The yeast Fun30 and human SMARCAD1 chromatinremodellers promote DNA end resectionThomas Costelloe1*, Raphael Louge2,3*, Nozomi Tomimatsu4*, Bipasha Mukherjee4, Emmanuelle Martini5, Basheer Khadaroo2,3,Kenny Dubois2,3, Wouter W. Wiegant1, Agnes Thierry6, Sandeep Burma4, Haico van Attikum1 & Bertrand Llorente2,3
Several homology-dependent pathways can repair potentiallylethal DNA double-strand breaks (DSBs). The first step commonto all homologous recombination reactions is the 59–39 degrada-tion of DSB ends that yields the 39 single-stranded DNA requiredfor the loading of checkpoint and recombination proteins. In yeast,the Mre11–Rad50–Xrs2 complex (Xrs2 is known as NBN or NBS1in humans) and Sae2 (known as RBBP8 or CTIP in humans)initiate end resection, whereas long-range resection depends onthe exonuclease Exo1, or the helicase–topoisomerase complexSgs1–Top3–Rmi1 together with the endonuclease Dna2 (refs 1–6).DSBs occur in the context of chromatin, but how the resectionmachinery navigates through nucleosomal DNA is a process that isnot well understood7. Here we show that the yeast Saccharomycescerevisiae Fun30 protein and its human counterpart SMARCAD1(ref. 8), two poorly characterized ATP-dependent chromatin remo-dellers of the Snf2 ATPase family, are directly involved in the DSBresponse. Fun30 physically associates with DSB ends and directlypromotes both Exo1- and Sgs1-dependent end resection througha mechanism involving its ATPase activity. The function of Fun30in resection facilitates the repair of camptothecin-induced DNAlesions, although it becomes dispensable when Exo1 is ectopicallyoverexpressed. Interestingly, SMARCAD1 is also recruited toDSBs, and the kinetics of recruitment is similar to that of EXO1.The loss of SMARCAD1 impairs end resection and recombinationalDNA repair, and renders cells hypersensitive to DNA damage result-ing from camptothecin or poly(ADP-ribose) polymerase inhibitortreatments. These findings unveil an evolutionarily conservedrole for the Fun30 and SMARCAD1 chromatin remodellers incontrolling end resection, homologous recombination and genomestability in the context of chromatin.
Fun30 possesses intrinsic ATP-dependent chromatin-remodellingactivity8, which is required to promote gene silencing in hetero-chromatin9. FUN30 deletion renders cells hypersensitive to camptothecin(CPT)9, whereas overexpression results in genomic instability10. A rolefor Fun30 in the DSB response, however, remains enigmatic. Whileperforming a genomic screen using a plasmid-based assay, we discoveredthat the fun30D mutant shows an increased efficiency in one-endedhomologous recombination or break-induced replication (BIR) (Fig. 1,Supplementary Fig. 1 and Supplementary Table 1). We also found thatgap repair, which is a two-ended homologous recombination reaction,is increased in the fun30D mutant (Supplementary Fig. 2). This showsthat Fun30 affects a step common to all homologous recombinationreactions. Interestingly, the fun30D mutant shares this phenotype withthe resection mutants sgs1D and exo1D1,2, in which impaired resectionslows down the degradation of transformed plasmids, favouringplasmid-based recombination11 (Fig. 1 and Supplementary Fig. 2).Altogether, this indicates that Fun30 promotes DNA end processing.
To test whether Fun30 contributes to 59–39 DNA end resection, weanalysed single-stranded (ss)DNA formation at an HO endonuclease-induced DSB at the MAT locus12. Because ssDNA is resistant tocleavage by restriction enzymes, 59–39 resection at the DSB generatesa ladder of ssDNA bands after restriction digestion of the genomicDNA and electrophoresis under alkaline conditions. In the absenceof Fun30, the shortest ssDNA intermediate is formed with normalkinetics, but the formation of longer ssDNA intermediates is eitherdelayed or abolished (Fig. 2a and Supplementary Fig. 3a). Chromatinimmunoprecipitation (ChIP) of the ssDNA-binding replicationprotein A (RPA) complex at the HO-induced DSB confirmed theseresults (Supplementary Fig. 3c, d). Notably, we detected a similarresection defect at an I-SceI endonuclease cut site inserted at theHIS3 locus (Fig. 2c), ruling out a locus-specific effect. Our resultsindicate that Fun30 facilitates long-range end resection, which isfurther supported by a delay in the kinetics of DSB repair by single-strand annealing (SSA) in the fun30D mutant (Supplementary Fig. 4).
In the combined absence of Fun30 and either Sgs1 or Exo1, theresection defect was stronger than the defects in the correspondingsingle mutants (Fig. 2b and Supplementary Fig. 3b), leading to a morepronounced defect in RPA loading at the HO-induced DSB(Supplementary Fig. 3c). This correlated with higher plasmid-based
*These authors contributed equally to this work.
1Department of Toxicogenetics, Leiden University Medical Center, Einthovenweg 20, 2333 ZC Leiden, The Netherlands. 2Centre National de la Recherche Scientifique, Unite Mixte de Recherche 7258,Centre de Recherche en Cancerologie de Marseille, Marseille F-13009, France. 3Aix-Marseille University, Unite Mixte de Recherche 7258, Marseille F-13284, France. 4Division of Molecular RadiationBiology, Department of Radiation Oncology, University of Texas Southwestern Medical Center, 2201 Inwood Road, NC7.214E, Dallas, Texas 75390, USA. 5CEA-DSV-Institut de Radiobiologie Cellulaire etMoleculaire, 92265 Fontenay aux Roses, France. 6Institut Pasteur, Unite de Genetique Moleculaire des Levures, Centre National de la Recherche Scientifique and University Pierre and Marie Curie-Paris, 25rue du Docteur Roux, F75724 Paris Cedex 15, France.
