Molecular Cell, Vol. 16, 1003–1015, December 22, 2004, Copyright ©2004 by Cell Press
Postreplicative Recruitment of Cohesin
to Double-Strand Breaks Is Required for DNA Repair
hesion, which is a prerequisite for correct chromosome
segregation (reviewed in Haering and Nasmyth ).
The evolutionarily conserved core cohesin complex
consists of Scc1 (Mcd1/Rad21), Scc3, and two struc-
tural maintenance of chromosomes (SMC) proteins,
Guacci et al., 1997; Hartman et al., 2000; Losada et al.,
1998; Michaelis et al., 1997; Vass et al., 2003). Cohesin
is loaded onto chromosomes before DNA replication in
a process that requires the Scc2/Scc4 complex in yeast
mediated connections are formed at centromeres and
at regular intervals along chromatid arms (Laloraya et
al., 2000; Megee et al., 1999; Tanaka et al., 1999) in a
process involvingthe Eco1/Ctf7 protein (Skibbenset al.,
1999; Toth et al., 1999). In S. cerevisiae, sister chroma-
tids stay linked by cohesin along their entire length until
brate cells, arm-cohesin is released in prophase in a
Polo-like kinase 1 (PLK1)-promoted process (Losada et
al., 1998, 2002; Sumara et al., 2002; Waizenegger et al.,
2000). At anaphase, remaining cohesion is dissolved
through cleavage of Scc1 by the protease Separase,
which until then is kept inactive through its interaction
with its inhibitor Securin (Pds1) (Ciosk et al., 1998; Uhl-
mann et al., 1999, 2000).
The central role of cohesin in DNA repair becomes
that interference with cohesin function also impedes
DNA repair and recombination (Birkenbihl and Subra-
mani, 1992; Sjo ¨gren and Nasmyth, 2001; Sonoda et al.,
2001). The most straightforward interpretation of this is
that sister chromatid cohesion is required for DNA re-
pair. This is supported by the finding that damage of
cation is not repaired, regardless of the presence of
functional, chromatin bound cohesin (Sjo ¨gren and Nas-
myth, 2001). The G2 DNA damage checkpoint, partly
working through the inhibition of PLK1 (Smits et al.,
2000) and the stabilization of Securin (Cohen-Fix and
Koshland, 1997; Yamamoto et al., 1996), also argues for
a central role of cohesion in DNA damage repair. Even
though the requirement of sister chromatid cohesion for
toa morecomplex roleforcohesin inthe repairprocess.
Human cohesin subunits are involved in DNA damage-
induced cell cycle checkpoint activation (Kim et al.,
2002b; Yazdi et al., 2002) and interact physically with
Rad50 (Kim et al., 2002a). Rad50 is one of three compo-
nents of the Mre11/Rad50/Nbs1(Xrs2) (MRN) complex
(van den Bosch et al., 2003). Moreover, the cohesin
complex has been shown to accumulate in regions of
damaged DNA in human cells in a Rad50/Mre11-depen-
dent manner (Kim et al., 2002a). Despite these connec-
tions, the interplay between cohesin, sister chromatid
cohesion, and DNA repair is poorly understood. To shed
light on this matter, we have further investigated the
DNA repair role of cohesin in S. cerevisiae. We show
here that cohesin localizes at DSBs induced in the G2
Lena Stro ¨m,1Hanna Betts Lindroos,1
Katsuhiko Shirahige,2and Camilla Sjo ¨gren1,*
1Department of Cell and Molecular Biology
Berzelius va ¨g 35
171 77 Stockholm
2Gene Research Center
Tokyo Institute of Technology
4259 Nagasuta, Midori-Ku
Chromosome stability depends on accurate chromo-
some segregation and efficient DNA double-strand
break (DSB) repair. Sister chromatid cohesion, estab-
lished during S phase by the protein complex cohesin,
is central to both processes. In the absence of cohe-
sion, chromosomes missegregate and G2-phase DSB
repair fails. Here, we demonstrate that G2-phase re-
pair also requires the presence of cohesin at the dam-
age site. Cohesin components are shown to be re-
DNA breaks induced during G2. We find that in the
absence of functional cohesin-loading proteins (Scc2/
Scc4), the accumulation of cohesin at DSBs is abol-
ished and repair is defective, even though sister chro-
matidsare connectedby Sphase generatedcohesion.
Evidence is also provided that DSB induction elicits
establishment of sister chromatid cohesion in G2, im-
plicating that damage-recruited cohesin facilitates
DNA repair by tethering chromatids.
DNA DSB repair is fundamental for cell survival and
genomic stability. Survival after DSB induction varies
between organisms and depends on the cell cycle
phase, often being lowest in G1 and reaching a maxi-
mum in late S/G2. A prominent example of this is the
budding yeast Saccharomyces cerevisiae (S. cerevis-
iae), which shows a ?10 000-fold increased resistance
to X-rays in G2 as compared to G1 (Brunborg and Wil-
liamson, 1978). The increase in cell survival during DNA
replication has been attributed to the formation of two
identical copies of each chromosome, the sister chro-
matids, which serve as templates for DSB repair by
homologous recombination (Kadyk and Hartwell, 1992).
The physical linkage of sister chromatids by the protein
complex cohesin is also vital for DSB repair during this
period of the cell cycle (Sjo ¨gren and Nasmyth, 2001);
however, the precise role of cohesin remains unclear.
As indicated by its name, cohesin is required for the
establishment and maintenance of sister chromatid co-
ment abolishes DNA repair. We provide evidence that
DSB generation allows the establishment of de novo
sister chromatid cohesion in G2 cells, implicating dam-
age-recruited cohesin in holding the broken chromatid
near its undamaged sister template. Our findings raise
the possibility that cohesin recruited to replication-
induced DNA lesions could contribute to sister chroma-
tid cohesion in normal cell cycle progression.
complexes in a process similar to the one occurring
before replication in each cell cycle, such a situation
could be created by destroying cohesin chromatin load-
ing factors (i.e., Scc2/Scc4) in the G2 phase. We there-
fore initially tested if cohesin accumulated close to the
HO-induced DSB in G2 and, subsequently, whether this
process was Scc2 dependent.
