Locally, Meiotic Double-Strand Breaks Targeted by Gal4BD-Spo11 ...

MOLECULAR AND CELLULAR BIOLOGY, July 2009, p. 3500–3516 0270-7306/09/$08.00⫹0 doi:10.1128/MCB.00088-09 Copyright © 2009, American Society for Microbiology. All Rights Reserved.

Vol. 29, No. 13

Locally, Meiotic Double-Strand Breaks Targeted by Gal4BD-Spo11 Occur at Discrete Sites with a Sequence Preference䌤 Hajime Murakami† and Alain Nicolas* Recombination and Genetic Instability, Institut Curie Centre de Recherche, UMR3244 CNRS, Universite´ Pierre et Marie Curie, 26 Rue d’Ulm, 75248 Paris Cedex 05, France Received 20 January 2009/Returned for modification 18 February 2009/Accepted 10 April 2009

Meiotic recombination is initiated by DNA double-strand breaks (DSBs) that are catalyzed by the type II topoisomerase-like Spo11 protein. Locally, at recombination hot spots, Spo11 introduces DSBs at multiple positions within ⬃75 to 250 bp, corresponding to accessible regions of the chromatin. The molecular basis of this multiplicity of cleavage positions, observed in a population of meiotic cells, remains elusive. To address this issue, we have examined the properties of the Gal4BD-Spo11 fusion protein, which targets meiotic DSBs to regions with Gal4 binding sites (UAS). By single-nucleotide resolution mapping of targeted DSBs, we found that DSB formation was restricted to discrete sites approximately 20 nucleotides from the UAS, defining a “DSB targeting window.” Thus, the multiplicity of cleavage positions at natural Spo11 hot spots likely represents binding of Spo11 to different distinct sites within the accessible DNA region in each different meiotic cell. Further, we showed that mutations in the Spo11 moiety affected the DSB distribution in the DSB targeting window and that mutations in the DNA at the Spo11 cleavage site affected DSB position. These results demonstrate that Spo11 itself has sequence preference and contributes to the choice of DSB positions. that the distribution of meiotic DSBs is not random along chromosomes and displays multiple levels of spatial organization. At the genome-wide view, there are hot and cold domains that span large portions of chromosomes (on the order of 100 kb) (2, 9, 29). DSB formation is repressed in the chromosomal regions proximal to centromeres (ca. 8 to 10 kb) and chromosome ends (⬃20 kb) (5, 6). In addition, it is known that DSBs are preferentially formed in chromatin within loops as opposed to chromatin bound to chromosome axes where sister chromatid cohesion proteins are located (4). Locally, DSBs are preferentially introduced within regions called hot spots (22) separated by several kilobases of DNA in which few, if any, DSBs are formed. Most hot spots in S. cerevisiae are located within transcriptional promoter regions (2). Because hot spots are located in nuclease-hypersensitive regions (i.e., regions with an “open” chromatin configuration), local accessibility of chromatin is considered to be a prerequisite for the location of DSB formation sites at this level (20, 27). DSB mapping at high resolution has revealed that each hot spot consists of multiple DSB sites clustered within regions of ⬃75 to 250 bp (7, 17, 28). Spo11 cleaves the DNA to yield a two-nucleotide 5⬘ overhang and that some positions within hot spots are cut more frequently than others (17). Mutations at the putative DNA-binding surface of Spo11 alter the distribution of DSBs within a hot spot, suggesting that Spo11 itself retains the function to select the binding and the cleavage positions at this level of DSB distribution (8). However, no obvious DNA sequence preference has yet emerged. Hence, the molecular basis of how Spo11 introduces DSBs at multiple sites within each hot spot is still elusive. Fusion of Spo11 to the Gal4 DNA-binding domain showed that the Gal4BD-Spo11 protein is functional and introduces DSBs not only in natural regions but also in new regions containing consensus-binding sites of the Gal4 protein (UAS) (21, 24). Thus, the normal recruitment of Spo11 to chromatin

