collaborative roles of γh2ax and the rad51 paralog xrcc3 in homologous recombinational repair
TRANSCRIPT
d n a r e p a i r 6 ( 2 0 0 7 ) 280–292
avai lab le at www.sc iencedi rec t .com
journa l homepage: www.e lsev ier .com/ locate /dnarepai r
Collaborative roles of �H2AX and the Rad51 paralogXrcc3 in homologous recombinational repair
Eiichiro Sonodaa,1, Guang Yu Zhaoa,1, Masaoki Kohzakia,b,3, Pawan Kumar Dhara,2,Koji Kikuchia, Christophe Redonc, Duane R. Pilchc, William M. Bonnerc,Atsushi Nakanoa,d, Masami Watanabeb,3, Tatsuo Nakayamae,f,Shunichi Takedaa, Yasunari Takamif,∗
a CREST Laboratory, Department of Radiation Genetics, Kyoto University Graduate School of Medicine, Kyoto 606-8501, Japanb Department of Radiology and Radiation Biology, Graduate School of Biomedical Sciences, Nagasaki University,Nagasaki 852-8521, Japanc NIH/NCI/CCR/LMP, Building 37, Room 5050A MSC 4255, 9000 Rockville Pike, Bethesda, MD 20892, USAd Department of Cardiovascular Medicine, Kyoto University Graduate School of Medicine, Kyoto 606-8501, Japane Department of Life Science, Frontier Science Research Center, Miyazaki Medical College, University of Miyazaki,5200 Kihara, Kiyotake, Miyazaki 889-1692, Japanf Section of Biochemistry and Molecular Biology, Department of Medical Sciences, Miyazaki Medical College,University of Miyazaki, 5200 Kihara, Kiyotake, Miyazaki 889-1692, Japan
a r t i c l e i n f o
Article history:
Received 16 August 2006
Received in revised form
29 September 2006
Accepted 19 October 2006
Published on line 22 November 2006
Keywords:
�H2AX
Rad51 paralog
a b s t r a c t
One of the earliest events in the signal transduction cascade that initiates a DNA damage
checkpoint is the phosphorylation on serine 139 of histone H2AX (�H2AX) at DNA double-
strand breaks (DSBs). However, the role of �H2AX in DNA repair is poorly understood. To
address this question, we generated chicken DT40 cells carrying a serine to alanine mutation
at position 139 of H2AX (H2AX−/S139A) and examined their DNA repair capacity. H2AX−/S139A
cells exhibited defective homologous recombinational repair (HR) as manifested by delayed
Rad51 focus formation following ionizing radiation (IR) and hypersensitivity to the topoiso-
merase I inhibitor, camptothecin (CPT), which causes DSBs at replication blockage. Deletion
of the Rad51 paralog gene, XRCC3, also delays Rad51 focus formation. To test the interaction
of Xrcc3 and �H2AX, we disrupted XRCC3 in H2AX−/S139A cells. XRCC3−/−/H2AX−/S139A mutants
Homologous recombination
Genome stability
Double-strand break repair
were not viable, although this synthetic lethality was reversed by inserting a transgene that
conditionally expresses wild-type H2AX. Upon repression of the wild-type H2AX transgene,
XRCC3−/−/H2AX−/S139A cells failed to form Rad51 foci and exhibited markedly increased lev-
els of chromosomal aberrations after CPT treatment. These results indicate that H2AX and
XRCC3 act in separate arms of a branched pathway to facilitate Rad51 assembly.
∗ Corresponding author. Tel.: +81 985 85 3127; fax: +81 985 85 6503.E-mail address: [email protected] (Y. Takami).
1 These authors contributed equally to this work.2 Systems Biology Group, Bioinformatics Institute, Biopolis Street, 133 Research Reactor Institute, Kyoto University, Osaka 590-0494, Japan
1568-7864/$ – see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.dnarep.2006.10.025
© 2006 Elsevier B.V. All rights reserved.
8671 Singapore, Singapore..
( 2 0 0
1
TtachtkameifacM(a
thhttoatdsraipgmsD
g(dsNfActtitDwRyRcoT
d n a r e p a i r 6
. Introduction
he DNA double-strand break (DSB) poses a serious challengeo genome stability. Within a few minutes of its formation,
DSB triggers various checkpoint switches and repair pro-esses. One of the earliest responses is the phosphorylation ofistone H2AX at serine 139 to form �H2AX. The phosphoryla-ion is carried out by the phosphatidylinositol 3-kinase familyinases, DNA-PK, Atr, and Atm [1–3]. The response is highlymplified with several hundred to several thousand H2AXolecules being phosphorylated in the chromatin around
ach DSB. A role for �H2AX in a damage checkpoint signal-ng cascade has been suggested by the findings that �H2AXacilitates the recruitment and retention of a number of dam-ge checkpoint proteins, such as the Mre11/Nbs1/Rad50 (MRN)omplex, Brca1, 53BP1, and NFBD1/MDC1 to the DSB sites [4–6].ice lacking H2AX are hypersensitive to ionizing radiation
IR), and display a dramatic increase in cancer incidence inp53-deficient background [7,8].
Genetic studies in yeast demonstrated a causal rela-ionship between defective damage checkpoints and IRypersensitivity [9]. However, it is an open question whetherypersensitivity to IR in vertebrate cells is due to defec-ive cell cycle arrest. The cause of radio-sensitivity in ataxiaelangiectasia (A-T) cells remains controversial, being vari-usly attributed to compromised cell cycle arrest at the G1/Snd G2/M boundaries, loss of apoptosis, and repair abnormali-ies [10–17]. Likewise, damage checkpoint dysfunction in cellseficient in Nbs1 does not necessarily account for their IR sen-itivity, as reconstitution with a hypomorphic mutant Nbs1everses their IR sensitivity, but not their defective cell cyclerrest [18]. A similar hypomorphic mutation has also beendentified in the BRCA1 gene [18]. Thus, the defective check-oint of H2AX deficient mice does not necessarily explain theirenome instability and IR sensitivity. Alternatively, �H2AXight directly affect DSB repair, as accumulating evidence
uggests that the damage checkpoint pathways can controlSB repair pathways [16,17].
There are two major DSB repair pathways, non-homolo-ous end joining (NHEJ) and homologous recombinationHR), which appear to contribute differentially to DSB repair,epending on the origin of the DSBs. “Accidental” DSBs,uch as those induced by IR are preferentially repaired byHEJ in mammalian cells. On the other hand, DSBs resulting
rom blocked replication are repaired primarily by HR [19,20].ccordingly, HR is essential for genome stability in cyclingells, and plays a critical role in cellular tolerance to the DNAopoisomerase I inhibitor, CPT, which causes replication blockshat ultimate lead to DSBs [21]. HR is a multi-step processnvolving a number of repair proteins. During its early steps,he MRN complex and other unknown nucleases resect theNA at DSB sites to generate 3′ single strand (ss) overhangs,hich associate with Rad51, a RecA homolog [22]. Althoughad51, Rad52, and Rad54 similarly contribute to HR in buddingeast, Rad51 plays a considerably more important role than
ad52 or Rad54 in vertebrate cells. Accordingly, vertebratesells posses a number of Rad51 cofactors that control activityf Rad51, including the five Rad51 paralogs, Brca1, and Brca2.he five Rad51 paralogs, which include XRCC3, appear to act7 ) 280–292 281
as a functional unit to promote Rad51 assembly at DSB sites[23–25]. The resulting Rad51–ssDNA filaments invade otherintact homologous sequences to form a D-loop. Finally, DNAsynthesis from the invading strand results in gene conversion[26] (reviewed in Ref. [27]). In the budding yeast, histone H2Aat serine 129 is quickly phosphorylated upon DSB formation.However, the role of �H2A in DSB repair remains to be eluci-dated, because an alanine substitution at serine 129 has noeffect on HR or NHEJ in yeast.
