Tài liệu Báo cáo khoa học: Roles of the human Rad51 L1 and L2 loops in DNA binding doc

Roles of the human Rad51 L1 and L2 loops in DNA binding
Yusuke Matsuo
1
, Isao Sakane
2
, Yoshimasa Takizawa
1
, Masayuki Takahashi
3
and Hitoshi Kurumizaka
1,2
1 Graduate School of Science and Engineering, Waseda University, Tokyo, Japan
2 Institute for Biochemical Engineering, Waseda University, Tokyo, Japan
3 UMR 6204 Biocatalyse-Biotechnologie-Bioregulation, Centre National de la Recherche Scientifique, and University of Nantes, France
The Rad51 proteins are the eukaryotic orthologs of
the bacterial RecA protein [1], which promotes key
steps in homologous recombination [2–5]. A RAD51
null mutation causes severe defects in meiotic homol-
ogous recombination and mitotic recombinational
repair of double strand breaks (DSBs) in Saccharomy-
ces cerevisiae [1]. Rad51 is thus required for both the
meiotic and mitotic homologous recombination pro-
cesses, while another ortholog, Dmc1, is specific to
meiotic homologous recombination [6–8]. In higher
eukaryotes, Rad51 is even essential for cell survival:
disruption of the RAD51 gene in mice results in early
embryonic lethality [9,10] and the RAD51 gene knock-
out in chicken DT40 cells causes cell death, with the
accumulation of spontaneous chromosomal breaks
[11].
Rad51 and RecA apparently use similar mechanisms

to promote homologous recombination [12–15]. Dur-
ing the homologous recombination process, Rad51 is
thought to bind single-stranded tails produced at DSB
sites, and to form a helical nucleoprotein filament. The
single-stranded DNA (ssDNA) and double-stranded
Keywords
DNA binding; DNA repair; Rad51; Rad51
mutant; recombination
Correspondence
H. Kurumizaka, Graduate School of Science
and Engineering, Waseda University, 3-4-1
Okubo, Shinjuku-ku, Tokyo 169-8555, Japan
Fax: +81 3 5292 9211
Tel: +81 3 5286 8189
E-mail:
(Received 12 November 2005, revised
3 April 2006, accepted 16 May 2006)
doi:10.1111/j.1742-4658.2006.05323.x
The human Rad51 protein, a eukaryotic ortholog of the bacterial RecA
protein, is a key enzyme that functions in homologous recombination and
recombinational repair of double strand breaks. The Rad51 protein con-
tains two flexible loops, L1 and L2, which are proposed to be sites for
DNA binding, based on a structural comparison with RecA. In the present
study, we performed mutational and fluorescent spectroscopic analyses on
the L1 and L2 loops to examine their role in DNA binding. Gel retarda-
tion and DNA-dependent ATP hydrolysis measurements revealed that the
substitution of the tyrosine residue at position 232 (Tyr232) within the L1
loop with alanine, a short side chain amino acid, significantly decreased the
DNA-binding ability of human Rad51, without affecting the protein fold-
ing or the salt-induced, DNA-independent ATP hydrolysis. Even the

conservative replacement with tryptophan affected the DNA binding,
indicating that Tyr232 is involved in DNA binding. The importance of the
L1 loop was confirmed by the fluorescence change of a tryptophan residue,
replacing the Asp231, Ser233, or Gly236 residue, upon DNA binding. The
alanine replacement of phenylalanine at position 279 (Phe279) within the
L2 loop did not affect the DNA-binding ability of human Rad51, unlike
the Phe203 mutation of the RecA L2 loop. The Phe279 side chain may not
be directly involved in the interaction with DNA. However, the fluores-
cence intensity of the tryptophan replacing the Rad51-Phe279 residue was
strongly reduced upon DNA binding, indicating that the L2 loop is also
close to the DNA-binding site.
Abbreviations
DSB, double strand break; dsDNA, double-stranded DNA; HsRad51, Homo sapiens Rad51; RPA, replication protein A; ScRad51,
Saccharomyces cerevisiae Rad51; ssDNA, single-stranded DNA; SSB, single stranded DNA-binding protein.
3148 FEBS Journal 273 (2006) 3148–3159 ª 2006 The Authors Journal compilation ª 2006 FEBS
DNA (dsDNA) molecules bind within the Rad51
nucleoprotein filament along the helical axis, thus
forming the ternary complex containing ssDNA,
dsDNA, and Rad51. In the ternary complex, the
homologous sequence between ssDNA and dsDNA is
aligned, and the ssDNA forms a heteroduplex with a
complementary strand of dsDNA (homologous pair-
ing). The heteroduplex region produced by homolog-
ous pairing is then extended by the Rad51-mediated
strand exchange. Therefore, Rad51 should have at
least two DNA-binding sites, as in RecA [16], and the
identification of these sites is important for under-
standing the reaction mechanism and the regulation of
homologous recombination.
So far, the crystal structures of bacterial RecA, ar-

chaeal Rad51 (RadA), yeast Rad51 (ScRad51), human
Rad51 (HsRad51), and human Dmc1 (HsDmc1) have
been solved [17–22]. These structural analyses revealed
that these proteins have highly conserved three-dimen-
sional structures, especially in their ATPase domains.
Two flexible loops, L1 and L2, which are involved in
DNA binding by Escherichia coli RecA [17], have also
been identified in these eukaryotic and archaeal pro-
teins (Fig. 1A). Like the case of bacterial RecA, the L1
and L2 loops of the eukaryotic Rad51 proteins face
inside of their helical filaments, where the DNA should
be located, and are not found at the ATP-binding site
or the subunit–subunit interface of the Rad51 filament.
Therefore, the Rad51 L1 and L2 loops may also be
involved in DNA binding.
In the present study, we performed mutational and
fluorescence spectroscopic analyses on HsRad51 to
examine whether these loops are actually involved in
DNA binding. Because aromatic residues are involved
in the ssDNA binding by bacterial single stranded
DNA-binding protein (SSB) and human replication
protein A (RPA) [23,24], we performed mutational
analyses on Tyr232 in the L1 loop and Phe279 in the
L2 loop of HsRad51. We also performed tryptophan-
scanning mutagenesis across the HsRad51-L1 loop and
measured the fluorescence changes of the tryptophan
residues upon DNA binding.
Results
Strategy of mutational analysis
In order to study the functions of the L1 and L2