Meiosis comprises one round of DNA replication and two successive rounds of cell division, allowing the reduction of a diploid genome to produce haploid gametes. Recombination between homologous chromosomes (homologs) takes place during prophase of the first meiotic division. Meiotic recombination is crucial not only for creating genetic diversity of gametes but also for providing physical connections between homologs. These connections (i.e., chiasmata) are absolutely required to ensure proper alignment of chromosomes on the spindle so that homologs segregate accurately during the first meiotic division. Meiotic recombination is initiated by programmed DNA double-strand breaks (DSBs), a subset of which, upon their repair, generates the crossovers (COs). These DSBs are generated by the evolutionarily conserved Spo11 protein (14). Besides Spo11, at least nine additional proteins (DSB proteins) are required for DSB formation in Saccharomyces cerevisiae (14). DSB proteins interact with each other to load Spo11 onto the chromosome (1, 23). Spo11 introduces DSBs by forming an intermediate in which Spo11 is covalently linked to 5⬘ strand termini of DSB ends (15). Spo11 is then removed by endonucleolytic cleavage (19), and broken ends are processed by 5⬘ to 3⬘ resection to generate 3⬘ single-stranded DNA regions and engage in repair by homologous recombination. Genome-wide map of meiotic recombination closely parallels these maps of DSBs (18). Extensive studies performed with S. cerevisiae have shown

* Corresponding author. Mailing address: Recombination and Genetic Instability, Institut Curie Centre de Recherche, UMR3244 CNRS, Universite´ Pierre et Marie Curie, 26 Rue d’Ulm, 75248 Paris Cedex 05, France. Phone: 33(0)1 56 24 65 20. Fax: 33(0)1 56 24 66 44. E-mail: [email protected]. † Present address: Molecular Biology Program, Memorial SloanKettering Cancer Center, 1275 York Ave., New York, NY 10065. 䌤 Published ahead of print on 20 April 2009. 3500

VOL. 29, 2009

MEIOTIC RECOMBINATION INITIATION

3501

TABLE 1. Yeast strainsa Strain

Genotype

ORD8050 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 HIS4/his4 ARG4/arg4 ORD8421 ...................................trp1 rad50S::LEU2 HIS4/his4 ARG4/arg4 ORD8704 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 L-GAL2::KanMX/o o/R-GAL2::HphMX HIS4/his4 arg4/arg4 ORD8705 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 L-GAL2::KanMX/o o/R-GAL2::HphMX HIS4/ his4 arg4/arg4 ORD8706 ...................................trp1 L-GAL2::KanMX/o o/R-GAL2::HphMX HIS4/his4 ARG4/ARG4 ORD8707 ...................................trp1 rad50S::LEU2 L-GAL2::KanMX/o o/R-GAL2::HphMX HIS4/HIS4 ARG4/arg4 ORD8713 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 L-GAL2::KanMX/o R-GAL2::HphMX/o HIS4/his4 arg4/arg4 ORD8714 ...................................trp1 L-GAL2::KanMX/o R-GAL2::HphMX/o HIS4/his4 ARG4/arg4 ORD8715 ...................................trp1::pADH1-GAL4BD-SPO11 spo11::URA3 L-GAL2::KanMX/o o/R-GAL2::HphMX HIS4/his4 ARG4/arg4 ORD8723 ...................................trp1::pADH1-GAL4BD L-GAL2::KanMX/o o/R-GAL2::HphMX HIS4/his4 ARG4/arg4 ORD8731 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_aBCDE HIS4/his4 arg4/arg4 ORD8741 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_ABCdE HIS4/his4 ARG4/arg4 ORD8758 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_AbCDE HIS4/HIS4 arg4/arg4 ORD8759 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_ABcDE HIS4/his4 ARG4/ARG4 ORD8760 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_ABCDe HIS4/his4 ARG4/arg4 ORD8796 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_abcde HIS4/HIS4 