Analyses of the yeast Rad51 recombinase
A265V mutant reveal different in vivo roles
of Swi2-like factors
Peter Chi1, YoungHo Kwon1, Mari-Liis Visnapuu2, Isabel Lam3, Sergio R. Santa Maria3,
Xiuzhong Zheng3, Anastasiya Epshtein3, Eric C. Greene2,4, Patrick Sung1and
Hannah L. Klein3,*
1Molecular Biophysics and Biochemistry, Yale University School of Medicine, 333 Cedar Street, New Haven,
CT 06520,2Department of Biochemistry and Molecular Biophysics, Columbia University Medical Center,
3Department of Biochemistry and NYU Cancer Institute, New York University School of Medicine, 550 First
Avenue, New York, NY 10016 and4Howard Hughes Medical Institute, 650 West 168th Street, New York,
NY 10032 USA
Received March 11, 2011; Revised April 11, 2011; Accepted April 13, 2011
The Saccharomyces cerevisiae Swi2-like factors
Rad54 and Rdh54 play multifaceted roles in homolo-
gous recombination via their DNA translocase
activity. Aside from promoting Rad51-mediated
DNA strand invasion of a partner chromatid, Rad54
and Rdh54 can remove Rad51 from duplex DNA for
properties of the two proteins are similar, differ-
ences between the phenotypes of the null allele
mutants suggest that they play different roles
in vivo. Through the isolation of a novel RAD51
allele encoding a protein with reduced affinity for
DNA, we provide evidence that Rad54 and Rdh54
have different in vivo interactions with Rad51. The
mutant Rad51 forms a complex on duplex DNA that
is more susceptible to dissociation by Rdh54. This
Rad51 variant distinguishes the in vivo functions of
Rad54 and Rdh54, leading to the conclusion that
two translocases remove Rad51 from different sub-
strates in vivo. Additionally, we show that a third
Swi2-like factor, Uls1, contributes toward Rad51
clearance from chromatin in the absence of Rad54
and Rdh54, and define a hierarchy of action of the
Swi2-like translocases for chromosome damage
Althoughthe in vitro
Rad54 and Rdh54 are members of the Swi2 protein
family. These evolutionarily conserved proteins possess
dsDNA-dependent ATPase activity that fuels their trans-
location on dsDNA, resulting in DNA supercoiling and
transient strand unwinding. Both proteins physically
interact with the recombinase Rad51 and synergize with
the Rad51–ssDNA nucleoprotein filament to promote
D-loop formation, DNA branch migration and chromatin
remodeling, all of which are essential steps in homologous
recombination (HR) (1). Interestingly, Rad54 and Rdh54
both can remove Rad51 from dsDNA in vitro. The ability
to dissociate the Rad51–dsDNA complex has been
postulated to be important for releasing Rad51 from
bulk chromatin, to ensure that a sufficient pool of free
recombinase is available for repair and to prevent the ac-
Moreover, removal of Rad51 by Rad54 and Rdh54 may
be necessary to allow access of a DNA polymerase to the
primer terminus in the newly made D-loop during HR.
RAD54 and RDH54 likely serve distinct functions in
mitotic and meiotic recombination, as mutants have
distinct phenotypes (2–4). rad54? mutants are sensitive
to DNA damaging agents and have significant reduction
in mitotic recombination whereas rdh54D are only slightly
sensitive to DNA damage and have a modest reduction in
interchromosomal recombination, but are not affected in
intrachromosomal recombination. rdh54D diploids have
*To whom correspondence should be addressed. Tel: +1 212 263 5778; Fax: +1 212 263 8166; Email: firstname.lastname@example.org
Peter Chi, Institute of Biochemical Sciences, National Taiwan University, No. 1, Sec. 4, Roosevelt Road, Taipei, 10617 Taiwan (R.O.C.).
Isabel Lam, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10065, USA.
Mari-Liis Visnapuu, McKinsey & Co., Inc., 55 East 52nd Street, New York, NY 10022, USA.
The authors wish it to be known that, in their opinion, the first four authors should be regarded as joint First Authors.
Published online 9 May 2011 Nucleic Acids Research, 2011, Vol. 39, No. 156511–6522
? The Author(s) 2011. Published by Oxford University Press.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/
by-nc/3.0), which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
significant meiotic recombination defects and are delayed
in the repair of meiotic double strand breaks whereas
rad54D diploids does not show a delay in the repair of
meiotic double strand breaks, although spore viability is
Even though both Rad54 and Rdh54 can dissociate
Rad51 from dsDNA in vitro (5,6), whether these
proteins remove Rad51 from chromatin in vivo and the
functional significance and relative contributions of
Rad54 and Rdh54 toward Rad51 clearance from chroma-
tin remain unanswered. In vivo, RAD54 has a more im-
portant role than RDH54 in the recombinational repair of
methyl methanesulfonate (MMS) damaged DNA and
gaps that occur from replication across a damaged DNA
template, as seen by the strong MMS sensitivity of rad54
mutants, while rdh54 mutants show only a modest sensi-
tivity (3,4). In a recent study we have found that Rdh54
has a critical role in removing Rad51 from chromatin
when Rad51 is expressed in excess, while Rad54 has
only a minor role in this regard (7). In contrast, following
irradiation damage, Rad51 foci persist in rad54 mutants,
but not in rdh54 mutants (7). Likewise, we do not yet
know whether Uls1, another Swi2 family member that
was originally identified based on its two-hybrid inter-
action with the meiotic recombinase Dmc1, also plays a
role in Rad51 clearance from chromatin (8). ULS1 has no
known role in DNA repair, as deletion of the gene does
not render cells sensitive to DNA damaging agents,
although it appears to play a role in removing excess
Rad51 from chromatin (7). Here, we show synthetic
growth deficiency and DNA damage sensitivity of
double and triplemutants
Swi2-like factors that can be efficiently suppressed by
deleting RAD51. In congruence with this, a genomic sup-
pressor of the rad54D uls1D mutant defects is shown to
harbor a mutation, A265V, in RAD51. We provide
evidence that the suppressor activity of rad51A265V
stems from the combined effect of this mutation on the
affinity of Rad51 for DNA and the accelerated removal of
the mutant rad51 protein by Rdh54 from dsDNA.
Taken together, our results provide compelling evidence
for a cytotoxic effect of gratuitous Rad51–dsDNA
complexes and suggest that Rad54, Rdh54 and Uls1 all
contribute toward clearance of these toxic nucleoprotein
complexes in a hierarchical fashion. Additionally, our
results suggest that Rad54 and Rdh54 recognize different
Rad51–dsDNA complexes in vivo.
MATERIALS AND METHODS
Spot assays on MMS-supplemented plates
Yeast cultures were incubated overnight at 30?C in YPD
medium. After determining cell density, the cultures were
adjusted to 107cells/ml and then serially diluted. Aliquots
of 4ml from the serial dilutions were spotted onto SC or
SC-containing MMS at the indicated concentration. SC
plates containing MMS were made directly before use.
The plates were then incubated at 30?C for 5–6 days.