Smc1 and Scc1 Are Loaded at DSB in G2/M
To determine whether cohesin components localize to
a DSB induced during the G2 phase of the cell cycle,
the cells with Myc-tagged Smc1 or Scc1, GAL:HO, and
the HO recognition site on ChrV were again utilized. In
addition, their sole CDC20 gene was now regulated by
growth in minimal media lacking methionine, cells were
transferred to rich media containing methionine to re-
press CDC20 expression. FACS-scan analysis of the
cells showed that a 3.5 hr repression of CDC20 resulted
in the expected accumulation of cells in G2/M (Figure
2A). At this point, the culture was divided and galactose
added to one half, thereby inducing the HO endonucle-
ase. ChIP analysis was performed on cells collected
from both cultures before, 2, and 4 hr after the split.
Both Scc1 and Smc1 accumulated near the DSB in cells
expressing the HO endonuclease but remained low in
noninduced cells (Figure 2B). Break formation was de-
in cells where the endonuclease had been induced (Fig-
Scc1 and Smc1 Accumulate at HO-Induced DSB
If cohesin has DNA-repair roles distinct from its function
in sister chromatid cohesion, it is conceivable that the
complex is to be found at the site of damage. We used
Scc1 and Smc1 as markers for the entire cohesin com-
plex and investigated their binding close to a DSB by
chromatin immunoprecipitation (ChIP). Myc-tagged ver-
sions of Smc1 or Scc1 were crosslinked to chromosomes
in vivo. After chromatin shearing, immunoprecipitation,
and reversal of crosslinks, interacting chromosomal re-
gions were analyzed by PCR. DSBs were generated by
(GAL:HO), and cohesin distribution was analyzed before
and after DSB induction in a chromosomal region where
cohesin binding in undamaged cells had already been
determined (Tanaka et al., 1999). By using the same
PCR primer pairs as Tanaka and coworkers, DNA frag-
ments in a region spanning 534–558 103bases (kb) from
(Tanaka et al., 1999 and Figure 1A). Galactose induction
of HO in logarithmically growing cells that lack an HO
recognition site in this region caused no change in the
distribution pattern of Smc1 compared to noninduced
cells (compare Figures 1B and 1C, ?HO cut site) and
ever, when an HO cut site was introduced 541 kb from
the left telomere, in a region that normally binds little or
no Smc1 or Scc1, HO expression caused the accumula-
tion of both proteins at 534 and 542 kb from the left
telomere; i.e. 7 kb downstream and 1 kb upstream of
the DSB (Figures 1C–1E). To correlate the increase in
Smc1 and Scc1 binding with DSB formation, cleavage
of ChrV was analyzed. Whole yeast chromosomes were
prepared from samples withdrawn from the cultures
used for the ChIP analysis, and ChrV was detected by
Southern blot after separation by pulse field gel electro-
phoresis (PFGE).This confirmed thatonly incells having
an HO recognition site on ChrV did endonuclease ex-
580 kbchromosome byapproximately 40kb (Figure1F).
If the physical accumulation of cohesin components in
the vicinity of a DSB plays a role in DNA repair, inhibition
of this accumulation would impede the repair process.
Studies of this issue encounter an important obstacle:
the central role of sister chromatid cohesion in DNA
repair. As the destruction of cohesin’s function abol-
ishes S phase-generated cohesion that, in turn, is re-
quired for DNA repair, a situation has to be created in
which chromatid cohesion remains intact, but the build
up of cohesin near the break is perturbed. If cohesin
accumulation atDSBsisduetoloading ofsolublecohesin
Loading of Smc1 at DSBs in G2/M
Requires Scc2 Function
The S. cerevisiae Scc2 and Scc4 proteins are required
for loading of the cohesin complex onto chromatin in
late G1/early S phase (Ciosk et al., 2000). Consequently,
progression from G1 through S phase in the absence
of Scc2 or Scc4 function is a lethal event due to the
of cohesion during G2 (Ciosk et al., 2000), making it
possible to test the role of the loading factors in cohesin
accumulation at DSB in this cell cycle phase. Wild-type
(wt) or scc2-4 cells, both containing GAL:HO, the HO
repressible CDC20 gene, were first arrested in G2/M
by Cdc20 depletion at permissive temperature (25?C).
Complete arrest was confirmed by FACS (Figure 3A). In
an initial experiment, the temperature was increased
to destroy Scc2 function before induction of the HO
endonuclease. However, activation of the HO endonu-
clease appeared to be ineffective at the elevated tem-
perature, and no breakage of ChrV was detected by
PFGE (data not shown). Galactose was therefore added
to the G2-arrested cells 1 hr before increasing the tem-
perature to 35?C, and samples for ChIP were withdrawn
at this time point, 1, 2, and 4 hr after HO induction.
Although these conditions rendered breakage of ChrV
less efficient than in previous experiments (compare
Figures 3C and 2C), Smc1 still accumulated near the
HO cut in wt cells (Figure 3B). In Scc2 mutated cells,
however, no accumulation was detected, showing that
Scc2 function is indeed required for efficient loading of
DSB Repair Requires Cohesin at the Break
Figure 1. Cohesin Subunits Smc1 and Scc1 Localize to an HO-Induced DSB
(A) Schematic representation of the ChrV region investigated by ChIP. Positions of primer pairs (below) and sizes of corresponding PCR
products (above) are indicated. Arrow denotes the HO cut site.
(B) ChIP analysis of logarithmically growing wt (CB117) and Smc1-Myc cells (CB127) grown in glucose-containing media at 25?C. ChrV regions,
as outlined in (A), were amplified by PCR. Resulting PCR products were separated on a 2.3% agarose gel and stained with ethidium bromide.
Each band corresponds to a specific region of ChrV as indicated. Both strains contain the galactose-inducible GAL:HO. Abbreviations: WCE,
whole-cell extract; IP, immunoprecipitated material.
(C) Smc1-Myc cells containing GAL:HO alone (CB127, ?HO cut site) or in combination with an HO recognition site at position 541 on ChrV
(CB166, ?HO cut site) were first grown in raffinose at 25?C and then subjected to ChIP analysis 0 and 5 hr after induction of GAL:HO. ChrV
regions, as outlined in (A), were amplified by PCR. Reaction products in the linear range of serial dilutions of WCE and IP were separated
and stained as in (B).
(D) Quantification of results obtained in (C). The ethidium bromide staining intensity of each band was quantified with Image Gauge. Left
graph, the percentage of IP compared to the material in WCE is shown for each chromosomal region. Right graph, intensity of bands relative
to that representing the 549.7 region, known to bind cohesin in undamaged cells.
