Báo cáo khoa học: A DNA-binding surface of SPO11-1, an Arabidopsis SPO11 orthologue required for normal meiosis docx

A DNA-binding surface of SPO11-1, an Arabidopsis SPO11 orthologue required for normal meiosis Yoshinori Shingu 1 , Tsutomu Mikawa 2 , Mariko Onuma 1 , Takashi Hirayama 3 and Takehiko Shibata 1 1 Cellular & Molecular Biology Laboratory, RIKEN Advanced Science Institute, 2-1 Hirosawa, Wako-shi, Saitama, Japan 2 Biometal Science Laboratory, RIKEN Spring-8 Center, Mikazuki cho, Hyogo, Japan 3 Molecular Membrane Biology Laboratory, RIKEN Advanced Science Institute, 2-1 Hirosawa, Wako-shi, Saitama, Japan Introduction In eukaryotes, homologous recombination contributes to genetic diversity, assists with the meiotic segregation of homologous chromosomes and promotes the repair of double-strand breaks to maintain genome stability. Numerous studies have used Saccharomyces cerevisiae to demonstrate that meiotic recombination is initiated by DNA double-stranded breaks [1], which are intro- duced by the SPO11 protein [2,3]. The SPO11 gene was first isolated from S. cerevisiae [4]. SPO11 is found in virtually all eukaryotes and shares homology with the A-subunit (i.e. Top6A) of topoisomerase VI in the archaeon, Sulfolobus shibatae [3]. SPO11 orthologues Keywords AtSPO11-1; homology modelling; meiotic recombination; topoisomerase VI; transgenic Arabidopsis Correspondence T. Shibata, Cellular & Molecular Biology Laboratory, RIKEN Advanced Science Institute, 2-1 Hirosawa, Wako-shi, Saitama 351-0198, Japan Fax: +81 48 462 1227 Tel: +81 48 467 9528 E-mail: tshibata@postman.riken.go.jp (Received 28 November 2009, revised 9 March 2010, accepted 15 March 2010) doi:10.1111/j.1742-4658.2010.07651.x Meiotic recombination is initiated by DNA double-stranded breaks intro- duced by the SPO11 protein. Despite a decade of research, the biochemical functions of SPO11 remain largely unknown, perhaps because of difficulties in studying the functionally active SPO11. Arabidopsis thaliana encodes three SPO11-related proteins, two of which (SPO11-1 and SPO11-2) are required for, and cooperate in, meiosis. We isolated soluble SPO11-1, fused with or free of a trigger factor-tag at its N terminus. The tag-free SPO11-1 needed to interact physically with soluble SPO11-1 to maintain its solubil- ity, suggesting a multimeric active form including a solubilizing protein cofactor. An N-terminal fragment of PRD1, a SPO11-1-interacting protein required for normal meiosis, but not SPO11-2, forms a soluble complex with trigger factor-tagged SPO11-1, but the trigger factor-tag was required for the solubility. Formation of the complex is not sufficient to express endonuclease activity. Trigger factor-tagged SPO11-1 exhibited DNA-bind- ing activities: Glu substitutions of the invariant Gly215 and Arg222 and of the nonconserved Arg223 and Arg226 in a conserved motif (G215E, R222E, R223E, R226E) reduced the DNA-binding ability in vitro, but sub- stitutions of the conserved Arg130 and invariant Tyr103 (a residue in the putative endonuclease-active center) and of Arg residues outside conserved motifs by Glu or Phe (R130E, Y103F, R207E and R254E), did not. Tests for the ability of mutant spo11-1 proteins to complement the silique-defec- tive phenotype of a spo11-1-homozygous mutant in vivo revealed that R222E and G215E induced serious deficiencies, while R130E caused a partial defect in silique formation. Thus, the Gly215, Arg222 and Arg223 residues of SPO11-1 form a DNA-binding surface that is functional in meiosis. Abbreviations CV1, co-expression verctor 1; CV2, co-expression vector 2; EV-S1, expression vector of Spo11-1; IPTG, isopropyl thio-b- D-galactoside; PRD1N, N-terminal domain of PRD1; TF, trigger factor; TF-SPO11, SPO11 fusion protein with a TF tag attached to its N terminus. 2360 FEBS Journal 277 (2010) 2360–2374 ª 2010 The Authors Journal compilation ª 2010 FEBS responsible for the initiation of meiotic recombination or recombination-related events have been identified in Schizosaccharomyces pombe (i.e. the Rec12 protein) [2,5], Drosophila melanogaster (i.e. the mei-W68 protein) [6], Coprinus cinereus [7], mouse [8–13] and Arabidopsis thaliana [14,15]. Yeast, flies, nematodes and mammals encode a single SPO11; however, plants (e.g. Arabidopsis and rice Oryza sativa) encode at least three SPO11 paralogues [16,17]. In Arabidopsis, SPO11-1 and SPO11-2 are considered to be the ‘true’ SPO11 proteins that function in the initiation of mei- otic recombination, as demonstrated by the sterile and reduced meiotic recombination phenotypes of spo11-1 and spo11-2 homozygous mutants, respectively [14,15]. SPO11-3, the third Arabidopsis SPO11 paralogue, is highly expressed in somatic cells and contributes to somatic development (i.e. endoreduplication) by forming a complex with Top6B, a homologue of the B-subunit of topoisomerase VI [18–20]. SPO11-2 and SPO11-3 interacted with Top6B in a yeast two-hybrid system, while SPO11-1 did not [16]. These findings imply that SPO11-1 and SPO11-2 perform different functions. In S. cerevisiae, meiotic double-strand break forma- tion to initiate meiotic recombination requires the products of at least nine genes (i.e. REC102, SKI8, REC104, REC114, MEI4, MER2, MRE11, RAD50 and XRS2) in addition to that of the SPO11 gene (see refs. 21,22 for review). Among these, only three homologous genes (SKI8, MRE11 and RAD50) have been identified; however, their corresponding proteins are not required for meiotic double-strand break for- mation in Arabidopsis (23,24, see ref. 25 for review). By contrast, in Arabidopsis, the PRD1 protein, which is unique to plants, is required for meiotic double- strand break formation [26]. These observations and the presence of two or more SPO11 paralogues suggest that plants and yeast differ in their regulation of mei- otic double-strand break formation. In Arabidopsis, a genetic study revealed a genetic interaction between SPO11-1 and SPO11-2 [15], while another study showed that spo11-1 and spo11-2 were epistatic with respect to meiotic functions and that a spo11-1 or a spo11-2 mutation abolished meiotic double-strand cleavage [27]. These studies suggested that SPO11-1 and SPO11-2 function as members of the same com- plex for the double-strand cleavage. Although various studies have concluded that SPO11 plays important roles in meiotic double-strand break formation and the initiation of meiotic recombi- nation, there is little biochemical evidence of the role of this protein in meiotic double-strand break forma- tion. The lack of biochemical data probably stems from the difficulty in isolating the active form of SPO11 by a method avoiding a renaturation process or a refolding process. Only one study has described the purification of SPO11 from S. pombe (i.e. Rec12) by refolding a denatured protein that had been expressed in Escherichia coli [28]. A number of mutations have been reported in SPO11, and certain amino acid residues have been suggested to play a role in the biochemical functions of SPO11 [3,29,30]. However, for the same reason mentioned above, the biochemical defects caused by these mutations have not yet been determined. Func- tional characterization of a purified and active form of SPO11 and the complexes of SPO11 with regulatory proteins would greatly enhance our understanding of the molecular functions and regulation of SPO11 pro- teins. This would be especially helpful in plants, because genetic studies in plants are more time-con- suming than those in yeast. Here, we developed a method for purifying the active form of SPO11-1, and a complex of SPO11-1 and the N-terminal domain of PRD1, from E. coli cells without the need for a rena- turation process, and studied a DNA-binding surface that contributes to the meiotic function of SPO11-1 in vivo. We found that the meiotic defects in vivo caused by single amino acid-substitution mutations in SPO11-1 were correlated with defects in the in vitro DNA-binding activities of the mutant spo11-1 proteins (hereafter denoted as spo11-1 to discriminate them from wild-type SPO11-1). Using this information, we identified a DNA-binding surface on SPO11-1. To our knowledge, this is the first report of the characteriza- tion of a soluble and active form of SPO11 in vitro,in a decade since the discovery of the function of SPO11 in the meiotic double-stranded breakage [2,3], and of the correlation between the in vivo defects caused by mutations in SPO11 and the in vitro biochemical activ- ities of SPO11 mutant proteins purified in an active form. Results Maintenance of soluble forms of SPO11-1 and a SPO11-1-associated protein, PRD1 After making various attempts to express and isolate the Arabidopsis SPO11-1 protein in a soluble form, we tried to use the trigger factor (TF), which reportedly helps with the folding of co-expressed proteins [31,32]. Thus, we constructed the co-expression vector 1 (CV1)–SPO11-1 (Fig. 1A). Protein expression in trans- formed E. coli BL21 cells was induced by exposure to cold shock and isopropyl thio-b-d-galactoside (IPTG), Y. Shingu et al. DNA-binding surface of Arabidopsis SPO11 protein FEBS Journal 277 (2010) 2360–2374 ª 2010 The Authors Journal compilation ª 2010 FEBS 2361 A B C Fig. 1. Expression of soluble TF-tagged SPO11-1 and co-expression of tag-free SPO11-1 and related proteins. (A) Expres- sion vectors. TF-SPO11-1 and free TF were expressed under the control of the cold shock promoter (CsP), while TF-tag-free SPO11-1, SPO11-2 and PRD1N were under the control of the T7 promoter (T7P). Note that the His-tag of the TF-SPO11-1 on CV2-PR (PRDIN) was removed in order to avoid detection of TF-SPO11, because the band of PRD1N (92 kDa) and that of TF-SPO11-1 (94 kDa) would overlap. (B) Immunoblotting analyses of TF-tag-free proteins co-expressed with free TF or TF-SPO11-1. Proteins were detected with an anti-His-tag IgG. All proteins have a His-tag at the N terminus unless otherwise noted (see Fig. 1A). Cell-free lysates obtained from cells expressing each expres- sion vector were separated by centrifugation into supernatant fractions (sup) and precipi- tates (ppt). The samples were then subjected to PAGE under denaturing condi- tions. Arrows indicate TF-SPO11-1 (94 kDa), TF (53 kDa), SPO11-1 (44 kDa), SPO11-2 (45 kDa), TF-SPO11-2 (95 kDa) and PRD1N (92 kDa). (C) Quantification of the results shown in panel B. The percentages of solu- ble