Screen for suppressors of rad54 uls1 MMS sensitivity
The rad54D uls1D strain used for EMS mutagenesis is
leu2-3, 112 his3-11, 15 ade2-1 ura3-1 trp1-1 can1-100
hom3-10 RAD5. For EMS mutagenesis, cells were grown
overnight at 30?C, collected and washed twice with water
and resuspended in an equal volume of 0.1M sodium
phosphate buffer (pH 7). The exact cell density was
determined with a hemacytometer and adjusted to
2?108cells/ml. Two 1-ml aliquots were made, and
0.5ml EMS (Sigma) was added to one aliquot, while the
other aliquot served as control. The tubes were vortexed
vigorously before incubating for 1h at 30?C with agita-
tion. After incubation, the cells were collected and washed
three times with 8ml 5% sodium thiosulfate, and once
with sterile distilled water. The cells were then resuspended
in 5ml YPD, and incubated for 3h to allow cells to
express the mutant proteins. After outgrowth, 106
EMS-treated and control cells were plated onto YPD
plates containing 0.004% MMS. In addition, 104, 102
and 101EMS-treated and control cells were plated onto
YPD to estimate the percentage cell death resulting from
the EMS mutagenesis protocol. Colonies growing on YPD
were counted after 36h of incubation at 30?C, and percent
survival was calculated to be ?25%.
Colonies growing on 0.004% MMS after 7 days of in-
cubation at 30?C were streaked onto fresh 0.004% MMS
plates to yield single colonies and compared to growth of
rad54D, uls1D and rad54D uls1D strains. Mutants that
were less sensitive to 0.004% MMS than the rad54D
uls1D double mutant were screened for growth in
0.0025% MMS to confirm resistance to levels of MMS
at which rad54D uls1D is not viable. Colonies with signifi-
cant growth in 0.0025% MMS were chosen as putative
Suppressor-containing strains were crossed several times
to a wild-type strain to recover the suppressor mutation in
a RAD54 ULS1 background and to eliminate any
unlinked mutations that might have arisen during the
Sequencing of RAD51
Primers 50CATATCCCACGACTAGGCCA30and 50CAT
GGGTGACAGACAATACG30were used to amplify the
RAD51 gene from yeast strains containing putative sup-
pressors of rad54D uls1D. The PCR product was
sequenced and a mutation at base 794 changing C to T
or amino acid residue 265 changing alanine to valine was
found in RAD51.
Reconstruction of the rad51 A265V mutation
The rad51 A265V mutant allele was introduced into an
unmutagenized wild-type strain to replace the endogenous
RAD51 allele. A 1.7-kb DNA fragment including the
rad51 A265V open reading frame was amplified using
primers SSM93 (50CGGGGTACCCGGGGATCCCAC
GACTAGGCCACAC) and SSM94 (50GCCAAGCTTG
genomic DNA of a mutant yeast strain. The PCR
6512Nucleic Acids Research, 2011,Vol.39, No. 15
product was then introduced into vector YIplac211 as
BamHI–PstI fragment to yield the integration construct
pHK440. pHK440 was then cut with the SpeI restriction
endonuclease and used to transform yeast cells to Ura+.
Positive transformants were then grown on 5-FOA-
containing medium and the resulting clones confirmed
for the rad51 A265V mutation by DNA sequencing.
Recombination assays for haploid intragenic recom-
bination and diploid intragenic heteroallelic recombin-
Recombination rates were determined by the method of
the median (9).
Rad51 and rad51 A265V proteins were expressed in the
rad51D strain LSY411 by the use of the PGK promoter in
plasmid pMA91 (2m, PGK, leu-2d) (10) and were purified
to near homogeneity, following our published protocol
(11). His6-tagged Rad54 or Rdh54 was expressed in
Escherichia coli and purified to near homogeneity as
The fX replicative form I DNA was purchased from
prepared by treatment with calf thymus topoisomerase I
(Invitrogen), as described previously (13). For DNA
mobility shift assay, the 80-mer Oligo 1–50TTATATCC
CAT-30was 50-end labeled with [g-32P] ATP (Amersham
Bioscience) and polynucleotide kinase (Roche) and
separated from the unincorporated ATP using a Spin 6
column (Bio-Rad). To generate radiolabeled dsDNA,
radiolabeled Oligo 1 was annealed to its exact comple-
ment, Oligo 2. The resulting duplex was purified from a
10% polyacrylamide gel by overnight diffusion at 4?C into
TE buffer (20mM Tris–HCl, pH 7.5, 0.5mM EDTA).
For the D-loop assay, the 90-mer oligonucleotide D1
(14), being complementary to pBluescript SK DNA
from position 1932 to 2022, was 50-end labeled and then
purified using the MERmaid Spin Kit (Bio101). The 600
base pairs biotinylated dsDNA used for monitoring
Rad51 removal was prepared by PCR amplification of
pBluscript SK DNA using the 50-biotinylated primer 1
and non-biotinylated primer 2, as described previously
(5). The amplified DNA was deproteinized by phenol–
chloroform extraction, ethanol precipitated and dissolved
in TE. The biotinylated dsDNA was immobilized on
Biochemicals), as described previously, to give 50ng of
the biotinylated DNA per microliter of suspended
volume (5). The 83-mer 50- TTTATATCCTTTACTTTA
ssDNA was used as the Rad51 trap.
DNA mobility shift assay
dsDNA (4.5mM base pairs) was incubated with the
indicated amounts of Rad51 or rad51A265V for 5min at
37?C in 10ml of buffer B (35mM Tris–HCl at pH 7.5,
1mM DTT, 50mM KCl, 100mg/ml BSA) containing
2mM ATP and the indicated concentration of MgCl2.
The reaction mixtures were resolved in 10% polyacryl-
amide gels in TA buffer (40mM Tris–acetate, pH 7.5) at
4?C. The gels were dried onto a sheet of DEAE paper and
then subject to phosphorimaging analysis.
32P-labeled 80-mer ssDNA (4.5mM nucleotides) or
DNA topology modification reaction
Topologically relaxed fX dsDNA (10mM based pairs)
was incubated with the indicated amount of Rad51 or
rad51A265V in 9.6ml buffer B containing 2mM ATP
and the indicated concentration of MgCl2 for 5min,
followed by the addition of 3 units of calf thymus topo-
isomerase I (Invitrogen) in 0.4ml. Reaction mixtures were
incubated for 15min at 37?C and then deproteinized with
0.5% SDS and 0.5mg/ml proteinase K for 10min.
Samples were resolved in 0.8% agarose gels run in TAE
buffer (40mM Tris–acetate, pH 7.5, 0.5mM EDTA) at
23?C, and the DNA species were stained with ethidium
bromide (2mg/ml in water) for 1h. After being destained
in water at 4?C for 24h, the gels were analyzed in a gel
documentation station (Bio-Rad).
The32P-labeled 90-mer ssDNA (2.4mM nucleotides) was
incubated with the indicated amounts of Rad51 or rad51
A265V in 10.5ml buffer B containing an ATP-regenerating
system and 5mM MgCl2 for 5min at 37?C. Rad54
(140nM) or Rdh54 (300nM) was then added in 1ml,
followed by a 1-min incubation at 30?C. The D-loop
reaction was initiated by adding pBluescript replicative
form I DNA (35mM base pairs) in 1ml. The reaction
mixtures were incubated for 5min at 30?C, deproteinized
and processed for electrophoresis in 0.9% agarose gels in
TAE buffer (40mM Tris–acetate, pH 7.5 0.5mM EDTA).