(E) The HO endonuclease was induced in logarithmically growing Scc1-Myc cells (CB178) containing the HO cut site at 541 on ChrV. ChIP
analysis was performed on samples collected 0 and 4 hr after HO induction. Quantification was carried out as in (C). Left graph, percentage
of IP as in (D). Right graph, intensity of bands relative to that representing the 549.7 region, as in (D).
(F) PFGE and Southern blot analysis of ChrV from Scc1-Myc cells without (?HO cut site) or with (? HO cutsite) HO recognition site on ChrV.
Chromosomes were prepared from samples withdrawn 0, 2, and 4 hr after HO induction. After blotting, a ChrV-specific probe was used for
detection of both the uncleaved and the shorter, cleaved form of ChrV (ChrV, arrows).
Figure 2. Cohesin Subunits Smc1 and Scc1 Localize to a DSB in G2/M-Arrested Cells
(A) FACS-scan analysis of logarithmically growing Scc1-Myc (CB202) and Smc1-Myc (CB188) (left and right, respectively) cells containing an
HO cut site on ChrV, GAL:HO, and with their sole CDC20 gene under control of the MET3 promoter were arrested in G2/M by growth in
methionine-containing media (?Met) at 25?C. After complete arrest, the HO endonuclease was induced by galactose (?Gal) in half of the
(B) ChIP analysis and quantification of chromatin-immunoprecipitated material was performed 0, 2, and 4 hr after HO induction by galactose
addition (?Gal), or at corresponding time points in noninduced, control cells (?Gal). Band detection and quantification was done as in Figure
1C. Left graphs, percentage of IP compared to the material in WCE as in Figure 1. Right graphs, intensity of bands relative to that representing
the 549.7 region.
(C) Southern blot detection of ChrV. Simultaneously with sample-withdrawal for ChIP analysis (B), samples were collected for chromosome
preparation and analyzed by PFGE and Southern blot as in Figure 1F. Arrows indicate the uncleaved and the shorter, cleaved ChrV.
Smc1 to the break site (Figure 3B). This made it possible
to test whether DSB repair in G2 necessitates cohesin
loading onto the damaged DNA.
mutant ssc2-4 and scc4-4 cells (generously provided
by K. Nasmyth) was compared. In these cells, sister
35 kb from the centromere of ChrV, which is marked
with tetracycline (Tet) operators that bind GFP-linked
Tet-repressors. In addition, the strains had their sole
Cells were arrested in G2/M by Cdc20 depletion at 21?C
DSB Repair in G2/M Requires Scc2
and Scc4 Functions
To investigate whether cohesin loading factors are re-
quired for DNA DSB repair in G2, DNA repair in wt and
DSB Repair Requires Cohesin at the Break
Figure 3. Localization of Smc1 to a DSB Induced in G2/M Is Dependent on the Cohesin Loading Factor Scc2
(A) Wt (CB188) or scc2-4 (CB187) cells with Smc1-Myc, GAL:HO, and an HO cut site on ChrV were synchronized in G2/M as in Figure 2. Time
points for addition of methionine-containing media (?Met), galactose (?Gal) and temperature increase (35?C) are indicated.
(B) Samples for ChIP analysis were collected at 0, 1, 2, and 4 hr after addition of galactose. Quantification of ChIP PCR products was performed
as in Figure 1C.
(C) Samples for detection of the HO-induced break on ChrV were taken at indicated time points. Chromosomes were prepared and analyzed
with PFGE and Southern blot as in Figure 1F. Note that as compared to other samples on the scc2-4 gel, 30% less material was loaded in
the lane representing the 2 hr time point.
for 3.5 hr, and complete arrest was verified by FACS
(data not shown). The temperature was subsequently
increased to 35?C to destroy Scc2 and Scc4 functions
in the mutant cells, and 30 min later cultures were ?-ray
irradiated to induce DSBs. Samples for chromosome
aration were collected before irradiation and during a
2.5 hr recovery period, during which the G2 arrest was
maintained. As shown previously (Ciosk et al., 2000),
cells throughout the course of experiment (Figure 4A).
Figure 4. Cohesin Loading Factors Scc2 and Scc4 Are Necessary for DSB Repair in G2/M
(A) Wt (K8465), scc2-4 (K8466), scc4-4 (K8467), or scc1-73 (K8468, included as an example of cells with temperature-sensitive cohesion) cells
with their sole CDC20 gene under the control of the GAL1-10 promoter were arrested in G2/M by growth in glucose-containing media at 21?C.
Temperature was increased to 35?C at ?0.5 hr, and cells were ?-ray irradiated at 0 hr with 150 Gy. Percentage of sister chromatid separation
at the URA3 locus was determined immediately before (?0.5 hr) and every 30 min after temperature increase by using the Tet-repressor-
GFP/ Tet-operator system.
(B) Southern blots of PFGE on DNA repair samples collected immediately before, after (0 hr), and every 30 min subsequent to irradiation for
2.5 hr from wt, scc2-4, or scc4-4 cells. Hybridization was performed with a radioactive probe recognizing both ChrXVI and a loading control
(C) Quantification of ChrXVI signals in (B). ChrXVI values were normalized to loading control chromosome signals. Signals corresponding to
samples taken immediately before irradiation were set to 100% arbitrary units.
blot of chromosome 16 and an internal control chromo-
some (Sjo ¨gren and Nasmyth, 2001). In all three strains,
the chromosome 16 signal decreased to 20%–30% of
the initial value after irradiation and was subsequently
restored to about 80% during the recovery period in wt
cells (Figures 4B and 4C). In scc2-4 and scc4-4 mutants,
cates defective DNA repair (Figures 4B and 4C). This
demonstrates that although the functions of Scc2 and
matid cohesion in G2 (Ciosk et al., 2000), they play a
crucial role in DNA repair at this stage of the cell cycle.
of cohesin at DSBs is Scc2 dependent, this argues that
loading of cohesin onto damaged DNA is central to DSB
repair in G2.
links between sister chromatids during S phase, it is
chromatids together and thereby facilitate the repair
process. If true, these complexes could interfere with
sister chromatid separation; i.e., they would form novel
sister chromatid cohesion sites. This can be tested in
are present on the chromatids. To achieve this, we used
temperature sensitive smc1-259 cells (Michaelis et al.,
1997) containing the wt SMC1 gene under the control
of the GAL1-10 promoter. The strain also contained the
URA3 Tet-operon/Tet-repressor-GFP system and a sin-
gle CDC20 gene regulated by the MET3 promoter. The
experimental setup is outlined in Figure 5A.