loops of HsRad51, we examined the effect of replacing
the aromatic residues in the L1 and L2 loops with
A
B
C
Fig. 1. HsRad51 and the Rad51 mutants. (A) Alignment of the HsRad51 domains to those of the Methanococcus voltae RadA (MvRadA),
Saccharomyces cerevisiae Rad51 (ScRad51), and Escherichia coli RecA (EcRecA) domains. The N-terminal domains, the conserved ATPase
domains, and the C-terminal domain are indicated by shaded boxes. The L1 and L2 loops are indicated by black boxes. (B) Alignment of the
HsRad51 sequence to those of MvRadA, Pyrococcus furiosus Rad51 (PfRad51), and ScRad51 around the L1 and L2 loops. The L1 and L2
loops, which are invisible in the crystal structure of the ATPase domain of HsRad51 [21], are represented by boxes, and the Y232 and F279
residues are indicated by shaded boxes. (C) Purified HsRad51 (lane 2), Y232A mutant (lane 3), F279A mutant (lane 4), Y232W mutant (lane
5), F279W mutant (lane 6), D231W mutant (lane 7), S233W mutant (lane 8), and G236W mutant (lane 9) were analyzed by 15% (w ⁄ v)
SDS ⁄ PAGE with Coomassie Brilliant Blue staining. Lane 1 indicates the molecular mass markers.
Y. Matsuo et al. Roles of the HsRad51 L1 and L2 loops
FEBS Journal 273 (2006) 3148–3159 ª 2006 The Authors Journal compilation ª 2006 FEBS 3149
alanine. If the aromatic side chain is involved in the
interaction with DNA, then its replacement with
alanine, a short side chain amino acid residue,
should affect the DNA binding of the protein. The
DNA-binding ability of these alanine-substituted
Rad51 mutants was evaluated by a gel retardation tech-
nique and by measurements of the DNA-dependent
ATPase activity. To ensure that the defect in DNA
binding is not due to incorrect folding of the Rad51
mutants or a loss of binding cooperativity by changing
the subunit–subunit contacts, we measured their CD
spectra and carried out gel filtration chromatography.
The DNA binding by Rad51 is highly cooperative, with
strong subunit–subunit contacts, and the protein can
form a polymer even in the absence of DNA [20,25].

We also prepared another type of Rad51 mutant, in
which one of these aromatic residues was replaced by
tryptophan, a fluorescent probe. If the residue is within
or close to the DNA-binding site, then we would
expect a large change in its fluorescence upon DNA
binding. Using such an approach, we previously
showed that Phe203 in the L2 loop of RecA is close to
the DNA-binding site [26]. Furthermore, we performed
the tryptophan-scanning mutagenesis across the
HsRad51-L1 loop, and tested the interaction between
the L1 loop and DNA.
Involvement of the L1 loop-Tyr232 residue in
DNA binding
The L1 loop of HsRad51 contains an aromatic residue
(Tyr232) that is highly conserved among the eukaryotic
and archaeal Rad51 proteins (Fig. 1B). We prepared
the Rad51-Y232A and Rad51-Y232W mutants, in
which the Tyr232 residues were replaced by alanine
(Y232A) and tryptophan (Y232W), respectively, by site
directed mutagenesis, and purified them to near homo-
geneity by a four-step purification method based on
HsRad51 purification, including nickel-nitrilotriacetic
acid (Ni-NTA) agarose column chromatography,
removal of the hexahistidine tag from HsRad51 with
thrombin protease, spermidine precipitation, and
MonoQ column chromatography (Fig. 1C).
Rad51-Y232A yielded a CD spectrum similar to that
of HsRad51, indicating that the mutation did not
affect either the folding or global structure of the
protein (Fig. 2A,B). Gel filtration chromatography

revealed that Rad51-Y232A formed polymers by self
association, like HsRad51: the protein eluted in the
void volume from the Superdex 200 gel filtration
column (data not shown). Therefore, the mutation did
not appear to affect the subunit–subunit contact in
the Rad51 polymer. Rad51-Y232A also exhibited
ABC
FED
Fig. 2. Circular dichroism analysis and ATPase activities of the Rad51 mutants. (A–C) CD spectra of HsRad51 (6.7 lM) and the Rad51 mutant
(6.7 l
M) were recorded at 25 °C. HsRad51 (A); Y232A mutant (B); and F279A mutant (C). (D–F) The ATPase activities of the Rad51 mutants.
Time course experiments are shown. d, m,andn indicate experiments in the presence of NaCl (1.6
M), ssDNA (20 lM), and dsDNA
(20 l
M), respectively. s indicate experiments in the absence of NaCl and DNA. HsRad51 (D); Y232A mutant (E); F279A mutant (F).
Roles of the HsRad51 L1 and L2 loops Y. Matsuo et al.
3150 FEBS Journal 273 (2006) 3148–3159 ª 2006 The Authors Journal compilation ª 2006 FEBS
salt-induced ATPase activity very similar to that of
HsRad51 in the absence of DNA (Figs 2D,E). These
results suggest that the Tyr232 residue is within neither
the subunit–subunit interface nor the ATP-binding
site.
In contrast to the salt-induced ATPase activity,
Rad51-Y232A did not exhibit DNA-dependent
ATPase activity (Fig. 2E). Neither ssDNA nor dsDNA
induced the ATPase activity of Rad51-Y232A, while
the ATPase activity of HsRad51 was stimulated by
ssDNA and dsDNA (Fig. 2D). These results indicate
that Rad51-Y232A was defective in DNA binding.
Consistent with this finding, a gel retardation experi-

ment showed that Rad51-Y232A was defective in
ssDNA and dsDNA binding (Fig. 3A,B, lanes 5–8), as
compared to HsRad51 (lanes 1–4). These results dem-
onstrate the importance of the Rad51-Tyr232 residue
in DNA binding. The gel retardation experiments also
revealed that even the conservative replacement of
Tyr232 with tryptophan (Rad51-Y232W) caused signi-
ficant defects in dsDNA binding, although it possessed
the ssDNA-binding ability (Fig. 3A,B, lanes 9–12). As
expected from the DNA binding defect, neither Rad51-
Y232A nor Rad51-Y232W promoted the strand-
exchange reaction (Fig. 4B). These results suggest
that the Rad51-Tyr232 residue in the L1 loop is
involved in the functional DNA binding during strand
exchange.
Tryptophan-scanning mutagenesis of the
HsRad51-L1 loop
To gain further information about DNA binding by
the L1 loop, we performed tryptophan-scanning
mutagenesis across the L1 loop (from Thr230 to
Gly236). Five mutant genes corresponding to
the Rad51-D231W, Rad51-S233W, Rad51-G234W,
Rad51-R235W, and Rad51-G236W mutants, in which
Asp231, Ser233, Gly234, Arg235, and Gly236 were
replaced by tryptophan, respectively, were constructed,
and were expressed in E. coli cells. The Rad51-
D231W, Rad51-S233W, and Rad51-G236W mutants
were purified to near homogeneity by the same proto-
col employed with the wildtype HsRad51 (Fig. 1C,
lanes 7–9), while the Rad51-G234W and Rad51-

