD51 at the site-specific DSBsWe transfected HeLa cells with a plasmid coding for the homingendonuclease I-PpoI to generate site-specific DSBs and furtheranalyze the requirement of EXOSC10 for the recruitment of

RAD51 to sites of DNA damage. I-PpoI has a cleavage site in the28S rDNA (Monnat et al., 1999) and cells transfected with theI-PpoI expression plasmid often showed cH2AX-rich foci

associated with the nucleolus (Fig. 7A). We used the I-PpoIsystem to determine whether EXOSC10 is recruited to DSBs, assuggested by the PLA experiment reported in Fig. 5C. HeLa cells

were transfected with the I-PpoI expression plasmid and analyzed

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 1097–1107 doi:10.1242/jcs.158733

1101

Jour

nal o

f Cel

l Sci

ence

Fig. 4. The ribonucleolytic activity of RRP6 is required for the efficient recruitment of RAD51 to DSBs. S2 cells that overexpress the catalytically inactiveRRP6-Y361A–V5 mutant were used to analyze whether the role of RRP6 in homologous recombination involved RNA degradation. (A) The figure shows theconserved Tyr361 in the sequences of three RRP6 orthologs. (B) The overexpression of RRP6–V5 and RRP6-Y361A–V5 was analyzed with the anti-RRP6antibody. Overexpression of RRP6–V5 or RRP6-Y361A–V5 did not affect the levels of RAD51. ‘Empty’ indicates control S2 cells. (C) S2 cells thatoverexpressed either RRP6–V5 or RRP6-Y361A–V5 were irradiated and allowed to recover for 1 h or 20 h before fixation and immunofluorescence with anti-cH2Av antibody as in Fig. 3B, to analyze the effect of the overexpression on the disassembly of DDR foci. The histogram shows the mean6s.d. from threeindependent experiments. (D) S2 cells that overexpressed either RRP6–V5 or RRP6-Y361A–V5 were irradiated and analyzed by immunofluorescence 1 h later.The percentages of cells with cH2Av foci and RAD51 foci were quantified in three independent experiments. The histogram shows the mean6s.d. (E) Thegrowth rates of S2 cells that overexpressed either RRP6–V5 or RRP6-Y361A–V5 before and after irradiation were analyzed as in Fig. 2D. (F) Nuclear proteinextracts were prepared from irradiated S2-RRP6-Y361A–V5 cells and used for immunoprecipitation using the anti-RAD51 antibody as in Fig. 3D. Theimmunoprecipitated proteins were analyzed by western blotting using the anti-V5 antibody. *P,0.05, **P,0.01 (Student’s t-test).

Fig. 5. cIR-induced relocalization ofEXSOC10 in HeLa cells. (A) HeLa cells wereirradiated (5 Gy) and fixed at different timesafter irradiation, as indicated. Non-irradiatedcontrol cells were also fixed and stained inparallel. (B) The histogram shows theEXOSC10 intensity ratios between nucleoplasmand nuclei at different times. *P,0.05, **P,0.01(Student’s t-test). (C) HeLa cells were fixedbefore or 10 min after irradiation (5 Gy) andPLA staining was carried out with antibodiesagainst RAD51 and EXOSC10. The graphshows the number of PLA dots per cell. Themean is indicated. ***P,0.001 (Student’s t-testwith correction for multiple comparisons; n5120cells analyzed in each condition). Scale bars:10 mm.

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 1097–1107 doi:10.1242/jcs.158733

1102

Jour

nal o

f Cel

l Sci

ence

by chromatin immunoprecipitation followed by real-time

quantitative PCR (ChIP-qPCR) using the anti-EXOSC10antibody and PCR primers designed to amplify two regionslocated at each side of the I-PpoI-induced DSBs. The level ofEXOSC10 detected at the DSBs in cells transfected with I-PpoI

was significantly higher than that of mock-transfected cells. This

increase was specific for the I-PpoI target site as shown by thefact that an unrelated genomic region was unaffected (Fig. 7B).In summary, these results show that EXOSC10 is recruited to I-PpoI-induced DSBs in HeLa cells.