ARG4/arg4 ORD8797 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 pGAL2_abcde L-GAL2::KanMX/o R-GAL2::HphMX/o HIS4/ his4 ARG4/arg4 ORD8799 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_abcDe HIS4/his4 ARG4/arg4 ORD9200 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_Abcde HIS4/his4 arg4/arg4 ORD9201 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_aBcde HIS4/his4 arg4/arg4 ORD9207 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_abcdE HIS4/his4 arg4/arg4 ORD9208 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_abCde HIS4/his4 arg4/arg4 ORD9209 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_ab2cde his4/his4 arg4/arg4 ORD9210 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_abcd2e HIS4/HIS4 arg4/arg4 ORD9216 ...................................trp1::pREC8-gal4bdC14Y-SPO11-TRP1 spo11::URA3 rad50S::LEU2 HIS4/his4 arg4/arg4 ORD9217 ...................................trp1::pREC8-gal4bdL32P-SPO11-TRP1 spo11::URA3 rad50S::LEU2 HIS4/his4 arg4/arg4 ORD9220 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_ab2cd2e HIS4/his4 arg4/arg4 ORD9221 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pYNL328c_Gal4UASm his4/his4 ARG4/ARG4 ORD9225 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 YAL048c_Gal4UASm his4/his4 arg4/arg4 ORD9226 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 pGAL2_ab2cd2e L-GAL2::KanMX/o R-GAL2::HphMX/o his4/ his4 ARG4/arg4 ORD9228 ...................................trp1::pREC8-GAL4BD105-SPO11-TRP1 spo11::URA3 rad50S::LEU2 HIS4/his4 ARG4/arg4 ORD9229 ...................................trp1::pREC8-GAL4BD105-SPO11-TRP1 spo11::URA3 HIS4/his4 arg4/arg4 ORD9233 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_abcdE_csm1 HIS4/his4 ARG4/arg4 ORD9234 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_abcdE_csm2 HIS4/his4 ARG4/arg4 ORD9235 ...................................trp1::pREC8-GAL4BD-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_abcdE_csm3 HIS4/his4 ARG4/arg4 ORD9240 ...................................trp1::pREC8-GAL4BD-spo11D290A-TRP1 spo11::URA3 rad50S::LEU2 HIS4/his4 ARG4/arg4 ORD9241 ...................................trp1::pREC8-GAL4BD-spo11D290A-TRP1 spo11::URA3 his4/his4 arg4/arg4 ORD9258 ...................................trp1::pREC8-GAL4BD-spo11F260R-TRP1 spo11::URA3 rad50S::LEU2 HIS4/his4 arg4/arg4 ORD9259 ...................................trp1::pREC8-GAL4BD-spo11F260R-TRP1 spo11::URA3 HIS4/his4 ARG4/arg4 ORD9260 ...................................trp1::pREC8-GAL4BD-spo11R131K-TRP1 spo11::URA3 rad50S::LEU2 HIS4/his4 ARG4/arg4 ORD9261 ...................................trp1::pREC8-GAL4BD-spo11R131K-TRP1 spo11::URA3 HIS4/his4 ARG4/arg4 ORD9262 ...................................trp1::pREC8-GAL4BD-spo11E235A-TRP1 spo11::URA3 rad50S::LEU2 HIS4/his4 arg4/arg4 ORD9263 ...................................trp1::pREC8-GAL4BD-spo11E235A-TRP1 spo11::URA3 HIS4/his4 ARG4/arg4 ORD9270 ...................................trp1::pREC8-GAL4BD105-SPO11-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_abcdE HIS4/his4 ARG4/arg4 ORD9271 ...................................trp1::pREC8-GAL4BD-spo11R131K-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_abcdE HIS4/HIS4 arg4/arg4 ORD9272 ...................................trp1::pREC8-GAL4BD-spo11F260R-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_abcdE his4/his4 ARG4/ARG4 ORD9273 ...................................trp1::pREC8-GAL4BD-spo11D290A-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_abcdE HIS4/his4 ARG4/arg4 ORD9274 ...................................trp1::pREC8-GAL4BD-spo11E235A-TRP1 spo11::URA3 rad50S::LEU2 pGAL2_abcdE HIS4/his4 ARG4/arg4 a