R235W mutants could not be purified because they
formed insoluble aggregates. In contrast to the Rad51-
Y232W mutant, the Rad51-D231W, Rad51-S233W,
and Rad51-G236W mutants did not cause significant
defects in ssDNA binding and dsDNA binding
A
B
Fig. 3. The DNA binding activities of the Rad51 mutants. (A) The ssDNA binding experiments. The /X174 circular ssDNA (40 lM) was incu-
bated with HsRad51 or the Rad51 mutants at 37 °C for 10 min. (B) The dsDNA binding experiments. Linearized /X174 DNA (20 l
M)was
incubated with HsRad51 or the Rad51 mutants at 37 ° C for 10 min. The samples were analyzed by 0.8% (w ⁄ v) agarose gel electrophoresis
in 1 · TAE buffer. Lanes 1, 5, 9, 13, 17, 21, 25, and 29 indicate control experiments without HsRad51. The bands were visualized by ethi-
dium bromide staining. The protein concentrations used in the ssDNA binding experiments were 1 l
M (lanes 2, 6, 10, 14, 18, 22, 26, and
30), 2 l
M (lanes 3, 7, 11, 15, 19, 23, 27, and 31), and 4 lM (lanes 4, 8, 12, 16, 20, 24, 28, and 32).
Y. Matsuo et al. Roles of the HsRad51 L1 and L2 loops
FEBS Journal 273 (2006) 3148–3159 ª 2006 The Authors Journal compilation ª 2006 FEBS 3151
(Fig. 3A,B). These results suggest that the Asp231,
Ser233, and Gly236 residues of HsRad51 are not in
direct contact with DNA. As expected, the Rad51-
S233W and Rad51-G236W mutants were proficient
in strand exchange (Fig. 4B). However, the Rad51-
D231W mutant was defective in strand exchange
(Fig. 4B, lanes 17–19), suggesting that this acidic resi-
due (Asp231) may have some role in this process.
Therefore, the Rad51-S233W and Rad51-G236W
mutants are suitable for the fluorescent spectroscopic
analysis, in contrast to the Rad51-Y232W mutant,
which is significantly defective in dsDNA binding and

strand exchange.
Fluorescent spectroscopic analysis of the
HsRad51-L1 mutants
The fluorescence change of the Rad51-D231W, Rad51-
S233W, and Rad51-G236W mutants upon DNA bind-
ing was examined, to confirm that the L1 loop is in
the DNA-binding site. HsRad51 has no tryptophan
residue, and therefore, the fluorescence of these
mutants corresponded to that of the inserted trypto-
phan residue. The fluorescence peaked at 341, 343 and
347 nm for the tryptophan residues inserted at posi-
tions 231, 233, and 236 of HsRad51, respectively. The
peak positions indicate that residues 231 and 233 are
in a rather nonpolar environment (only partly exposed
to the solvent), while residue 236 is more exposed to
the solvent. The fluorescence intensity decreased by
about 30, 50 and 60% for the tryptophan 231, 233 and
236 residues, respectively, in the presence of poly(dT),
a model ssDNA, with or without ATP (Table 1).
These results confirm that the residues are close to the
DNA-binding site.
We then examined if these fluorescence changes
occurred by the binding of the first or second DNA,
by titrating these modified Rad51 proteins with
poly(dT). In the presence of ATP, Rad51 can bind at
least two DNA strands, each with a stoichiometry of 3
bases per monomer. However, the fluorescence changes
of these proteins were almost saturated at 3 bases per
monomer of poly(dT), showing that the changes were
mainly due to the binding of the first DNA, while the

binding of the second DNA had less influence
(Fig. 5A). To ensure that the estimation of the pro-
tein:poly(dT) ratio was correct, we performed the titra-
tion in the absence of ATP, where Rad51 binds only
one DNA strand, with a stoichiometry of about 4–5
bases per monomer [27]. The titration of these proteins
revealed that the fluorescence change was saturated at
about 4–5 bases per monomer of poly(dT) for all of
the Rad51 mutants (data not shown), as expected.
By contrast, the change in fluorescence upon dissoci-
ation of the Rad51 filament to monomers, by adding
2.5 m urea [28], was slight (Table 1) and shifted the
fluorescence peak to a shorter wavelength. We expected
the peak position to move to a longer wavelength if the
residue is involved in the subunit–subunit interaction,
because of its exposure to solvent upon the disso-
Y232W F279W
Y232A F279AWT
D231W G236W
S233W
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
joint molecule
(jm)
nicked circular DNA
(nc)
ssDNA
dsDNA
jm
nc
A

B
Fig. 4. The strand-exchange activities of the Rad51 mutants. (A) A schematic diagram of the strand-exchange assay. (B) The Rad51 concen-
trations were 1 l
M (lanes 2, 5, 8, 11, 14, 17, 20, and 23), 2 lM (lanes 3, 6, 9, 12, 15, 18, 21, and 24), and 4 lM (lanes 4, 7, 10, 13, 16, 19,
22, and 25). Lane 1 indicates a negative control experiment without the Rad51 protein. Joint molecules and nicked circular DNA are indica-
ted by jm and nc, respectively.
Roles of the HsRad51 L1 and L2 loops Y. Matsuo et al.
3152 FEBS Journal 273 (2006) 3148–3159 ª 2006 The Authors Journal compilation ª 2006 FEBS
ciation, like the case of Tyr188 of Xenopus laevis
Rad51 [28]. These results indicate that these residues in
the modified Rad51 proteins are not strongly involved
in the subunit–subunit contacts. However, the signifi-
cant change in the peak position suggests that the L1
loop is close to the subunit–subunit interface.
The binding of ATP only minimally affected the
fluorescence intensity in the modified Rad51 proteins:
only about a 5% increase was detected for Trp231 and
Trp233, and a 10% decrease was found for Trp236.
The peak position is also only slightly affected by ATP
(less than 1 nm). The results confirm the conclusion
obtained from the mutational analysis that the Rad51-
L1 loop is not directly involved in ATP binding or
ATP hydrolysis. The slight change in the fluorescence
intensity upon nucleotide binding indicates that some
environmental change occurs around the L1 loop. This
nucleotide-induced allosteric effect on the L1 loop may
explain the mechanism of DNA-binding regulation by
nucleotide binding.
The L2 loop is close to the DNA-binding site
The L2 loop of HsRad51 contains only one aromatic

