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: kurumizaka@waseda.jp (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 [...]... 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, Schaffer-Judge C & Camerini-Otero... Voloshin ON, Stasiak A & Camerini-Otero RD (1998) Homologous DNA pairing domain peptides of RecA protein: intrinsic propensity to form beta-structures and filaments J Mol Biol 277, 1–11 49 Selmane T, Wittung-Stafshede P, Maraboeuf F, Voloshin ON, Norden B, Camerini-Otero DR & Takahashi M (1999) The L2 loop peptide of RecA stiffens and Roles of the HsRad51 L1 and L2 loops restricts base motions of single-stranded... cofactor-dependent structural change of Xenopus laevis Rad51 protein filament detected by small-angle neutron scattering measurements in solution Biochemistry 36, 13524–13529 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–... characterization of a mutant RecA protein altered in DNA- binding loop 1 Biochemistry 42, 5945–5954 35 Nastri HG & Knight KL (1994) Identification of residues 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) Photocrosslinks 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.. .Roles of the HsRad51 L1 and L2 loops 22 23 24 25 26 27 28 29 30 31 32 33 Y Matsuo et al recombination from the structure of a RAD51- BRCA2 complex Nature 420, 287–293 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... (2000) Structure of the DNA binding domain of E coli SSB bound to ssDNA Nat Struct Biol 7, 648–652 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 Ellouze C, Kim HK, Maeshim K, Tuite E, Morimatsu K, Horii T, Mortensen K, Norden B & Takahashi M (1997) Nucleotide cofactor-dependent... (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 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 protein: similarity to RecA J Mol Biol... promotes its DNA annealing activity Genes Cells 7, 1125–1134 42 Takahashi M & Norden B (1993) Accessibility to modification 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 residues required for Rad52 binding J Mol Biol 291, 537– 548 44 Fortin GS & Symington LS (2002)... 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 promotes 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 Rad51single stranded DNA complex and . actually involved in DNA binding. Because aromatic residues are involved in the ssDNA binding by bacterial single stranded DNA- binding protein (SSB) and human. 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