All strains are MATa/MAT␣ ho::LYS2 lys2 ura3 leu2. All alleles are homozygous unless otherwise specified.

can be bypassed by the Gal4BD moiety of the fusion protein. However, DSBs targeted by Gal4BD-Spo11 still require all of the DSB proteins (21). The simplest interpretation of this targeting capacity is that the binding of the Gal4BD moiety to its UAS brings Spo11 to that site. If so, and if the multiplicity of DSBs observed within a natural hot spot is based on a feature of Spo11 binding to DNA, then we would expect that, upon DSB targeting, Spo11 would bind and introduce DSBs in a limited region near the UAS. Here, we tested this prediction by mapping the DSB sites to single-nucleotide resolution in various protein and DNA constructs. The results show that the DSBs targeted by Gal4BD-Spo11 requires specific binding of

the Gal4BD moiety to UAS and that the targeted DSBs are limited to a small window in the vicinity of the UAS. Then, taking advantage of this simplified experimental system, we show a heretofore unseen DNA sequence preference of Spo11.

MATERIALS AND METHODS Strains and plasmids. The strains used in the present study are described in Table 1 and are all derivatives of the SK1 background. A PCR fragment, which contains the GAL2 promoter sequence between positions ⫺743 and ⫹816 relative to the GAL2 translation start, was cloned into a TA cloning vector pGEM-T to produce pGAL2pr#1. Mutations at UASs were introduced by site-directed mutagenesis to produce pmGAL2pr⬙aBCDE⬙, pmGAL2pr⬙AbCDE⬙, pmGA

3502

MURAKAMI AND NICOLAS

L2pr⬙ABcDE⬙, pmGAL2pr⬙ABCdE⬙, pmGAL2pr⬙ABCDe⬙, pmGAL2pr⬙ Abcde⬙, pmGAL2pr⬙aBcde⬙, pmGAL2pr⬙abCde⬙, pmGAL2pr⬙abcDe⬙, pmGA L2pr⬙abcdE⬙, pmGAL2pr⬙abcde⬙, pmGAL2pr⬙ab2cde⬙, pmGAL2pr⬙abcd2e⬙, and pmGAL2pr⬙ab2cd2e⬙. Cleavage site mutations were further introduced on pmGAL2pr⬙abcdE⬙ to produce pmGAL2pr⬙abcdE_csm1⬙, pmGAL2pr⬙abcdE_ csm2⬙, and pmGAL2pr⬙abcdE_csm3⬙. A SacI fragment containing the KanMX cassette was removed from the original pAP1 (21), producing pAP11, which bears pADH1-GAL4BD-SPO11. A 708-bp fragment containing the AHD1 promoter was removed from pAP11 by using EcoRV and HindIII sites and an EcoRV-HindIII-flanked PCR fragment, which contains the REC8 promoter from positions ⫺461 to ⫺1 relative to the REC8 translation start, was inserted via subcloning, producing pAP118 (pREC8-GAL4BD-SPO11). The TG nucleotides of pAP118, located at position ⫹40 relative to the GAL4BD-SPO11 translation start, were changed to AC, producing pAP118C14Y (pREC8-gal4bdC14YSPO11). The T nucleotide of pAP118, located at position ⫹95, was changed to a C, producing pAP118L32P (pREC8-gal4bdL32P-SPO11). A 126-bp sequence between positions ⫹316 and ⫹441 of pAP118, relative to the GAL4BD-SPO11 translation start, was removed to produce pAP118-105 (pREC8-GAL4BD105SPO11). A 621-bp sequence surrounding UAS of the YAL048c ORF, i.e., positions ⫹1110 to ⫹1730 relative to the YAL048c translation start, was cloned to pCR2.1, producing pYAL048c-seq. A 615-bp sequence surrounding UAS of the YNL328c promoter, positions ⫺826 to ⫺212 relative to the YNL328c translation start, was cloned into pCR2.1, producing pYNL328c-seq. Introduction of mutations at UASs and at cleavage site in the GAL2 promoter. Mutations were introduced by two-step replacement. The URA3 gene was inserted in the GAL2 promoter, which involves a deletion of the GAL2 promoter from positions ⫺580 to ⫺1 relative to the GAL2 translation start site. Subsequently, the URA3 gene was replaced by PCR fragments amplified using the above-described plasmids, which bear the GAL2 promoter sequence with a series of mutations, as a template and the following primer pair: TGAGCATAACGG GCTGT and ATAACGTGGGGATTCAGGA. Mutations were verified by Southern analysis and sequencing. Introduction of mutations at UAS in YAL048c and YNL328c. The URA3 gene was amplified from pRS416 by using primers which have 50 bp of flanking sequence containing the mutated UAS sequence of YAL048c or YNL328c as with the following primers: primer pair for YAL048c mutations, GCCTTCTTGGGCAGATC TTTCTCtGAGGAGTATTCTCCtACAATCAAACCAAGAATTGCAGTCAAT AGTTTAGAACTCCGGTGATGACGGTGAAAACCTC and CTCTAGCAAA GAGCTTTTGCCGCAACATGGCTTTCCAATGACAAAGCAATTGAACAC TTTTCTGTCATTGATATTGGAACAACACTCAACCCTATCTCGG; and primer pair for YNL328c mutations, GCCTGCTGTTTTTCCTCACGGCGCtG AGCCAGCGCGTCtGAATTAGGAAAATAATAGAGCCAGCAAGAACCA TTTCACCGGTGATGACGGTGAAAACCTC and TGACAAGGGACTAAT TTACTGACACTTTCGGCTGACACTTCCGAAAAGGTACTCAAGCTTTT ATGAGTAAGATGCTGAACAACACTCAACCCTATCTCGG (mutations introduced at the UAS are indicated in lowercase letters; the annealing sequence to pRS416 is underlined). The PCR fragments were transformed into ura3 haploid yeast, and URA⫹ transformants were selected. Subsequently, from nonselectively cultured transformants, URA⫺ species (popout) were selected and verified by sequencing. Sporulation conditions. Cells were grown up to an optical density at 600 nm of 1.4 in SPS (0.5% yeast extract, 1% peptone, 0.17% yeast nitrogen base [without amino acid and ammonium sulfate], 1% potassium acetate, 0.5% ammonium sulfate, 0.05 M potassium biphthalate [pH 5.5]). Cells then were washed with water and transferred to double volume of sporulation medium (SPM; 1% potassium acetate) at 30°C to induce sporulation. DNA preparation. Genomic DNA was extracted from cells at 0 h and at several hours after transfer to sporulation medium as described elsewhere (25) with some modifications. Briefly, from ⬃8 ⫻ 108 cells, spheroplasts were produced by using 1 mg of Zymolyase-20T (MP Biomedicals, Inc.) in 1 ml of spheroplast buffer (1% 2-mercaptoethanol, 1 M sorbitol, 0.1 M EDTA [pH 8.0]) for 45 min at 30°C. The spheroplasts were lysed, and proteins were digested in 500 ␮l of lysis buffer (50 mM EDTA [pH 8.0], 50 mM Tris [pH 8.0], 0.5% sodium dodecyl sulfate, 200 ␮g of proteinase K). Proteins were removed by adding 200 ␮l of 5 M potassium acetate and subsequent 15,000 rpm centrifugation for 20 min at 4°C. After the digestion of RNA, DNA was ethanol precipitated and suspended in 100 ␮l of TE buffer (10 mM Tris-Cl [pH 7.5], 1 mM EDTA [pH 8.0]). Construction of flanking markers. The drug-resistant cassettes, KanMX and HphMX, were amplified from plasmids pFA6a (26) and pMJ696 (identical to pAG32 in reference 10) using the following sequences: CCTCACAGACGC GTTG and GACCAGCATTCACATACG. Since the primers have 50 bp of flanking sequence, the drug resistance cassette was introduced upstream (between positions 287725 and 287726: chromosome XII SGD) and downstream