BIR
effi
cie
ncy r
ela
tive t
o W
T
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
Transformations of mutant pools Individual transformations5.0
0
psy4Δ (7
,3)
yap3Δ (6
,4)
bud3Δ (5
,8)
ard1Δ (1
9,11
)
est3
Δ (2,8
3)
rad9Δ (1
0,41
)
est2
Δ (28,
38)
cac2
Δ (46,
122)
rlf2Δ (7
4,87
)
pol32
Δ (81,
130)
msi1Δ (1
35,1
12)
sgs1
Δ (382
,128
)
WT
diplo
id
rad24
Δ (61,
35)
ddc1Δ (1
5,11
8)
rad17
Δ (11,
99)
mec
3Δ (33,
27)
fun3
0Δ (16,
3)
exo1
Δ (13,
21)
BUD3
CEN4
CEN3
BIR assay
URA3
Figure 1 | fun30D and DNA end-resection mutants show high BIRefficiencies. BIR efficiencies of selected homozygous diploid null mutantsrelative to wild type (WT; BY4743). Mutants have been ranked according totheir BIR efficiencies. Two BIR experiments using transformations of mutantpools were performed (Supplementary Fig. 1). The rank of each mutant in thesetwo BIR experiments is given in parentheses. This rank is bottom-up formutants with BIR efficiencies lower than wild type, and top-down otherwise. Aschematic of the BIR assay is provided in the box. Error bars denote 6 meanabsolute deviation of two independent experiments.
2 7 S E P T E M B E R 2 0 1 2 | V O L 4 8 9 | N A T U R E | 5 8 1
BIR efficiencies and stronger delays in the kinetics of SSA (Supplemen-tary Figs 2 and 4). These results demonstrate that Fun30 promotesboth Sgs1- and Exo1-dependent resection of DSBs. Interestingly, weobserved smeared cut fragments in the SSA assay in the fun30D exo1Ddouble mutant (Supplementary Fig. 4b). These indicate severelyimpaired long-range resection1, which suggests that the Sgs1 resectionpathway depends more strongly on Fun30 than the Exo1 pathway.
The ATPase activity of Fun30 is essential for its chromatin-remodelling activity8. Expression of wild-type Fun30, but notATPase-deficient Fun30(K603R), restored end resection to wild-typelevels in fun30D mutants (Fig. 2c). This indicates that chromatinremodelling driven by Fun30 facilitates long-range resection, eitherdirectly or indirectly. After induction of an HO endonuclease-inducedDSB at the MAT locus, Fun30 accumulated at sites near the DSB within60 min and spread away at later time points (Fig. 2d), as previouslyobserved for Sgs1, Dna2 and Exo1 (refs 2, 13). This supports a directrole for Fun30 in long-range resection, acting together with the Exo1and Sgs1 resection machineries. However, Fun30 could affect endresection indirectly by regulating gene transcription or by establishingan abnormal chromatin structure. The loss of Fun30 neither led to anymajor change in transcript accumulation of end-resection factors(Supplementary Fig. 5), nor affected nucleosome positioning at theHIS3 locus used to monitor resection (Supplementary Fig. 6).Together, these results implicate Fun30 in directly promoting long-range resection at DSBs. This is further supported by the fact that acuteloss of Fun30 led to a long-range resection defect at the I-SceI breakinduced at the HIS3 locus (Supplementary Fig. 7). Interestingly, ChIPanalysis of histone H3 and H2B occupancy around an HO-inducedDSB at MAT showed that the loss of the histone ChIP signal is coupledto long-range resection in both wild-type and fun30D cells(Supplementary Figs 8 and 9)14. This indicates that Fun30 does not
facilitate long-range resection by modulating histone occupancy, butrather by increasing access to DNA within DSB-associated chromatin8.
We next investigated the physiological role of the resection functionof Fun30. Gene conversion at a single HO-induced DSB at MAT isnormal in a fun30D mutant, both in the presence and absence of Sgs1or Exo1 (data not shown). This shows that long-range resection is notessential for efficient gene conversion1,3. We confirmed that thefun30D mutant is hypersensitive to the topoisomerase I inhibitorCPT, but not to the ribonucleotide reductase inhibitor hydroxyureaor ultraviolet light (Supplementary Fig. 10)9. Expression of wild-type,but not ATPase-deficient Fun30(K603R), restored CPT resistance infun30D mutants (Supplementary Fig. 10a), indicating that resectiondriven by Fun30 ATPase activity protects cells against CPT-induced DNA damage. To show directly that the resection functionof Fun30 is responsible for CPT resistance, we ectopically expressedExo1 in a fun30D mutant. Expression of wild-type Exo1, but notthe Exo1(D173A) nuclease-deficient mutant, suppressed both theresection defect and the CPT hypersensitivity of the fun30D mutant(Fig. 2e and Supplementary Fig. 11). This confirms that the resectionfunction of Fun30 is required for the repair of CPT-induced DNAdamage. Interestingly, the fun30D exo1D and fun30D sgs1D mutantsare more sensitive to CPT, but not to hydroxyurea, than are thefun30D, exo1D and sgs1D mutants (Supplementary Fig. 10b), whichcorroborates their stronger resection defects. However, the combinedabsence of Fun30 and Sae2 led to a synergistic hypersensitivity to bothCPT and hydroxyurea (Supplementary Fig. 10b), despite a resectiondefect that is comparable to that in the fun30D mutant (Fig. 2b), indi-cating that the roles of Fun30 and Sae2 in genome maintenance do notrely exclusively on facilitating resection15.