Cells were first arrested in G2/M by Cdc20 depletion
at 21?C, and arrest maintenance was verified by FACS
(Figure 5B). Because cells entered the arrest under non-
inducing conditions for wt SMC1, the sister chromatid
cohesion created during replication was mediated by a
temperature-sensitive cohesin complex containing the
Smc1 mutant. When all cells had reached G2, SMC1
DSB-Recruited Cohesin Mediates Sister
What is the molecular function of cohesin recruited to
DSBs? Because DNA repair requires cohesin-mediated
DSB Repair Requires Cohesin at the Break
Figure 5. DSB-Induced Loading of Cohesin
onto Chromatin Generates Cohesion
(A) Schematic outline of the experimental
(B) FACS-scan analysis of temperature-sen-
sitive smc1-259 cells containing wt SMC1-
MYC controlled by the GAL1-10 promoter,
the Tet-repressor-GFP/ Tet-operator system,
and a single CDC20 gene regulated by the
MET3 promoter (CB198). Cells were arrested
in G2 at permissive temperature (21?C) by
then the culture was split and galactose
(?Gal) was added to one half in order to in-
duce expression of wt SMC1-MYC. The cul-
tures were then divided a second time and
one half was irradiated (?IR) with 450 Gy
whereas the other was left as unirradiated
control (“C”). After irradiation, cells were left
at permissive temperature for 1.5 hr before
the temperature was raised to 35?C. ?Gal, no
induction of wt SMC1-MYC. ?Gal, induction
at the URA3 locus, as determined by the Tet-
repressor-GFP/ Tet-operator system. Cell
samples were withdrawn every 30 min during
the 2 hr period at 35?C. See (A) and (B) for de-
(D) Percentage of cells with divided nuclei
as determined by microscopic observation of
four 6-diamidino-2-phenylindole (DAPI)-stained
cells collected at the same time points as in
(C). See (A) and (B) for details.
Figures 5C and 5D show the mean of three
experiments with SD included.
expression was induced in half of the culture by galac-
toseaddition. After60min, bothinduced anduninduced
cultures were split again, and half of each was treated
with ?-rays at 450 Gy. From this moment on, chromatid
separation at the URA3 locus was determined in cells
from all cultures every 30 min. After irradiation, cells
were cultured at 21?C for 1.5 hr to allow loading of
cohesinat irradiation-inducedDSBs.Here, onlytemper-
ature-sensitive cohesin complexes would be loaded at
treated cells a mixture of wt and mutated complexes
would be recruited to the damaged regions. After the
loading period, the temperature was raised to 35?C to
render the mutated smc1-259 nonfunctional, thereby
destroying both the cohesin complexes loaded during
S phase and the smc1-259-containing complexes re-
cruited to DNA breaks. Before irradiation and during the
incubation at 21?C, sister chromatids remained associ-
ated in all four cultures (data not shown). The upshift in
temperature caused sister chromatid separation in both
irradiated and unirradiated cells in the absence of wt
Smc1, which was also the case in unirradiated cells
expressing the functional SMC1 gene. In contrast, sis-
ter chromatids remained together in irradiated SMC1-
expressing cells (Figure 5C). To exclude the possibility
that sister chromatid separation detected in irradiated
mentation of the G2 DNA damage checkpoint, nuclear
division was scored in cells from each culture. Nuclear
age signaling was intact in both cell types (Figure 5D).
In unirradiated cells, the amount of divided nuclei in-
creased during the course of the experiment, potentially
reflecting the presence of small amounts of Cdc20 pro-
duced due to incomplete repression of the MET3 pro-
moter. Another explanation for the absence of sister
separation in irradiated, wt SMC1-expressing cells
could be that DNA repair was dependent on fully func-
tional Smc1, and that repair-created DNA links able to
prevent sister separation would thus only form in the
presence of wt Smc1. However, DNA repair as scored
by PFGE and Southern blot was equally efficient during
the incubation at 21?C, both in the presence and ab-
sence of functional Smc1 (data not shown), arguing
against such an interpretation. Taken together, these
data show that DSB-recruited cohesin is able to estab-
lishsister chromatidcohesionand suggestthat theDNA
repair role of these cohesin complexes is to hold a dam-
aged chromatid in proximity to its undamaged sister
reflect the fact that HO cleavage at this site induces a
recombination process with the HML? and HMRa loci
present on the same chromosome (reviewed in Haber
Cohesin Is Recruited to ?50 kb Regions
Surrounding HO-Induced DSBs
The irradiation dose utilized in the above experiment
(Contopoulou et al., 1987; Lisby et al., 2001; Potter and
detected at a single position in the genome, the forma-
tion of cohesion suggested major changes in cohesin
ization on entire chromosomes (Chr) III, V, and VI was
investigated by using a ChIP on chip method based on
a high-resolution tiling array (Katou et al., 2003). Cells
containing GAL:HO with its recognition site on ChrV and
Flag-marked Scc1 were grown in raffinose-containing
media at 23?C and arrested in G2 by benomyl treatment
(data not shown). Samples were subsequently with-
drawn from cultures before and 4 hr after induction of
HO expression by the addition of galactose. Ensuing
plification, labeling, and hybridization to ChrIII, ChrV,
and ChrVI arrays were performed as described pre-
viously (Katou et al., 2003). Control cells expressing un-
tagged Scc1 gave few specific signals, as shown for
ChrVI (Figure 7B). ChIP analysis on extracts from SCC1-
FLAG-expressing cells, however, showed that Scc1 lo-
calized at centromeres and at restricted regions on the
arms of all three chromosomes (Figures 6A–6C, 0 hr).