The gels were dried onto a sheet of DEAE paper and the
radiolabeled D-loop was visualized and quantified in the
Rad51 or rad51 A265V (4mM each) was incubated with
100mM ATP, 0.1mCi/ml [g-32P] ATP at 37?C in a 10ml
buffer containing 40mM Tris–HCl pH 7.5, 1mM DTT,
100ng/ml BSA, 50mM KCl, 2 or 5mM MgCl2 (as
indicated) in the presence of fX 174 viral (+) strand
(30mM nucleotides) or replicative form I DNA (30mM
base pairs) or in the absence of DNA. Aliquots (2ml)
were drawn at the indicated times and mixed with an
equal volume of 500mM EDTA to stop the reaction.
After thin layer chromatography in polyethyleneimine
sheets (J.T. Baker Inc.), the level of ATP hydrolysis was
determined by phosphorimaging analysis of the chroma-
tography plate (14).
Nucleic Acids Research, 2011,Vol.39, No. 15 6513
Assay to monitor Rad51 removal from DNA
Rad51 or rad51 A265V (3.7mM) was incubated with
magnetic beads containing biotinylated dsDNA (15mM
base pairs) in 18ml buffer B containing 2mM ATP,
5mM MgCl2and an ATP-regeneration system for 5min
at 37?C. After the incorporation of the indicated amounts
of Rad54 or Rdh54 in 1ml and a 3-min incubation at 23?C,
the reaction was completed by adding 83-mer ssDNA
(150mM nucleotides), as Rad51 trap, in 1ml. Following a
10-min incubation at 30?C, the beads were captured with
the Magnetic Particle Separator (Boehringer Mannheim),
and the supernatant was set aside. Bound proteins were
eluted from the beads with 20ml of 2% SDS. The various
supernatant and SDS eluate (8ml each) were analyzed by
SDS–PAGE and Coomassie blue staining to determine
their content of proteins.
Rad51 or rad51 A265V (5mg each) was incubated with
His6-tagged Rad54 or Rdh54 (5mg each) in 30ml buffer
D (20mM KH2PO4, pH 7.4, 75mM KCl, 10% glycerol,
0.5mM EDTA, 0.01% Igepal, 1mM 2-mercaptoethanol)
for 30min on ice. The reaction mixture was incubated with
8ml of Ni2+-NTA agarose beads for 30min on ice, with
gentle mixing every 30s. The beads were pelleted by cen-
trifugation, and the supernatant was removed. After being
washed twice with 30ml buffer D containing 10mM imid-
azole, the beads were treated with 20ml of 2% SDS to
elute bound proteins. The supernatant (8ml), wash (8ml)
and SDS eluate (8ml) were subject to SDS–PAGE and
Coomassie blue staining to determine their protein
A Sephacryl S-400 column (30?1cm) was used to analyze
Rad51 and rad51 A265V with buffer C (25mM Tris, pH
7.5, 10% glycerol, 0.5mM EDTA, 1mM DTT, 0.01%
Igepal and 150mM KCl) as the eluent and collecting
0.2-ml fractions. The indicated column fractions were sub-
jected to SDS–PAGE and Coomassie blue staining.
Thyroglobulin (669kDa), and catalase (232kDa) were
used for calibrating the column.
TIRFM measurements of Rad51 nucleoprotein filament
Flowcells and DNA curtains were prepared essentially as
described (15). The DNA substrate was a 23-kb segment
of the human b-globin locus that was amplified with the
Expand 20kBPLUSPCR system from human genomic
DNA using forward and reverse primers that are covalent-
ly linked to biotin and digoxigenin, as described elsewhere
(16). The DNA was labeled with quantum dots covalently
conjugated to anti-digoxygenin Fab-fragments (Roche).
Rad51 (700mL of 1mM, either wild-type or A265V
mutant) was injected at 37?C in buffer containing 1mM
ATP and either 10 or 2mM of MgCl2, as indicated, and
reactions were chased with buffer lacking Rad51. Images
were recorded for 15min at 1-s intervals. The DNA length
was measured by tracking the location of the quantum
dot (17). The data corresponding to filament disassembly
are fit with a sigmoidal curve the slope of which is used to
determine the disassembly rate, and all reported rates rep-
resent the mean obtained from three independent experi-
ments, as described previously (16).
Genetic interaction of uls1 with rad54 and rdh54
The ULS1 gene was identified in a two-hybrid screen for
yeast proteins that interact with Dmc1, a meiosis-specific
Rad51 homolog (8). Uls1 protein (also known as Ris1 or
Tid4) is a member of the Swi2 family. Interestingly, the
Uls1 protein possesses a RING finger domain suggestive
of an ubiquitin ligase activity (18). The most closely
related Saccharomyces cerevisiae proteins are Rad5 and
Rad16, which are involved in post-replication repair
(PRR) and nucleotide excision repair (NER), respectively.
Mutations in the homologous recombination factors
Rad54 and Rdh54, which are also Swi2 family members,
result in DNA damage sensitivity and reduced homolo-
gous recombination, although the rad54 mutant is much
more affected than the rdh54 mutant. The rad54D rdh54D
haploid double mutant grows at the same rate as the
haploid rad54? strain, but diploid growth of rad54?
rdh54? is greatly impaired, with clonal lethal sectors (3)
(Figure 1A). Deletion of RAD51 in the rad54D rdh54D
diploids restored normal growth (3) (Figure 1A), suggest-
ing that Rad51 may form toxic intermediates in these
diploids. To determine if additional SWI2 genes could
be involved in avoidance of the Rad51 toxic intermediates,
we focused on ULS1, based on its reported interaction
with the Rad51-related recombinase protein Dmc1 (8).
We found that a homozygous mutation of ULS1 in the
rad54D rdh54D diploids further aggravates growth defi-
ciency (Figure 1A), while the haploid triple mutant
strain rad54D rdh54D uls1D grows slower than single or
double mutants, but without clonal lethality (7). As
expected, the poor diploid growth of the triple mutant
can be completely overcome by deleting the RAD51
gene. Importantly, these results implicate Uls1 in the regu-
lation of HR, possibly via the clearance of Rad51 from
chromatin (see below). Indeed, we have found that rad54?
rdh54? uls1? strains spontaneously accumulate Rad51
To further explore a possible overlap in function
between Uls1 and either Rad54 or Rdh54, haploid
double mutants rad54D uls1D and rdh54D uls1D were
examined for sensitivity to DNA damage. uls1D did not
show a significant increase in DNA damage sensitivity as a
single mutant, nor did it enhance the DNA damage sen-
sitivity of a rdh54D mutant. However, uls1D did slightly
elevate the MMS sensitivity of rad54D at low MMS doses,
providing further support for the idea that Uls1 can par-
tially substitute for Rad54 in DNA damage repair and
recombination (Figure 1B).