Resection mutants are known to affect the type of yeast survivorsthat form by different recombination mechanisms in the absence of
a b
Uncut 0.9 kb
Cut 0.7 kb
r1 (1.6 kb)
r2 (2.7 kb)
r3 (5 kb)
r4 (5.7 kb) r5 (6.1 kb)
r6 (6.8 kb)
r7 (10 kb)
Probe
S HO S B B S S B S/B S
Time (min)
fun30Δ rad51Δ rad51Δ
24
0
0
30
60
9
0
12
0
18
0
0
30
60
9
0
12
0
18
0
24
0
r3
r2
r1 Uncut
Cut
MATα locus
exo1Δ fun30Δ exo1Δ rad51Δ rad51Δ
24
0
0
30
60
90
12
0
18
0
24
0
0
30
60
90
12
0
18
0
sae2Δ fun30Δ sae2Δ rad51Δ rad51Δ
24
0
0
30
60
90
1
20
18
0
24
0
0
30
6
0
90
12
0
18
0
sgs1Δ fun30Δ sgs1Δ rad51Δ rad51Δ
24
0
0
30
60
9
0
12
0
18
0
24
0
0
30
60
90
12
0
18
0
MATa locus MATa locus MATa locus
r4–r7
c
r1 (2.4 kb)
r2 (4 kb)
r3 (5.7 kb)
r4 (6.9 kb)
r7 (12 kb)
WT
pRS316
fun30ΔpRS316
fun30ΔpRS316-FUN30
fun30ΔpRS316-fun30(K603R)
0
30
6
0
90
1
20
1
50
2
10
0
30
6
0
90
1
20
1
50
2
10
0
30
6
0
90
1
20
1
50
2
10
0
30
6
0
90
1
20
1
50
2
10
Cut (0.4 kb)
Uncut (1 kb)
r5–6
(8.5–8.6 kb)
HIS3 locus HIS3 locus
1 20 4
Time in galactose (h)
WT
Fun3
0–M
yc WT
Fun3
0–M
yc WT
Fun3
0–M
yc WT
Fun3
0–M
yc
0.4 kb7
6
5
4
3
2
1
0
1.6 kb
5.3 kb
9.8 kb
25.3 kb
Ab
so
lute
enrich
men
t
d
CPT (30 μM) No CPT
WT + pRS424
fun30Δ + pRS424
WT + pExo1
WT + pExo1(D173A)
fun30Δ + pExo1
fun30Δ + pExo1(D173A)
e
Figure 2 | Fun30 promotes long-range 59239 DNA end resection and isrecruited to DSBs. a, Southern blot analysis of StyI (S) and BstXI (B)-digestedgenomic DNA after alkaline gel electrophoresis. r1–r7 denote partially ssDNAfragments. b, As in a, except that exo1Dmutants were MATa strains, showing alonger uncut fragment (1.9 kb). c, Southern blot analysis of StyI-digested
genomic DNA after alkaline gel electrophoresis to monitor ssDNA formation(r1–r7 fragments) at an I-SceI DSB generated at the HIS3 locus. d, Fun30–Myclevels at MAT before and after galactose-mediated HO induction measured byChIP coupled to quantitative PCR. Error bars denote 6 s.e.m. of threeindependent experiments. e, Tenfold serial dilutions of yeast cultures.
RESEARCH LETTER
5 8 2 | N A T U R E | V O L 4 8 9 | 2 7 S E P T E M B E R 2 0 1 2
functional telomerase16,17. Under liquid culture conditions, cells lackingthe Est2 subunit of telomerase accumulate mostly type II survivors.However, we detected almost equal proportions of type I and type IIsurvivors in a fun30D est2Dmutant, similar to what is observed in otherresection-defective mutants (rad24D, rad17D (ref. 17) and exo1D(ref. 16)) (Supplementary Fig. 12). Introduction of the cdc13-1 mutationthat induces the formation of long ssDNA tracts at telomeres18
suppresses the fun30D est2D phenotype as it suppresses the phenotypeof a rad17D est2D mutant17. Therefore, Fun30 affects recombination atunprotected telomeres, probably because of its role in resection.