Fig. 6. Depletion of EXOSC10 in HeLa cells impairs the assembly of RAD51 in cIR-induced DDR foci. (A) HeLa cells were transfected with scrambled(Scr), EXOSC10 or EXOSC2 siRNAs (KD), and the depletion was monitored by quantitative RT-PCR. The relative EXOSC10 mRNA levels were normalized tothe levels of ARPP mRNA. (B) EXOSC10-depleted cells and control Scr cells were irradiated, fixed 1 h after irradiation and double stained with anti-RAD51and anti-cH2AX antibodies. Colocalized pixels are shown in white. Scale bar: 10 mm. (C) The histogram shows the percentage of cells with cH2AX foci at 1, 6and 24 h after irradiation. (D) HeLa cells were irradiated and fixed 1 h after irradiation. The histogram shows the number of RAD51 foci per cell (black bars)and the number of RAD51 foci that overlapped with cH2AX foci (gray bars). The data represent the mean6s.d. from three independent experiments. (E) Thelevels of EXOSC10 and RAD51 were analyzed by western blotting in control Scr cells and in EXOSC10-depleted cells. (F) Control Scr cells and EXOSC10-depleted cells were irradiated with 2, 4 or 6 Gy, and diluted to single cell per well in 96-well plates. Cell growth was analyzed 7 days after the irradiation.The relative growth rates for each condition are expressed as fraction of wells with growth relative to non-irradiated cultures. The error bars represent s.d. fromone experiment with four technical replicates. (G) The histogram shows the number of 53BP1-rich foci per cell in control Scr cells and in EXOSC10-depletedcells. The cells were irradiated with 5 Gy and fixed 1 h later. The error bars represent s.d. from two different experiments with two technical replicates each.*P,0.05, **P,0.01 (Student’s t-test).

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 1097–1107 doi:10.1242/jcs.158733

1103

Jour

nal o

f Cel

l Sci

ence

Next we depleted EXOSC10 by siRNA as in Fig. 6 andanalyzed the phosphorylation of H2AX and the recruitment of

RAD51 to the I-PpoI-induced foci by immunofluorescence. In63% of the control (Scr) cells, RAD51 was recruited to the I-PpoI-induced DSBs, whereas only 27% of the EXOSC10-

depleted cells showed association of RAD51 with cH2AX-richfoci (Fig. 7C,D). We also analyzed the recruitment of RAD51 toI-PpoI-induced DSBs by ChIP-qPCR. In control (Scr) cells, thelevels of RAD51 detected in the vicinity of the DSBs were

between two and five times higher in cells transfected with I-PpoIthan in mock-transfected cells, whereas depletion of EXOSC10abolished the recruitment of RAD51 to the I-PpoI-induced DSBs

(Fig. 7E). These results confirmed that EXOSC10 is required forthe recruitment of RAD51 to DSBs.

DISCUSSIONRRP6 and EXOSC10 are predominantly nuclear proteins that

participate in a plethora of RNA metabolic processes to whichthey contribute with their 39R59 ribonucleolytic activity(reviewed by Chlebowski et al., 2013). We show that RRP6

and EXOSC10 are redistributed in response to DNA damage cuesin insect and mammalian cells, respectively, and that a fraction ofRRP6/EXOSC10 is rapidly recruited to DSBs. This finding isconsistent with data from proteomics studies in which it has been

shown that exosome components interact physically with factorsinvolved in the DDR. A systematic analysis of protein–proteininteractions involving BRCT-domain-containing proteins, for

example, has identified exosome subunits among the interactionpartners of mediator of DNA-damage checkpoint 1 (MDC1)

Fig. 7. Depletion of EXOSC10 in HeLa cells impairsthe assembly of RAD51 at site-specific DSBs. HeLacells were treated with siRNAs to knockdown (KD)EXOSC10 as in Fig. 5, and were transfected withpOPRSVI/MCS-IPpoI to express I-PpoI and producesite-specific DSBs. (A) Transfected cells were fixed24 h after transfection with the pOPRSVI/MCS-IPpoIplasmid and stained with anti-cH2AX. An antibodyagainst MYBBP1 was used as a marker for thenucleolus. (B) Cells transfected with the pOPRSVI/MCS-IPpoI plasmid and mock-transfected cells wereanalyzed by ChIP-qPCR using the anti-EXOSC10antibody. The histogram shows qPCR results withprimers designed to amplify DNA sequences located atpositions 2128 to 243 and +108 to +195 relative to anI-PpoI cleavage site. An unrelated region was alsoquantified in parallel. The ChIP-qPCR results areexpressed relative to the input standard curve andnormalized to ARPP. (C) Control Scr cells andEXOSC10-depleted cells were transfected with thepOPRSVI/MCS-IPpoI plasmid, fixed 24 h afterpOPRSVI/MCS-IPpoI transfection, and stained withantibodies against RAD51 and cH2AX. Thearrowheads point at colocalized RAD51 and cH2AXfoci typical of control Scr cells. (D) The histogramshows the percentage of cells with cH2AX-RAD51colocalization after pOPRSVI-IPpoI transfection in Scrand EXOSC10-depleted HeLa cells. Error barsrepresent the s.d. from two different experiments withtwo technical replicates each. (E) Chromatin wasprepared from control Scr cells and EXOSC10-depleted cells transfected with the pOPRSVI/MCS-IPpoI plasmid as above. Mock-transfected (without I-PpoI) cells were processed in parallel. ChIP-qPCRexperiments were carried out using an anti-RAD51antibody. The figure shows qPCR results obtained withprimer pairs designed to amplify DNA sequences in thevicinity of the I-PpoI cleavage site, as in B. The ChIP-qPCR signals were calculated using the input standardcurve and normalized to Arpp, and the histogramshows the fold between I-PpoI-transfected cells andmock-transfected cells. The error bars represent s.d.from two different experiments with two technicalreplicates each. *P,0.05, **P,0.01 (Student’s t-test).Scale bars: 10 mm.