residue, Phe279, which is highly conserved among the
eukaryotic and archaeal Rad51 proteins (Fig. 1B). For
this residue, we performed analyses similar to those for
Tyr232 of the L1 loop, to examine its role in DNA
binding, by preparing two mutant proteins, Rad51-
F279A and Rad51-F279W, in which the Phe279
residue was replaced by alanine and tryptophan,
respectively (Fig. 1C, lanes 3 and 5). The CD spectrum
(Fig. 2C), elution pattern from the gel filtration col-
umn (data not shown), and salt-induced ATPase activ-
ity (Fig. 2F) of the purified Rad51-F279A were all
similar to those of HsRad51, indicating that the muta-
tion did not affect the global structure, the polymer
formation, and the ATPase activity. The mutation also
did not affect the DNA-dependent ATPase, DNA
binding, and strand-exchange activities of HsRad51
(Figs 2F, 3, and 4).
Because the Rad51-F279A mutant did not show a
deficiency in DNA binding, we next tested the fluores-
cence changes of Rad51-F279W upon DNA binding.
Rad51-F279W was confirmed to bind DNA like
HsRad51, according to the gel retardation experiments
(Fig. 3A,B, lanes 17–20). The fluorescence of Rad51-
F279W peaked at 340 nm, with a rather large intensity
(Table 1). This fluorescence feature indicates that the
residue is only partly exposed to the solvent or exists
in a rather nonpolar environment, like Trp231 and
Trp233, but does not strongly contact other residues.
The fluorescence of Rad51-F279W was strongly
Table 1. Changes in the fluorescence of tryptophan probes inserted in the L1 and L2 loop upon binding of nucleotides and DNA. The fluorescence of modified Rad51 was measured after

the addition of the indicated element. The wavelength (k
max
) and the intensity (I
max
) at the maximum emission are noted. The intensities were normalized to that of free tryptophan as
100, and were determined with a precision of 3%. The changes occurring upon the addition of the element are noted in parentheses. ND, not determined.
D231W (L1 loop) S233W (L1 loop) G236W (L1 loop) F279W (L2 loop)
k
max
I
max
k
max
I
max
k
max
I
max
k
max
I
max
Protein alone 341 nm 82 343 nm 88 347 nm 69 340 nm 68
+ ATP 341 (no change) 85 (+4%) 343 (no change) 92 (+4%) 347 (no change) 63 ()11%) 340 (no change) 68 (no change)
+ poly(dT) 340 (no change) 62 ()25%) 343 (no change) 53 ()53%) 347 (no change) 32 ()55%) 340 (no change) 44 ()35%)
+ poly(dT) to Rad51-ATP 340 (no change) 56 ()32%) 343 (no change) 48 ()45%) 347 (no change) 28 ()60%) 340 (no change) 46 ()32%)
+ poly(dT) to
Rad51-ATP-poly(dT)
340 (no change) 50 ()7%) 343 (no change) 44 ()4%) 347 (no change) 24 ()5%) 340 (no change) 44 ()3%)

+ poly(dT):poly(dA) to
Rad51-ATP-poly(dT)
ND ND ND 340 (no change) 45 ()1%)
+2.5
M urea 339 ()2 nm) 53 (+4%) 341 ()2 nm) 70 ()12%) 344 ()3 nm) 74 (+7%) 342 (+2 nm) 70 (+2%)
Y. Matsuo et al. Roles of the HsRad51 L1 and L2 loops
FEBS Journal 273 (2006) 3148–3159 ª 2006 The Authors Journal compilation ª 2006 FEBS 3153
decreased (more than 30% decrease) upon poly(dT)
binding, without a change in peak position, in both
the presence and absence of ATP (Table 1). Although
a smaller change was observed upon dsDNA binding
(15%), it was accompanied by a change in the peak
position ()1 nm). These results suggest that the residue
is close to the DNA-binding site. The titration of
Rad51-F279W with poly(dT) revealed that the fluores-
cence change became saturated at about 4.5 bases per
monomer of poly(dT) in the absence of ATP (data not
shown), as observed for Xenopus Rad51 [27]. In the
presence of ATP, the fluorescence change was almost
saturated with 3 bases per monomer of poly(dT), the
amount needed to saturate the first DNA-binding site
(Fig. 5B). The binding of the second poly(dT) did not
change the fluorescence of Rad51-F279W, in contrast
to the results obtained with RecA-F203W, with a tryp-
tophan inserted in the L2 loop of RecA [26], which
displayed changes in fluorescence with the second
poly(dT). The addition of poly(dA):poly(dT) duplex
DNA to the preformed Rad51-Y279W–ATP–poly(dT)
complex also did not affect the fluorescence (Table 1),
suggesting that residue 279 is not close to the second

DNA.
The fluorescence was not significantly changed by
the addition of nucleotides (ATP and ADP) or by the
addition of 2.5 m urea, which dissociates the protein to
monomers, confirming the conclusion from the muta-
tional analysis that the residue is involved in neither
the subunit–subunit contacts nor ATP hydrolysis
(Table 1).
Discussion
We have investigated the DNA-binding sites of
HsRad51 to understand the mechanism of homologous
pairing and strand exchange catalyzed by this protein
for homologous recombination. Because several rela-
tionships exist between homologous recombination
and cancer [29–32], and Rad51 is thus a potential tar-
get for anticancer treatment [33], this study would also
contribute to its development. Our mutational and
fluorescent spectroscopic analyses indicated the
involvement of the HsRad51 L1 and L2 loops in DNA
binding, like the case of RecA. However, there could
be some mechanistic differences between the proteins.
The Rad51-L1 loop
In the present study, we found that the replacement of
Tyr232 with alanine strongly reduced the ssDNA- and
dsDNA-binding abilities of HsRad51. The fact that
even the conservative replacement with tryptophan
affects the dsDNA binding clearly indicated that
Tyr232 is involved in DNA binding by HsRad51. The
tryptophan-scanning mutagenesis suggested that other
residues, such as Asp231, Ser233 and Gly236, within

the L1 loop are less important for DNA binding,
although the D231W mutation affects the strand-
exchange reaction. The proximity of the L1 loop to
the DNA-binding site was also verified by fluorescent
spectroscopic analyses of tryptophan residues inserted
in this loop. The importance of the L1 loop for DNA
binding by RecA ⁄ Rad51 family proteins was observed
by mutational and photocrosslinking analyses of RecA
[34–37] and a mutational analysis of HsDmc1 [22].
The Dmc1-F233A mutation, which corresponds to the
Y232A mutation of HsRad51, also affected DNA
binding. However, interestingly, the Dmc1-F233A
mutation affected only ssDNA binding, but not
dsDNA binding [22], in contrast to the case of the
A
B
Fig. 5. Fluorescence changes of tryptophan probes inserted in the
L1 and L2 loops of Rad51 upon poly(dT) binding. The fluorescence
intensities of 1 l
M modified Rad51, in which a tryptophan probe
was inserted within the L1 (A) or L2 loop (B), were measured at
350 nm, after each stepwise addition of poly(dT) in the presence of
1m
M ATP. The intensity was normalized to that of the correspond-
ing protein without poly(dT), and is presented as a function of the
poly(dT):protein ratio. (A) Graphic representation of fluorescence
intensities with Rad51: Rad51-D231W (d), Rad51-S233W (m) and
Rad51-G236W (n). (B) Graphic representation of the fluorescence
intensities with Rad51-F279W (s).
Roles of the HsRad51 L1 and L2 loops Y. Matsuo et al.