SMARCAD1 is the human Snf2 family member that has thehighest sequence similarity with Fun30. SMARCAD1 may functionin the DNA-damage response because it is phosphorylated atcanonical (Ser/Thr-Gln) ATM/ATR phosphorylation sites, as well asat non-canonical sites, in response to genotoxic insults19,20. Weexamined whether SMARCAD1 also promotes DNA end resection.SMARCAD1 knockdown reduced the accumulation of RPA intoionizing radiation-induced foci (IRIF) (Fig. 3a), as well as that ofgreen fluorescent protein (GFP)-tagged RPA at laser micro-irradiation-induced DSBs in U2OS cells21 (Supplementary Fig. 13a).Accordingly, we found that SMARCAD1 knockdown reduced ssDNAformation as determined by directly staining ssDNA-associated5-bromo-29-deoxyuridine (BrdU) IRIF (Supplementary Fig. 13b).These phenotypes are similar to those seen after knockdown ofEXO1, a major resection enzyme in human cells21, indicating thatthe absence of SMARCAD1 impairs resection. In accord with aresection defect, we found that the loss of SMARCAD1 also impairedrecombinational DSB repair. SMARCAD1 knockdown cells (1) weredefective in the repair of an I-SceI-induced DSB by gene conversion inthe direct-repeat (DR)-GFP reporter22 (Fig. 3b); (2) showed a majorreduction in the repair of CPT-induced DSBs as monitored by thedisappearance of TP53BP1 foci in S/G2 phase cells (SupplementaryFig. 13c); and (3) were hypersensitive to DNA damage resulting fromCPT or poly(ADP-ribose) polymerase (PARP) inhibitor (ABT-888)treatments (Fig. 3c). In addition, SMARCAD1 co-localized with
c
Surv
ival
1
0.1
0.01 0 20 40 60 80 100
CPT (nM)
0.001
Surv
ival
1
0.1
0.01
0 20 40 60 80 100
ABT-888 (μM)
siSMARCAD1 siScr
siEXO1
siSMARCAD1 siScr
siEXO1
a
RP
A+ c
ells
(%
)
siScr siSMARCAD1 siEXO1
80
60
40
20
0 1 2 3 Pool 1 2 3
DA
PI
RP
A
siScr siSMARCAD1 siEXO1
Mock Irradiated Mock Irradiated Mock Irradiated
EXO1
KU80
SMARCAD1
siScr siSMARCAD1 siEXO1
1 2 3 Pool 1 2 3
1 2 3 Pool 1 2 3
b
siSMARCAD1
Untransfected
siScr
siEXO1
0
50
100
Ho
mo
log
ous
reco
mb
inatio
n (%
)
1 2 3 Pool 1 2 3
SMARCAD1
EXO1
Actin
siScr siSMARCAD1 siEXO1
75
25
d
Fo
kI W
T
Fo
kI(D
450A
)
mCherry– LacI–FokI SMARCAD1 MergeγH2AXDAPI
0
40
80
FokI WT
Fo
kI–γH
2A
X–S
MA
RC
AD
1-
po
sitiv
e c
ells
(%
)
20
10
70
50
60
30
FokI(D450A)
GFP–SMARCAD1
GFP–RPA GFP–EXO1
Rela
tive fl
uo
rescence
1.0
0.5
0.0 0 50 100 Time (min)
e
Figure 3 | SMARCAD1 promotes end resection, homologousrecombination and cell survival after genotoxic insults in U2OS cells.a, Immunodetection (top) and quantification (bottom right) of RPA foci 3 hafter 6 Gy of ionizing radiation. Scale bar, 10mm. Western blot analysis ofSMARCAD1 in cells transfected with individual or pooled short interferingRNAs (siRNAs) (bottom left). Knockdown of EXO1 serves as a control. Nucleiwith more than ten RPA foci were scored. b, Western blot analysis ofSMARCAD1 (left) and quantification of homologous recombinationfrequencies using a DR-GFP assay (right). c, Clonogenic survival ofSMARCAD1 knockdown cells treated with CPT or the PARP inhibitor ABT-888. d, Immunofluorescence staining of SMARCAD1 and cH2AX at DSBsinduced by mCherry–LacI–FokI (FokI WT) at a 2563 Lac operator genomicarray (top). Nuclease-deficient mCherry–LacI–FokI(D450A) was used as acontrol. Quantification of cells showing co-localization of SMARCAD1 andcH2AX at FokI-induced DSBs (bottom). Scale bar, 5mm. e, Quantification ofGFP–SMARCAD1, GFP–EXO1 and GFP–RPA accumulation at sites of lasermicro-irradiation in live cells. DAPI, 49,6-diamidino-2-phenylindole; siScr,scrambled siRNA. Error bars denote s.e.m. (a, b and d) and 6 s.e.m. (c and e) ofthree independent experiments for all plots in a, b, c and e.
DSB
resection tracts
Exo1/STR–Dna2 Fun30/SMARCAD1
WT fun30Δor
siSMARCAD1
fun30Δ + pExo1or
siSMARCAD1 + pEXO1
Reduced extension of resection tracts
Extension ofresection tracts
Extension of
Figure 4 | Model for Fun30 and SMARCAD1 control of end resectionthrough DSB-associated nucleosomes. Fun30 and SMARCAD1 weakenhistone–DNA interactions in nucleosomes flanking DSBs, which facilitatesssDNA production by the Exo1- and Sgs1–Top3–Rmi1 (STR)–Dna2 resectionmachineries. In the absence of Fun30 and SMARCAD1, histone–DNAinteractions limit the extent of resection, but plasmid-based overexpression ofExo1 (pExo1) or EXO1 (pEXO1) bypasses this impediment in yeast and U2OScells, respectively.
LETTER RESEARCH
2 7 S E P T E M B E R 2 0 1 2 | V O L 4 8 9 | N A T U R E | 5 8 3
cH2AX at laser-induced DNA damage and at DNA breaks generatedby the FokI nuclease (Fig. 3d and Supplementary Fig. 13d), demon-strating that SMARCAD1 is recruited to DSBs. Importantly, GFP-tagged SMARCAD1 was recruited to laser micro-irradiation-inducedlesions before GFP-tagged RPA and with kinetics similar to that ofGFP-tagged EXO1 (Fig. 3e)21, as expected for a factor that promotesresection. Finally, the defect in RPA IRIF formation in SMARCAD1-depleted cells could be partially rescued by overexpression of EXO1(Supplementary Fig. 13e), indicating that SMARCAD1, like Fun30, hasa direct role in DNA end resection and recombinational DSB repair.
Recent reports from budding yeast9, fission yeast23 and humancells24 have shown that the Fun30 or SMARCAD1 Snf2 familymembers have related roles in promoting heterochromatinization.We show that Fun30 and SMARCAD1 are new DNA-damage-response proteins that facilitate DNA end resection and DSB repairin chromatin (Fig. 4). Their precise modes of action and the extent oftheir functional conservation remain to be determined.