This is in line with earlier studies of the Scc1 distribution
pattern (Lalorayaet al., 2000;Megee et al.,1999; Tanaka
et al., 1999) and identical to what has been determined
recently by Uhlmann, Shirahige, and coworkers (Len-
gronne et al., 2004). When the endonuclease was in-
duced in cells lacking the HO recognition site on ChrV,
no change in distribution pattern was detected (data
not shown). In contrast, induction of a DSB at position
541 on ChrV caused recruitment of Scc1 to a ?50 kb
region spanning the break site, leaving the distribution
of Scc1 in other parts of ChrV and on ChrVI largely
unchanged (Figures 6A and 6B, 4 hr). This confirmed
our results from conventional ChIP that demonstrated
Notably in the 542 kb region, 1 kb from the break site,
no enrichment of Scc1 was detected (Figure 6A). Low
accumulation of cohesin subunits proximal to the break
was also observed in some experiments by using con-
ventional ChIP (Figure 2B, Smc1Myc) and could either
reflect an experimental variation or, possibly, a variable
degree of DNA resection close to the damage. To exam-
ine if Scc1 also was recruited close to breaks at other
positions in the genome, we investigated Scc1 localiza-
tion in the vicinity of a DSB at the natural HO recognition
site in the MAT locus on ChrIII (Figure 6C, 4 hr) and at
an inserted site on the middle of the ChrVI arm (Figure
7A, 4 hr). Also here, Scc1 was localized to the sur-
in the closest proximity to the DSB in the MAT locus
(Figure 6C, 4 hr). The reason for this is unclear but may
There is substantial evidence that the DNA repair func-
tion of cohesin could extend beyond the maintenance
of sister chromatid cohesion established during DNA
replication. We show here that the budding yeast
cohesin components Smc1 and Scc1 localize around
a DSB in cells arrested in G2 in chromosome regions
normally unoccupied by the complex (U¨nal et al., 2004
[this issue of Molecular Cell]; this study). Loading of
Scc1 during G2 has been shown previously (Uhlmann
varies only modestly between late G1/S and G2 phases
in undamaged cells (Laloraya et al., 2000), indicating
that DNA damage initiates a so far uncharacterized pro-
cess that allows cohesin binding in new chromosomal
regions. Cohesin localization close to the break could
in principle be due to the reorganization of complexes
already bound to chromatin, but the finding that the
localization depends on the cohesin-loading protein
Scc2 argues that soluble complexes are recruited to
chromatin (U¨nal et al., 2004; this study). Also, the dam-
age-induced formation of cohesion in G2 in mutated
on postreplicative expression of wt SMC1, supports the
idea that soluble cohesin is recruited to broken chromo-
somes (this study).
The damage-elicited loading of soluble cohesin com-
plexes allowed investigation of the DNA repair role of
DSB-recruited cohesin. Elimination of Scc2 function in
tion at DSBs (U¨nal et al., 2004; this study). Furthermore,
abolished repair of DNA damage induced by ionizing
irradiation, signifying that cohesin loading onto DSBs is
required for DNA repair. Notably, DNA repair in Scc2-
or Scc4-deficient cells is as low as it is in cells lacking
the damaged DNA is destroyed (Sjo ¨gren and Nasmyth,
2001). This indicates that sister chromatid cohesion is
necessary, but not sufficient, for DNA repair; i.e., in the
absence of additional cohesin complexes recruited to
the damaged DNA, little or no repair takes place regard-
less of chromatid linkage by S phase-generated cohe-
sion. Inline witha crucial rolefor thecohesin complexes
at the break site, an excellent study performed by the
Koshland laboratory shows that cohesin loading de-
pends on central DNA damage-signaling kinases and a
fundamental damage-dependent histone modification
(U¨nal et al., 2004).
Considering cohesin’s recognized role in sister chro-
matid cohesion, it is conceivable that the molecular
matids together close to the break and thereby assist
repair by homologous recombination. This is supported
by our finding that G2 expression of wt cohesin, com-
DSB Repair Requires Cohesin at the Break
Figure 6. Cohesin Localizes to Extended Regions around DSBs
(A) Scc1 is recruited to a ?50 kb region surrounding a DSB at the end of the right ChrV arm. Scc1-Flag (CB281) cells containing an HO cut
site at position 541 on ChrV and GAL:HO were arrested in G2 by the addition of benomyl at 23?C. HO was subsequently induced by the
addition of galactose. Samples for ChIP on chip were collected before (0 hr) and 4 hr after the addition of galactose. ChIP was performed
with anti-Flag antibodies. G2 arrest was maintained until the end of the experiment. The vertical axis scale is log2 and the horizontal axis
shows kilobase units (kb).
(B) Scc1 distribution on ChrVI in cells containing an HO recognition site on ChrV is not altered in response to HO induction. See (A) for details.
(C) Scc1 is recruited to a ?50 kb region surrounding an HO-induced DSB in the MAT locus on chromosome III. See (A) for details.
The HO cleavage sites on ChrV and at the MAT locus are indicated by arrows.
Figure 7. Scc1 Is Recruited to a ?50 kb Region Surrounding a DSB at the Central Part of the Right ChrVI Arm
(A) Scc1-Flag cells with GAL:HO and an HO cleavage site at position 206 on ChrVI (CB 310) were arrested in G2, and ChIP on chip was
performed as in Figure 6. Induction of HO for 4 hr caused accumulation of Scc1 in the region around the break site (compare 0 hr with 4 hr).
(B) Control ChIP on chip performed on untagged cells containing GAL:HO and the HO recognition site on ChrVI (CB306). Cells were arrested
in G2 by benomyl treatment and subsequently grown in the presence of galactose for 4 hr before collecting a sample for ChIP on chip, which
was performed as in Figure 6.
bined with irradiation, prevented sister chromatid sepa-
ration elicited by a temperature increase in smc1-259
cells. Furthermore,we have found thatdestroying Eco1/
Ctf7 function in G2 abolishes DNA repair (data not
shown).As Eco1/Ctf7isrequiredfor cohesionestablish-
that sister chromatids must indeed be held together
by cohesin for DNA repair to take place. However, the
temperature sensitive eco1-1 mutant cells used in these
experiments showed relatively high levels of chromatid
separation at the permissive temperature, rendering
these results ambiguous.
The inhibitoryeffects of SMC1 expressionand irradia-
tion on chromatid separation in smc1-259 cells also
show that cohesion can be established in G2-phase
cells in the presence of DSBs. This is in contrast to
undamaged cells in which cohesion can only be estab-
lished during replication (Uhlmann and Nasmyth, 1998;
Haering et al., 2004). The ?-ray dose used in the experi-
ments has been estimated to generate approximately
20–60 DSBs in a Saccharomyces genome (Contopoulou
This gives a probability of 0.64–0.95 that ChrV, which
represents approximately 5% of the total genome and
chromatid separation, suffers at least one DSB. To un-
derstand how this relatively low number of breaks could
affect chromatid cohesion to the extent observed, we
investigated cohesin localization over entire chromo-
somes. Scc1 was shown to localize to ?50 kb regions
around single DSBs localized at different positions in
the genome,with littleor nochange inundamaged parts
of the chromosomes. Even though Scc1 is recruited to
an extended region around the breaks, it is still unclear
in the genome could completely block separation of
is recruitedto othertypes of DNAlesions inducedby the
?-rays. A second interpretation is that DSB generation
elicits a more general cohesion response. Conceivably,
DNA damage may affect the cohesin complexes nor-
mally loaded onto chromatin in G2, without linking chro-
matids, in such a way that cohesion can be established.