3154 FEBS Journal 273 (2006) 3148–3159 ª 2006 The Authors Journal compilation ª 2006 FEBS
Rad51-Y232A mutant. Therefore, the DNA-binding
mode of HsRad51 should be somewhat different from
that of HsDmc1. The octameric ring form [38,39] and
the helical filament form [40] of HsDmc1 are capable
of binding DNA, but Rad51 does not form such an
octameric ring with DNA. These differences in the
higher ordered structures of Rad51 and Dmc1 with the
DNA may reflect the different DNA binding modes of
these proteins.
Interestingly, even the conservative replacement of
Rad51-Tyr232 by tryptophan affects the dsDNA bind-
ing, indicating direct contact between Tyr232 and
DNA. Aromatic residues, such as Tyr, Phe, and Trp,
can stack with the base moieties of ssDNA for the inter-
action. Such contacts have been observed in the interac-
tions of the SSB and RPA proteins with ssDNA [23,24].
The Rad51-Tyr232 side chain could thus stack with
DNA bases for its interaction with DNA. A reduction
in the tyrosine fluorescence intensity of human Rad51
upon DNA binding was reported [41]. The fluorescence
of Tyr232 could be the source of this fluorescence
change. Consistent with its importance, the Tyr232 resi-
due is highly conserved as an aromatic residue among
the eukaryotic and archaeal Rad51 and Dmc1 proteins.
In contrast to Rad51, Tyr232 is not conserved in the
L1 loop of E. coli RecA. This fact suggests that the L1
loop of Rad51 interacts with DNA in a different man-
ner from that of RecA. His163 is the only residue with
a ring structure similar to that of tyrosine in the

L1 loop of RecA. A chemical interference analysis
revealed the protection of one of the two histidine resi-
dues, His97 and His163, of RecA by DNA binding
[42]. His163 could be the protected histidine residue.
However, its chemical modification did not affect the
DNA binding by RecA [42], and the residue appar-
ently could be replaced with another amino acid [35],
unlike the Rad51-Tyr232 residue. Therefore, the
His163 residue of RecA is not functionally equivalent
to the Tyr232 residue of HsRad51.
The Rad51-L2 loop
In contrast to the Tyr232 residue, the direct involve-
ment of the Phe279 residue within the HsRad51-L2
loop in DNA binding is less evident, because the
F279A mutation in the L2 loop did not reduce the
DNA-binding ability of HsRad51. It has been reported
that some other mutations of residues in the ScRad51
and HsRad51 L2 loops did not affect the DNA-bind-
ing abilities [43,44]. In addition, most of the Rad51
mutants with a mutation in the L2 loop displayed
enhanced DNA-binding abilities [43,44]. This enhance-
ment may be caused by an allosteric effect induced by
mutations on the DNA-binding site of Rad51, suggest-
ing that the L2 loop of Rad51 is not far from the
DNA-binding site. Consistent with this idea, the fluor-
escence of the tryptophan inserted in the place of
Rad51-Phe279 strongly decreased upon poly(dT)
binding, suggesting that this residue is close to the
DNA-binding site. The fluorescence change upon
DNA binding is not strong enough for a stacking

interaction of the residue with a DNA base, but is
large enough to indicate DNA binding in its proximity.
Several experimental methods, including photocros-
slinking, mutational analysis, and fluorescence meas-
urements, have been used to show that the RecA-L2
loop is involved in DNA binding [26,36,37,45]. Satura-
tion mutagenesis of the RecA-L2 loop revealed that
mutations in the L2 amino acid residues result in
recombination defects in vivo [46]. In addition, 20 resi-
due peptides that comprise the L2 loop region can bind
DNA by forming filamentous beta structures [47–49].
In the L2 peptide, an aromatic residue, which corres-
ponds to Phe203, was found to be absolutely required
for the DNA binding [47]. Therefore, the L2 loop
may be a functional DNA-binding site among the
RecA ⁄ Rad51 class of proteins, although its DNA-bind-
ing mode differs somewhat between RecA and Rad51.
Experimental procedures
Preparation of the human Rad51 mutants
The Rad51 mutant genes, inserted at the NdeI site of
the pET15b expression vector (Novagen, Darmstadt,
Germany), were constructed using a Quik-ChangeÒ kit
(Stratagene, La Jolla, CA, USA). The hexahistidine-tagged
HsRad51 and Rad51 mutants were expressed in the E. coli
JM109(DE3) strain, which also carries an expression vector
for the minor tRNAs (Codon(+)RILÒ, Novagen). The
proteins were purified on nickel-nitrilotriacetic acid (Ni-
NTA) agarose (Qiagen, Hilden, Germany). The hexahisti-
dine tag was then removed from the Rad51 portion with
thrombin protease (Amersham Biosciences, Piscataway, NJ,

USA). Then, the HsRad51 and the Rad51 mutants without
the hexahistidine tag were dialyzed against 100 mm
Tris ⁄ acetate buffer (pH 7.5), containing 7 mm spermidine
and 5% (v ⁄ v) glycerol. During this dialysis step, the
HsRad51 and the Rad51 mutants were precipitated (sper-
midine precipitation) [50], and the proteins were dissolved
in 100 mm potassium phosphate buffer (pH 7.0) containing
150 mm NaCl, 1 mm EDTA, 2 m m 2-mercaptoethanol, and
10% (v ⁄ v) glycerol. The HsRad51 and the Rad51 mutants
were further purified by chromatography on a MonoQ col-
umn (Amersham Biosciences). The purified HsRad51 and
Rad51 mutants were dialyzed against 20 mm Hepes ⁄ NaOH
Y. Matsuo et al. Roles of the HsRad51 L1 and L2 loops
FEBS Journal 273 (2006) 3148–3159 ª 2006 The Authors Journal compilation ª 2006 FEBS 3155
buffer (pH 7.5), containing 150 mm NaCl, 0.1 mm EDTA,
2mm 2-mercaptoethanol, and 10% (v ⁄ v) glycerol. Protein
concentrations were determined using the Bio-Rad
(Hercules, CA, USA) protein assay kit with bovine serum
albumin as the standard protein.
DNAs
The /X174 phage ssDNA and dsDNA used in the DNA
binding and ATPase assays were purchased from New
England Biolabs (Ipswich, MA, USA). Poly(dT) and
poly(dA):poly(dT) were obtained from Amersham Bio-
sciences. All of the DNA concentrations are expressed in
moles of nucleotides.
Assays for DNA binding
The /X174 circular ssDNA (40 lm) or the PstI-digested
linear /X174 dsDNA (20 lm) was mixed with the Rad51
protein or the Rad51 mutants in 10 lLof25mm Hepes