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 1097–1107 doi:10.1242/jcs.158733

1104

Jour

nal o

f Cel

l Sci

ence

(Woods et al., 2012), a protein that interacts with cH2AX nearDSBs to promote the recruitment of the repair machinery

(Stewart et al., 2003). The results presented here are consistentalso with data from a genome-wide siRNA screen in whichdepletion of EXOSC10 inhibited homologous recombination in ahuman cell line (Adamson et al., 2012).

We have focused our analysis on RRP6/EXOSC10, but themultisubunit structure of the exosome makes it difficult toestablish whether the entire exosome has a role in DNA repair.

We could not detect any relocalization of the core exosomesubunit RRP4 in response to cIR, and RRP4 depletion did notaffect the recruitment of RAD51 to cIR-induced DSBs in S2

cells. These observations suggest that, at least in Drosophila, theinvolvement of RRP6 in homologous recombination does notinvolve the exosome core. This possibility would be compatible

with previous reports on the core-independent function of RRP6in cell cycle progression and error-free mitosis in Drosophila

(Graham et al., 2009).RRP6 and EXOSC10 interact with RAD51, as shown by co-

immunoprecipitation and PLA experiments. Moreover, depletionof RRP6 in S2 cells or EXOSC10 in HeLa cells prevents therecruitment of RAD51 to DSBs, regardless of whether the

damage is induced by radiation or by endonuclease cleavage. Bycontrast, depletion of EXOSC10 in HeLa cells did not affect therecruitment of the NHEJ factor 53BP1 to the damaged sites.

These findings link RRP6/EXOSC10 specifically to thehomologous recombination pathway.

The most interesting issue that arises from the results presented

here is the molecular mechanism by which RRP6/EXOSC10contributes to the recruitment of RAD51 to DSBs. RRP6 andEXOSC10 are distributive RNase D-like exonucleases with39R59 activity (reviewed by Houseley et al., 2006; Lykke-

Andersen et al., 2011). We constructed a catalytically inactiveRRP6 mutant, RRP6-Y361A–V5, and we show thatoverexpression of RRP6-Y361A–V5 in S2 cells reproduces the

effects of RRP6 depletion: delayed repair of cIR-induced DSBs,as inferred from clearance of cH2Av foci; impaired recruitmentof RAD51 to DSBs; and increased sensitivity to cIR.

Overexpression of RRP6-Y361A–V5 did not affect the cellularlevel of RAD51 and did not disrupt the interaction of RRP6 withRAD51. Taken together, these observations strongly support theconclusion that the 39R59 ribonucleolytic activity of RRP6 is

needed for the recruitment of RAD51 to DSBs.DNA damage induces bidirectional transcription and formation

of short RNAs in the vicinity of the damaged sites (reviewed by

d’Adda di Fagagna, 2014; Sharma and Misteli, 2013). These shortRNAs, referred to as DDR RNAs (‘DDRNAs’) or ‘diRNAs’(Francia et al., 2012; Wei et al., 2012), are DROSHA- and

DICER-dependent RNA products that act together with AGO2 tofacilitate the recruitment of RAD51 to DSBs (Gao et al., 2014).The prevalent roles of RRP6/EXOSC10 in the processing of other

non-coding RNAs suggest that RRP6 and EXOSC10 mightparticipate in (DDRNA) biogenesis, which would link RRP6/EXOSC10 to DSB repair. However, inhibition of DDRNAbiogenesis in mammalian cells affects not only the homologous

recombination pathway but also the recruitment of early NHEJfactors such as 53BP1 to DSBs (Francia et al., 2012), whereas ourexperiments with EXOSC10 depletion failed to reveal defects in

the NHEJ pathway. This difference argues against a mechanismin which EXOSC10 is necessary for the biogenesis of DDRNAs.