Recent findings provide evidence that cohesin loading
onto its more permanent positions occurs through slid-
ing of the complexes from other regions of the chromo-
some in a Scc2/Scc4 dependent manner (Lengronne et
al., 2004). DSB generation possibly triggers loading of
cohesion-competent complexes that will slide both to
the site of damage and to the normal cohesin associa-
DSB Repair Requires Cohesin at the Break
position as shown by PCR were assayed for efficient HO cleavage
(see the ChIP section). SMC1-MYC13 was cloned by PCR and in-
serted downstream of the GAL1-10 promoter in a YIPLac vector
(Gietz and Sugino, 1988). When introduced into the LEU2 locus of
smc1-259 cells, the GAL:SMC1-MYC13 vector rendered the nor-
mally temperature-sensitive cells resistant to high temperatures
when grown in the presence of galactose, thereby demonstrating
the functionality of the Smc1-13Myc protein (data not shown). The
induction of Smc1 was also verified by Western blot (data not
shown). For introduction of the MET3 promoter upstream of CDC20,
the MET3:HA:CDC20 construct in YCp22 (kindly provided by F. Uhl-
mann) was linearized and transformed for integration.
tion sites on damaged as well as undamaged chromo-
Formation of sister chromatid cohesion in response
to ionizing irradiation raises the possibility that cohesin
recruited to DNA lesions induced by replication failures
(Zou and Rothstein, 1997) could contribute to cohesion
in normal cell cycle progression. Investigations of the
interplay between cohesin and the MRN complex have
revealed data that support such an idea. First, human
cohesin interacts with Rad50 and accumulates in re-
gions of laser-induced DNA damage during S and G2
phases in a Rad50/Mre11-dependent manner (Kim et
al., 2002a). Second, the Mre11 protein has been shown
to be required for loading of cohesin to DSB sites in
S. cerevisiae (U¨nal et al., 2004). Third, the MRN proteins
were recently shown to play a role in sister chromatid
cohesion, apparently operating through cohesin (War-
ren et al., 2004). Together with data presented here,
which indicate that damage-recruited cohesin is able to
link sister chromatids, this argues that cohesin com-
plexes recruited to replication-induced DNA damage in
a Rad50/Mre11-dependent manner may contribute to
sister chromatid cohesion in normal cell cycle progres-
sion. DNA polymerase ? (Pol2) provides another link
between replication-generated damage and sister chro-
matid cohesion. Mutations in Pol2 interfere with both
cohesion (Edwards et al., 2003) and the formation of
X-DNA molecules, during repair of replication-induced
lesions (Zou and Rothstein, 1997). Break induction and/
or the formation of X molecules may be important for
the association of cohesin with chromatin. This is in
line with recent suggestions that cohesion is primed
by X-DNA chromatid junctions that form early during
replication, apparently independently of DNA damage
repair proteins (Lopes et al., 2003). Together with the
results presented here, this suggests that damage-
induced recruitment of cohesin could be similar to nor-
mal, DNA-damage independent, S phase establishment
of sister chromatid linkages. Future exploration of DSB-
damage repair function in more detail, clarify how sister
chromatid cohesion is established during DNA repli-
ChIP was performed after a 25 min fixation with 1% formaldehyde
by using monoclonal antibodies against the Myc epitope (9E10,
Roche), essentially as described (Strahl-Bolsinger et al., 1997; Ta-
naka et al., 1999). Precipitated DNA was amplified by using the
primers outlined in Figure 1A, and PCR products were analyzed on
2.3% agarose gels. The intensities of the respective bands were
quantified with Image Gauge V3.41 (Fuji Film Science lab 99). For
somes were prepared and then separated by PFGE (Biorad, Chef
DRIII) at 14?C in a 1% agarose gel for 33.2 hr at 6 V/cm with a
35.4–83.55 s switch time at an included angle of 120?. A radioactive
probe hybridizing to the 5? region of the ChrV was used for Southern
blot detection of the undamaged and the cleaved, 40 kb shorter,
chromosome. All experiments were repeated two to three times
with little variation in the results.
ChIP on Chip Analysis
ChIP on chip DNA purification and amplification were performed as
described, except that 10 ?g of amplified DNA was used for diges-
tion and labeling before hybridization to the oligonucleotide chips
trix) (Iyer et al., 2001; Katou et al., 2003; Ren et al., 2000). Scc1 was
tagged at the C terminus with 6HIS-3xFLAG, and monoclonal anti-
Flag antibody M2 (Sigma-Aldrich) was used for IP. All experiments
were repeated twice.
Irradiation of cells was performed by using a Cs137?-ray source at
a dose rate of 10 Gy/min, and DNA repair was determined by PFGE
andSouthern blotof ChrXVI(Sjo ¨gren andNasmyth, 2001).Detection
sor-GFP/Tet-operator system, and FACS-scan analysis were per-
formed as described previously (Michaelis et al., 1997).
We thank F. Uhlmann and C. Ho ¨o ¨g for inspiring discussions and
critical comments on the manuscript and K. Nasmyth, F. Uhlmann,
and P. Ljungdahl for generous gifts of plasmids and strains. C.S.
and L.S. were supported by the Swedish Research Council and L.S.
by the Japanese Society for Promotion of Science.