buffer (pH 7.5), containing 75 mm NaCl, 1 mm MgCl
2
,
0.1 mm EDTA, 1 mm 2-mercaptoethanol, 1 mm dithiothrei-
tol, 0.2 mgÆmL
)1
BSA, and 1 mm ATP. The reaction mix-
tures were incubated at 37 °C for 10 min, and were then
analyzed by 0.8% (w ⁄ v) agarose gel electrophoresis in
1· TAE buffer (40 mm Tris ⁄ acetate and 1 mm EDTA) at
3.3 VÆcm
)1
for 3 h. The bands were visualized by ethidium
bromide staining.
Assays for strand exchange
The /X174 circular ssDNA (40 lm) was incubated with the
Rad51 protein or the Rad51 mutants at 37 °C for 15 min,
in 10 lLof20mm potassium phosphate buffer (pH 7.4),
containing 50 mm NaCl, 1 mm dithiothreitol, 100 lgÆmL
)1
BSA, 1 mm MgCl
2
,2%(v⁄ v) glycerol, 1 mm ATP, 1 mm
CaCl
2
,2mm creatine phosphate, and 75 lgÆmL
)1
creatine
kinase. After this incubation, 2 lm RPA and 0.2 m KCl
were added to the reaction mixture, which was incubated at

37 °C for 15 min. Then, the reactions were initiated by the
addition of 20 lm /X174 linear dsDNA, and were contin-
ued for 1 h. The reactions were stopped by the addition of
0.5% (w ⁄ v) SDS, and 1.82 mgÆmL
)1
proteinase K (Roche
Applied Science, Basel, Switzerland), and were further incu-
bated at 37 °C for 15 min. After adding the six-fold loading
dye, the deproteinized reaction products were separated by
1% (w ⁄ v) agarose gel electrophoresis in 1· TAE buffer at
3.3 VÆcm
)1
for 2 h. The products were visualized by SYBR
gold (Invitrogen, Carlsbad, CA, USA) staining.
Gel filtration
Rad51 (150 lg) and Rad51 mutants (150 lg) were analyzed
by Superdex 200 HR 10 ⁄ 30 (Amersham Biosciences) gel
filtration chromatography. The elution buffer contained
20 mm Hepes ⁄ NaOH (pH 7.5), 150 m m NaCl, 0.1 mm
EDTA, 2 mm 2-mercaptoethanol, and 10% (v ⁄ v) glycerol,
and the flow rate was 0.5 mLÆmin
)1
.
CD measurements
The CD spectrum of a 0.25 mgÆmL
)1
solution of HsRad51
or the Rad51 mutants was measured on a JASCO J-820
spectropolarimeter (Japan Spectroscopic Co., Ltd, Tokyo,
Japan) using a 1 cm pathlength quartz cell. All of the CD

experiments were performed in 20 mm Hepes ⁄ NaOH buffer
(pH 7.5), containing 150 mm NaCl, 0.1 mm EDTA, 2 mm
2-mercaptoethanol, and 10% (v ⁄ v) glycerol.
Fluorescence measurements
Fluorescence was measured with an FP-6500 spectrofluo-
rometer (Japan Spectroscopic Co., Ltd), in 20 mm potas-
sium phosphate buffer (pH 7.4), containing 50 mm NaCl,
1mm dithiothreitol, 1 mm MgCl
2
, and 2% (v ⁄ v) glycerol,
in the presence or absence of 1 mm ATP. The emission
spectra were measured (bandwidth: 3 nm; response time:
0.5 s; scan rate: 100 nmÆmin
)1
)ina1· 1 cm quartz cell
with continuous stirring (300 r.p.m. per min), or in a
0.2 · 1 cm mini cell (Hellma, Mu
¨
llheim, Germany). The
excitation wavelength was 295 nm (bandwidth: 3 nm) for
selective excitation of the tryptophan residue. The spectra
were measured at least twice to verify the absence of signifi-
cant photobleaching, and were averaged to increase the sig-
nal to noise ratio. All of the spectra were corrected for the
Raman signal and background by subtracting the spectrum
of the buffer.
ATPase activity
Rad51 (5 lm) or a Rad51 mutant (5 lm) was incubated
with 1 mm ATP (Roche, ATP sodium salt) in 25 mm Hepes
buffer (pH 7.5), containing 75 mm NaCl, 1 mm MgCl

2
,
0.1 m m EDTA, 1 mm 2-mercaptoethanol, 1 mm dithiothrei-
tol, and 0.2 mgÆmL
)1
BSA, in the presence or absence of
ssDNA or dsDNA. In the ssDNA-dependent reaction, the
/X174 circular ssDNA (20 lm) was used as a substrate. In
the dsDNA-dependent reaction, the /X174 RF I DNA
(20 lm) (supercoiled dsDNA) was used as a substrate. In
the high salt conditions, the reaction mixture contained
1.58 m NaCl. The reaction was performed at 37 °C. After a
10 min preincubation in the absence of ATP, the reaction
was initiated by adding 1 mm ATP. Then, a 20 lL aliquot
of the reaction mixture was mixed with 30 lL of 100 mm
EDTA, to quench the reaction at the indicated time. The
amount of inorganic phosphate released was determined by
a colorimetric assay [51,52]. Briefly, 500 lL of a malachite
green solution [0.034% (w ⁄ v) malachite green oxalate,
Roles of the HsRad51 L1 and L2 loops Y. Matsuo et al.
3156 FEBS Journal 273 (2006) 3148–3159 ª 2006 The Authors Journal compilation ª 2006 FEBS
1.05% (w ⁄ v) hexaammonium heptamolybdate tetrahydrate,
and 0.1% (w ⁄ v) polyvinyl alcohol in 1 m HCl] was mixed
with 50 lL of sample solution (i.e., the reaction mixture
quenched with EDTA). After 1 min, 50 lL of 34% (w ⁄ v)
sodium citrate dihydrate was added to stop further color
development. The absorbance at 655 nm was measured
with a 96-well micro plate reader (Bio-Rad). A 1 mgÆmL
)1
phosphate ion standard solution (Wako Pure Chemicals,