tag-free SPO11-1, SPO11-2 and PRD1N were calculated by densitometric measure- ments of the photographs shown in Fig. 1B. Each bar represents the mean value ± stan- dard deviation from three independent experiments. DNA-binding surface of Arabidopsis SPO11 protein Y. Shingu et al. 2362 FEBS Journal 277 (2010) 2360–2374 ª 2010 The Authors Journal compilation ª 2010 FEBS and the cell lysates were fractionated by centrifugation into supernatants (soluble form) and precipitates (insoluble form). Co-expression with TF greatly improved the recovery of the expressed SPO11-1 in the soluble form (i.e. under these conditions, 43% of the total expressed SPO11-1 was obtained in a soluble form, but the majority of SPO11-1 remained in the precipitates) (Fig. 1B, lane 1 versus lane 2, and Fig. 1C). Next, we tried to express SPO11-1 as a fusion pro- tein with a TF tag attached at the N terminus (i.e. TF-SPO11-1) by the use of the expression vector of SPO11-1, EV-S1 (Fig. 1A), and found that fusion to TF greatly improved the recovery of SPO11-1 in a sol- uble fraction (Fig. 1B,C). Then, we successfully estab- lished a method to isolate an active and soluble form of TF-SPO11-1 in a sufficient amount for biochemical studies, without the need for a refolding step during the purification process, as described in the Materials and methods. The heparin column facilitated the separation of TF-SPO11-1 (94 kDa) from the partially degraded TF-SPO11-1 and free TF, which were pro- duced, along with the intact fusion protein, in E. coli cells (Fig. 2A). Most of the purified TF-SPO11-1 was retained by the Superdex 200 column and eluted as multimers with heterogeneous sizes larger than dimers (Fig. 2C). We obtained  1 mg of purified TF-SPO11-1 from a 2 L culture. To remove the TF tag after purification, a throm- bin-digestion site was inserted between TF and SPO11- 1 in the fusion protein (EV-S1; Fig. 1A). After purifi- cation, the 94 kDa protein was treated with thrombin, and denatured samples were separated by SDS ⁄ PAGE. We detected a fragment of  55 kDa and a weak sig- nal from a protein of  40 kDa (Fig. 2B), while the expected sizes of free TF and free SP11-1 were 53 and 42 kDa, respectively. The 40 kDa fragment was identi- fied, by N-terminal sequence analyses, as SPO11-1, and this result further confirmed that the 94 kDa protein was TF-SPO11-1. While the TF-tag-free A B C Fig. 2. Purification of TF-tagged SPO11-1 and co-expression of tag-free SPO11-1. (A) Purification of TF-SPO11-1. The samples from each of the purification steps were analyzed by PAGE under denaturing conditions. Lanes from left to right: molecular mass markers (M), supernatant (sup) and insoluble fractions (ppt) of the cell-free lysates from Escherichia coli cells expressing TF-SPO11-1, the fraction from a TALON column (TALON) and the fraction from a heparin column (Heparin). (B) Polyacrylamide gel electrophoretic analysis of thrombin-treated TF-SPO11-1 under denaturing conditions. TF-SPO11-1 treated with thrombin was centrifuged to obtain a supernatant (sup) and a precipitate (ppt). Arrows indicate TF-SPO11-1, free TF and TF-tag-free SPO11-1, from top to bottom. (C) The gel-filtration profiles of purified TF-SPO11-1 alone and the complex of TF-SPO11-1 and SPO11-1. The gray and black lines indicate absorbance at 280 nm in the gel filtration of TF-SPO11-1 and that of the complex of TF-SPO11-1 and SPO11-1, respectively. Y. Shingu et al. DNA-binding surface of Arabidopsis SPO11 protein FEBS Journal 277 (2010) 2360–2374 ª 2010 The Authors Journal compilation ª 2010 FEBS 2363 SPO11-1 was obtained in a soluble form through partial digestion with thrombin (Fig. 2B; 3 h thrombin cleavage), all of the tag-free SPO11-1 became insoluble upon complete digestion (Fig. 2B; overnight thrombin cleavage). A possible explanation for the results described above is that TF-tag-free SPO11 is maintained as a soluble form by direct interactions with a soluble form of SPO11-1 (i.e. TF-SPO11-1). To test this possibility, we modified CV1 to co-express TF-SPO11-1 and the TF-tag-free SPO11-1 (CV2-S1; Fig. 1A), and the expression experiments revealed that the majority of the TF-tag-free SPO11-1 was in a soluble form (i.e. 79% of the total protein; Fig. 1B, lane 3 versus lane 4, and Fig. 1C). Moreover, Superdex 200 gel filtration of the cell-free extracts prepared from cells co-expressing TF-SPO11-1 and TF-tag-free SPO11-1 on CV2-S1 showed that almost all of the TF-SPO11-1 and TF-tag-free SPO11-1 were in a complex in the void volume fraction (Fig. 2C). These results, and those following solubilization of the N-terminal domain of PRD1 (PRD1N) co-expressed with TF-SPO11-1, support the explanation about the solubilization by a direct interaction with a soluble form. Spo11-2 and PRD1 co-operate with SPO11-1 in mei- osis in Arabidopsis (see Introduction). A yeast two- hybrid analysis showed that PRD1 physically interacts with SPO11-1 at its N-terminal domain (802 residues: ref. 26), while it is not clear whether SPO11-2 interacts directly with SPO11-1 [15,27]. To test for the possible direct interactions of these two proteins with SPO11-1 in vitro, we co-expressed SPO11-2 or PRD1N with TF-SPO11-1 or TF on CV2 