Recent studies have shown that accumulation of dsRNAs in the

chromatin has deleterious effects and leads to genomic instability

(Bhatia et al., 2014; White et al., 2014). In the DNA repair context,it is easy to envision that the presence of complementary RNA

strands derived from bidirectional transcription near the DSBscould compromise the assembly of RAD51 nucleofilaments on theDNA. Our results are consistent with a hypothetical mechanism inwhich the 39R59 ribonucleolytic activity of RRP6/EXOSC10 is

required in the vicinity of the DSBs to degrade RNA molecules thatcould inhibit the recruitment of homologous recombination factors.This hypothesis is compatible with recent observations on the role

of Rrp6 and Trf4 in the loading of RPA to single-stranded DNA(ssDNA) tails generated at DSBs and replication forks in S.

cerevisiae (Manfrini et al., 2014).

MATERIALS AND METHODSCell culture and c-irradiationCulture of D. melanogaster S2 cells and induction of protein expression

with CuSO4 were carried out according to the instructions of the

Drosophila Expression System manual from Invitrogen. S2-RRP6–V5

cells were characterized by Hessle et al. (Hessle et al., 2009). HeLa cells

were grown at 37 C in Dulbecco’s modified Eagle’s medium (HyClone,

Thermo Scientific) supplemented with 10% fetal bovine serum and 1%

penicillin-streptomycin in a humidified incubator at 10% CO2. DSBs

were induced in HeLa cells using a 137Cs Gammacell 1000 radiation

source (Nordion) at a photon dose rate of 6.4 Gy/min, and in S2 cells

using a Scanditronix chamber (Radiation Products Design) at a photon

dose rate of 0.4 Gy/min. S2 cells and HeLa cells were irradiated with

10 Gy and 5 Gy, respectively. In some experiments, S2 cells were treated

overnight with 25 mg/ml CGK733 (Sigma Aldrich) before irradiation.

Cell growth analysisS2 cells were knocked down using dsRNA against RRP6 or GFP as a

control, exposed to 10 Gy of cIR, left in culture for 72 h, and counted.

Non-irradiated samples were cultured in parallel as the control. The cell

growth rate in each plate was calculated by dividing the number of cells

in the plate at 72 h by the number of cells at 0 h. The growth rate for each

condition (RRP6-KD and GFP-KD) was obtained by dividing the growth

rate of the irradiated samples by that of the non-irradiated samples. For

cell cycle analysis, HeLa cells were stained with 7-aminoactinomycin D

and analyzed in a FACSCalibur (BD Biosciences).

Site-directed mutagenesisA single amino-acid mutation in RRP6-V5 was made by oligonucleotide-

directed site-specific mutagenesis using the site-directed mutagenesis kit

(Invitrogen) on the pMT-Rrp6 plasmid (Hessle et al., 2009). The

resulting protein was given the name RRP6-Y361A–V5. The sequences

of the oligonucleotides used were 59-CAGCTGGTGGACGCCGCTCG-

TCAGGATAC-39 (oligonucleotide Ae125) and 59-GTATCCTGACGA-

GCGGCGTCCACCAGCTG-39 (oligonucleotide Ae126). The mutation

was verified by DNA sequence analysis. The plasmids for the expression

of RRP6–V5 and RRP6-Y361A–V5 were co-transfected into S2 cells

together with a hygromycin selection plasmid (Invitrogen), and stable

transfectants were selected with 0.3 mg/ml hygromycin. ‘Empty’ control

cells were transfected with the hygromycin resistance plasmid alone.

AntibodiesA rabbit anti-V5 antibody (ab9116, Abcam) was used for

immunoprecipitation and a mouse anti-V5 (p/n 46-0705, Invitrogen) was

used for immunofluorescence and western blotting. The anti-cH2Av

antibody was from Rockland, and the anti-AGO2 and anti-RPA were from

Abcam (ab57113 and ab97338, respectively). The rabbit antibodies against

Drosophila RAD51 (Chiolo et al., 2011) and histone H1 were kindly

provided by James Kadonaga (Division of Biological Sciences, University

of California San Diego, CA). The anti-RRP6 antibody was a serum

from rabbits immunized with the peptide CLVQMSTRSKDYIFDT.