METHODS SUMMARYThe yeast strains used are derivatives of S288C, W303 and JKM179 (seeSupplementary Table 2). Details of their construction are provided in Methods.The BIR genomic screen was adapted from ref. 25, except that pADW17 and pLS192were used11. Tag arrays were from C.-Y Ho (Samuel Lunenfeld Research Institute,Toronto, Canada). The gap-repair assay used pSB110 (ref. 26), which contains anautonomous replication sequence but no centromere. Detection of ssDNA inter-mediates, SSA assays and ChIP experiments were performed as in refs 1 and 27.Transfection of U2OS cells, quantification of RPA foci after c-irradiation, co-immunostaining for SMARCAD1 and cH2AX after laser micro-irradiation, andlive-cell imaging of GFP-tagged proteins to laser-induced breaks were carried out asdescribed21,28. SMARCAD1 localization studies at FokI-induced DSBs and DR-GFPassays were performed as previously reported22,29. Survival of U2OS cells after CPTor ABT-888 treatment was quantified by the standard colony-formation assay.
Full Methods and any associated references are available in the online version ofthe paper.
Received 7 October 2011; accepted 28 June 2012.
Published online 9 September; corrected online 26 September 2012 (see full-text
HTML version for details).
1. Mimitou, E. P. & Symington, L. S. Sae2, Exo1 and Sgs1 collaborate in DNA double-strand break processing. Nature 455, 770–774 (2008).
2. Zhu, Z., Chung, W. H. S. h. i. m. E. Y., Lee, S. E. & Ira, G. Sgs1 helicase and twonucleases Dna2 and Exo1 resect DNA double-strand break ends. Cell 134,981–994 (2008).
3. Gravel, S., Chapman, J. R., Magill, C. & Jackson, S. P. DNA helicases Sgs1 and BLMpromote DNA double-strand break resection. Genes Dev. 22, 2767–2772 (2008).
4. Cejka, P. et al. DNA end resection by Dna2-Sgs1-RPA and its stimulation by Top3-Rmi1 and Mre11-Rad50-Xrs2. Nature 467, 112–116 (2010).
5. Nicolette, M. L. et al. Mre11–Rad50–Xrs2 and Sae2 promote 59 strand resection ofDNA double-strand breaks. Nature Struct. Mol. Biol. 17, 1478–1485 (2010).
6. Niu, H. et al. Mechanism of the ATP-dependent DNA end-resection machineryfrom Saccharomyces cerevisiae. Nature 467, 108–111 (2010).
7. Sinha, M. & Peterson, C. L. Chromatin dynamics during repair of chromosomalDNA double-strand breaks. Epigenomics 1, 371–385 (2009).
8. Awad, S., Ryan, D., Prochasson, P., Owen-Hughes, T. & Hassan, A. H. The Snf2homolog Fun30 acts as a homodimeric ATP-dependent chromatin-remodelingenzyme. J. Biol. Chem. 285, 9477–9484 (2010).
9. Neves-Costa, A., Will, W. R., Vetter, A. T., Miller, J. R. & Varga-Weisz, P. The SNF2-familymemberFun30promotesgenesilencing inheterochromatic loci.PLoSONE4, e8111 (2009).
10. Ouspenski, I. I., Elledge, S. J. & Brinkley, B. R. New yeast genes important forchromosome integrity and segregation identified by dosage effects on genomestability. Nucleic Acids Res. 27, 3001–3008 (1999).
11. Marrero, V. A. & Symington, L. S. Extensive DNA end processing by Exo1 and Sgs1inhibits break-induced replication. PLoS Genet. 6, e1001007 (2010).
12. White, C. I. & Haber, J. E. Intermediates of recombination during mating typeswitching in Saccharomyces cerevisiae. EMBO J. 9, 663–673 (1990).
13. Shim, E. Y. et al. Saccharomyces cerevisiae Mre11/Rad50/Xrs2 and Ku proteinsregulate association of Exo1 and Dna2 with DNA breaks. EMBO J. 29, 3370–3380(2010).
14. Chen, C. C. e. t. a. l. Acetylated lysine 56 on histone H3 drives chromatin assemblyafter repair and signals for the completion of repair. Cell 134, 231–243 (2008).
15. Clerici, M., Mantiero, D., Lucchini, G. & Longhese, M. P. The Saccharomycescerevisiae Sae2 protein negatively regulates DNA damage checkpoint signalling.EMBO Rep. 7, 212–218 (2006).
16. Zubko, M. K. Exo1 and Rad24 differentially regulate generation of ssDNA attelomeres of Saccharomyces cerevisiae cdc13–1 mutants. Genetics 168, 103–115(2004).
17. Grandin, N. & Charbonneau, M. Control of the yeast telomeric senescence survivalpathwaysof recombinationby theMec1andMec3DNAdamagesensorsandRPA.Nucleic Acids Res. 35, 822–838 (2007).
18. Garvik, B., Carson, M. & Hartwell, L. Single-stranded DNA arising at telomeres incdc13 mutants may constitute a specific signal for the RAD9 checkpoint. Mol. Cell.Biol. 15, 6128–6138 (1995).
19. Matsuoka, S. et al. ATM and ATR substrate analysis reveals extensive proteinnetworks responsive to DNA damage. Science 316, 1160–1166 (2007).
20. Beli, P. et al. Proteomic investigations reveal a role for RNA processing factorTHRAP3 in the DNA damage response. Mol. Cell 46, 212–225 (2012).
21. Tomimatsu, N. et al. Exo1 plays a major role in DNA end resection in humans andinfluences double-strand break repair and damage signaling decisions. DNARepair 11, 441–448 (2012).
22. Bolderson, E. et al. Phosphorylation of Exo1 modulates homologousrecombination repair of DNA double-strand breaks. Nucleic Acids Res. 38,1821–1831 (2010).