and CV1, respectively, considering that these proteins were mostly expressed in an insoluble form in E. coli cells (Fig. 1A). We have not solved the problem with the expression of the full- length PRD1 in E. coli cells and thus were not able to study the full-length PRD1. As expected by the known interactions that would maintain its own solubility, PRD1N showed a significant improvement in the recovery (91%) as a soluble form when co-expressed with TF-SPO11-1 on CV2 (Fig. 1B,C). By contrast, the solubility of Spo11-2 was not improved by the co-expression of TF-SPO11-1 on CV2 (Fig. 1B,C), while TF-tagged SPO11-2 was soluble (Fig. 1B). These results suggest that SPO11-2 does not directly interact with SPO11-1. DNA-binding activities of TF-SPO11-1 As described above, TF-tag-free SPO11 is not soluble without the associated TF-SPO11-1 and therefore we mainly used TF-SPO11-1 in the subsequent biochemical analyses. We then assessed the DNA-bind- ing activities of SPO11-1 using a native gel-mobility assay. Negatively supercoiled pUC18 closed-circular dsDNA was incubated with TF-SPO11-1 for 5 min at 37 °C and then analyzed by agarose-gel electrophore- sis. A slower mobility shift was observed in the pUC18 dsDNA as the amount of protein was increased, while TF alone caused only a slight change in the DNA mobility (Fig. 3A). DNA-binding analyses using M13mp18 ssDNA indicated that TF-SPO11-1 bound to the ssDNA with a slightly lower affinity (approxi- mately twofold) than to dsDNA, as judged by the amounts of proteins giving the same extents of band- shift (Fig. 3B). These results indicate that the soluble form of SPO11-1 is active in DNA binding. DNA binding was indicated by a broadened band that gradually decreased in mobility; thus, we had to assess the DNA binding by monitoring the decreases in the mobility and amount of DNA in the bands. This made it difficult to quantify the DNA-binding ability of mutant SPO11-1. When we tested the 180 bp dsDNA as a substrate for DNA binding, we found that the free 180 bp dsDNA and the Spo11-1-bound DNA formed discrete bands (Fig. 3C), and this allowed us to quantify the DNA-binding ability clearly (see Fig. 5B below). Again, TF alone did not bind to the 180 bp dsDNA molecule (Fig. 3C). We conclude, from these results, and from those described above, that the SPO11 domain of TF-SPO11-1 binds to dsDNA. The binding of TF-SPO11-1 to DNA remained sta- ble in the presence of up to 150 mm NaCl, as little free DNA was detected (Fig. 3D). When the NaCl concen- tration was increased above 150 mm, another signal for DNA binding, appearing between the fully bound form and the free DNA, became more significant. In the presence of 600 mm NaCl, about half of the DNA still remained in the bound forms, and the amount of the second signal for the DNA binding was compara- ble to that of the fully bound form (Fig. 3D). The sec- ond signal for the DNA binding was weak at a low concentration of SPO11-1, even without a high concen- tration of NaCl (Fig. 3C; see Fig. 5A). This result shows that there are two distinct forms of the com- plexes of DNA and SPO11-1. We did not detect DNase or topoisomerase activities on the dsDNAs of pUC18, pBR322 or E. coli phage lambda, or on the dsDNA containing a topoisomerase target sequence (Pryv, TopoGen), a yeast hypersensi- tive sequence to yeast SPO11 [33] or an Arabidopsis hot-spot for meiotic recombination [34], when these DNA species were incubated with TF-SPO11-1 under these and modified conditions (Fig. S1). We did not detect any activity for covalent attachment to the DNA-binding surface of Arabidopsis SPO11 protein Y. Shingu et al. 2364 FEBS Journal 277 (2010) 2360–2374 ª 2010 The Authors Journal compilation ª 2010 FEBS DNA terminus (Figs 3C and S1; with SDS treatment). These results suggest that either the endonucleolytic activity of SPO11-1 is not expressed by itself or the attachment of the TF-tag at the N terminus prevents the expression of the activity. Genetic complementa- tion tests using the cDNA encoding TF-SPO11-1 were unsuccessful, but this does not mean that all biochemi- cal functions of SPO11-1 are inactivated by the attach- ment of a TF-tag. Amino acid residues required for DNA binding by SPO11-1 in vitro We tried to identify a surface on SPO11-1, involved in DNA binding, by analyzing the effects of a series of single amino acid replacements. SPO11-1 shares  30% amino acid sequence homology with Metha- nococcus jannaschii Top6A, suggesting structural simi- larity between these proteins [35]. Therefore, we performed homology modelling of SPO11-1, based on the crystal structure of Top6A [36], to consider candidates of basic amino acid residues composing a DNA-binding surface. The modelled SPO11-1 struc- ture without the 70 N-terminal residues is shown in Fig. 4A, in which the basic amino acid residues are shown in blue. Gly202 is reportedly essential for DNA binding by S. pombe Spo11 (Rec12); however, no biochemical data have been described to support this observation (see ref. 30). Gly215 (i.e. equivalent to the Gly202 residue of S. pombe Rec12) was found in the AB CD Fig. 3. DNA-binding activities of TF-SPO11-1. (A) Binding to pUC18 negatively supercoiled