Immunization was carried out at Agrisera AB. For detection of NONA in

co-immunoprecipitation experiments we used the monoclonal antibody

Bj6, kindly provided by Harald Saumweber (Institute of Biology, Humboldt

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 1097–1107 doi:10.1242/jcs.158733

1105

Jour

nal o

f Cel

l Sci

ence

University Berlin, Germany). The following antibodies were used against

human proteins: anti-cH2AX from Cell Signaling Technology, mouse

and rabbit anti-human-RAD51 from Abcam (ab88572 and ab63801,

respectively), anti-RAD51 (ChIP grade; ab76458), anti-EXOSC10

from Santa Cruz Biotechnology (sc-374595), anti-53BP1 from Novus

Biologicals (NB100-904), and anti-MYBBP1 from Abcam (ab89121). The

non-specific control antibody used for ChIP is an anti-IgG (ab46540).

Secondary antibodies conjugated to FITC and Alexa Fluor 594 were from

Jackson ImmunoResearch. Horseradish peroxidase (HRP)-conjugated

antibodies (Dako) were used for western blotting.

ImmunoprecipitationCells were homogenized in PBS containing 0.2% NP-40 substitute

(Sigma), Protease Inhibitor-Complete EDTA Free (Roche), 25 mM b-

glycerophosphate (Sigma) and Phostop-Easy Pack (Roche) and the

homogenate was centrifuged at 1500 g. The pellet was resuspended in

detergent-free buffer, sonicated and supplemented with 0.1% NP-40

substitute (Roche) and 0.1 mg/ml RNase A (Sigma). The sample was

centrifuged at 16,000 g and the supernatant was used as input to the

immunoprecipitation. Primary antibodies were used at 1–3 mg/ml and the

bound proteins were isolated on Sepharose beads.

SDS-PAGE and western blottingProteins were separated by SDS-PAGE using a Mini-Protean II system

(BioRad), and transferred to polyvinylidene fluoride membranes

(Millipore) in Tris-glycine buffer with 0.02% SDS and 4 M urea using

a semi-dry electrophoretic transfer cell (BioRad). The membranes were

probed with antibodies following standard procedures. Chemiluminiscent

detection was carried out using the ECL kit (GE Healthcare).

Liquid chromatography and mass spectrometryThe liquid chromatography and mass spectrometry analysis was carried

out at the Proteomics Karolinska (PK/KI) facility. Selected lanes from

SDS-PAGE gels were cut out, and the proteins were in-gel reduced,

alkylated and digested with modified sequence-grade trypsin (Promega).

The resulting peptides were extracted, resolved in a Nano-Easy-HPLC

(Thermo Scientific) and analyzed using a Q-Exactive instrument

(Thermo Scientific) modified with a nanoelectrospray ion source

(Thermo Scientific). The proteins contained in each sample were

identified against the SwissProt database (Concat_2012.062). The

proteins with a coverage value above 2% were selected and are

presented in supplementary material Tables S1 and S2.

RNAiLong double-stranded RNAs (dsRNAs) were used for RNAi in S2 cells

as described by Hessle et al. (Hessle et al., 2009). HeLa cells were seeded

in 60-mm plates 1 day before transfection, with a confluence of 30–50%,

in 4 ml DMEM supplemented with 10% FBS without penicillin-

streptomycin. Transfection was performed with 8 ml of Lipofectamine

RNAiMAX (Invitrogen) and 10 nM of siRNA in 1 ml Opti-MEM

(Gibco) per plate. The siRNA oligonucleotides used were 59-

GUGCGAGGGGGUUGUAAUCTT-39 (Scrambled control) and 59-

GCUGCAGCAGAAGAGGCCATT-39 (EXOSC10-KD).

ImmunofluorescenceS2 cells were plated onto polylysine-coated slides (Menzel-Glasser) and

allowed to attach before irradiation. HeLa cells were seeded on

coverslips. The cells were fixed with 3.7% formaldehyde in PBS for

10 min at room temperature, permeabilized in 0.1% Triton X-100 for

16 min, and washed with PBS before incubation in a solution of 5% BSA

in PBS for 40 min. Immunostaining was carried out following standard

procedures. The slides were mounted with Vectashield containing 1.5 mg/

ml DAPI (Vector Laboratories) and analyzed either in an epifluorescence

(Axioplan2, Carl Zeiss) or confocal microscope (LSM780, Carl Zeiss).