The DT40 cell line provides a unique opportunity for dis-secting the mechanism of HR, because its highly efficienttargeted integration makes it possible to construct a widerange of HR mutants [28,29]. In attempt to identify a role for�H2AX in HR-mediated DSB repair, we constructed a chickenDT40 cell line, in which one H2AX allele carried a muta-tion changing serine 139 to alanine, the other allele beingcompletely deleted (H2AX−/S139A). H2AX−/S139A cells grew withnearly normal kinetics and exhibited a modest defect in HR,as previously reported [30]. However, deletion of the XRCC3gene in H2AX−/S139A cells resulted in loss of Rad51 foci for-mation at IR-induced DSBs, extensive chromosomal breaks,and subsequent cells death, as observed in Rad51 depletedcells [31] Our data unmask a critical role for �H2AX in prevent-ing chromosomal breaks by partially substituting for a Rad51paralog. Our findings suggest a novel function for histonemodification acting with Rad51 cofactors in facilitating homol-ogous recombination by facilitating Rad51 polymerizationat DSBs.
2. Materials and methods
2.1. Plasmid construction
A 9 kb genomic fragment containing the chicken H2AX genewas isolated from a DT40 genomic library [32]. pBSK con-taining the 9 kb fragment was recircularized at EcoRI sitesto remove a 3 kb fragment containing the single H2AX exonand 3′ sequence (pH2AX5′). A PCR-amplified 1.3 kb 3′ arm wasinserted into the pH2AX5′ containing 6 kb 5′ arm (pH2AX5′3′).To construct the H2AX knockout vector, a bsr resistant cas-sette was inserted into unique BamHI site between 5′ and 3′
arms of pH2AX5′3′. To construct the H2AX knock in constructs,a hisD resistant cassette was inserted into a unique BamHIsite, followed by inserting PCR-amplified genomic EcoRI–XhoIfragments containing wild-type H2AX or H2AXS139A, gener-ated by site-directed mutagenesis, at EcoRI (in the genome)and XhoI site (in the hisD cassette). To construct the H2AXexpression vector, chicken H2AX cDNA was inserted into anexpression vector driven by tet-repressible promoter followedby IRES-luciferase (tet-H2AX-IRES-luciferase) [33].
2.2. Generation of H2AX mutants
DT40 cells were sequentially transfected with the H2AX-bsrknockout vector, and subsequently the wild-type H2AX-
hisD, H2AXS139A-hisD or H2AXnull-hisD knockin vectorto obtain H2AX−/wild, H2AX−/S139A or H2AX−/− cells, respec-tively. For the generation of H2AX−/S139A/XRCC3−/− mutant,H2AX−/S139A cells were transfected with the XRCC3-hyg
6 ( 2
282 d n a r e p a i rknockout vector. The resulting H2AX−/S139A/XRCC3+/− cloneswere co-transfected with tet-H2AX-IRES-luciferase and ptTA-neo (Clontech). The expression of H2AX was monitored by theexpression of luciferase. The H2AX−/S139A/XRCC3+/− +H2AXclones were subsequently transfected with XRCC3-puro vec-tor to obtain H2AX−/S139A/XRCC3−/− cells. Three-independentclones were obtained.
2.3. Transcriptome analysis
Total RNAs were isolated from wild-type and H2AX mutantcells using the Trizol reagent (GibcoBRL), and purified by theRNAeasy mini kit (Qiagen) to remove contaminating genomicDNA. All analysis was done using the Affymetrix GeneChip®
System. In brief, to generate cRNA probes, 5 �g of total RNAwas reverse transcribed into double-strand cDNA using T7-oligo primers. The resulting products were used as templatesto synthesize biotinylated cRNA probes using GeneChip IVTlabeling kit. The probes were purified, fragmented, and thenhybridized to the GeneChip® Chicken Genome Array at 45 ◦Cfor 16 h. After hybridization, the GeneChips were washed andthen stained with streptavidin phycoerythrin. The stainedGeneChips was then scanned by a GeneChip Scanner 3000 andexpression values for each probe set and background bindinglevels were determined using the GeneChip GCOS program.The expression levels of several mRNAs were also confirmedby northern blot analysis (data not shown).
2.4. Western blot analysis
Histone extractions were performed as follows. Frozen cellpellets from approximately 1 × 107 cells were rapidly thawed,resuspended in 1 ml lysis buffer (10 mM Tris–HCl, pH 7.5, 1 mMMgCl2, 0.5% NP40). After centrifugation, pellets were resus-pended in 400 �l of ice-cold salt wash buffer (0.4 M NaCl inlysis buffer) and incubated on ice for 30 min. Pellets were thenspun down (10 min, 15,000 rpm) and resuspended in ice-cold75 �l extraction buffer (0.2 M H2SO4, 1% 2-mercaptoethanol)for 60 min and spun down. The supernatant was precipitatedby adding TCA to a final concentration of 25% and left on icefor 60 min. Precipitated histone was washed twice by adding1 ml of ice-cold acetone and incubating 10 min on ice. Aftercentrifugation, pellets were dissolved in 20 �l of 4 M urea. ForWestern blotting, 2 �l of the histone extracts were loaded perlane and electrophoresed by 12.5% SDS-PAGE gels. �H2AX wasvisualized by mouse anti-�H2AX antibodies (Upstate; BJW103),followed by incubation with HRP-conjugated secondary anti-bodies. Other histones were visualized by Coomassie BrilliantBlue (CBB) staining as a loading control. To examine thetotal amount of H2AX proteins, two-dimensional AUT-AUC gelanalysis was performed as described previously [34]. After gelswere stained with CBB, intensities of spots corresponding toH2AX and H2A were quantified by Image Gauge version 3.3(Fuji Film).
2.5. Analysis of cell-cycle progression and checkpoint
activationThe experimental methods for cell counting and cell-cycleanalysis were performed as described previously [35]. To
0 0 7 ) 280–292
monitor the S phase checkpoint, cells were either untreatedor irradiated with 4 Gy �-rays, then incubated for 1 h.[3H]thymidine (20 �Ci/ml) was pulsed for the last 15 min. Inhi-bition of DNA synthesis, which indicates the activation of Sphase checkpoint, was monitored by [3H]thymidine incorpo-ration. To monitor the G2/M checkpoint, cells were irradiatedwith 2 Gy �-rays, and then incubated with colcemid. Aliquotsof the culture were taken every hour. Cells in mitosis wereidentified by co-staining with propium iodine (PI) and antibodyagainst anti-phospho-histone H3 (Upstate, NY).
2.6. Indirect immunofluorescence
Staining and visualization of Rad51 foci were performed asdescribed previously [23]. Anti-Rad51 antibody was raisedagainst human Rad51 protein and affinity purified with Rad51-bound nitrocellulose membrane. For �H2AX foci visualization,cells were fixed by 4% paraformaldehyde (PFM), followedby permeabilization with 100% methanol for 5 min. Fixedcells were then stained with monoclonal anti-�H2AX (1:500,Upstate; BJW103) for 1 h, followed by anti-mouse Ig Alexa 488(Molecular Probes).
2.7. Measurement of induced recombination frequency
DSB-induced recombination frequency was measured asdescribed previously [36,37].
2.8. Colony formation assay
Colony formation assay was performed as described previ-ously [38]. To measure the sensitivity to CPT, we plated cellsinto methylcellulose medium containing CPT.