The above observations are reminiscent of the synthetic
lethality that results from combining mutations in genes
whose products act to remove Rad51 from DNA
6514Nucleic Acids Research, 2011,Vol.39, No. 15
combined with other DNA repair factors. The Srs2 DNA
helicase can disrupt Rad51 filaments on ssDNA, and mu-
tations in this gene show synthetic lethality with a variety
of other mutations, which arise because of the accumula-
tion of toxic recombination intermediates, likely filaments
of Rad51 on ssDNA (19). Since Rad54 and Rdh54 remove
Rad51 from dsDNA (5,6), we suspected that the poor
growth and DNA damage sensitivity of various double
and triple mutants, seen in Figure 1, were the result of
toxic intermediates generated by persistent association of
Rad51 with DNA. We validated this hypothesis via the
isolation and characterization of a suppressing mutation
Isolation of rad51A265V as a suppressor mutation
The enhanced MMS sensitivity of the haploid rad54D
uls1D double mutant at low MMS doses provided a
means for isolating recessive suppressors, which we
reasoned, could be hypomorphic loss-of-function alleles
in HR factors. Complete loss-of-function mutations
should not rescue the DNA damage sensitivity phenotype,
as the HR factors are needed for repairing chromosome
damage induced by MMS.
Following EMS mutagenesis of the haploid rad54D
uls1D strain, variants of the reduced MMS sensitivity
were sought. One such variant displayed enhanced MMS
SC 0.0025% MMS 0.005% MMS
rad54 uls1 rad51A265V
rad54 rdh54 uls1
rad54 rdh54 uls1
rad54 rdh54 uls1
rdh54 uls1 rad51A265V
Figure 1. rad51A265V suppresses the deleterious effects of multiple Snf/Swi mutants. (A) Growth of diploid strains from newly formed zygotes is
shown. Pictures were taken after growth on YPD medium at 30?C for 4 days. (B) Serial dilution of the indicated haploid strains on SC or
SC-containing MMS is shown. Pictures were taken after growth at 30?C for 6 days. (C) Gene conversion of an intrachromosomal direct repeat
is shown. The recombination system contains LEU2 heteroalleles separated by the URA3 marker. Gene conversion events were selected as Leu+
Ura+segregants. Rates were determined by the median method of the Lea and Coulson fluctuation test. Each test was done of three strains of each
genotype. The mean and standard deviation (SD) of these rates are shown. (D) Serial dilution of the indicated haploid strains on SC or
SC-containing MMS is shown. Pictures were taken after growth at 30?C for 6 days.
Nucleic Acids Research, 2011,Vol.39, No. 156515
segregants from genetic crosses. Since the suppressor
strain had a slight MMS sensitivity in an otherwise
wild-type background, we reasoned that it harbors a
hypomorphic mutation in a HR gene, possibly RAD51.
Indeed, sequencing of the chromosomal RAD51 locus in
the suppressor strain revealed the alteration of alanine 265
to valine (change of nucleotide C at position 794 to T).
This mutation lies near the subunit interface between
adjacent monomers in the Rad51 filament (Figure 2) (20).
To further study the rad51A265V mutation, we replaced
the wild-type allele with this mutant version in an
unmutagenized strain, and used this strain in crosses to
generate all of the strain combinations with rad51A265V
shown in Figure 1A, B and D. Importantly, the
rad51A265V mutation overcomes the growth deficiency
of the rad54D rdh54D uls1D diploid (Figure 1A). This sup-
pression is semidominant, as suppression of the growth
deficiency of the rad54D rdh54D uls1D diploid is seen in
RAD51/rad51A265V, but the suppression is stronger when
rad51A265V is homozygous. Thus, the rad51A265V
mutation alleviates the DNA damage sensitivity of the
rad54D uls1D double mutant and the growth deficiency
of the rad54D rdh54D uls1D triple mutant. At low MMS
doses, 0.0025 and 0.005%, the rad51A265V mutation is
epistatic to the rad54D mutation and suppresses the
rad54D mutation. At higher MMS doses of 0.01 and
0.015%, the rad51A265V mutation does not suppress the
rad54D mutation (data not shown).
Genetic assays revealed that the rad51A265V strain is
reduced in mitotic intrachromosomal gene conversion
compared to wild type, although it is still 10-fold higher
between homologous chromosomes is not significantly
reduced in the rad51A265V mutant, whereas it decreases
>100-fold inthe rad51D
rad51A265V mutation does not suppress the recombin-
ation gene conversion defect of rad54D strains, showing
that the rad51A265V mutation does not bypass the need
for the Rad54 protein in homologous recombination
(Supplementary Figure S1).
uls1D and rdh54D show no genetic interaction in a
RAD51 background (Figure 1D). However, the uls1D
rad51A265V double mutant has an enhanced MMS sensi-
tive phenotype. One explanation for this is that Rad54 and
Rdh54 compete for the rad51 A265V protein and may
prematurely remove the rad51 A265V protein from
damage repair sites, thus leading to increased damage sen-
sitivity. To test this, the triple mutant rdh54D uls1D rad51
A265V was examined. Loss of RDH54 restored MMS re-
sistance to the uls1D rad51A265V strain. This is consistent
with a model in which Rdh54 acts inappropriately on
Rad51 dsDNA nucleofilaments to remove Rad51, pre-
venting Rad54 from performing its function in DNA
repair. To provide further support for this model, the bio-
chemical properties of the rad51 A265V protein were
Expression and purification of rad51 A265V mutant
To determine if rad51A265V encodes a hypomorphic
mutant protein that either forms unstable filaments on
DNA (hence its suppression of the poor diploid growth
of the rad54D rdh54D uls1D triple mutant) and/or is more
readily removed from dsDNA by Rdh54 (hence its sup-
pression of the DNA damage sensitivity of the rad54D
uls1D mutant), we expressed the rad51 A265V mutant
protein in yeast cells and purified it (see ‘Materials and
Methods’ section) to near homogeneity (Supplementary
below). During purification, the rad51 A265V protein ex-
hibited the same chromatographic properties as the
wild-type Rad51 protein, and a yield of the mutant
protein very similar to that of the wild-type counterpart
Characterization of the rad51 A265V mutant protein
Even though the A265V mutation lies near the subunit
interface between adjacent monomers in the Rad51
filament (Figure 2), gel filtration analysis showed that
the rad51 A265V mutant protein has the same oligomeric
structure as wild-type Rad51 in the absence of DNA and
ATP (Supplementary Figure S2B). We next tested the
purified rad51 A265V mutant alongside wild-type Rad51
for DNA binding, using radiolabeled ssDNA and dsDNA
as substrates and mobility shift of these substrates in poly-
acrylamide gels as assay. Since Rad51 needs ATP to bind
DNA, the experiments were conducted in the presence of
2mM ATP with either 2 or 5mM MgCl2. As shown in
Figure 3A, at 2mM MgCl2, rad51 A265V is less capable
than the wild-type protein in binding ssDNA and dsDNA.
This DNA binding deficiency is alleviated to a significant
Figure 2. The location of A265V mutation in the Rad51 filament.
Rad51 filament structure (20) is represented in blue and green
indicating the six monomers that make up a full turn of the filament.
Residue A265 is indicated in magenta and the orientation of the residue
in the two adjacent monomers is highlighted.