closed-circular dsDNA. pUC18 closed-circular dsDNA (7.5 n M) was incubated with free TF or with TF-SPO11-1 at 37 °C for 5 min. The products were separated by electrophoresis at 8.3 V ⁄ cm for 1 h, at room temperature through a 0.8% agarose gel, then stained with ethidium bromide. Lane M contains a smart ladder DNA marker (Nippon Gene). The amounts of proteins are 0, 0.24, 0.47 and 0.71 l M (from left to right). (B) Binding to ssDNA. M13mp18 ssDNA (2.2 n M) was incubated with free TF or with TF-SPO11-1 at 37 °C for 5 min. The products were analyzed as described above, for panel A. The amounts of proteins are 0, 0.24, 0.47 and 0.71 l M (from left to right). (C) Binding to the 180 bp dsDNA. The dsDNA (9 nM) was incubated with free TF or TF-SPO11-1 at 37 °C for 5 min. Lane S indicates the addition of SDS (final concentration 1%) after the binding reaction with 0.71 l M TF-SPO11-1. The products were separated by electrophoresis through a 0.8% agarose gel at 8.3 V ⁄ cm for 1 h, at room temperature, and were detected by Southern hybridization using the 180 bp [ 33 P]DNA probe. The amounts of proteins are 0, 0.24, 0.47 and 0.71 l M (from left to right). (D) The NaCl sensitivity of the DNA binding of TF-SPO11-1. The 180 bp dsDNA (9 nM) was incubated in the presence of NaCl with 0.47 l M TF-SPO11-1 at 37 °C for 5 min. The NaCl concentrations were 0, 150, 300, 450 and 600 mM (from left to right). After the agarose gel electrophoresis, the products were detected by Southern hybridization, as described above for panel C. 2nd protein-bound DNA is a minor species of protein-bound DNA. Y. Shingu et al. DNA-binding surface of Arabidopsis SPO11 protein FEBS Journal 277 (2010) 2360–2374 ª 2010 The Authors Journal compilation ª 2010 FEBS 2365 vicinity of a basic surface (Fig. 4A). Proteins in the SPO11 family contain five conserved amino acid sequence motifs (i.e. motifs I to V; Fig. 4B). Consider- ing these two series of structural features, we con- structed a series of DNA fragments that each encoded a mutant spo11-1 with a single amino acid substitu- tion, wherein a conserved amino acid residue was replaced with an acidic residue. These mutations included R130E (where the first letter followed by the number designates the original amino acid residue replaced with the amino acid residue indicated by the last letter) in motif II, and G215E and R222E in motif IV (Fig. 4B). As controls for the in vitro DNA-binding activity of these mutant spo11-1 proteins, to exclude possible non-specific negative effects caused by the replacement of a basic amino acid residue with an acidic amino acid residue, we constructed a series of coding DNAs in which the non-conserved basic amino acid residues surrounding the putative DNA-binding surface (i.e. the Arg223 and Arg226 residues in motif IV, and the Arg207 and Arg254 residues outside the conserved motifs) were substituted with acidic Glu res- idues (Fig. 4B). In addition, the substitution, by Phe, of the invariant Tyr103 residue (Y103F) in motif I of the putative endonuclease-active center of SPO11-1 was constructed as a negative control [3,27,29]. All of the constructed mutants (i.e. R130E, R207E, G215E, R222E, R223E, R226E, R254E and Y103F) of TF-SPO11-1 were expressed in E. coli in soluble forms, and were purified from the cell-free lysates (Fig. S2A). Quantitative analyses of the DNA-binding activities of the mutant spo11-1 proteins to the 180 bp dsDNA molecule revealed that the mutations of invariant amino acid residues (i.e. G215E and R222E) and non- conserved residues (i.e. R223E and R226E) in motif IV reduced the DNA-binding ability of SPO11-1 (Fig. 5A,B). Substitution of the invariant Tyr103 residue in the putative endonuclease-active centre of SPO11-1 (i.e. Y103F) did not alter the DNA- binding ability of SPO11-1 (Fig. 5A,B). The remaining A B Fig. 4. Amino acid residues of SPO11-1 substituted in the mutant spo11-1 proteins. (A) A 3D structural model of SPO11-1 obtained by homology modelling. This SPO11-1 model consists of amino acid residues 71–363. The electrostatic potential was calculated using MolFeat. Positive and negative potentials are indicated in blue and red, respectively. The Gly215 , Arg222, Arg223 and Arg226 residues that were found to function in DNA binding are shown in light blue (see Fig. 5). The Tyr103, Arg130, Arg207 and Arg254 residues that are not involved in DNA binding are shown in white. (B) Amino acid alignments of regions containing mutated amino acid residues. Numbers indicate the positions of amino acid residues. The SPO11-1 (accession no.: AJ251989) and SPO11-2 (accession no.: AJ251990) proteins from Arabidopsis thaliana, DmSpo11 from Drosophila melanogaster (mei-W68; accession no.: AAC61735), HsSpo11 from Homo sapiens (accession no.: AAD52562), ScSpo11 from Saccharomyces cerevisiae, SpSpo11 from Schizosaccharomyces pombe (rec12; accession no.: CAB11511) and MjTop6A from Methanococcus jannaschii (accession no.: AAB98358) are shown. Amino acids that were conserved among at least six of the seven homologues are shaded in gray. DNA-binding surface of Arabidopsis SPO11 protein Y. Shingu et al. 2366 FEBS Journal 277 (2010) 2360–2374 ª 2010 The Authors Journal