Strains and Plasmids
All strains used are haploid and of W303 origin (ade2-1, trp1-1,
can1-100, leu2-3, 112, his3-11 and -15, and ura3). Modifications and
names of the individual strains are indicated in the text and figure
legends. All strains are RAD5 except K8465, K8466, K8467, and
K8468 (generously provided by K. Nasmyth [Ciosk et al., 2000]),
which contain the rad5-535 mutation (Fan et al., 1996). DNA repair
in G2 cells is not affected by this mutation, as scored by PFGE
(Sjo ¨gren and Nasmyth, 2001). The YIPadeHO plasmid (kindly pro-
vided by N. Sugawara, described in Sandell and Zakian ) was
used for insertion of GAL:HO into the ADE3 locus, resulting in an
ade3 strain. For insertion of the HO cleavage site, we used the PET
vector designed by Schneider et al. (1995) in which the sequences
coding for the HA epitopes were replaced by a 117 bp sequence
containing the HO cut site. For integration at position 541 on ChrV
and 206 on ChrVI, the HO-URA3-HO construct was amplified by
PCR (primer sequences used are available upon request). The PCR
products were used for transformation, and the resulting URA colo-
nies wereplated on 5-fluoro-orotic acid-containingplates. Surviving
Received: June 23, 2004
Revised: October 4, 2004
Accepted: November 5, 2004
Published: December 21, 2004
tion of rad21 an essential gene of Schizosaccharomyces pombe
involved in DNA double-strand-break repair. Nucleic Acids Res.
Brunborg, G., and Williamson, D.H. (1978). The relevance of the
nuclear division cycle to radiosensitivity in yeast. Mol. Gen. Genet.
Chan, R.C., Chan, A., Jeon, M., Wu, T.F., Pasqualone, D., Rougvie,
A.E., and Meyer, B.J. (2003). Chromosome cohesion is regulated by
a clock gene paralogue TIM-1. Nature 423, 1002–1009.
Ciosk, R., Zachariae, W., Michaelis, C., Shevchenko, A., Mann, M.,
and Nasmyth, K. (1998). An ESP1/PDS1 complex regulates loss of
sister chromatid cohesion at the metaphase to anaphase transition
in yeast. Cell 93, 1067–1076.
Ciosk, R., Shirayama, M., Shevchenko, A., Tanaka, T., Toth, A., and
Nasmyth, K. (2000). Cohesin’s binding to chromosomes depends
on a separate complex consisting of Scc2 and Scc4 proteins. Mol.
Cell 5, 243–254.
Cohen-Fix, O., and Koshland, D. (1997). The anaphase inhibitor of
Saccharomyces cerevisiae Pds1p is a target of the DNA damage
checkpoint pathway. Proc. Natl. Acad. Sci. USA 94, 14361–14366.
Contopoulou, C.R., Cook, V.E., and Mortimer, R.K. (1987). Analysis
of DNA double strand breakage and repair using orthogonal field
alternation gel electrophoresis. Yeast 3, 71–76.
Darwiche, N., Freeman, L.A., and Strunnikov, A. (1999). Character-
izationof thecomponentsof theputativemammalian sisterchroma-
tid cohesion complex. Gene 233, 39–47.
Edwards, S., Li, C.M., Levy, D.L., Brown, J., Snow, P.M., and Camp-
and polymerase sigma interact physically and functionally, sug-
gesting a role for polymerase epsilon in sister chromatid cohesion.
Mol. Cell. Biol. 23, 2733–2748.
Fan, H.Y., Cheng, K.K., and Klein, H.L. (1996). Mutations in the RNA
polymerase II transcription machinery suppress the hyperrecombi-
nation mutant hpr1 delta of Saccharomyces cerevisiae. Genetics
vectors constructed with in vitro mutagenized yeast genes lacking
six-base-pair restriction sites. Gene 74, 527–534.
Guacci, V., Koshland, D., and Strunnikov, A. (1997). A direct link
between sister chromatid cohesion and chromosome condensation
revealed through the analysis of MCD1 in S. cerevisiae. Cell 91,
Haber, J.E. (1998). Mating-type gene switching in Saccharomyces
cerevisiae. Annu. Rev. Genet. 32, 561–599.
between sister chromatids. Bioessays 25, 1178–1191.
Haering, C.H., Schoffnegger,D., Nishino, T., Helmhart,W., Nasmyth,
K., and Lowe, J. (2004). Structure and stability of cohesin’s Smc1-
kleisin interaction. Mol. Cell 15, 951–964.
Hartman, T., Stead, K., Koshland, D., and Guacci, V. (2000). Pds5p is
anessential chromosomalproteinrequired forboth sisterchromatid
cohesion and condensation in Saccharomyces cerevisiae. J. Cell
Biol. 151, 613–626.
Iyer, V.R., Horak, C.E., Scafe, C.S., Botstein, D., Snyder, M., and
Brown, P.O. (2001). Genomic binding sites of the yeast cell-cycle
transcription factors SBF and MBF. Nature 409, 533–538.
over homologs as substrates for recombinational repair in Sacchar-
omyces cerevisiae. Genetics 132, 387–402.
Sugimoto, K., and Shirahige, K. (2003). S-phase checkpoint proteins
Tof1 and Mrc1 form a stable replication-pausing complex. Nature
Kim, J.S., Krasieva, T.B., LaMorte, V., Taylor, A.M., and Yokomori,
K. (2002a). Specific recruitment of human cohesin to laser-induced
DNA damage. J. Biol. Chem. 277, 45149–45153.
Kim, S.T., Xu, B., and Kastan, M.B. (2002b). Involvement of the
cohesin protein, Smc1, in Atm-dependent and independent re-
sponses to DNA damage. Genes Dev. 16, 560–570.
Laloraya, S., Guacci, V., and Koshland, D. (2000). Chromosomal
Lengronne, A., Katou, Y., Mori, S., Yokobayashi, S., Kelly, G.P., Itoh,
T., Watanabe, Y., Shirahige, K., and Uhlmann, F. (2004). Cohesin
relocation from sites of chromosomal loading to places of conver-
gent transcription. Nature 430, 573–578.
Lisby, M., Rothstein, R., and Mortensen, U.H. (2001). Rad52 forms
DNA repair and recombination centers during S phase. Proc. Natl.
Acad. Sci. USA 98, 8276–8282.
Lopes, M., Cotta-Ramusino, C., Liberi, G., and Foiani, M. (2003).
Branch migrating sister chromatid junctions form at replication ori-
gins through Rad51/Rad52-independent mechanisms. Mol. Cell
Losada, A., Hirano, M., and Hirano, T. (1998). Identification of Xeno-
pus SMC protein complexes required for sister chromatid cohesion.
Genes Dev. 12, 1986–1997.