2.9. Chromosomal aberration analysis
Preparation of chromosome spreads and karyotype analysiswere performed as described previously [31,38]. For CPT-induced chromosomal aberrations, cells were incubated with10 nM camptothecin for 5.5 h, add colcemid, and incubated foranother 2.5 h before fixation.
3. Results
3.1. Generation and properties of H2AX mutant DT40cells
To investigate the role of the DNA damage-induced phospho-rylation of H2AX at Ser-139, we wished to generate cells thatexpressed only H2AX S139A mutant protein. To this end, wefirst deleted the whole coding sequence (null mutation) inone H2AX allele to generate H2AX−/+ cells and subsequentlyinserted the S139A mutation to create H2AX−/S139A cells. Togenerate control cells, we inserted the null mutation and wild-type H2AX construct into the remaining wild-type allele of
H2AX−/+ cells to obtain H2AX−/− (H2AX-null) and H2AX−/wildcells, respectively (Fig. 1A and B). We chose to knock-in theS139A mutation into the endogenous H2AX gene in DT40cells instead of expressing a H2AX S139A transgene under
d n a r e p a i r 6 ( 2 0 0 7 ) 280–292 283
Fig. 1 – Gene targeting of the H2AX locus. (A) Schematic representation of a part of the H2AX locus, and the targeted allele.The H2AX replacement constructs, which contains a BSR or HisD selection marker gene was used to target the H2AX locus.The BSR construct was used to generate H2AX−/+ cells. By transfecting the three different HisD constructs (shown in thethird and fourth rows) into the H2AX−/+ cells, H2AX−/−, H2AX−/wild, and H2AX−/S139A cells were generated. B, H, and RIindicate relevant BamHI, HindIII and EcoRI sites, respectively. The filled box represents the single H2AX exon. (B) Southernblot analysis using a probe indicated in (A). (C) Expression levels of H2AX. Two-dimensional analysis of H2A fraction (upperpanel). Western blot analysis of �H2AX in an acid-extracted histone fraction using a monoclonal anti-�H2AX antibody( ells
tegyHw
lower panel). (D) Immunocytochemical analysis of �H2AX. C
he control of strong promotor in H2AX-null cells because anxcess of H2AX proteins from the transgene may alter the
lobal expression pattern of genes. Two-dimensional gel anal-sis revealed that the levels of H2AX protein in H2AX−/wild and2AX−/S139A are comparable; the ratio of H2AX against H2Aas 0.098 and 0.092 in H2AX−/wild and H2AX−/S139A, respec-were analyzed at 20 min after 2 Gy �-rays. Scale bar = 10 �m.
tively (Fig. 1C, upper panel). The absence of IR-induced �H2AXwas verified by western blot analysis (Fig. 1C, lower panel) as
well as by immunofluorescence, using anti-�H2AX antibody(Fig. 1D). The level of �H2AX in H2AX−/wild cells was ∼four-foldlower than that of wild-type (H2AX+/+) cells (Fig. 1D), presum-ably because H2AX−/wild cells expressed a four-fold lower level
6 ( 2
284 d n a r e p a i rof H2AX protein due to the deletion of one allelic H2AX geneand integration of the marker gene in the other allele (Fig. 1C).
The proliferative properties of wild-type and the H2AX tar-geted clones were monitored by growth curves and by cellcycle analysis. H2AX−/wild cells showed the same growth rateas H2AX+/+ cells whereas the proliferation of H2AX−/S139A andH2AX−/− cells was slightly retarded due to apoptosis in a frac-tion of the cells (Fig. 2A and B). To assess the role of �H2AX inthe G2 checkpoint, we exposed H2AX mutant cells to 2 Gy of
ionizing radiation (IR), and subsequently monitored cell cycleprogression from G2 to M phase by measuring the mitoticindex using flow cytometry (Fig. 2C and supplementary Fig. 1).At 1 h after 2 Gy, the mitotic index was reduced to ∼30% rela-Fig. 2 – Growth properties and IR-induced checkpoint of H2AX mindicated genotypes. Each value represents the mean of the resuRepresentative cell-cycle distribution of the indicated cell cultureDNA-content (PI staining) in flow cytometric analysis. The gatesleftmost correspond to cells incorporating BrdU (∼S phase), G1 cesub-G1 fraction represent cells undergoing apoptosis. Numbers sIR-induced G2/M phase checkpoint. Cells were irradiated with 2 Gharvested at one and 2 h post-IR. The extent of cell cycle delay isthat of unirradiated cells. The mitotic index of unirradiated cellsthe mean for three-independent experiments. (D) �-Ray-inducedcultured for 1 h. [3H]thymidine was pulsed for the last 15 min. Thdefined as 100%. Error bars show the standard error of the meanrepresentative result from two separate experiments.
0 0 7 ) 280–292
tive to non-irradiated cells in H2AX+/+ cells and only to ∼50%in H2AX−/S139A and H2AX−/− cells, indicating a mild G2 check-point defect, as previously observed [4]. On the other hand,radio-resistant DNA synthesis, which reflects the functional-ity of S phase checkpoint during replication, was not impaired(Fig. 2D).
3.2. Defective double-strand break repair inH2AX−/S139A cells
H2AX−/S139A cells were slightly sensitive to IR but not toUV (Fig. 3A and B). H2AX−/S139A as well as H2AX−/− cellsshowed a significantly higher sensitivity to CPT (Fig. 3C) than
utant cells. (A) The relative growth rate is plotted from thelts from two-independent clones for each genotype. (B)s as measured by pulsed BrdU incorporation andat the upper-half, the lower-left, the lower-right and thells, G2/M cells and sub-G1 cells, respectively. The cells inhow the percentage of cells falling in each gate. (C)y �-rays, then continuously exposed to colcemid, andshown by the mitotic index of irradiated cells relative tois defined as 100%. Error bars show the standard error ofS phase checkpoint. Cells were irradiated with �-rays, ande [3H]thymidine incorporation of unirradiated control isfor triplicate culture. The data shown are the
d n a r e p a i r 6 ( 2 0 0 7 ) 280–292 285
Fig. 3 – Sensitivities of H2AX mutants to various genotoxic stresses. (A–C) Colony survival assay after the treatment of UV�-rays, and camptothecin (CPT). The doses of various genotoxic agents are displayed on the X-axis on a linear scale, whilethe percentile fractions of surviving colonies are displayed on the Y-axis on a logarithmic scale. Error bars show thestandard error of the mean for at least three-independent experiments. Plating efficiencies were 100% for wild-type and theH2AX mutant cells. (D) Level of chromosomal aberrations induced by IR. For each preparation, 100 mitotic cells wereanalyzed 3 h after 2 Gy �-rays. The average number of aberrations per cell ± S.E. is calculated as x/N ± √
x/N, based on thePoisson distribution of chromosomal aberrations. The numbers of cells analyzed and total aberrations are defined as N andx oso
Htret(
, respectively, 0 Gy indicates the level of spontaneous chrom
2AX+/+ cells. These results suggest that �H2AX contributeso cellular tolerance to DSBs, particularly those induced by
eplication blocks. Of note, H2AX−/wild cells showed a slightlylevated CPT sensitivity than H2AX+/+ cells, presumably dueo the decreased protein level of H2AX (∼25% of H2AX+/+ cells)Fig. 1C). To examine IR sensitivity at the G2 phase, we exposedmal aberrations.
an asynchronous population of cells to 2 Gy, harvested cellsand monitored chromosomal breaks in mitotic cells at 3 h
post-IR (Fig. 3D). Cells that were exposed to IR during theG2 phase are expected to enter mitosis within 3 h [39]. Thus,by measuring induced chromosomal breaks at 0–3 h post-IR, one can assess DSB repair kinetics during the G2 phase,
6 ( 2 0 0 7 ) 280–292
Table 1 – Targeted integration frequencies
Genotype analyzed Targeted loci
XRCC3 OVALBUMIN ATM
Wild-type 22/31 (71%) 22/25 (86%) n.d.H2AXwild 6/31 (20%) 19/23 (83%) 12/46 (26%)H2AX−/S139A 2/46 (4.3%) 9/23 (39%) 1/46 (2%)H2AX−/− 1/52 (1.9%) 19/72 (26%) n.d.XRCC3−/− n.d. 0/36 (0%) n.d.