Confocal images were acquired using a 636oil immersion objective. The

lasers used were 405 nm for DAPI staining, 488 nm argon laser for FITC

and 561 nm for Alexa Fluor 594. The thickness of the optical sections

was 0.9 mm for Hela and 1.1 mm for S2 cells. Quantitative analyses of the

number of cells with foci or the number of foci per cell were carried out

in random areas of the preparations. Statistical significance was assessed

using paired Student’s t-tests. The ‘Colocalization finder’ and ‘Plot

profile’ functions of ImageJ 1.42q (Schneider et al., 2012) were used for

colocalization analysis and measurement of profile intensities,

respectively.

Proximity ligation assayCells were fixed 10 min after irradiation with 3.7% formaldehyde in PBS

for 10 min and permeabilized with 0.1% Triton X-100 for 15 min at

room temperature. A blocking solution of 5% BSA was added for 30 min

followed by 1 h incubation of the primary antibodies diluted in 1% BSA.

The proximity ligation assay (PLA) was carried out using the Duolink

PLA in situ kit (Olink) following the manufacturer’s protocol. PLA

experiments in S2-RRP6–V5 cells were carried out with antibodies

against V5 and RAD51. PLA experiments in HeLa cells were carried out

with antibodies against endogenous EXOSC10 and RAD51. The

preparations were analyzed in an LSM 780 confocal microscope (Carl

Zeiss). The ‘In cell’ module of Imaris v.7.7.2 was used for automated

quantification of PLA results.

Site-specific DNA double-strand breaksA cDNA corresponding to the open reading frame (ORF) of

endonuclease I-PpoI was amplified by PCR using as a template the

pICE-HA-NLS-I-PpoI plasmid (Addgene) and oligonucleotides 59-ATA-

GGTACCTATGTACCCCTACGACGTGCCC-39 and 59-TATCTCGA-

GTGTTGCACCACGAAGTGGGAGC-39, and cloned between the KpnI

and XhoI sites of the pOPRSVI/MCS vector (Agilent Technologies). The

resulting pOPRSVI/MCS-IPpoI plasmid was transfected to HeLa cells

using a Lipofectamine 3000 transfection kit (Invitrogen) and the cells

were harvested 24 h after the transfection.

Chromatin immunoprecipitationHeLa cells were transfected with the pOPRSVI/MCS-IPpoI plasmid and

harvested 24 h after transfection. Chromatin was extracted, sonicated in a

Bioruptor (Diagenode) and immunoprecipitated with an antibody against

EXOSC10 or RAD51 essentially as described by Eberle et al. (Eberle

et al., 2010). The DNA in input and immunoprecipitated samples was

purified using Zymo-ChIP Clean & Concentrator (Zymo Research) and

analyzed by qPCR in a RotorGene (Qiagen) using Power SYBR Green

PCR Master Mix (Kapa Biosystems). Two primer-pairs were used to

amplify sequences in the proximity of I-PpoI cleavage sites: 59-AATCA-

GCGGGGAAAGAAGAC-39 and 59-GGGGCCTCCCACTTATTCTA-39

amplified a sequence located at 2128 to 243 bp relative to a I-PpoI

cleavage site in chromosome 21, 59-GTCTTCTTTCCCCGCTGATT-39

and 59-CGAGATTCCCACTGTCCCTA-39 amplified a sequence located

at +108 to +195 bp from a I-PpoI cleavage site in chromosome 8. An

unrelated region (Rad50) was analyzed in parallel as a control using

primers 59-CACACCCTGTGAGAAACAC-39 and 59-CAAGCCACTG-

CCATCTCTAA-39. ARPP primers 59-GCACTGGAAGTCCAACTACT-

39 and 59-TGAGGTCCTCCTTGGTGAACAC-39 were used to normalize

the samples. Data from two technical replicates of two independent

experiments were analyzed. The data from each sample was related to the

input of that same experiment and normalized to ARPP.

AcknowledgementsWe thank J. Kadonaga for providing antibodies, S. Bohm and S. Haghdoost fortechnical help, A. K. Ostlund Farrants for constructive discussions and G. W.Farrants for language editing. We are also grateful to D. Rutishauser (PKKI), A.S.Hoglund (IFSU) and the Severinson’s laboratory (MBW, Stockholm University) forassistance with proteomics, imaging, and FACS, respectively.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsC.M.-V. and J.D.-P. designed and executed the experiments, and interpreted thedata being published. N.V. contributed to the design of the experiments andinterpretation of the data. A.B.E. produced cell lines for expression of wild-typeand mutant RRP6 proteins. All authors contributed to the writing of the article.