compilation ª 2010 FEBS mutations (i.e. R130E, R207E and R254E) outside motif IV, including the substitution of the conserved Arg130 residue in motif II, exhibited normal DNA- binding activity (Fig. 5A,B). As the concentration of DNA was very low relative to the concentration of protein, the apparent K d values, which are equal to the protein concentration at which 50% DNA is bound, were estimated to be  0.35 lm for wild-type, Y103F, R130E, R207E and R254E,  0.8 lm for G215E and R226E and much larger than 0.8 lm (generally esti- mated to be  1.5 lm) for R222E and R223E. These results revealed that the reduced DNA binding of mutant spo11-1 proteins (i.e. G215E, R222E, R223E and R226E) observed in vitro was not the result of a global change in the net charge (i.e. basic or neutral to acidic) of the mutant proteins. Thus, Gly215, Arg222, Arg223 and Arg226, but not Arg130, Arg207, Arg254 and Tyr103, play an important role in the DNA-bind- ing activity of SPO11-1 in vitro. We confirmed that this deficiency was not caused by misfolding of the mutant proteins. First, the spo11-1 mutants and the wild-type SPO11-1 showed the same changes in florescence emission from aro- matic residues (generally reflecting a change in protein tertiary structure) when they were subjected to dena- turation (data not shown). We further analyzed the far-UV CD spectra (in the region between 200 and 250 nm) for the most severely DNA-binding defective mutant, R222E, in comparison with the wild-type, and found that both showed negative peaks at around 210 and 220 nm (Fig. 5C), which are charac- teristic of the secondary structure of proteins. The signal intensity of the R222E mutant spo11-1 was almost identical to that of the wild-type (Fig. 5C), indicating that the spo11-1-R222E exhibits normal protein folding. In addition, we compared the gel- filtration profile of the spo11-1-R222E with that of the wild-type SPO11-1, and both profiles were identi- cal (Fig. S2). Thus, the higher-order structures of the spo11-1-R222E mutant are normal, and the DNA- binding deficiency is quite likely to be a direct effect of the R222E substitution. A B C Fig. 5. dsDNA-binding activities of mutant TF-spo11-1 proteins. (A) The TF-spo11-1 mutants were incubated with 9 n M 180 bp dsDNA at 37 °C for 5 min. The amounts of proteins are 0, 0.24, 0.47 and 0.71 l M (from left to right). After the reaction, DNA binding was analyzed as described in Fig. 3C. The symbols * and ** indicate protein-bound DNA and total input DNA, respectively. (B) Quantification of the DNA-binding experiments described above, in the legend to panel A. The wild-type and mutant proteins are indicated within the figure. (C) Far-UV CD spectra of TF-SPO11-1 and TF-spo11-1-R222E. The spectra of TF-SPO11-1 and TF-spo11-1- R222E (both 10 l M) are represented as black and gray lines, respectively, and were measured using a 1 mm cell in 50 m M sodium phosphate buffer, at 25 °C. Y. Shingu et al. DNA-binding surface of Arabidopsis SPO11 protein FEBS Journal 277 (2010) 2360–2374 ª 2010 The Authors Journal compilation ª 2010 FEBS 2367 Effects of single amino acid substitutions on in vivo functions of SPO11-1 We sought to address whether the Gly and Arg resi- dues responsible for the DNA-binding activity of SPO11-1 in vitro were also required for the in vivo functions of the protein. First of all, we expressed wild- type SPO11-1 cDNA under the control of the SPO11-1 promoter in a spo11-1-3 homozygous A. thaliana mutant. This construct complemented spo11-1-3,as evidenced by the formation of normal siliques in all five transformants (Fig. 6A). By contrast, the expres- sion of spo11-1–Y103F cDNA did not complement spo11-1-3 in any of the four transformants tested (Fig. 6A), as described previously [27]. Then, using a common vector, we expressed spo11-1 cDNAs harbor- ing the glutamic acid substitution of the invariant Gly215 (G215E) as a partially defective DNA-binding mutant, the glutamic acid-substitution of invariant Arg222 (R222E) as a DNA-binding defective mutant, and the glutamic acid-substitution of invariant Arg130 (R130E), or of the non-conserved Arg207 (R207E), as DNA-binding proficient variants (Table 1). We observed no complementation in the three G215E transformants and the three R222E transformants, and observed full complementation in two of the three R207E transformants (Fig. 6A). As no SPO11-1-spe- cific antiserum is currently available to test for the expression of these proteins, we analyzed the expres- sion and the sequence of each mutant spo11-1 cDNA transgene by RT-PCR, and found that the amounts of SPO11-1 mRNA were not affected by any of the single substitution mutations (Fig. 6B). We then determined the number of seeds produced by 20 randomly selected siliques from the transformants. The number of seeds produced by the G215E and R222E transformants was similar to that produced by the spo11-1-3 homozygous mutant (Table 1). A B Fig. 6. Complementation of the meiotic defects of an Arabidopsis spo11-1-3 homo- zygous mutant by the expressed wild-type and mutant SPO11-1 cDNAs. (A) Panels show the flowering stems of the wild-type strain, the spo11-1-3 homozygous mutant and spo11-1-3 mutants transformed