6516 Nucleic Acids Research, 2011,Vol.39, No. 15
degree upon increasing the MgCl2concentration to 5mM
Duplex DNA becomes extended (by ?50% relative to B
form DNA) when bound by Rad51 (1). The use of calf
thymus topoisomerase I allows one to register the DNA
extension as a topological change; the product of this
topoiosmerase-linked reaction is a negatively supercoiled
species, called Form UW (Supplementary Figure S3A).
In congruence with the results from the DNA mobility
shift assay, rad51 A265V was significantly less effective
than Rad51 in generating Form UW DNA at 2mM
MgCl2(Supplementary Figure S3B), whereas a much less
pronounced deficit was seen upon increasing the MgCl2
concentration to 5mM (Supplementary Figure S3C).
Collectively, the results from the DNA mobility shift
Figure 3. Biochemical attributes of the rad51 A265V mutant protein. (A) Rad51 and rad51 A265V proteins (0.8, 1.2, 1.6 and 2.0mM) were examined
for their ability to bind32P-labeled ssDNA or dsDNA in buffer containing either 2mM (panel i) or 5mM (panel iii) MgCl2. The percent DNA
bound (complex) was plotted in panel ii (2mM MgCl2) and panel iv (5mM MgCl2). Error bars show SD of three experiments. (B) Rad51 or rad51
A265V was incubated either alone or with Rad54 (panel i) or Rdh54 (panel ii) and protein complexes were captured on Ni2+NTA-agarose. The
beads were washed and then treated with SDS. The supernatant (S), wash (W) and SDS-eluate (E) were analyzed by SDS–PAGE and Coomassie
Blue staining. (C) Rad51 or rad51 A265V (0.3, 0.6, 1.2 and 1.8mM) was examined in conjunction with Rad54 (panel i) or Rdh54 (panel ii) for the
ability to catalyze the D-loop reaction. Rad51 or rad51 A265V alone (1.8mM) was also examined (in panel i). The results were plotted. Error bars
show SD of three experiments.
Nucleic Acids Research, 2011,Vol.39, No. 156517
Supplementary Figure S3), revealed a DNA binding defi-
ciency in rad51 A265V. Additional evidence, presented
below, further indicated that rad51 A265V dissociates at
a faster rate from dsDNA.
Protein–protein interactions are not affected by the
rad51 A265V mutation
In fulfilling its biological role, Rad51 needs to physically
and functionally interact with several other factors,
including Rad54 and Rdh54. We tested the rad51
A265V mutant for physical interaction with Rad54 and
Rdh54 in a pulldown assay that made use of the (His)6
tag on the latter two proteins and nickel NTA agarose
beads to capture protein complexes. As shown in
Figure 2B, rad51 A265V was just as proficient in interact-
ing with Rad54 or Rdh54 as wild-type Rad51. We further
asked whether rad51 A265V retains the ability to func-
tionally synergize with Rad54 or Rdh54 in the D-loop
reaction. To do this, a
incubated with Rad51 or rad51 A265V before adding
Rad54 or Rdh54, followed by the incorporation of the
32P-labeled 90-mer ssDNA was
negatively supercoiled homologous duplex target to
complete the reaction. Pairing of the ssDNA and homolo-
gous duplex yields a32P-labeled D-loop. Rad54 or Rdh54
greatly enhanced the ability of Rad51 or rad51 A265V to
catalyze D-loop formation (Figure 3C).
The above results demonstrated that the A265V
mutation has little or no negative impact on Rad510s inter-
action with Rad54 or Rdh54 or on its ability to function-
ally synergize with these Swi2-like factors in catalyzing
The rad51A265V mutation destabilizes the
To ask whether the A265V mutation affects the stability of
the Rad51–dsDNA filaments, total internal reflection
fluorescence microscopy (TIRFM) was used to monitor
the dissociation of the rad51 A265V mutant protein
from dsDNA (Figure 4A) (16). To do this, nucleoprotein
filaments of either Rad51 or rad51 A265V were assembled
on single molecules of dsDNA within the context of a
DNA curtain (17). As expected, Rad51 and rad51
DNA length (μm)
DNA length (mm)
wtRad51, 10mM MgCl2
A265VRad51, 10mM MgCl2
wtRad51, 2mM MgCl2
A265VRad51, 2mM MgCl2
Figure 4. Single-molecule assay of Rad51 and rad51 A265V nucleoprotein filament stability. (A) Describes the experimental set-up used to visualize
nucleoprotein filament assembly and disassembly. Quantum dot labeled DNA molecules are confined within the evanescent field by application of
buffer flow. Rad51 is injected and the quantum dot position is tracked over time. The DNA lengthens (?x) as Rad51 binds and assembles into a
nucleoprotein filament. Disassembly of the filament is then initiated by rinsing with buffer that lacks free Rad51, and contains 2mM ATP and either
2 or 10mM MgCl2, as indicated. (B) Shows an example of a kymogram with a single DNA molecule as it forms a nucleoprotein filament with rad51
A265V at 10mM MgCl2. The injection of the protein initiates the rapid lengthening of the DNA. Free protein and ATP are then flushed from the
sample chamber at the 200-s mark and the filament begins to disassemble, as indicated by a gradual shortening of the DNA. (C) Shows represen-
tative tracking data used to quantitate Rad51 filament stability. Each trace represents data obtained from one experiment (?15–20 DNA molecules
each), and the disassembly rates reported in the text represent the mean and SD from three independent measurements (?45–60 DNA molecules
total). The colors corresponding to different buffer conditions, and either Rad51 or rad51 A265V are indicated.
6518 Nucleic Acids Research, 2011,Vol.39, No. 15
A265V both extended the length of the DNA by ?68%
upon nucleoprotein filament assembly (16,17), and, im-
portantly, rad51 A265V extended the DNA to the same
degree as the wild-type protein. This greater than expected
increase in the apparent length of the DNA can be
attributed to the behavior of the naked DNA versus the
nucleoprotein complexes in shear flow; there is an increase
in persistence length as Rad51 binds the DNA (21–23),
which makes the Rad51-bound DNA easier to stretch at
a given flow rate (16,17). These results are in complete
agreement with the topological assays presented above
(Supplementary Figure S3) confirming that rad51 A265V
forms an elongated nucleoprotein filament on dsDNA.
The presence of a fluorescent quantum dot at the end of
the DNA molecule allowed us to accurately measure the
DNA length and quantitate the filament dissociation rate
using an automated particle-tracking algorithm, as previ-
ously described (24). As shown in Figure 4, the TIRFM
analysis provided evidence for a reduced stability of the
rad51 A265V nucleoprotein filaments compared to Rad51
filaments. Specifically, at 10mM MgCl2wild-type Rad51
filaments disassembled at a rate of 5.6±0.8nm/s, whereas
a more rapid disassembly rate of 7.0±1.1nm/s was seen
for rad51 A265V under the same buffer conditions. At
2mM MgCl2, the difference in stability was even more
pronounced, with an observed value of 5.8±1.1nm/s
for wild type Rad51 compared to 9.9±1.1nm/s for the
mutant protein. These results thus demonstrate that the
rad51 A265V mutant is more prone to spontaneously
dissociating from dsDNA than wild-type Rad51. Since
the turnover of Rad51 from DNA is coupled to ATP hy-
drolysis, we also tested whether the rad51 A265V mutant
has an enhanced ability to hydrolyze ATP. However,
results from ATPase assays done in the absence or
presence of DNA revealed that the rad51 mutant in fact
(Supplementary Figure S4).