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 1097–1107 doi:10.1242/jcs.158733

1106

Jour

nal o

f Cel

l Sci

ence

FundingOur research is financed by The Swedish Research Council; and The SwedishCancer Society.

Supplementary materialSupplementary material available online athttp://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.158733/-/DC1

References

Adamson, B., Smogorzewska, A., Sigoillot, F. D., King, R. W. and Elledge, S. J.(2012). A genome-wide homologous recombination screen identifies the RNA-binding protein RBMX as a component of the DNA-damage response. Nat. CellBiol. 14, 318-328.

Bhatia, V., Barroso, S. I., Garcıa-Rubio, M. L., Tumini, E., Herrera-Moyano, E.and Aguilera, A. (2014). BRCA2 prevents R-loop accumulation and associateswith TREX-2 mRNA export factor PCID2. Nature 511, 362-365.

Chiolo, I., Minoda, A., Colmenares, S. U., Polyzos, A., Costes, S. V. andKarpen, G. H. (2011). Double-strand breaks in heterochromatin move outside ofa dynamic HP1a domain to complete recombinational repair. Cell 144, 732-744.

Chlebowski, A., Lubas, M., Jensen, T. H. and Dziembowski, A. (2013). RNAdecay machines: the exosome. Biochim. Biophys. Acta 1829, 552-560.

d’Adda di Fagagna, F. (2014). A direct role for small non-coding RNAs in DNAdamage response. Trends Cell Biol. 24, 171-178.

Eberle, A. B. and Visa, N. (2014). Quality control of mRNP biogenesis:networking at the transcription site. Semin. Cell Dev. Biol. 32, 37-46.

Eberle, A. B., Hessle, V., Helbig, R., Dantoft, W., Gimber, N. and Visa, N.(2010). Splice-site mutations cause Rrp6-mediated nuclear retention of theunspliced RNAs and transcriptional down-regulation of the splicing-defectivegenes. PLoS ONE 5, e11540.

Francia, S., Michelini, F., Saxena, A., Tang, D., de Hoon, M., Anelli, V., Mione,M., Carninci, P. and d’Adda di Fagagna, F. (2012). Site-specific DICER andDROSHA RNA products control the DNA-damage response. Nature 488, 231-235.

Gao, M., Wei, W., Li, M. M., Wu, Y. S., Ba, Z., Jin, K. X., Li, M. M., Liao, Y. Q.,Adhikari, S., Chong, Z. et al. (2014). Ago2 facilitates Rad51 recruitment and DNAdouble-strand break repair by homologous recombination. Cell Res. 24, 532-541.

Graham, A. C., Kiss, D. L. and Andrulis, E. D. (2009). Core exosome-independentroles for Rrp6 in cell cycle progression. Mol. Biol. Cell 20, 2242-2253.

Hessle, V., Bjork, P., Sokolowski, M., Gonzalez de Valdivia, E., Silverstein, R.,Artemenko, K., Tyagi, A., Maddalo, G., Ilag, L., Helbig, R. et al. (2009). Theexosome associates cotranscriptionally with the nascent pre-mRNP throughinteractions with heterogeneous nuclear ribonucleoproteins. Mol. Biol. Cell 20,3459-3470.

Hieronymus, H., Yu, M. C. and Silver, P. A. (2004). Genome-wide mRNAsurveillance is coupled to mRNA export. Genes Dev. 18, 2652-2662.

Houseley, J., LaCava, J. and Tollervey, D. (2006). RNA-quality control by theexosome. Nat. Rev. Mol. Cell Biol. 7, 529-539.

Januszyk, K. and Lima, C. D. (2014). The eukaryotic RNA exosome. Curr. Opin.Struct. Biol. 24, 132-140.

Kiss, D. L. and Andrulis, E. D. (2011). The exozyme model: a continuum offunctionally distinct complexes. RNA 17, 1-13.

Krejci, L., Altmannova, V., Spirek, M. and Zhao, X. (2012). Homologousrecombination and its regulation. Nucleic Acids Res. 40, 5795-5818.

Lieber, M. R. (2008). The mechanism of human nonhomologous DNA end joining.J. Biol. Chem. 283, 1-5.

Lorentzen, E., Basquin, J. and Conti, E. (2008). Structural organization of theRNA-degrading exosome. Curr. Opin. Struct. Biol. 18, 709-713.