with a vector carrying wild-type SPO11-1 or a mutant spo11-1 (i.e. Y103F, R130E, R207E, G215E, or R222E). Arrows indicate siliques. The number of complementation-positive lines and the total number of lines analyzed are denoted under each picture. (B) RT-PCR analyses of the spo11-1 transgenes. Lane 1, wild type strain; lane 2, spo11-1-3 homozy- gous mutant; lanes 3–8, the spo11-1-3 homozygous transformants containing the transgene of SPO11-1, Y103F, R130E, R207E, G215E and R222E, respectively. DNA-binding surface of Arabidopsis SPO11 protein Y. Shingu et al. 2368 FEBS Journal 277 (2010) 2360–2374 ª 2010 The Authors Journal compilation ª 2010 FEBS We did not initially observe complementation in five R130E transformants, but closer examination revealed that the siliques of the R130E mutants were slightly thicker than those of the spo11-1-3 homozy- gous mutant (Fig. 6A). The R130E transformants produced a much smaller number of seeds than the wild-type and R207E transformants, but they produced a significantly larger number of seeds in each silique, as compared with the Y103F, G215E or R222E transformants and the spo11-1-3 homozygous mutant (Table 1). From these results, we concluded that the invariant Gly215 and Arg222 residues, the conserved Arg130 residue, as well as the invariant Tyr103 residue, are required for the in vivo function of SPO11-1 during meiosis, and that a spo11-1 mutant containing the R130E substitution retained partial in vivo function. Discussion In this study, we isolated an active form of SPO11 with functional DNA-binding activity. TF-SPO11-1 is able to bind to dsDNA and to ssDNA (Fig. 3A,B); however, it possesses approximately twofold higher binding affinity towards dsDNA than towards ssDNA. As free TF did not bind to either form of DNA, the DNA-binding activities of TF-SPO11-1 reside on the SPO11-1-domain. SPO11 is a homologue of the type II topoisomerases [3], which change the topological states of closed-circular dsDNA through transient double- stranded cleavage. Thus, it is intriguing that SPO11-1 can bind to ssDNA. However, it was shown previously that the type II topoisomerase of E. coli phage T4, the prototype type II topoisomerase, bound and cleaved ssDNA [37]. The biological significance of this charac- teristic remains to be studied. We detected two discrete signals for TF-SPO11-1- bound 180-bp dsDNA: a major signal of a larger com- plex that was detected in all experiments; and a minor signal of a smaller complex that was detected only in the presence of a certain amount of NaCl (Fig. 3D) or in the presence of a smaller amount of TF-SPO11-1 (Figs 3C and 5A). The DNA-binding surface of the homology modelled SPO11-1 is a channel of  20 A ˚ in diameter and  30 A ˚ in width, and therefore each SPO11-1 would accommodate eight to nine base pairs of dsDNA (Fig. 4A). As the DNA-binding experi- ments were performed in the presence of a molar excess of the protein (Fig. 3C), and TF-SPO11 forms multimers consisting of various numbers of protein molecules (Fig. 2C), it is likely that a number of SPO11-1 proteins can be simultaneously bound to a 180 bp dsDNA molecule as a multimer. The observed two discrete bands of TF-SPO11-1-bound dsDNA might reflect two distinct states of the multimer. One may speculate that each TF-SPO11 multimer binds to one side of a dsDNA molecule, but we have no further results with which to discuss each state of the protein– DNA complex. The solubility of SPO11-1 was greatly enhanced by using TF, especially when TF was attached to the N terminus of SPO11-1 (TF-SPO11-1, Fig. 1). TF has been shown to associate with the ribosomal 50S sub- unit, and to catalyze the proline-limited refolding of an RNase T1 variant and other proteins in vitro [31,32]. The refolding activity of TF and the solubility of TF itself (especially in fused SPO11-1) might enhance the solubility of the attached SPO11-1. Almost all of the TF-tag-free SPO11-1 was expressed in a soluble form via co-expression with TF-SPO11-1 (Fig. 1B,C), and the co-expressed TF-SPO11 and TF-tag-free SPO11-1 form much larger complexes than TF-SPO11-1 alone (Fig. 2C). Thus, the solubility of SPO11-1 is maintained via the direct interaction with a soluble form of SPO11-1 (i.e. the SPO11-1 domain of TF-SPO11-1). This large complex of the co-expressed TF-SPO11 and TF-tag-free SPO11-1 may exist as soluble aggregates of Spo11-1 simply coated by soluble TF-SPO11-1. However, it is more likely that the forma- tion of the large soluble complexes is explained by co-operativity in the proper folding of SPO11-1, and this suggests that a solubilizing protein cofactor is Table 1. Effect of amino acid substitutions within Spo11-1 upon DNA-binding affinity and seed production. ND, not determined. Strain Transgene a DNA binding b Seeds c Wild-type ++++ 36.9 ± 2.7 spo11-1-3 None 1.9 ± 0.5 Y103F ++++ 2.0 ± 0.6 R130E ++++ 5.3 ± 1.6 R207E ++++ 34.5 ± 3.5 G215E ++ 1.9 ± 0.8 R222E + 2.0 ± 0.8 R223E + ND R226E ++ ND R254E ++++ ND a spo11-1 containing the indicated amino acid substitution. b The DNA-binding ability in vitro of the proteins containing the indicated amino acid-substitution: ++++, as good as that of the wild-type pro- tein (K d ’ 0.35 