Cells were transfected with targeting constructs of the indicatedloci. The data shown are the number of targeted clones at eachlocus divided by the number of drug-resistant clones analyzed. The
286 d n a r e p a i r
where HR is preferentially used over NHEJ for DSB repair inDT40 cells [35]. H2AX−/S139A mitotic cells showed a two-foldincrease in the number of induced chromosome breaks incomparison with H2AX−/wild cells. This observation, togetherwith the increased sensitivity to CPT, supports the idea thatH2AX−/S139A has a defect in HR-dependent DSB repair.
Intriguingly, H2AX−/wild cells showed higher CPT sen-sitivity in comparison with H2AX+/+ cells. Furthermore,over-expression of H2AX appeared to be toxic to the cells(data not shown). These observations imply that H2AX maychange the expression of some genes and thereby affect cel-lular the response to DNA damage. The level of H2AX could notbe directly measured due to possible cross reactivity of anti-body towards H2A. Instead, we investigated the effect of H2AXexpression on global gene expression by employing a chickengenome array carrying ∼30,000 genes. 2.5% of the total geneson the genome array exhibited significantly different expres-sion levels (<1/2 or <2) between H2AX+/+ and H2AX−/− cells.In contrast, the gene expression pattern between H2AX−/wild
and H2AX−/S139A were virtually identical (supplementary Fig.2). Thus, the phenotypic difference between H2AX−/wild andH2AX−/S139A should be attributed to H2AX phosphorylationstatus.
3.3. �H2AX facilitates homologous recombination
To evaluate HR, we measured the efficiencies of gene tar-geting, HR-dependent repair of DSB generated by the I-SceIrestriction enzyme, and Rad51 foci formation following �-rayirradiation [24,28,40]. The ratio of targeted to random integra-tion at the OVALBUMIN, ATM and XRCC3 loci was up to 10-foldlower in H2AX−/S139A and H2AX−/− cells than in H2AX−/wild
cells (Table 1). The targeting frequency at the XRCC3 and ATMloci might be further compromised by the genetic interactionsbetween H2AX and these genes, since the OVALBUMIN locus
showed a less dramatic reduction of targeting frequency. Inthe second experiment, we inserted an artificial HR substrate,SCneo [36,37] into the OVALBUMIN locus. In the resulting cells,we transiently expressed I-SceI, and measured the numberTable 2 – DSB-induced recombination frequencies
Genotype Plasmid transfected No. of neocolo
Experiment 1H2AX−/wild I-SceI 106,0H2AX −/S139A I-SceI 34,0
Experiment 2H2AX−/wild I-SceI 246,0H2AX−/S139A I-SceI + pBS 88,0
I-SceI + H2AX 89,0
Experiment 3Wild-type I-SceI 259,0XRCC3−/− I-SceI 24,5
SCneo was stably integrated into the OVALBUMIN locus of H2AX−/wild and Htransfected with 20 �g of the I-SceI expression vector (pCBASce) either alopBS, and subsequently plated in semi-solid medium with or without neomthe numbers of neo resistant colony by the numbers of the total colonies.
percent frequency is in parentheses. n.d. means experiments notdone.
of cells that successfully reconstituted the neomycin resis-tant gene by HR. The efficiency of HR-dependent DSB repairwas decreased to 55–60% (experiments 1 and 2 in Table 2)in H2AX−/S139A cells relative to the level in H2AX−/wild cells.Additionally, the reduction was reversed by co-transfection ofwild-type H2AX transgene together with I-SceI expression plas-mid (experiment 2 in Table 2). To investigate at which stepof HR �H2AX plays a role, we examined IR-induced nuclearRad51 focus formation, which reflects the assembly of Rad51at damage sites. In H2AX−/S139A cells, Rad51 focus formationwas significantly decreased 1 h post-IR treatment (Fig. 4A andB). Similarly, the reduction of IR-induced Rad51 foci formationat early time points, e.g. at 45 min post-IR, was observed inH2AX−/− murine ES cells (supplementary Fig. 3). In summary,�H2AX may be directly involved in HR, possibly at an early stepof HR.
3.4. Defects in �H2AX and XRCC3 have synergisticeffect on both genome instability and hypersensitivity to
camptothecinDiminished Rad51 focus formation is shared by several otherHR mutants, including each five Rad51 paralog deficient cells.
resistantnies
Total no. ofcolonies
No. of neoR/no.of total (%)
00 1,680,000 6.300 900,000 3.8
00 5,100,000 4.800 3,240,000 2.700 1,510,000 5.9
00 4,270,000 6.100 5,310,000 0.46
2AX−/S139A cells by gene targeting. 5 × 106 cells of each genotype werene or with wild-type H2AX expression vector or with control vectorycin. The ratio of recombination frequency is calculated by dividing
d n a r e p a i r 6 ( 2 0 0 7 ) 280–292 287
Fig. 4 – Delay of �-ray induced Rad51 subnuclear foci formation in H2AX−/S139A cells. (A) Micrographs showing Rad51 focusformation of the indicated genotypes at indicated times after 2 Gy IR. Scale bar = 10 �m. (B) Average number of Rad51 foci percell of the indicated genotypes at indicated times, following irradiation with 4 Gy. Two-independent experiments ares
TfrsbimmmlcvH(iii
hown. At lest 50 cells were analyzed per each data point.
his phenotypic similarity prompted us to investigate theunctional interaction between �H2AX and XRCC3, a rep-esentative Rad51 paralog. If both proteins function in theame pathway in the Rad51 focus formation, then the dou-le mutant should exhibit deficits similar to those foundn either single mutant. On the other hand, if the two
utants function in different pathways, then the doubleutant should exhibit further reductions in Rad51 focus for-ation. Unexpectedly, we were unable to construct a cell
ine deficient in both �H2AX and Xrcc3, suggesting that theombined deficiencies are synthetically lethal. To circum-ent this lethality, we inserted a transgene for wild-type2AX under the control of a tetracycline repressible promoter
tet-H2AX) into the H2AX−/S139A/XRCC3−/+, and subsequentlysolated H2AX−/S139A/XRCC3−/−/tet-H2AX cells. This line grewn the absence of tetracycline (tet), but ceased growing with anncrease in the G2/M fraction 5 days after the addition of tet to
the culture (Fig. 5A and data not shown). Further investigationshowed that �H2AX focus formation was absent in cultures4 days after tet addition (data not shown). Six days after tetaddition, the cells exhibited a greatly increased number ofspontaneous chromosomal breaks (Fig. 5B).