23. Stralfors, A., Walfridsson, J., Bhuiyan, H. & Ekwall, K. The FUN30 chromatinremodeler, Fft3, protects centromeric and subtelomeric domains fromeuchromatin formation. PLoS Genet. 7, e1001334 (2011).
24. Rowbotham, S. P. et al. Maintenance of silent chromatin through replicationrequires SWI/SNF-like chromatin remodeler SMARCAD1. Mol. Cell 42, 285–296(2011).
25. Ooi, S. L., Shoemaker, D. D. & Boeke, J. D. DNA microarray-based genetic screen fornonhomologous end-joining mutants in Saccharomyces cerevisiae. Science 294,2552–2556 (2001).
26. Bartsch, S., Kang, L. E. & Symington, L. S. RAD51 is required for the repair ofplasmid double-stranded DNA gaps from either plasmid or chromosomaltemplates. Mol. Cell. Biol. 20, 1194–1205 (2000).
27. van Attikum, H., Fritsch, O. & Gasser, S. M. Distinct roles for SWR1 and INO80chromatin remodelingcomplexes at chromosomaldouble-strandbreaks.EMBOJ.26, 4113–4125 (2007).
28. Tomimatsu, N., Mukherjee, B. & Burma, S. Distinct roles of ATR and DNA-PKcs intriggering DNA damage responses inATM-deficient cells. EMBORep. 10, 629–635(2009).
29. Shanbhag, N. M., Rafalska-Metcalf, I. U., Balane-Bolivar, C., Janicki, S. M. &Greenberg, R. A. ATM-dependent chromatin changes silence transcription in cis toDNA double-strand breaks. Cell 141, 970–981 (2010).
Supplementary Information is linked to the online version of the paper atwww.nature.com/nature.
Acknowledgements We thank G. Ira for sharing unpublished data and S. Janicki,R. Greenberg, L. Symington and all the laboratories from the Centre National de laRecherche Scientifique (CNRS) UPR3081 for providing reagents. We thank S. Coulonfor help in the analysis of the fun30 repressible allele, I. Lafontaine for support instatistical analyses, C. V. Camacho for generating the V5-EXO1 constructs andA. Guenole, R. Srivas, T. Ideker, K. Vreeken and M. Vermeulen for help in searching forFun30 interactors. B.L. is grateful to B. Dujon for hosting him and providing theopportunity to perform the BIR screen. S.B. is supported by grants from the NationalInstitutes of Health (RO1 CA149461), National Aeronautics and Space Administration(NNX10AE08G) and the Cancer Prevention and Research Institute of Texas(RP100644). H.v.A. receives funding from the Netherlands Organization for ScientificResearch (NWO-VIDI grant) and Human Frontiers Science Program (HFSP-CDA grant).B.L. is supported by grants from the CNRS (ATIP) and the Agence Nationale de laRecherche (ANR-10-BLAN-1606-03).
Author Contributions B.L. andA.T. performed the genetic screen andB.L. identified theresection defect of fun30D. T.C. constructed yeast strains and plasmids and performedthe yeast ChIP experiments. R.L. constructed yeast strains and performed ssDNAanalysis by alkaline gels, BIR and gap-repair assays. R.L. and T.C. analysed SSA defects.N.T. and B.M. performed all of the SMARCAD1 knockdown experiments in human cellsand the DR-GFP assays. E.M. designed and built the strain containing the inducibleI-SceI cut site at HIS3, performed the micrococcal nuclease assay and contributed todata analysis. B.K. performed the analysis of survivors in the absence of telomerase.K.D. assisted R.L. and K.D., R.L. and T.C. performed fun30 DNA-damage-sensitivityassays. W.W.W. examined the localization of SMARCAD1 at FokI-induced DSBs. T.C.,S.B., H.v.A. and B.L. designed the experiments and analysed the data. H.v.A. and B.L.wrote the manuscript.
Author Information Microarray datahave been deposited in the NCBIGene ExpressionOmnibus and are accessible through accession numbers GSE38715 (BIR screen) andGSE38735 (fun30D transcriptome). Reprints and permissions information is availableat www.nature.com/reprints. The authors declare no competing financial interests.Readers are welcome to comment on the online version of the paper. Correspondenceand requests for materials should beaddressed to B.L. ([email protected]) orH.v.A. ([email protected]).
RESEARCH LETTER
5 8 4 | N A T U R E | V O L 4 8 9 | 2 7 S E P T E M B E R 2 0 1 2
METHODSYeast strains and plasmids. The yeast strains used are derivatives of S288C, W303and JKM179 and are listed in Supplementary Table 2. Details of the primers used forgene disruption and confirmation are available on request. FUN30 was disrupted byone-step gene replacement using the hphNT1, KanMX4 or 13MYC::KanMX4cassettes30. Using such a strategy, BLY187 and BLY188 were obtained fromLSY1709-9D, BLY185 and BLY186 from LSY1709-4A, BLY189 from LSY1983-16B, BLY031 from BY4741, BLY033 from BY4742, BLK020 from W303, BLY137from EM111, yHA629 from JKM179 and yHA630 from JKM179. SAE2 wasdisrupted by one-step gene replacement using the HIS3MX6 cassette to generateBLY195 and BLY196 from LSY1709-9D, and to generate BLY197 and BLY198 fromBLY187. rad51D yeast strains used for CPT- and hydroxyurea-sensitivity assays inSupplementary Fig. 10b are of the W303 background and were obtained by crossingthe corresponding rad51D mutants with the parental W303 reference strain.