Accelerated removal of rad51 A265V from dsDNA by
The biochemical experiments described earlier (Figure 3)
have provided evidence that rad51 A265V has a lower
affinity for DNA, which, as indicated from TIRFM ex-
periments, likely stems from a heightened tendency of the
mutant protein to dissociate from DNA (Figure 4).
Previous studies have found an ability of Rad54 and
Rdh54 to remove Rad51 from dsDNA (5,6) and, as
noted above, our genetic data have hinted at the possibil-
ity that rad51 A265V may be removed from dsDNA at a
faster rate by Rdh54. To directly test this premise, we used
a biochemical assay, devised previously (5) (Figure 5A), to
monitor the dissociation of the Rad51–dsDNA nucleopro-
tein filament. Briefly, filaments of Rad51 or rad51 A265V
were assembled on a biotinylated dsDNA fragment fixed
to streptavidin magnetic beads, followed by the addition
of Rad54 or Rdh54 together with a non-biotinylated
ssDNA molecule to trap the Rad51 molecules that have
been dislodged from dsDNA by either of the latter two
proteins. As shown before (5,6) and reiterated here, there
was a Rad54 or Rdh54 concentration-dependent transfer
of Rad51 protein from the magnetic bead-bound DNA to
the non-biotinylated ssDNA trap, indicative of dissoci-
ation of the Rad51–dsDNA nucleoprotein filament by
these Swi2-like factors. Interestingly, while rad51 A265V
was removed by Rad54 from the dsDNA at about the
same rate as wild-type Rad51 (Figure 5B), a significantly
accelerated rate of rad51 A265V dissociation was seen
with Rdh54 (Figure 5C). In particular, at the lowest con-
centration (70nM) of Rdh54 used, there was a nearly
3-fold difference in susceptibility of the mutant rad51
and wild-type Rad51 filaments to the dissociative action
of the former protein (Figure 5C). Taken together, these
results suggest that compared to wild-type Rad51 protein,
the mutant rad51 A265V protein is more prone to spon-
taneous dissociation from dsDNA and is also more readily
removed from dsDNA by Rdh54.
Accumulated Rad51–dsDNA complexes affect cell fitness
With the use of S. cerevisiae mutants ablated for the
Swi2-related factors Rad54, Rdh54 and Uls1, we have
furnished additional evidence that the inability to
properly regulate Rad51 nucleoprotein filaments can ad-
versely affect the fitness of mitotic cells. Specifically,
haploid and diploid strains deficient in these Swi2-like
proteins are sensitive to the DNA damaging agent MMS
or are growth impaired, phenotypes that can be overcome
via RAD51 deletion. That the toxicity of Rad51 in the
mutant cells stems from gratuitous Rad51–dsDNA
rad51A265V suppressor allele and the biochemical dem-
onstration of a reduced stability of the rad51 A265V–
dsDNA filament and its enhanced susceptibility to clear-
ance by Rdh54.
Rad51–dsDNA complexes can form either during DSB
repair or can arise from nonspecific binding of Rad51 to
complexes are created after DNA strand exchange where
the invading single strand coated with Rad51 protein pairs
and forms a heteroduplex joint with the homologous
duplex partner chromatid. Rad51 removal likely allows
DNA polymerases access to the primer terminus for com-
pletion of the repair reaction (1,2). Similar to the meiotic
recombinase Dmc1, Rad51 accumulates on chromatin and
requires translocases to promote its removal from
dsDNA, ensuring there is a pool of free protein available
when required for DSB repair (7,25). Unlike the bacterial
RecA protein, the yeast Rad51 protein displays little
binding preference for ssDNA over dsDNA, arguing
that recycling mechanisms must come into play if the
protein is to be targeted to the correct locations during
DSB repair (26–28).
Even though Rad54 and Rdh54 can both remove
Rad51 from dsDNA (5,6), unlike Rdh54, Rad54 is no
more capable of dissociating rad51 A265V from dsDNA
than does wild-type Rad51. Whether this distinction owes
to different epitopes on Rad51 being recognized by Rad54
and Rdh54, or to another reason, remains to be
delineated, although we suggest that Rad51 bound to
Nucleic Acids Research, 2011,Vol.39, No. 156519
dsDNA at a homologous recombination repair intermedi-
ate may differ from Rad51 bound to chromatin (7).
Regardless, the differential sensitivities of the rad51
A265V–dsDNA filament to Rad54 and Rdh54 provide
evidence for important mechanistic differences in the
way these Swi2-like factors act to prevent the non-specific
association of Rad51 with bulk chromatin, which is
further supported by studies of translocase mutant sensi-
tivity to Rad51 overexpression (7). The damage sensitivity
of the uls1 rad51A265V mutant may reflect an interference
of DNA repair by Rdh54 stemming from the accelerated
removal of rad51 A265V mutant protein from a recom-
bination intermediate by this motor protein. An alterna-
tive explanation for suppression phenotypes observed in
this study may be focused on the action of Rad54 after the
formation of a recombination intermediate. Rad54 can
promote dissociation of D-loops (29), and this function
may be easier to perform with the rad51 A265V protein.
In this manner, the rad51A265V mutation would prevent
the accumulation of recombination intermediates.
We note that genetic studies have revealed other import-
ant differences between RAD54 and RDH54 as well.
Specifically, RAD54 is more important for DSB repair,
DNA extension after strand invasion (30) and haploid
recombination than RDH54, while RDH54 makes a
more prominent contribution to inter-homologue recom-
bination than to intra-homologue recombination. This
suggests that for mitotic DSB repair and recombination,
RAD54 is the frontline translocase that is used in Rad51
removal from dsDNA. The observation that rad54 rdh54
diploids have a growth defect and are increased in DNA
damage sensitivity shows that in diploid mitotic situations,
RDH54 has a role in Rad51 removal, and suggests that it
is secondary to RAD54 for damage repair. In meiosis,
RAD54 and RDH54 also seem to fulfill distinct roles;
Rdh54 interacts with Dmc1 and dissociates Dmc1 from
bulk chromatin while Rad540s action appears to be
specific for Rad51 (25,31).
In the initial step of the recombination reaction, Rad51
polymerizes onto ssDNA to form the presynaptic
filament. Biochemical and genetic studies have provided
evidence that the Srs2 helicase regulates recombination
outcome and DNA checkpoint signaling by dismantling
the Rad51 presynaptic filament (19). Although we believe
Figure 5. Removal of Rad51 or rad51 A265V from DNA by Rad54 or Rdh54. (A) Schematic of the assay. Briefly, filaments of Rad51 or rad51
A265V are assembled on biotinylated dsDNA conjugated to streptavidin magnetic beads. Following an incubation with Rad54 or Rdh54, ssDNA is
added to trap Rad51 or rad51 A265V molecules that have been dissociated by Rad54 or Rdh54. The supernatant containing ssDNA and trapped
Rad51 or rad51 A265V and the beads containing Rad51 or rad51 A265V that remains on the biotinylated dsDNA are subjected to SDS–PAGE and
Coomassie Blue staining. (B) and (C) Reactions that contained Rad54 (0, 100, 200, 300 and 400nM in B) or Rdh54 (0, 70, 90, 110 and 140nM in C)
were analyzed and the results were plotted. Error bars show SD of three experiments.