To gain an insight into the nature of the severe chromo-somal instability of H2AX−/S139A/XRCC3−/− cells, we exposedthem to DNA damaging agents 4 days after the additionof tet, when �H2AX foci were undetectable, and evaluatedthe efficiency of DSB repair by measuring the level of chro-mosome aberrations in metaphase spreads (Table 3). Uponexposure to CPT, the H2AX−/S139A/XRCC3−/− (tet+) doublemutant cells exhibited a marked increase in chromosomal
aberrations compared to the relevant single mutants (com-pare 1.52 for H2AX−/S139A/XRCC3−/− with 0.36 for XRCC3−/−and 0.68 for H2AX−/S139A in Table 3). This synergistic increasein CPT-induced chromosomal aberrations, together with the
288 d n a r e p a i r 6 ( 2
Fig. 5 – The H2AX−/S139A mutation in an XRCC3−/−
background is synthetically lethal. (A) Growth curves ofH2AX−/S139A/XRCC3−/−/tet-H2AX cell cultures in the absence(+H2AX) and presence of tetracyclin, which inhibitswild-type H2AX transgene expression. Two independentlyisolated clones were analyzed. (B) Chromosomal breaks areaccumulated in dying H2AX−/S139A/XRCC3−/− cells at day 6following addition of tetracyclin. The average number ofaberrations per cell ± S.E. is calculated as in Fig. 3. Incomparison with XRCC3−/− cells,H2AX−/S139A/XRCC3−/−/tet-H2AX cells displayed higherlevels of spontaneous aberrations, presumably because ahigh level of H2AX protein from the tet-H2AX transgene is
toxic to the cells.synthetic lethality of H2AX−/S139A/XRCC3−/−, provides com-pelling evidence that �H2AX and Xrcc3 act in separate armsof a branched pathway to maintain the genome stability. Incontrast, this synergism was not observed with IR-induced
−/S139A −/−
DSBs. The H2AX /XRCC3 cells showed only a mod-est increase in the level of IR-induced chromosomal breaksover that of the XRCC3−/− cells (compare 1.24 with 1.00 inTable 3). The differential effect of the combined mutation of0 0 7 ) 280–292
H2AX−/S139A with XRCC3−/− after CPT and IR treatment maybe attributed to partial rescue of repair by NHEJ. Since, NHEJcontributes to repair of IR-induced DSBs, but not to repair ofCPT-induced ones [19], it could mask a severe defect in HR-dependent repair of IR-induced DSBs in H2AX−/S139A/XRCC3−/−
cells.Since, both the H2AX−/S139A and the XRCC3−/− single
mutant cells exhibited delay in Rad51 focus formation(Fig. 4B), we asked whether or not the synergistic effectwas also observed in IR-induced Rad51 focus formation.Both single mutants displayed clearly visible foci formation(Figs. 4A and 6A). In contrast, focus formation was virtuallyabsent in H2AX−/S139A/XRCC3−/− cells even at a later time point(4 h post-IR), when the intensity of Rad51 foci reaches theirmaximum level (Fig. 6A and B). In summary, the syntheticlethality associated with extensive chromosomal aberrationsand loss of Rad51 foci formation in the H2AX−/S139A/XRCC3−/−
cells indicates that �H2AX and Xrcc3 may have a functionaloverlap in the assembly of Rad51 in HR.
4. Discussion
This is the first genetic study that clearly shows a rolefor �H2AX in the recruitment of Rad51 in HR-mediatedrepair. H2AX−/S139A cells showed modest defects in bothHR-dependent repair of I-SceI induced DSBs and Rad51 fociformation after IR, and significant increase in CPT sensi-tivity. To manifest a HR defect associated with the loss of�H2AX, H2AX−/S139A was combined with XRCC3−/−, becauseboth mutants showed delayed Rad51 foci formation. Indeed,these mutations were synthetically lethal to the cells, exhibit-ing loss of Rad51 focus formation and a dramatic increasein the level of chromosomal breaks during the cell cycle andfollowing CPT treatment. Thus, �H2AX may play an impor-tant role in repairing DSBs in Xrcc3-deficient cells but notin wild-type cells. Since, cells deficient in the other fourRad51 paralogs have similar phenotypes as XRCC3-deficientcells with respect to Rad51 and Rad54 focus formation,CPT sensitivity and the efficiency of gene targeting [24,25],it seems reasonable to expect that the other Rad51 par-alogs and �H2AX may also have complementary functions inmaintaining chromosomal integrity and repairing IR-inducedDSBs.
A role for �H2AX in HR-mediated DSB repair was alreadydocumented by previous studies that investigated the impactof H2AX disruption in mouse, and analyzed H2AX−/− cellsreconstituted with a H2AX S139A transgene [5,30,41]. However,the observed reduction of HR is modest, as was the HR defectin H2AX−/S139A DT40 cells. Furthermore, an altered expres-sion levels of H2AX might affect the general phenotype ofcells by changing global chromatin structure, as exemplifiedby the male infertility of H2AX−/− mice [42]. In the currentstudy, this problem was solved by comparing H2AX−/S139A withH2AX−/wild cells, which display virtually identical gene expres-sion profile.
The effect of the H2AX mutation on genome stabilitywas dramatically increased in the absence of Xrcc3. Thisfinding is remarkable, because inactivation of Rad54 (unpub-lished), FancC [43], Rad18 (J.E. Sale, personal communication),
d n a r e p a i r 6 ( 2 0 0 7 ) 280–292 289
Table 3 – The frequency of induced-chromosomal aberrations is greatly increased in the H2AX−/S139A/XRCC3−/− doublemutant
Agent Cell type Aberration type 1 2 1 − 2
Chromatid Chromosome Exchange Total Spon Induced
CPT H2AX+/+ 0.20 0.12 0 0.32 ± 0.08 0.02 ± 0.02 0.30 ± 0.08H2AX−/wild 0.20 0.24 0 0.44 ± 0.09 0.02 ± 0.02 0.42 ± 0.10H2AX−/S139A 0.26 0.44 0.02 0.70 ± 0.12 0.02 ± 0.02 0.68 ± 0.12XRCC3−/− 0.20 0.26 0 0.46 ± 0.10 0.10 ± 0.04 0.36 ± 0.11H2AX−/S139A/XRCC3−/− 1.24 0.48 0.32 2.04 ± 0.20 0.52 ± 0.10 1.52 ± 0.23
�-Ray H2AX−/wild 0.12 0.04 0 0.16 ± 0.06 0.02 ± 0.02 0.14 ± 0.06H2AX−/S139A 0.18 0.04 0.02 0.24 ± 0.07 0.02 ± 0.02 0.22 ± 0.07XRCC3−/− 0.64 0.44 0.08 1.10 ± 0.15 0.10 ± 0.04 1.00 ± 0.15H2AX−/S139A/XRCC3−/− 0.80 0.84 0.04 1.76 ± 0.19 0.52 ± 0.10 1.24 ± 0.21
Cells were treated with 100 nM camptothecin for 7 h and harvested (above). Cells were irradiated with 0.3 Gy of �-ray and harvested at 3 h post-IR (below). All cultures were treated with colcemid during the last 3 h before analysis. Chromosome aberrations were scored in 50 metaphasespreads, and average values per cell are shown. The number of aberrations per cell ± S.E. is calculated as x/N ± √
x/N, based on the Poissondistribution of chromosomal aberrations (1, 2 and 1 −2). The H2AX−/S139A/XRCC3−/−/tet-H2AX cells were analyzed at day 4 after the addition of
quene freq
BnXRai�
brpbaDaaptdlTdctsca
hrNsftaytwo
tetracyclin to inhibit expression from the tet-H2AX transgene. The fresubtracting the frequency of spontaneous aberrations (spon) from th
loom helicase (M. Seki and T. Enomoto, personal commu-ication), or Brca2 (Yamazoe et al., in preparation) in anrcc3-deficient background does not cause lethality, thoughAD52−/−/XRCC3−/− cells are non-viable [33]. Thus, �H2AXnd Xrcc3 are substantially complementary to each othern the maintenance of chromosomal integrity. Presumably,H2AX may prevent chromosomal instability in an XRCC3−/−
ackground by facilitating HR-mediated DSB repair duringeplication. Alternatively, a �H2AX-dependent damage check-oint might play a critical role in preventing chromosomalreaks in XRCC3−/− cells. The latter possibly seems less likelys H2AX−/S139A cells exhibited normal levels of radio-resistantNA synthesis, which reflects the intact S-phase checkpoint,nd only a marginal defect in the G2 damage checkpoint. Inddition, caffeine treatment of XRCC3−/− cells did not com-romise their viability (data not shown) though it did abolishhe G2 damage checkpoints [44,45]. Likewise, a severe defect inamage checkpoint caused by Atm deficiency does not cause
ethality in a Rad54 deficient HR mutant background [16].hese observations argue against the idea that a combinedefect of HR and damage checkpoint in H2AX−/S139A/XRCC3−/−
ells accounts for their synthetic lethality. On the other hand,he synthetic lethality was well correlated with the synergisticuppression of Rad51 foci formation in H2AX−/S139A/XRCC3−/−
ells. Therefore, we conclude that severe defect of HR mainlyccounts for the synthetic lethality.