Yeast strains used for the analysis of survivors in the absence of telomerase werebuilt as follows: MNS961 was crossed with BLK020, the resulting diploid wasgrown on non-selective medium to lose pSD196 and obtain BLK018, andsenescence was analysed on BLK018 ascospores. To perform these experimentsin the presence of the cdc13-1 mutation, YE1553 was crossed with MNS961 andthe resulting diploid was sporulated to obtain BLK029. YE1552 was crossed withBLK020 and the resulting diploid was sporulated to obtain BLK026. BLK029 waseventually crossed with BLK026 and grown on non-selective medium to losepSD196 and to obtain BLK033. Senescence was analysed on BLK033 ascospores.cdc13-1 cells were grown at 25 uC, and the presence of the cdc13-1 mutation wasfollowed by its thermosensitivity at 30 uC. Senescence was analysed as described inthe Supplementary Fig. 12 legend.
EM111 was built as follows: the 59-ACATAATGAATTATACAT-39 sequenceupstream of the regulated TATA box of HIS3 from the JKM139 strain was replacedby the 59-TAGGGATAACAGGGTAAT-39 sequence containing the I-SceI cleavagesite. This modified HIS3 locus was then used to replace the HIS3 locus from Lev488(ref. 31), a W303 derivative strain containing the lys2::PGAL-I-SceI cassette.
A plasmid containing the wild-type Fun30-coding gene was generated by intro-ducing a 4.5 kb SpeI/NotI fragment containing FUN30 into pRS316. An ATPase-deficient version of the Fun30-coding gene, which contains a point mutation in theATPase domain that replaces a lysine with an arginine at position 603(Fun30(K603R)), was introduced by site-directed mutagenesis and confirmedby sequencing.
Exo1- and Exo1(D173A)-overexpressing plasmids are pSM502 and pSM638,respectively, described in ref. 32.Media, growth conditions and genetic methods. Media, growth conditions andgenetic methods are as described previously in ref. 33. Galactose induction of HO orI-SceI was performed as follows: one colony was grown overnight at 30 uC in eitheryeast extract peptone dextrose (YPD) or synthetic complete (SC) media. The nextday cells were diluted and grown for about 8 h using the same media to mid-logphase. Cells were washed with water and grown overnight in YPLGg (1% yeastextract, 2% peptone, 2% lactic acid, 3% glycerol, 0.05% glucose) or SC media, asindicated, to mid-log phase. HO or I-SceI expression was induced by adding 2%galactose, and switched off by adding 2% glucose. Note that HO expression in strainsfrom the W303 background was driven from the centromeric plasmid pGAL:HO34.BIR genomic screen. Details of the screen are given in the Supplementary Fig. 1legend. The pool of homozygous diploid null mutants was preferred instead of thepool of haploid null mutants because, in principle, the genetic quality of the diploiddeletion strain is higher25. Two independent experiments were carried out, eachone included one transformation with circular pADW17 and one transformationwith pLS192 linearized with SnaBI (Supplementary Table 3). For each mutant andfor each experiment, the median of the normalized hybridization values of thedifferent uptag and downtag spots was calculated35. The mean absolute deviationfrom the two experiments is given in Fig. 1. Transformation ratios correspondingto BIR efficiencies were derived by dividing the median hybridization signal fromthe linear mini-chromosome pLS192 by that of the circular mini-chromosomepADW17 (Supplementary Table 1). The BIR defect of each mutant was confirmedby one or two individual transformations, and at least four individual transforma-tions were performed for wild-type diploid, sgs1D and fun30D.Alkaline electrophoresis. ssDNA intermediates were analysed by alkaline gelelectrophoresis as described12, and the blots were hybridized with double-strandedprobes obtained by PCR. The coordinates of the MAT probe are 201176–201570on chromosome 3, and the coordinates of the HIS3 probe are 722025–722227 onchromosome 15.ChIP. ChIP was performed as previously described27. In brief, cells were grownovernight in YPAD, diluted in YPLGg and grown to mid-log phase. Glucose wasadded to a fraction of the cells to repress HO, whereas galactose was added to theremainder of the cells to induce HO. Cells were fixed using formaldehyde andcollected at 0, 1, 2, 4 and 6 h after galactose addition (cells grown in the presence of
glucose were collected at the 2-h time point). Extracts were prepared and subjectedto ChIP using the following antibodies: anti-Myc (9B11; Cell SignalingTechnology), rabbit anti-RPA (Agrisera) and rabbit anti-histone H2B or H3(Abcam). An aliquot of each extract was not immunoprecipitated and served asinput. Input and immunoprecipitated DNA were purified and analysed by quant-itative PCR. Absolute fold enrichment for RPA or Fun30–Myc at the HO DSB wascalculated as follows: for each time point, the signal from a site near the HO DSB atthe MAT locus was normalized to that from the non-cleaved SMC2 locus in ChIPand input DNA samples. For each time point and site, the normalized ChIP signalswere normalized to the normalized input DNA signals, because end resectioncan reduce the available DNA template. Finally, relative-fold enrichment wascalculated by dividing the absolute-fold enrichment from induced cells to thatof uninduced cells.