6520Nucleic Acids Research, 2011,Vol.39, No. 15
that the altered rad51A265V binding to dsDNA is respon-
sible for the suppression phenotypes reported in this
article, we found that rad51 A265V also has reduced
binding to ssDNA. Consistent with this altered property,
we have observed that the synthetic sickness/lethality of
srs2D sgs1D (32) and srs2D rad54D (32) haploids is sup-
pressed by the rad51A265V mutation (unpublished data).
That Rad51 binding to ssDNA may cause some of the
rad54 rdh54 uls1 growth problems remains a possibility,
although we note neither Rad54 nor Rdh54 can remove
Rad51 from ssDNA. However, it is possible that they
might remove Rad51 from secondary structures in
ssDNA that resemble dsDNA, and this problem is
alleviated by the rad51 A265V mutant.
Uls1 functions in DNA damage repair
That diploid growth becomes severely affected upon
introducing the uls1D deletion into rad54D rdh54D
mutant cells suggests that Uls1 protein can function to
minimize the toxic accumulation of Rad51 on dsDNA in
the absence of Rad54 and Rdh54. Uls1 possesses both
Swi2-like DNA translocase motifs and a RING finger
domain suggestive of a ubiquitin ligase activity. Whether
these activities are germane for Rad510s removal from
dsDNA represent important issues to address in future,
although preliminary experiments with the ATPase null
show that this mutant version behaves similar to the
complete deletion mutant in vivo.
Additional rad51 mutant alleles have been isolated as
suppressors of other genetic deficiencies. Some of these
rad51 mutations affect DNA binding or at least lie in
regions of Rad51 thought to be involved in DNA
binding (33–35). It may be informative to examine these
rad51 mutant proteins for their susceptibility to removal
from dsDNA by Rad54 and Rdh54. It remains possible
that the rad51 mutants that have been recovered from
screens for suppressors for DNA damage sensitivity
phenotypes and appear to have reduced DNA binding
capacity really allow secondary translocases to remove
Rad51 protein from DNA and thereby suppress DNA
damage sensitivity in the absence of a major factor for
DNA damage repair. Such mutant rad51 proteins would
be expected to support some degree of DNA repair and
recombination, and indeed rad51 A265V fulfills all of
Given the high degree of evolutionary conservation of
recombination genes (36), our findings could provide the
requisite framework for dissecting the role of Rad54 and
related proteins, such as Rad54B, in preventing the accu-
mulation of toxic Rad51–dsDNA complexes in human
cells. The human RAD54 gene has been found mutated
in tumors (37), and it will be interesting to determine
whether any of the tumor-associated hRad54 mutations
affect the ability of this protein to remove Rad51 from
dsDNA. Additionally, the closest human homolog to the
yeast Uls1 protein is SMARCA3/HTLF. Gene silencing
of SMARCA3/HTLF has been linked to gastric tumors
(38–40), but any connection between hRad54 and
SMARCA3/HTLF has not been explored yet. A synergis-
tic enhancement of hRad54 mutations by a SMARCA3/
HTLF deletion or knockdown would be indicative of
some overlap in function.
Supplementary Data are available at NAR Online.
The authors thank Lorraine Symington for the gift of
plasmids, and the members of the Klein, Sung and
Greene laboratories for helpful discussions.
GM057814 to P.S. and CA146940 to H.L.K, E.C.G. and
P.S.). Funding for open access charge: NIH GM053738.
of Health (ES007061 to P.S.,
H.L.K., GM074739to E.C.G.,
Conflict of interest statement. None declared.
1. San Filippo,J., Sung,P. and Klein,H. (2008) Mechanism of
eukaryotic homologous recombination. Ann. Rev. Biochem., 77,
2. Heyer,W.D., Li,X., Rolfsmeier,M. and Zhang,X.P. (2006) Rad54:
the Swiss Army knife of homologous recombination?
Nucleic Acids Res., 34, 4115–4125.
3. Klein,H.L. (1997) RDH54, a RAD54 homologue in
Saccharomyces cerevisiae, is required for mitotic diploid-specific
recombination and repair and for meiosis. Genetics, 147,
4. Shinohara,A., Gasior,S., Ogawa,T., Kleckner,N. and Bishop,D.K.
(1997) Saccharomyces cerevisiae recA homologues RAD51 and
DMC1 have both distinct and overlapping roles in meiotic
recombination. Genes Cells, 2, 615–629.
5. Chi,P., Kwon,Y., Seong,C., Epshtein,A., Lam,I., Sung,P. and
Klein,H.L. (2006) Yeast recombination factor Rdh54 functionally
interacts with the Rad51 recombinase and catalyzes Rad51
removal from DNA. J. Biol. Chem., 281, 26268–26279.
6. Solinger,J.A., Kiianitsa,K. and Heyer,W.D. (2002) Rad54, a Swi2/
Snf2-like recombinational repair protein, disassembles
Rad51:dsDNA filaments. Mol. Cell, 10, 1175–1188.
7. Shah,P.P., Zheng,X., Epshtein,A., Carey,J.N., Bishop,D.K. and
Klein,H.L. (2010) Swi2/Snf2-related translocases prevent
accumulation of toxic Rad51 complexes during mitotic growth.
Mol. Cell, 39, 862–872.
8. Dresser,M.E., Ewing,D.J., Conrad,M.N., Dominguez,A.M.,
Barstead,R., Jiang,H. and Kodadek,T. (1997) DMC1 functions
in a Saccharomyces cerevisiae meiotic pathway that is largely
independent of the RAD51 pathway. Genetics, 147, 533–544.
9. Lea,D.E. and Coulson,C.A. (1948) The distribution of the
numbers of mutants in bacterial populations. J. Genetics, 49,
10. Sung,P. and Stratton,S.A. (1996) Yeast Rad51 recombinase
mediates polar DNA strand exchange in the absence of ATP
hydrolysis. J. Biol. Chem., 271, 27983–27986.
11. Van Komen,S., Macris,M., Sehorn,M.G. and Sung,P. (2006)
Purification and assays of Saccharomyces cerevisiae homologous
recombination proteins. Methods Enzymol., 408, 445–463.
12. Raschle,M., Van Komen,S., Chi,P., Ellenberger,T. and Sung,P.
(2004) Multiple interactions with the Rad51 recombinase govern
the homologous recombination function of Rad54. J. Biol. Chem.,
13. Petukhova,G., Van Komen,S., Vergano,S., Klein,H. and Sung,P.
(1999) Yeast Rad54 promotes Rad51-dependent homologous
Nucleic Acids Research, 2011,Vol.39, No. 156521
DNA pairing via ATP hydrolysis-driven change in DNA double Download full-text
helix conformation. J. Biol. Chem., 274, 29453–29462.