The role of �H2AX or Xrcc3 in the activation of Rad51as remained elusive. �H2AX foci appear to contribute to theapid accumulation of DNA repair factors such as Brca1 andbs1 and may also facilitate the modification of chromatin
tructure near the DSB site. Since, Brca1 promotes Rad51ocus formation [5,46], �H2AX might activate Rad51 indirectlyhrough Brca1. Alternatively, since, Nbs1 together with Mre11nd Rad50 might promote 3′ overhang formation as does
east Mre11 complex, �H2AX could promote Rad51 assemblyhrough Mre11-dependent 3′ overhang formation. Recently, itas reported that yeast �H2A is responsible for recruitmentf Smc1, a component of the cohesin complex [47,48], whichcy of induced chromosomal aberrations (induced) was calculated byuency of all chromosomal aberrations (total).
appears to contribute to HR between sister chromatids in yeastand DT40 cells [49,50].
Accumulating evidence has suggested that the binding ofchromatin-modifying activities to �H2A or �H2AX at a DSB sitefacilitates DSB repair. For example, the Tip60 complex involv-ing histone acetyl transferase (HAT) might link �H2AX to DSBrepair. Drosophila melanogaster (Dm) Tip60 complex acceler-ates the turnover of a phosphorylated form of H2Av (�H2Av),the Drosophila homolog of H2AX [51]. A defect in Tip60 causeshypersensitivity to DSBs in mammalian cells, as well as inDrosophila presumably due to a reduction in DSB repair effi-ciency [52]. Other examples are the NuA4 HAT complex andINO80 complex, an ATP-dependent nucleosome remodelingcomplex, which are also associated with �H2A at DSB sites inbudding yeast [53,54]. Thus, �H2AX appears to be involved invarious aspects of local chromatin modification at DSBs. Ourpresent data support the idea that chromatin modificationand Rad51 cofactors may contribute cooperatively to Rad51assembly at DSBs.
The roles of damage checkpoints in cellular survival fol-lowing DNA damage appear to be distinctly different betweenyeast and vertebrate cells. In multi-cellular organisms, thedamage checkpoint not only transiently arrests the cell cycle,as it is the case in yeast, but under certain circumstances,can also trigger apoptosis. Thus, vertebrate damage check-points can have both a positive and a negative impact oncellular survival. Moreover, recent studies suggest that verte-brate DNA damage checkpoint proteins can directly facilitateDNA repair, including control of Rad54-dependent HR byATM [16], Atm-dependent phosphorylation of Artemis in NHEJ[17], Chk1-dependent activation of Rad51 [55], reduced genetargeting in Rad17-deficient cells [56] as well as a contri-bution of the Nbs1 damage checkpoint factor to HR [57].H2AX may constitute another means for cross-talk between
damage checkpoint proteins and DNA repair. In the signaltransduction of damage checkpoint, variable initial responsesat each lesion trigger signal amplification, which leads tothe same repair pathways. A defect in a single molecular
290 d n a r e p a i r 6 ( 2 0 0 7 ) 280–292
Fig. 6 – Loss of Rad51 focus formation in H2AX−/S139A/XRCC3−/− cells. (A) Rad51 focus formation 4 h after 4 Gy �-rayirradiation. Scale bar = 10 �m. (B) Frequencies of the cells with the indicated numbers of Rad51 foci are shown above each
istiche re
histogram. Sixty cells were analyzed for each genotype. Statnon-parametric U-test (p < 0.0001). The results show one of t
species, such as H2AX, involved in an early step of signaltransduction may not by itself lead to a prominent pheno-type in downstream DNA repair. As demonstrated by theremarkable phenotype of H2AX−/S139A/XRCC3−/− cells, genera-
al significance was calculated by Mann–Whitneypresentative results from two-independent clones.
tion of cell lines with multiple gene-disruptions may elucidatethe complicated functional redundancy between differentmolecules acting at a very early point in the DNA damageresponse.
( 2 0 0
A
WHaNcpTrReM
A
Si
r
d n a r e p a i r 6
cknowledgements
e thank Y. Sato for her technical assistance and D. Yabe and. Hashimoto (Kyoto University) for their help for microarraynalysis. We also acknowledge Drs. C. Morrison, R.T. Bree and. Lowndes (National University of Ireland, Galway) for criti-al reading and discussion. Financial support was provided inart by grant from Core Research for Evolutional Science andechnology (CREST) of Japan Science and Technology Corpo-ation, by the Center of Excellence (COE) grant for Scientificesearch from the Ministry of Education, Culture, Sports, Sci-nce and Technology of Japan, and by grants from The Ueharaemorial Foundation and The Naito Foundation.
ppendix A. Supplementary data
upplementary data associated with this article can be found,n the online version, at doi:10.1016/j.dnarep.2006.10.025.
e f e r e n c e s
[1] H. Wang, M. Wang, W. Bocker, G. Iliakis, Complex H2AXphosphorylation patterns by multiple kinases includingATM and DNA-PK in human cells exposed to ionizingradiation and treated with kinase inhibitors, J. Cell. Physiol.202 (2005) 492–502.
[2] T. Stiff, M. O’Driscoll, N. Rief, K. Iwabuchi, M. Lobrich, P.A.Jeggo, ATM and DNA-PK function redundantly tophosphorylate H2AX after exposure to ionizing radiation,Cancer Res. 64 (2004) 2390–2396.
[3] J.D. Friesner, B. Liu, K. Culligan, A.B. Britt, Ionizingradiation-dependent {gamma}-H2AX focus formationrequires ATM and ATR, Mol. Biol. Cell (2005).
[4] O. Fernandez-Capetillo, H.T. Chen, A. Celeste, I. Ward, P.J.Romanienko, J.C. Morales, K. Naka, Z. Xia, R.D.Camerini-Otero, N. Motoyama, P.B. Carpenter, W.M. Bonner,J. Chen, A. Nussenzweig, DNA damage-induced G2-Mcheckpoint activation by histone H2AX and 53BP1, Nat. CellBiol. 4 (2002) 993–997.
[5] A. Celeste, S. Petersen, P.J. Romanienko, O.Fernandez-Capetillo, H.T. Chen, O.A. Sedelnikova, B.Reina-San-Martin, V. Coppola, E. Meffre, M.J. Difilippantonio,C. Redon, D.R. Pilch, A. Olaru, M. Eckhaus, R.D.Camerini-Otero, L. Tessarollo, F. Livak, K. Manova, W.M.Bonner, M.C. Nussenzweig, A. Nussenzweig, Genomicinstability in mice lacking histone H2AX, Science 296 (2002)922–927.
[6] G.S. Stewart, B. Wang, C.R. Bignell, A.M. Taylor, S.J. Elledge,MDC1 is a mediator of the mammalian DNA damagecheckpoint, Nature 421 (2003) 961–966.