Lubas, M., Christensen, M. S., Kristiansen, M. S., Domanski, M., Falkenby, L. G.,Lykke-Andersen, S., Andersen, J. S., Dziembowski, A. and Jensen, T. H.(2011). Interaction profiling identifies the human nuclear exosome targetingcomplex. Mol. Cell 43, 624-637.

Lykke-Andersen, S., Tomecki, R., Jensen, T. H. and Dziembowski, A. (2011).The eukaryotic RNA exosome: same scaffold but variable catalytic subunits.RNA Biol. 8, 61-66.

Madigan, J. P., Chotkowski, H. L. and Glaser, R. L. (2002). DNA double-strandbreak-induced phosphorylation of Drosophila histone variant H2Av helpsprevent radiation-induced apoptosis. Nucleic Acids Res. 30, 3698-3705.

Manfrini, N., Trovesi, C., Wery, M., Martina, M., Cesena, D., Descrimes, M.,Morillon, A., d’Adda di Fagagna, F. and Longhese, M. P. (2014). RNA-processing proteins regulate Mec1/ATR activation by promoting generation ofRPA-coated ssDNA. EMBO Rep. 2014, embr.201439458.

Monnat, R. J., Jr, Hackmann, A. F. and Cantrell, M. A. (1999). Generation ofhighly site-specific DNA double-strand breaks in human cells by the homingendonucleases I-PpoI and I-CreI. Biochem. Biophys. Res. Commun. 255, 88-93.

Phillips, S. and Butler, J. S. (2003). Contribution of domain structure to the RNA39 end processing and degradation functions of the nuclear exosome subunitRrp6p. RNA 9, 1098-1107.

Polo, S. E. and Jackson, S. P. (2011). Dynamics of DNA damage responseproteins at DNA breaks: a focus on protein modifications. Genes Dev. 25, 409-433.

Preston, C. R., Engels, W. and Flores, C. (2002). Efficient repair of DNA breaksin Drosophila: evidence for single-strand annealing and competition with otherrepair pathways. Genetics 161, 711-720.

Preston, C. R., Flores, C. C. and Engels, W. R. (2006). Differential usage ofalternative pathways of double-strand break repair in Drosophila. Genetics 172,1055-1068.

Richard, P., Feng, S. and Manley, J. L. (2013). A SUMO-dependent interactionbetween Senataxin and the exosome, disrupted in the neurodegenerativedisease AOA2, targets the exosome to sites of transcription-induced DNAdamage. Genes Dev. 27, 2227-2232.

Schneider, C. A., Rasband, W. S. and Eliceiri, K. W. (2012). NIH Image toImageJ: 25 years of image analysis. Nat. Methods 9, 671-675.

Sharma, V. and Misteli, T. (2013). Non-coding RNAs in DNA damage and repair.FEBS Lett. 587, 1832-1839.

Stewart, G. S., Wang, B., Bignell, C. R., Taylor, A. M. and Elledge, S. J. (2003).MDC1 is a mediator of the mammalian DNA damage checkpoint. Nature 421,961-966.

Synowsky, S. A., van Wijk, M., Raijmakers, R. and Heck, A. J. (2009).Comparative multiplexed mass spectrometric analyses of endogenouslyexpressed yeast nuclear and cytoplasmic exosomes. J. Mol. Biol. 385, 1300-1313.

Wahba, L., Gore, S. K. and Koshland, D. (2013). The homologous recombinationmachinery modulates the formation of RNA-DNA hybrids and associatedchromosome instability. eLife 2, e00505.

Wei, W., Ba, Z., Gao, M., Wu, Y., Ma, Y., Amiard, S., White, C. I., RendtlewDanielsen, J. M., Yang, Y. G. and Qi, Y. (2012). A role for small RNAs in DNAdouble-strand break repair. Cell 149, 101-112.

White, E., Schlackow, M., Kamieniarz-Gdula, K., Proudfoot, N. J. andGullerova, M. (2014). Human nuclear Dicer restricts the deleteriousaccumulation of endogenous double-stranded RNA. Nat. Struct. Mol. Biol. 21,552-559.

Woods, N. T., Mesquita, R. D., Sweet, M., Carvalho, M. A., Li, X., Liu, Y.,Nguyen, H., Thomas, C. E., Iversen, E. S., Jr, Marsillac, S. et al. (2012).Charting the landscape of tandem BRCT domain-mediated protein interactions.Sci. Signal. 5, rs6.

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 1097–1107 doi:10.1242/jcs.158733

1107