14. Van Komen,S., Petukhova,G., Sigurdsson,S., Stratton,S. and
Sung,P. (2000) Superhelicity-driven homologous DNA pairing by
yeast recombination factors Rad51 and Rad54. Mol. Cell, 6,
15. Graneli,A., Yeykal,C.C., Prasad,T.K. and Greene,E.C. (2006)
Organized arrays of individual DNA molecules tethered to
supported lipid bilayers. Langmuir, 22, 292–299.
16. Robertson,R.B., Moses,D.N., Kwon,Y., Chan,P., Chi,P.,
Klein,H., Sung,P. and Greene,E.C. (2009) Structural transitions
within human Rad51 nucleoprotein filaments. Proc. Natl Acad.
Sci. USA, 106, 12688–12693.
17. Prasad,T.K., Yeykal,C.C. and Greene,E.C. (2006) Visualizing the
assembly of human Rad51 filaments on double-stranded DNA.
J. Mol. Biol., 363, 713–728.
18. Uzunova,K., Gottsche,K., Miteva,M., Weisshaar,S.R.,
Glanemann,C., Schnellhardt,M., Niessen,M., Scheel,H.,
Hofmann,K., Johnson,E.S. et al. (2007) Ubiquitin-dependent
proteolytic control of SUMO conjugates. J. Biol. Chem., 282,
19. Sung,P. and Klein,H. (2006) Mechanism of homologous
recombination: mediators and helicases take on regulatory
functions. Nat. Rev. Mol. Cell. Biol., 7, 739–750.
20. Conway,A.B., Lynch,T.W., Zhang,Y., Fortin,G.S., Fung,C.W.,
Symington,L.S. and Rice,P.A. (2004) Crystal structure of a Rad51
filament. Nat. Struct. Mol. Biol., 11, 791–796.
21. Bennink,M.L., Scharer,O.D., Kanaar,R., Sakata-Sogawa,K.,
Schins,J.M., Kanger,J.S., de Grooth,B.G. and Greve,J. (1999)
Single-molecule manipulation of double-stranded DNA using
optical tweezers: interaction studies of DNA with RecA and
YOYO-1. Cytometry, 36, 200–208.
22. Egelman,E.H. and Stasiak,A. (1993) Electron microscopy of
RecA–DNA complexes: two different states, their functional
significance and relation to the solved crystal structure. Micron,
23. Hegner,M., Smith,S.B. and Bustamante,C. (1999) Polymerization
and mechanical properties of single RecA–DNA filaments.
Proc. Natl Acad. Sci. USA, 96, 10109–10114.
24. Prasad,T.K., Robertson,R.B., Visnapuu,M.L., Chi,P., Sung,P. and
Greene,E.C. (2007) A DNA-translocating Snf2 molecular motor:
Saccharomyces cerevisiae Rdh54 displays processive translocation
and extrudes DNA loops. J. Mol. Biol., 369, 940–953.
25. Holzen,T.M., Shah,P.P., Olivares,H.A. and Bishop,D.K. (2006)
Tid1/Rdh54 promotes dissociation of Dmc1 from
nonrecombinogenic sites on meiotic chromatin. Genes Dev., 20,
26. Ogawa,T., Yu,X., Shinohara,A. and Egelman,E.H. (1993)
Similarity of the yeast RAD51 filament to the bacterial RecA
filament. Science, 259, 1896–1899.
27. Sung,P. and Robberson,D.L. (1995) DNA strand exchange
mediated by a RAD51–ssDNA nucleoprotein filament with
polarity opposite to that of RecA. Cell, 82, 453–461.
28. Zaitseva,E.M., Zaitsev,E.N. and Kowalczykowski,S.C. (1999) The
DNA binding properties of Saccharomyces cerevisiae Rad51
protein. J. Biol. Chem., 274, 2907–2915.
29. Bugreev,D.V., Hanaoka,F. and Mazin,A.V. (2007) Rad54
dissociates homologous recombination intermediates by branch
migration. Nat. Struct. Mol. Biol., 14, 746–753.
30. Li,X. and Heyer,W.D. (2009) RAD54 controls access to the
invading 30-OH end after RAD51-mediated DNA strand invasion
in homologous recombination in Saccharomyces cerevisiae.
Nucleic Acids Res., 37, 638–646.
31. Chi,P., Kwon,Y., Moses,D.N., Seong,C., Sehorn,M.G.,
Singh,A.K., Tsubouchi,H., Greene,E.C., Klein,H.L. and Sung,P.
(2009) Functional interactions of meiotic recombination factors
Rdh54 and Dmc1. DNA Repair, 8, 279–284.
32. Klein,H.L. (2001) Mutations in recombinational repair and in
checkpoint control genes suppress the lethal combination of
srs2Delta with other DNA repair genes in Saccharomyces
cerevisiae. Genetics, 157, 557–565.
33. Aboussekhra,A., Chanet,R., Adjiri,A. and Fabre,F. (1992)
Semidominant suppressors of Srs2 helicase mutations of
Saccharomyces cerevisiae map in the RAD51 gene, whose
sequence predicts a protein with similarities to procaryotic RecA
proteins. Mol. Cell. Biol., 12, 3224–3234.
34. Chanet,R., Heude,M., Adjiri,A., Maloisel,L. and Fabre,F. (1996)
Semidominant mutations in the yeast Rad51 protein and their
relationships with the Srs2 helicase. Mol. Cell. Biol., 16,
35. Zhang,X.P., Lee,K.I., Solinger,J.A., Kiianitsa,K. and Heyer,W.D.
(2005) Gly-103 in the N-terminal domain of Saccharomyces
cerevisiae Rad51 protein is critical for DNA binding. J. Biol.
Chem., 280, 26303–26311.
36. Krogh,B.O. and Symington,L.S. (2004) Recombination proteins in
yeast. Ann. Rev. Genet., 38, 233–271.
37. Hiramoto,T., Nakanishi,T., Sumiyoshi,T., Fukuda,T.,
Matsuura,S., Tauchi,H., Komatsu,K., Shibasaki,Y., Inui,H.,
Watatani,M. et al. (1999) Mutations of a novel human RAD54
homologue, RAD54B, in primary cancer. Oncogene, 18,
38. Hamai,Y., Oue,N., Mitani,Y., Nakayama,H., Ito,R.,
Matsusaki,K., Yoshida,K., Toge,T. and Yasui,W. (2003) DNA
hypermethylation and histone hypoacetylation of the HLTF gene
are associated with reduced expression in gastric carcinoma.
Cancer Sci., 94, 692–698.
39. Hibi,K., Nakayama,H., Kanyama,Y., Kodera,Y., Ito,K.,
Akiyama,S. and Nakao,A. (2003) Methylation pattern of HLTF
gene in digestive tract cancers. Int. J. Cancer, 104, 433–436.
40. Moinova,H.R., Chen,W.D., Shen,L., Smiraglia,D., Olechnowicz,J.,
Ravi,L., Kasturi,L., Myeroff,L., Plass,C., Parsons,R. et al. (2002)
HLTF gene silencing in human colon cancer. Proc. Natl Acad.
Sci. USA, 99, 4562–4567.
6522 Nucleic Acids Research, 2011,Vol.39, No. 15