[7] A. Celeste, S. Difilippantonio, M.J. Difilippantonio, O.Fernandez-Capetillo, D.R. Pilch, O.A. Sedelnikova, M.Eckhaus, T. Ried, W.M. Bonner, A. Nussenzweig, H2AXhaploinsufficiency modifies genomic stability and tumorsusceptibility, Cell 114 (2003) 371–383.
[8] C.H. Bassing, H. Suh, D.O. Ferguson, K.F. Chua, J. Manis, M.Eckersdorff, M. Gleason, R. Bronson, C. Lee, F.W. Alt, Histone
H2AX: a dosage-dependent suppressor of oncogenictranslocations and tumors, Cell 114 (2003)359–370.[9] C. Redon, D.R. Pilch, E.P. Rogakou, A.H. Orr, N.F. Lowndes,W.M. Bonner, Yeast histone 2A serine 129 is essential for the
7 ) 280–292 291
efficient repair of checkpoint-blind DNA damage, EMBO Rep.4 (2003) 678–684.
[10] M.N. Cornforth, J.S. Bedford, On the nature of a defect incells from individuals with ataxia-telangiectasia, Science227 (1985) 1589–1591.
[11] P.A. Jeggo, A.M. Carr, A.R. Lehmann, Splitting the ATM:distinct repair and checkpoint defects inataxia-telangiectasia, Trends Genet. 14 (1998)312–316.
[12] M.S. Meyn, Ataxia-telangiectasia and cellular responses toDNA damage, Cancer Res. 55 (1995) 5991–6001.
[13] M.S. Meyn, Ataxia-telangiectasia, cancer and thepathobiology of the ATM gene, Clin. Genet. 55 (1999) 289–304.
[14] R.B. Painter, B.R. Young, Radiosensitivity inataxia-telangiectasia: a new explanation, Proc. Natl. Acad.Sci. U.S.A. 77 (1980) 7315–7317.
[15] J. Thacker, Cellular radiosensitivity in ataxia-telangiectasia,Int. J. Radiat. Biol. 66 (1994) S87–S96.
[16] C. Morrison, E. Sonoda, N. Takao, A. Shinohara, K.Yamamoto, S. Takeda, The controlling role of ATM inhomologous recombinational repair of DNA damage, EMBOJ. 19 (2000) 463–471.
[17] E. Riballo, M. Kuhne, N. Rief, A. Doherty, G.C. Smith, M.J.Recio, C. Reis, K. Dahm, A. Fricke, A. Krempler, A.R. Parker,S.P. Jackson, A. Gennery, P.A. Jeggo, M. Lobrich, A pathway ofdouble-strand break rejoining dependent upon ATM,Artemis, and proteins locating to gamma-H2AX foci, Mol.Cell 16 (2004) 715–724.
[18] B. Xu, S.T. Kim, D.S. Lim, M.B. Kastan, Two molecularlydistinct G(2)/M checkpoints are induced by ionizingirradiation, Mol. Cell. Biol. 22 (2002) 1049–1059.
[19] H. Hochegger, E. Sonoda, S. Takeda, Post-replication repair inDT40 cells: translesion polymerases versus recombinases,Bioessays 26 (2004) 151–158.
[20] M. Yamazoe, E. Sonoda, H. Hochegger, S. Takeda, Reversegenetic studies of the DNA damage response in the chickenB lymphocyte line DT40, DNA Repair (Amst.) 3 (2004)1175–1185.
[21] J. Thacker, A.N. Ganesh, DNA-break repair, radioresistanceof DNA synthesis, and camptothecin sensitivity in theradiation-sensitive irs mutants: comparisons toataxia-telangiectasia cells, Mutat. Res. 235 (1990) 49–58.
[22] G. Ira, A. Pellicioli, A. Balijja, X. Wang, S. Fiorani, W.Carotenuto, G. Liberi, D. Bressan, L. Wan, N.M.Hollingsworth, J.E. Haber, M. Foiani, DNA end resection,homologous recombination and DNA damage checkpointactivation require CDK1, Nature 431 (2004) 1011–1017.
[23] M. Takata, M.S. Sasaki, E. Sonoda, T. Fukushima, C. Morrison,J.S. Albala, S.M. Swagemakers, R. Kanaar, L.H. Thompson, S.Takeda, The Rad51 paralog Rad51B promotes homologousrecombinational repair, Mol. Cell. Biol. 20 (2000) 6476–6482.
[24] M. Takata, M.S. Sasaki, S. Tachiiri, T. Fukushima, E. Sonoda,D. Schild, L.H.T. Thompson, S. Takeda, Chromosomeinstability and defective recombinational repair in knockoutmutants of the five Rad51 paralogs, Mol. Cell. Biol. 21 (2001)2858–2866.
[25] Y. Yonetani, H. Hochegger, E. Sonoda, S. Shinya, H.Yoshikawa, S. Takeda, M. Yamazoe, Differential andcollaborative actions of Rad51 paralog proteins in cellularresponse to DNA damage, Nucleic Acids Res. 33 (2005)4544–4552.
[26] T. Kawamoto, K. Araki, E. Sonoda, Y.M. Yamashita, K.Harada, K. Kikuchi, C. Masutani, F. Hanaoka, K. Nozaki, N.Hashimoto, S. Takeda, Dual roles for DNA polymerase eta inhomologous DNA recombination and translesion DNA
synthesis, Mol. Cell 20 (2005) 793–799.[27] L.S. Symington, Role of RAD52 epistasis group genes inhomologous recombination and double-strand break repair,Microbiol. Mol. Biol. Rev. 66 (2002) 630–670.
6 ( 2
[57] H. Tauchi, J. Kobayashi, K. Morishima, D.C. van Gent, T.
292 d n a r e p a i r
[28] J.M. Buerstedde, S. Takeda, Increased ratio of targeted torandom integration after transfection of chicken B cell lines,Cell 67 (1991) 179–188.
[29] E. Sonoda, C. Morrison, Y.M. Yamashita, M. Takata, S.Takeda, Reverse genetic studies of homologous DNArecombination using the chicken B-lymphocyte line, DT40,Philos. Trans. R. Soc. Lond. B. Biol. Sci. 356 (2001) 111–117.
[30] A. Xie, N. Puget, I. Shim, S. Odate, I. Jarzyna, C.H. Bassing,F.W. Alt, R. Scully, Control of sister chromatid recombinationby histone H2AX, Mol. Cell 16 (2004) 1017–1025.
[31] E. Sonoda, M.S. Sasaki, J.M. Buerstedde, O. Bezzubova, A.Shinohara, H. Ogawa, M. Takata, Y. Yamaguchi-Iwai, S.Takeda, Rad51-deficient vertebrate cells accumulatechromosomal breaks prior to cell death, EMBO J. 17 (1998)598–608.
[32] Y. Takami, M. Higashio, T. Fukuoka, S. Takechi, T. Nakayama,Organization of the chicken histone genes in a major genecluster and generation of an almost complete set of the corehistone protein sequences, DNA Res. 3 (1996) 95–99.
[33] A. Fujimori, S. Tachiiri, E. Sonoda, L.H. Thompson, P.K. Dhar,M. Hiraoka, S. Takeda, Y. Zhang, M. Reth, M. Takata, Rad52partially substitutes for the Rad51 paralog XRCC3 inmaintaining chromosomal integrity in vertebrate cells,EMBO J. 20 (2001) 5513–5520.
[34] E.P. Rogakou, C. Redon, C. Boon, K. Johnson, W.M. Bonner,Rapid histone extraction for electrophoretic analysis,Biotechniques 28 (2000) 38–40.