Osaka, Japan) was used to prepare phosphate standards.
Acknowledgements
We thank Dr Chantal Prevost (CNRS-UPR) and Mr
Sebastien Conilleau for discussions, and Dr Takashi
Kinebuchi (RIKEN) for the CD measurements. The
fluorometer was kindly provided by Jasco Interna-
tional. This work was partly supported by a grant
from the Association pour la Recherche contre le Can-
cer (No 4813) to MT, and Grants-in-Aid from the Jap-
anese Society for the Promotion of Science (JSPS), and
the Ministry of Education, Sports, Culture, Science,
and Technology, Japan to HK. HK and IS were
supported by the Consolidated Research Institute for
Advanced Science and Medical Care, Waseda Univer-
sity.
References
1 Shinohara A, Ogawa H & Ogawa T (1992) Rad51 pro-
tein involved in repair and recombination in S. cerevi-
siae is a RecA-like protein. Cell 69, 457–470.
2 Shibata T, DasGupta C, Cunningham RP & Radding
CM (1979) Purified Escherichia coli recA protein cata-
lyzes homologous pairing of superhelical DNA and sin-
gle-stranded fragments. Proc Natl Acad Sci USA 76,
1638–1642.
3 McEntee K, Weinstock GM & Lehman IR (1979)
Initiation of general recombination catalyzed in vitro by
the recA protein of Escherichia coli. Proc Natl Acad Sci
USA 76, 2615–2619.
4 Cox MM & Lehman IR (1981) recA protein of Escheri-
chia coli promotes branch migration, a kinetically dis-

tinct phase of DNA strand exchange. Proc Natl Acad
Sci USA 78, 3433–3437.
5 Cox MM & Lehman IR (1981) Directionality and
polarity in recA protein-promoted branch migration.
Proc Natl Acad Sci USA 78, 6018–6022.
6 Bishop DK, Park D, Xu L & Kleckner N (1992)
DMC1: a meiosis-specific yeast homolog of E. coli recA
required for recombination, synaptonemal complex for-
mation, and cell cycle progression. Cell 69, 439–456.
7 Pittman DL, Cobb J, Schimenti KJ, Wilson LA, Cooper
DM, Brignull E, Handel MA & Schimenti JC (1998)
Meiotic prophase arrest with failure of chromosome
synapsis in mice deficient for Dmc1, a germline-specific
RecA homolog. Mol Cell 1, 697–705.
8 Yoshida K, Kondoh G, Matsuda Y, Habu T, Nishi-
mune Y & Morita T (1998) The mouse RecA-like gene
Dmc1 is required for homologous chromosome synapsis
during meiosis. Mol Cell 1, 707–718.
9 Lim DS & Hasty P (1996) A mutation in mouse rad51
results in an early embryonic lethal that is suppressed
by a mutation in p53. Mol Cell Biol 16, 7133–7143.
10 Tsuzuki T, Fujii Y, Sakumi K, Tominaga Y, Nakao K,
Sekiguchi M, Matsushiro A, Yoshimura Y & Morita T
(1996) Targeted disruption of the Rad51 gene leads to
lethality in embryonic mice. Proc Natl Acad Sci USA
93, 6236–6240.
11 Sonoda E, Sasaki MS, Buerstedde JM, Bezzubova O,
Shinohara A, Ogawa H, Takata M, Yamaguchi-Iwai Y
& Takeda S (1998) Rad51-deficient vertebrate cells accu-
mulate chromosomal breaks prior to cell death. EMBO

J 17, 598–608.
12 Sung P (1994) Catalysis of ATP-dependent homologous
DNA pairing and strand exchange by yeast RAD51
protein. Science 265, 1241–1243.
13 Baumann P, Benson FE & West SC (1996) Human
Rad51 protein promotes ATP-dependent homologous
pairing and strand transfer reactions in vitro. Cell 87,
757–766.
14 Maeshima K, Morimatsu K & Horii T (1996) Purifica-
tion and characterization of XRad51.1 protein, Xenopus
RAD51 homologue: recombinant XRad51.1 promotes
strand exchange reaction. Genes Cells 1, 1057–1068.
15 Gupta RC, Bazemore LR, Golub EI & Radding CM
(1997) Activities of human recombination protein
Rad51. Proc Natl Acad Sci USA 94, 463–468.
16 Takahashi M & Norden B (1994) Structure of RecA-
DNA complex and mechanism of DNA strand
exchange reaction in homologous recombination. Adv
Biophys 30, 1–35.
17 Story RM, Weber IT & Steitz TA (1992) The structure
of the E. coli recA protein monomer and polymer.
Nature 355, 318–325.
18 Shin DS, Pellegrini L, Daniels DS, Yelent B, Craig L,
Bates DYuDS, Shivji MK, Hitomi C, Arvai AS, Volk-
mann N, Tsuruta H, Blundell TL, Venkitaraman AR &
Tainer JA (2003) Full-length archaeal Rad51 structure
and mutants: mechanisms for RAD51 assembly and
control by BRCA2. EMBO J 22, 4566–4576.
19 Wu Y, He Y, Moya IA, Qian X & Luo Y (2004) Crys-
tal structure of archaeal recombinase RADA: a snap-

shot of its extended conformation. Mol Cell 15, 423–
435.
20 Conway AB, Lynch TW, Zhang Y, Fortin GS, Fung
CW, Symington LS & Rice PA (2004) Crystal structure
of a Rad51 filament. Nat Struct Mol Biol 11, 791–796.
21 Pellegrini L, Yu DS, Lo T, Anand S, Lee M, Blundell
TL & Venkitaraman AR (2002) Insights into DNA
Y. Matsuo et al. Roles of the HsRad51 L1 and L2 loops
FEBS Journal 273 (2006) 3148–3159 ª 2006 The Authors Journal compilation ª 2006 FEBS 3157
recombination from the structure of a RAD51-BRCA2
complex. Nature 420, 287–293.
22 Kinebuchi T, Kagawa W, Enomoto R, Tanaka K,
Miyagawa K, Shibata T, Kurumizaka H & Yokoyama S
(2004) Structural basis for octameric ring formation and
DNA interaction of the human homologous-pairing
protein Dmc1. Mol Cell 14, 363–374.
23 Raghunathan S, Kozlov AG, Lohman TM & Waksman
G (2000) Structure of the DNA binding domain of E.
coli SSB bound to ssDNA. Nat Struct Biol 7, 648–652.
24 Bochkarev A, Pfuetzner RA, Edwards AM & Frappier
L (1997) Structure of the single-stranded-DNA-binding
domain of replication protein A bound to DNA. Nature
385, 176–181.
25 Ellouze C, Kim HK, Maeshim K, Tuite E, Morimatsu
K, Horii T, Mortensen K, Norden B & Takahashi M
(1997) Nucleotide cofactor-dependent structural change
of Xenopus laevis Rad51 protein filament detected by
small-angle neutron scattering measurements in solu-
tion. Biochemistry 36, 13524–13529.
26 Maraboeuf F, Voloshin O, Camerini-Otero RD &