Losada, A., Hirano, M., and Hirano, T. (2002). Cohesin release is
Megee, P.C., Mistrot, C., Guacci, V., and Koshland, D. (1999). The
centromeric sister chromatid cohesion site directs Mcd1p binding
to adjacent sequences. Mol. Cell 4, 445–450.
Michaelis, C., Ciosk, R., and Nasmyth, K. (1997). Cohesins: chromo-
somal proteins that prevent premature separation of sister chroma-
tids. Cell 91, 35–45.
Potter, T., and Kohnlein, W. (2001). Pulsed-field gel electrophoresis
to x-rays (30-50 keV) and neutrons (14 MeV). Radiat. Environ. Bio-
phys. 40, 39–45.
Ren, B., Robert, F., Wyrick, J.J., Aparicio, O., Jennings, E.G., Simon,
I., Zeitlinger, J., Schreiber, J., Hannett, N., Kanin, E., et al. (2000).
Genome-wide location and function of DNA binding proteins. Sci-
ence 290, 2306–2309.
Sandell, L.L., and Zakian, V.A. (1993). Loss of a yeast telomere:
arrest, recovery, and chromosome loss. Cell 75, 729–739.
Schneider, B.L., Seufert, W., Steiner, B., Yang, Q.H., and Futcher,
A.B. (1995). Use of polymerase chain reaction epitope tagging for
protein tagging in Saccharomyces cerevisiae. Yeast 11, 1265–1274.
Sjo ¨gren, C., and Nasmyth, K. (2001). Sister chromatid cohesion is
required for postreplicative double-strand break repair in Saccharo-
myces cerevisiae. Curr. Biol. 11, 991–995.
Skibbens, R.V., Corson, L.B., Koshland, D., and Hieter, P. (1999).
Ctf7p is essential for sister chromatid cohesion and links mitotic
chromosome structure to the DNA replication machinery. Genes
Dev. 13, 307–319.
Smits, V.A., Klompmaker, R., Arnaud, L., Rijksen, G., Nigg, E.A., and
Medema, R.H. (2000). Polo-like kinase-1 is a target of the DNA
damage checkpoint. Nat. Cell Biol. 2, 672–676.
Sonoda, E., Matsusaka, T., Morrison, C., Vagnarelli, P., Hoshi, O.,
Ushiki, T., Nojima, K., Fukagawa, T., Waizenegger, I.C., Peters, J.M.,
et al. (2001). Scc1/Rad21/Mcd1 is required for sister chromatid co-
hesion and kinetochore function in vertebrate cells. Dev. Cell 1,
Strahl-Bolsinger, S., Hecht, A., Luo, K., and Grunstein, M. (1997).
SIR2 and SIR4 interactions differ in core and extended telomeric
heterochromatin in yeast. Genes Dev. 11, 83–93.
Sumara, I., Vorlaufer, E., Stukenberg, P.T., Kelm, O., Redemann, N.,
Nigg, E.A., and Peters, J.M. (2002). The dissociation of cohesin from
chromosomes in prophase is regulated by polo-like kinase. Mol.
Cell 9, 515–525.
Tanaka, T., Cosma, M.P., Wirth, K., and Nasmyth, K. (1999). Identifi-
cation of cohesin association sites at centromeres and along chro-
mosome arms. Cell 98, 847–858.
Toth, A., Ciosk, R., Uhlmann, F., Galova, M., Schleiffer, A., and Nas-
myth, K. (1999). Yeast cohesin complex requires a conserved pro-
tein, Eco1p(Ctf7), to establish cohesion between sister chromatids
during DNA replication. Genes Dev. 13, 320–333.
Uhlmann,F., andNasmyth, K.(1998). Cohesionbetween sisterchro-
matids must be established during DNA replication. Curr. Biol. 8,
Uhlmann, F., Lottspeich, F., and Nasmyth, K. (1999). Sister-chroma-
tid separation at anaphase onset is promoted by cleavage of the
cohesin subunit Scc1. Nature 400, 37–42.
Uhlmann, F., Wernic, D., Poupart, M.A., Koonin, E.V., and Nasmyth,
DSB Repair Requires Cohesin at the Break Download full-text
K. (2000). Cleavage of cohesin by the CD clan protease separin
triggers anaphase in yeast. Cell 103, 375–386.
U¨nal, E., Arbel-Eden, A., Sattler, U., Shroff, R., Lichten, M., Haber,
J.E., and Koshland, D. (2004). DNA damage response pathway uses
histone modification to assemble a double-strand break-specific
cohesin domain. Mol. Cell 16, this issue, 991–1002.
van den Bosch, M., Bree, R.T., and Lowndes, N.F. (2003). The MRN
complex: coordinating and mediating the response to broken chro-
mosomes. EMBO Rep. 4, 844–849.
Vass, S., Cotterill, S., Valdeolmillos, A.M., Barbero, J.L., Lin, E.,
Warren, W.D., and Heck, M.M. (2003). Depletion of Drad21/Scc1 in
Drosophila cells leads to instability of the cohesin complex and
disruption of mitotic progression. Curr. Biol. 13, 208–218.
Waizenegger, I.C., Hauf, S., Meinke, A., and Peters, J.M. (2000). Two
distinct pathways remove mammalian cohesin from chromosome
arms in prophase and from centromeres in anaphase. Cell 103,
Warren, C.D., Eckley, D.M., Lee, M.S., Hanna, J.S., Hughes, A.,
Peyser, B., Jie, C., Irizarry, R., and Spencer, F.A. (2004). S-phase
Mol. Biol. Cell 15, 1724–1735.
tor of anaphase in budding yeast, plays a critical role in the APC
and checkpoint pathway(s). J. Cell Biol. 133, 99–110.
Yazdi, P.T., Wang, Y., Zhao, S., Patel, N., Lee, E.Y., and Qin, J.
(2002). SMC1 is a downstream effector in the ATM/NBS1 branch of
the human S-phase checkpoint. Genes Dev. 16, 571–582.
Yeong, F.M., Lim, H.H., Padmashree, C.G., and Surana, U. (2000).
Exit from mitosis in budding yeast: biphasic inactivation of the
Cdc28-Clb2 mitotic kinase and the role of Cdc20. Mol. Cell 5,
Zou, H., and Rothstein, R. (1997). Holliday junctions accumulate in
replication mutants via a RecA homolog-independent mechanism.
Cell 90, 87–96.
Raw data from shown experiments and duplicates are available at
http://www.ncbi.nlm.nih.gov/geo/, accession number GSE1905.