[35] M. Takata, M.S. Sasaki, E. Sonoda, C. Morrison, M.Hashimoto, H. Utsumi, Y. Yamaguchi-Iwai, A. Shinohara, S.Takeda, Homologous recombination and non-homologousend-joining pathways of DNA double-strand break repairhave overlapping roles in the maintenance of chromosomalintegrity in vertebrate cells, EMBO J. 17 (1998) 5497–5508.
[36] R.D. Johnson, N. Liu, M. Jasin, Mammalian XRCC2 promotesthe repair of DNA double-strand breaks by homologousrecombination, Nature 401 (1999) 397–399.
[37] T. Fukushima, M. Takata, C. Morrison, R. Araki, A. Fujimori,M. Abe, K. Tatsumi, M. Jasin, P.K. Dhar, E. Sonoda, T. Chiba, S.Takeda, Genetic analysis of the DNA-dependent proteinkinase reveals an inhibitory role of Ku in late S-G2 phaseDNA double-strand break repair, J. Biol. Chem. 276 (2001)44413–44418.
[38] T. Okada, E. Sonoda, Y.M. Yamashita, S. Koyoshi, S. Tateishi,M. Yamaizumi, M. Takata, O. Ogawa, S. Takeda, Involvementof vertebrate polkappa in Rad18-independentpostreplication repair of UV damage, J. Biol. Chem. 277(2002) 48690–48695.
[39] E. Sonoda, T. Okada, G.Y. Zhao, S. Tateishi, K. Araki, M.Yamaizumi, T. Yagi, N.S. Verkaik, D.C. van Gent, M. Takata, S.Takeda, Multiple roles of Rev3, the catalytic subunit ofpolzeta in maintaining genome stability in vertebrates,EMBO J. 22 (2003) 3188–3197.
[40] N. Takao, H. Kato, R. Mori, C. Morrison, E. Sonada, X. Sun, H.Shimizu, K. Yoshioka, S. Takeda, K. Yamamoto, Disruption ofATM in p53-null cells causes multiple functionalabnormalities in cellular response to ionizing radiation,Oncogene 18 (1999) 7002–7009.
[41] C.H. Bassing, K.F. Chua, J. Sekiguchi, H. Suh, S.R. Whitlow,J.C. Fleming, B.C. Monroe, D.N. Ciccone, C. Yan, K. Vlasakova,D.M. Livingston, D.O. Ferguson, R. Scully, F.W. Alt, Increasedionizing radiation sensitivity and genomic instability in theabsence of histone H2AX, Proc. Natl. Acad. Sci. U.S.A. 99(2002) 8173–8178.
[42] O. Fernandez-Capetillo, S.K. Mahadevaiah, A. Celeste, P.J.
Romanienko, R.D. Camerini-Otero, W.M. Bonner, K. Manova,P. Burgoyne, A. Nussenzweig, H2AX is required forchromatin remodeling and inactivation of sex chromosomesin male mouse meiosis, Dev. Cell 4 (2003) 497–508.0 0 7 ) 280–292
[43] S. Hirano, K. Yamamoto, M. Ishiai, M. Yamazoe, M. Seki, N.Matsushita, M. Ohzeki, Y.M. Yamashita, H. Arakawa, J.M.Buerstedde, T. Enomoto, S. Takeda, L.H. Thompson, M.Takata, Functional relationships of FANCC to homologousrecombination, translesion synthesis, and BLM, EMBO J. 24(2005) 418–427.
[44] A. Blasina, B.D. Price, G.A. Turenne, C.H. McGowan, Caffeineinhibits the checkpoint kinase ATM, Curr. Biol. 9 (1999)1135–1138.
[45] J.N. Sarkaria, E.C. Busby, R.S. Tibbetts, P. Roos, Y. Taya, L.M.Karnitz, R.T. Abraham, Inhibition of ATM and ATR kinaseactivities by the radiosensitizing agent, caffeine, Cancer Res.59 (1999) 4375–4382.
[46] A. Celeste, O. Fernandez-Capetillo, M.J. Kruhlak, D.R. Pilch,D.W. Staudt, A. Lee, R.F. Bonner, W.M. Bonner, A.Nussenzweig, Histone H2AX phosphorylation is dispensablefor the initial recognition of DNA breaks, Nat. Cell Biol. 5(2003) 675–679.
[47] L. Strom, H.B. Lindroos, K. Shirahige, C. Sjogren,Postreplicative recruitment of cohesin to double-strandbreaks is required for DNA repair, Mol. Cell 16 (2004)1003–1015.
[48] E. Unal, A. Arbel-Eden, U. Sattler, R. Shroff, M. Lichten, J.E.Haber, D. Koshland, DNA damage response pathway useshistone modification to assemble a double-strandbreak-specific cohesin domain, Mol. Cell 16 (2004)991–1002.
[49] E. Sonoda, T. Matsusaka, C. Morrison, P. Vagnarelli, O. Hoshi,T. Ushiki, K. Nojima, T. Fukagawa, I.C. Waizenegger, J.M.Peters, W.C. Earnshaw, S. Takeda, Scc1/Rad21/Mcd1 isrequired for sister chromatid cohesion and kinetochorefunction in vertebrate cells, Dev. Cell 1 (2001) 759–770.
[50] C. Sjogren, K. Nasmyth, Sister chromatid cohesion isrequired for postreplicative double-strand break repair inSaccharomyces cerevisiae, Curr. Biol. 11 (2001) 991–995.
[51] T. Kusch, L. Florens, W.H. Macdonald, S.K. Swanson, R.L.Glaser, J.R. Yates III, S.M. Abmayr, M.P. Washburn, J.L.Workman, Acetylation by Tip60 is required for selectivehistone variant exchange at DNA lesions, Science 306 (2004)2084–2087.
[52] T. Ikura, V.V. Ogryzko, M. Grigoriev, R. Groisman, J. Wang, M.Horikoshi, R. Scully, J. Qin, Y. Nakatani, Involvement of theTIP60 histone acetylase complex in DNA repair andapoptosis, Cell 102 (2000) 463–473.
[53] J.A. Downs, S. Allard, O. Jobin-Robitaille, A. Javaheri, A.Auger, N. Bouchard, S.J. Kron, S.P. Jackson, J. Cote, Binding ofchromatin-modifying activities to phosphorylated histoneH2A at DNA damage sites, Mol. Cell 16 (2004)979–990.
[54] H. van Attikum, O. Fritsch, B. Hohn, S.M. Gasser,Recruitment of the INO80 complex by H2A phosphorylationlinks ATP-dependent chromatin remodeling with DNAdouble-strand break repair, Cell 119 (2004) 777–788.
[55] C.S. Sorensen, L.T. Hansen, J. Dziegielewski, R.G. Syljuasen,C. Lundin, J. Bartek, T. Helleday, The cell-cycle checkpointkinase Chk1 is required for mammalian homologousrecombination repair, Nat. Cell Biol. 7 (2005) 195–201.
[56] M. Budzowska, I. Jaspers, J. Essers, H. de Waard, E. vanDrunen, K. Hanada, B. Beverloo, R.W. Hendriks, A. de Klein,R. Kanaar, J.H. Hoeijmakers, A. Maas, Mutation of the mouseRad17 gene leads to embryonic lethality and reveals a role inDNA damage-dependent recombination, EMBO J. 23 (2004)3548–3558.
Shiraishi, N.S. Verkaik, D. vanHeems, E. Ito, A. Nakamura, E.Sonoda, M. Takata, S. Takeda, S. Matsuura, K. Komatsu, Nbs1is essential for DNA repair by homologous recombination inhigher vertebrate cells, Nature 420 (2002) 93–98.