Takahashi M (1995) The central aromatic residue in
loop L2 of RecA interacts with DNA: quenching of
the fluorescence of a tryptophan reporter inserted in
L2 upon binding to DNA. J Biol Chem 270, 30927–
30932.
27 Maeshima K, Maraboeuf F, Morimatsu K, Horii T &
Takahashi M (1998) Nucleotide dependent structural
and kinetic changes in Xenopus rad51.1-DNA complex
stimulating the strand exchange reaction: destacking of
DNA bases and restriction of their local motion. J Mol
Biol 284, 689–697.
28 Selmane T, Camadro JM, Conilleau S, Fleury F, Tran
V, Prevost C & Takahashi M (2004) Identification of
the subunit–subunit interface of Xenopus Rad51.1 pro-
tein: similarity to RecA. J Mol Biol 335 , 895–904.
29 Bishop AJ & Schiestl RH (2001) Homologous recombi-
nation as a mechanism of carcinogenesis. Biochim Bio-
phys Acta 1471, M109–M121.
30 Pierce AJ, Stark JM, Araujo FD, Moynahan ME, Ber-
wick M & Jasin M (2001) Double-strand breaks and
tumorigenesis. Trends Cell Biol 11, S52–S59.
31 Henning W & Sturzbecher HW (2003) Homologous
recombination and cell cycle checkpoints: Rad51 in
tumour progression and therapy resistance. Toxicology
193, 91–109.
32 Richardson C, Stark JM, Ommundsen M & Jasin M
(2004) Rad51 overexpression promotes alternative dou-
ble-strand break repair pathways and genome instabil-
ity. Oncogene 23, 546–553.
33 Ohnishi T, Taki T, Hiraga S, Arita N & Morita T

(1998) In vitro and in vivo potentiation of radiosensitiv-
ity of malignant gliomas by antisense inhibition of the
RAD51 gene. Biochem Biophys Res Commun 245, 319–
324.
34 Mirshad JK & Kowalczykowski SC (2003) Biochemical
characterization of a mutant RecA protein altered in
DNA-binding loop 1. Biochemistry 42, 5945–5954.
35 Nastri HG & Knight KL (1994) Identification of resi-
dues in the L1 region of the RecA protein which are
important to recombination or coprotease activities.
J Biol Chem 269, 26311–26322.
36 Malkov VA & Camerini-Otero RD (1995) Photocross-
links between single-stranded DNA and Escherichia coli
RecA protein map to loops L1 (amino acid residues
157–164) and L2 (amino acid residues 195–209). J Biol
Chem 270, 30230–30233.
37 Wang Y & Adzuma K (1996) Differential proximity
probing of two DNA binding sites in the Escherichia
coli recA protein using photo-cross-linking methods.
Biochemistry 35, 3563–3571.
38 Passy SI, Yu X, Li Z, Radding CM, Masson JY, West
SC & Egelman EH (1999) Human Dmc1 protein binds
DNA as an octameric ring. Proc Natl Acad Sci USA 96,
10684–10688.
39 Masson JY, Davies AA, Hajibagheri N, Van Dyck E,
Benson FE, Stasiak AZ, Stasiak A & West SC (1999)
The meiosis-specific recombinase hDmc1 forms ring
structures and interacts with hRad51. EMBO J 18,
6552–6560.
40 Sehorn MG, Sigurdsson S, Bussen W, Unger VM &

Sung P (2004) Human meiotic recombinase Dmc1 pro-
motes ATP-dependent homologous DNA strand
exchange. Nature 429, 433–437.
41 Kim HK, Morimatsu K, Norden B, Ardhammar M &
Takahashi M (2002) ADP stabilises the human Rad51-
single stranded DNA complex and promotes its DNA
annealing activity. Genes Cells 7, 1125–1134.
42 Takahashi M & Norden B (1993) Accessibility to modi-
fication of histidine residues of RecA protein upon
DNA and cofactor binding. Eur J Biochem 217, 665–
670.
43 Kurumizaka H, Aihara H, Kagawa W, Shibata T &
Yokoyama S (1999) Human Rad51 amino acid resi-
dues required for Rad52 binding. J Mol Biol 291, 537–
548.
44 Fortin GS & Symington LS (2002) Mutations in yeast
Rad51 that partially bypass the requirement for Rad55
and Rad57 in DNA repair by increasing the stability of
Rad51-DNA complexes. EMBO J 21, 3160–3170.
45 Morimatsu K & Horii T (1995) DNA-binding surface
of RecA protein photochemical cross-linking of the first
DNA binding site on RecA filament. Eur J Biochem
234, 695–705.
46 Hortnagel K, Voloshin ON, Kinal HH, Ma N, Schaf-
fer-Judge C & Camerini-Otero RD (1999) Saturation
mutagenesis of the E. coli RecA loop L2 homologous
DNA pairing region reveals residues essential for recom-
bination and recombinational repair. J Mol Biol 286,
1097–1106.
Roles of the HsRad51 L1 and L2 loops Y. Matsuo et al.

3158 FEBS Journal 273 (2006) 3148–3159 ª 2006 The Authors Journal compilation ª 2006 FEBS
47 Voloshin ON, Wang L & Camerini-Otero RD (1996)
Homologous DNA pairing promoted by a 20-amino
acid peptide derived from RecA. Science 272, 868–
872.
48 Wang L, Voloshin ON, Stasiak A & Camerini-Otero
RD (1998) Homologous DNA pairing domain peptides
of RecA protein: intrinsic propensity to form beta-struc-
tures and filaments. J Mol Biol 277, 1–11.
49 Selmane T, Wittung-Stafshede P, Maraboeuf F, Volo-
shin ON, Norden B, Camerini-Otero DR & Takahashi
M (1999) The L2 loop peptide of RecA stiffens and
restricts base motions of single-stranded DNA similar
to the intact protein. FEBS Lett 446, 30–34.
50 Baumann P, Benson FE, Hajibagheri N & West SC
(1997) Purification of human Rad51 protein by selective
spermidine precipitation. Mutat Res 384, 65–72.
51 Lanzetta PA, Alvarez LJ, Reinach PS & Candia OA
(1979) An improved assay for nanomole amounts of
inorganic phosphate. Anal Biochem 100, 95–97.
52 Chang KM, Delfert D & Junger KD (1986) A Direct
Colorimetric Assay for Ca
2+
-Stimulated ATPase Activ-
ity. Anal Biochem 157, 375–380.
Y. Matsuo et al. Roles of the HsRad51 L1 and L2 loops
FEBS Journal 273 (2006) 3148–3159 ª 2006 The Authors Journal compilation ª 2006 FEBS 3159