lM); ++, partially defective (K d ’ 0.8 lM); +, almost defective (K d >> 0.8 lM;or 1.5 lM). See Fig. 5B for quantitative results. The K d values were estimated by the protein concentra- tions at which 50% of the DNA is or would be bound in the pres- ence of excessive amounts of the protein (see the text). c The average number of seeds per silique (i.e. among 20 randomly selected siliques from the transformants). Y. Shingu et al. DNA-binding surface of Arabidopsis SPO11 protein FEBS Journal 277 (2010) 2360–2374 ª 2010 The Authors Journal compilation ª 2010 FEBS 2369 [...].. .DNA-binding surface of Arabidopsis SPO11 protein Y Shingu et al required for the proper folding and maintenance of a soluble or active form of SPO11- 1 In Arabidopsis, no protein factors, except for PRD1, were shown to have functions to activate SPO11s in meiosis [24,26], and PRD1N alone was not soluble and did not help to solubilize SPO11- 1 (data not shown) Unlike yeast and other organisms, Arabidopsis. .. et al (2003) Genome-wide insertional mutagenesis of Arabidopsis thaliana Science 301, 653– 657 43 Clough SJ & Bent AF (1998) Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana Plant J 16, 735–743 Supporting information The following supplementary material is available: Fig S1 Tests for biochemical activities of SPO11- 1 Fig S2 Gel filtration analysis of the... dimerization places the Tyr residue adjacent to the active center [39] Thus, it is also likely that the binding of an interacting protein(s) to SPO11- 1 is required to place Tyr103 at the appropriate location for DNA cleavage, and this could provide an explanation for the absence of a detectable endonuclease activity in our preparation of TF -SPO11- 1 This study has identified a group of spo11- 1 mutants... Research Advancement Institution (BRAIN), and in part by a Grant in Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan of 0.5 m Na2HPO4 (pH 7.2), 1 mm EDTA and 7% SDS [40] at 65 °C for 16 h The membranes were washed twice with 0.2 · NaCl ⁄ Cit (SSC) containing 0.1% SDS at 65 °C and then exposed to an imaging plate (Fuji Film, Tokyo, Japan) The bound DNA was detected and... Copeland NG, Jenkins NA, Manova K & Jasin M (1999) A mouse homolog of the Saccharomyces cerevisiae meiotic recombination DNA transesterase Spo11p Genomics 61, 170–182 9 Romanienko PJ & Camerini-Otero RD (1999) Cloning, characterization, and localization of mouse and human SPO11 Genomics 61, 156–169 10 Shannon M, Richardson L, Christian A, Handel MA & Thelen MP (1999) Differential gene expression of mammalian... a putative DNAbinding surface of Arabidopsis SPO11- 1 With controls with the same type of amino acid-substitution mutant spo11- 1s, we showed that the Gly215, Arg222, Arg223 and Arg226 residues, but not Arg130, Arg207, Arg254 and Tyr103, are members of a DNA-binding surface of SPO11- 1 (Figs 4 and 5) Both Gly215 and Arg222 residues are required for in vivo meiotic function of SPO11- 1 (Fig 6A) Thus, our... mutants with defects in meiosis that specifically correlate with the defects in DNA binding at an identified DNAbinding surface on SPO11- 1 Materials and methods Nucleic acids A 180 bp dsDNA cation of pUC18 180R (Table S1) M13mp18 ssDNA (Shiga, Japan) fragment was obtained by PCR amplifidsDNA, using the primers 180L and Closed-circular pUC18 dsDNA and were purchased from TAKARA Bio Inc Detailed procedure to... mutant was transformed with these SPO11 expression vectors using the floral dip method in Agrobacterium tumefaciens [43] After screening for kanamycin resistance of the T1 seeds, the genotypes of the transformants were tested by PCR with primers for the T-DNA insertion and the SPO11- 1 intron regions Only plants that were homozygous for the spo11- 1-3 mutation were analyzed further Transgenes were analyzed... 206–216 16 Hartung F & Puchta H (2001) Molecular characterization of homologues of both subunits A (SPO11) and B of the archaebacterial topoisomerase 6 in plants Gene 271, 81–86 17 Jain M, Tyagi AK & Khurana JP (2006) Overexpression of putative topoisomerase 6 genes from rice confers stress tolerance in transgenic Arabidopsis plants FEBS J 273, 5245–5260 18 Hartung F, Angelis KJ, Meister A, Schubert... Zhao Y, Hu J, Mora-Garcia S & Chory J (2002) A crucial role for the putative Arabidopsis topoisomerase VI in plant growth and development Proc Natl Acad Sci USA 99, 10191– 10196 21 Keeney S (2001) Mechanism and control of meiotic recombination initiation Curr Top Dev Biol 52, 1–53 22 Keeney S & Neale MJ (2006) Initiation of meiotic recombination by formation of DNA double-strand breaks: mechanism and . A DNA-binding surface of SPO11- 1, an Arabidopsis SPO11 orthologue required for normal meiosis Yoshinori Shingu 1 , Tsutomu Mikawa 2 , Mariko Onuma 1 , Takashi Hirayama 3 and Takehiko Shibata 1 1. flies, nematodes and mammals encode a single SPO11; however, plants (e.g. Arabidopsis and rice Oryza sativa) encode at least three SPO11 paralogues [16,17]. In Arabidopsis, SPO11- 1 and SPO11- 2 are. gray and black lines indicate absorbance at 280 nm in the gel filtration of TF -SPO11- 1 and that of the complex of TF -SPO11- 1 and SPO11- 1, respectively. Y. Shingu et al. DNA-binding surface of Arabidopsis