DNA end resection and loss of histone H2B and H3 ChIP signals were calcu-lated as follows: to measure end resection using input DNA from ChIP experi-ments, the input DNA value for each site near the DSB was normalized to that ofthe input DNA value from the non-cleaved SMC2 locus. The ratios obtained foreach time point after HO induction were normalized to that in uninduced cells.Similarly, to determine the loss of H2B and H3 ChIP signal, the ChIP value foreach site near the DSB was normalized to that of the ChIP value from the non-cleaved SMC2 locus. The ratios obtained for each time point after HO inductionwere normalized to that in uninduced cells. ChIP results are presented as the meanof at least two experiments 6 s.e.m.Micrococcal nuclease digestion of chromatin. Micrococcal nuclease digestion ofchromatin was performed as described in ref. 36. After purification, the DNA wasdigested by HindIII to show nucleosome positioning by indirect end-labelling. TheHIS3 proximal probe used for Southern blot hybridization is the same as foralkaline electrophoresis experiments. The coordinates of the DED1 proximalprobe are 722441–722809 on chromosome 15. Coordinates of the NOC2 proximalprobe are 727786–728097 on chromosome 15.Irradiation of U2OS cells, SMARCAD1 knockdown, immunofluorescencestaining and live-cell imaging. U2OS cells were irradiated with c-rays from acaesium source (J.L. Shepherd & Associates) or were micro-irradiated with apulsed nitrogen laser (Spectra-Physics; 365 nm, 10 Hz) with output set at 75% ofthe maximum, as described in ref. 28. SMARCAD1 was depleted using siRNA159-AGGAUGCAUCUUGUCUGAAUUGAAA-39, siRNA2 59-GGGACGAUUGAAGAAUCCAUGCUAA-39 and siRNA3 59-GAGAUGUAGUUAUAAGGCUUAUGAA-39. EXO1 was depleted using siRNA1 59-UGCCUUUGCUAAUCCAAUCCCACGC-39, siRNA2 59-UAGUGUUUCAGGAUCAACAUCAUCU-39
and siRNA3 59-UUUGUUAGUAGGUCCAUUUACCAGG-39. The efficiencyof knockdown for every experiment was verified by western blotting as describedin ref. 28; cells transfected with scrambled siRNA (Invitrogen) served as controls.Immunofluorescence staining for cH2AX and SMARCAD1 accumulation afterlaser irradiation or FokI expression, or for RPA or TP53BP1 foci after c-irradiationwas carried out as described previously in ref. 28. For demarcating cells inS/G2 phase, nuclei were co-immunostained with anti-cyclin A antibody37.Quantification of foci was done as described in ref. 22. The following antibodieswere used for immunofluorescence staining or western blotting: RPA (Abcam),SMARCAD1 (Bethyl Laboratories), EXO1 (Thermo Fisher), cyclin A, TP53BP1(Santa Cruz Biotechnology), actin (Sigma) and cH2AX (Upstate Biotechnology).For live-cell imaging combined with laser micro-irradiation, U2OS cells weretransfected with GFP–RPA38, GFP–EXO1 (ref. 21) or GFP–SMARCAD1(OriGene), laser micro-irradiated and time-lapse imaged, and fluorescenceintensities of micro-irradiated areas relative to non-irradiated areas calculated asdescribed previously in ref. 21. To stain for BrdU and ssDNA foci, U2OS cells weregrown in the presence of 10mM BrdU (Sigma) for 16 h, irradiated with 10 Gy ofc-rays, fixed at the indicated times and immunofluorescence-stained with an anti-BrdU antibody (BD Biosciences) under non-denaturing conditions to detect BrdUincorporated into ssDNA, as described in ref. 21. For ectopic expression of EXO1in U2OS cells, the gene coding for the b isoform of EXO1 was sequentially clonedinto pLenti6.3/V5-DEST by BP and LR clonase reactions (Invitrogen) usingpEGFP-C1-Exo1b22 as a template. The final vector, pLenti6.3/Exo1b-V5/DEST,was confirmed by sequencing. The accumulation of SMARCAD1 at FokI-inducedDSBs was monitored in U2OS cells containing 256 Lac operator repeats integratedinto the genome as described in ref. 29. Vectors expressing wild-type mCherry–LacI–FokI or nuclease-deficient mCherry–LacI–FokI(D450A) were transfected into thesecells, and 24 h later the cells were immunostained for SMARCAD1 and cH2AX.
30. Janke, C. et al. A versatile toolbox for PCR-based tagging of yeast genes: newfluorescent proteins, more markers and promoter substitution cassettes. Yeast21, 947–962 (2004).
31. Pardo, B., Ma, E. & Marcand, S. Mismatch tolerance by DNA polymerase Pol4 inthe course of nonhomologous end joining in Saccharomyces cerevisiae. Genetics172, 2689–2694 (2006).
32. Moreau, S., Morgan, E. A. & Symington, L. S. Overlapping functions of theSaccharomyces cerevisiae Mre11, Exo1 and Rad27 nucleases in DNAmetabolism. Genetics 159, 1423–1433 (2001).
33. Burke, D. & Strathern, J. in Methods in Yeast Genetics: a Cold Spring HarborLaboratory Course Manual (eds Amberg, D. C., Burke, D. & Strathern, J. N.) (2005).
34. Herskowitz, I. & Jensen, R. E. Putting the HO gene to work: practical uses formating-type switching. Methods Enzymol. 194, 132–146 (1991).
35. Decourty, L. et al. Linking functionally related genes by sensitive and quantitativecharacterization of genetic interaction profiles. Proc. Natl Acad. Sci. USA 105,5821–5826 (2008).
36. Martini, E. M. D., Keeney, S. & Osley, M. A. A role for histone H2B during repair ofUV-induced DNA damage in Saccharomyces cerevisiae. Genetics 160,1375–1387 (2002).
37. Bekker-Jensen, S. et al. Spatial organization of the mammalian genomesurveillance machinery in response to DNA strand breaks. J. Cell Biol. 173,195–206 (2006).
38. Sporbert, A., Gahl, A., Ankerhold, R., Leonhardt, H. & Cardoso, M. C. DNApolymerase clamp shows little turnover at established replication sites butsequentialdenovo assemblyat adjacent origin clusters. Mol. Cell10, 1355–1365(2002).