MOL. CELL. BIOL. (between positions 292068 and 292069) of GAL2, respectively. Transformants were confirmed to have the flanking markers at the correct locus and for whether any new DSB sites ware created at the junction or inside the markers (data not shown). Genetic assay of meiotic recombination. Diploids bearing flanking markers were sporulated and collected 24 h after transfer to the SPM. Tetrads were dissected and checked for the phenotype of resistance on YPD plates (1% yeast extract, 2% peptone, 2% dextrose, 2% agar) with either G418 (final concentration, 200 ␮g/ml) or hygromycin B (final concentration, 300 ␮g/ml). According to the phenotypes of dissected tetrads, the numbers of parental ditypes (PD), tetratypes (T), and nonparental ditypes (NPD) were counted to calculate the centimorgans: 100(T ⫹ 6 NPD)/2(PD ⫹ T ⫹ NPD). Standard and high-resolution mapping of meiotic DSB. For detection and quantification of DSBs, genomic DNA was prepared from meiotic cultures and subjected to Southern analysis. For standard and high-resolution mapping, 0.8 to 1.0% and 1.7% agarose gels were used, respectively. Parental and DSB bands were visualized by using a phosphorimager and quantified by using ImageQuant (Molecular Dynamics) as described previously (25). High-resolution mapping of meiotic DSB on polyacrylamide gels. DSB mapping in a polyacrylamide gel was performed with the procedure used for nucleotide resolution level mapping (see below), except for the following: the DNA fragments were separated in 6% polyacrylamide gels containing 8 M urea (8 cm long, 1 mm thick) and 1⫻ Tris-borate-EDTA (TBE) buffer at 240 V for 45 min. Nylon membranes (GeneScreen; Perkin-Elmer) were used for semidry blotting. DNA sequence standards. Sequence standards were prepared by using a SequiTherm EXEL II DNA sequencing kit (Epicentre) using template DNA and the primers listed in Table 2. Sequence gels. For the sequence gel analyses, 6% polyacrylamide (19:1) sequencing gels (0.4 mm thick) containing 8 M urea and 1⫻ TBE were used. Sample preparation for nucleotide resolution level mapping of DSBs. Genomic DNA was extracted from cells 9 h after meiosis induction. Approximately 1 ␮g of DNA was digested with the appropriate restriction enzyme, which cut DNA 200 to 250 nucleotides (nt) from the UAS. For the detection of the 5⬘ DSB end, DSB ends were filled in by using Phusion High-Fidelity DNA polymerase (Finnzymes), which produces a blunt end. All samples were purified by using a NucleoSpin Extract II kit (Macherey-Nagel) and resuspended in 4 ␮l of TE buffer and 2 ␮l of loading buffer (95% formamide, 10 mM EDTA, 0.1% xylene cyanol, 0.1% bromophenol blue). For the DNA sequence standard and the molecular weight marker, ⬃1 ␮g of mitotic DNA was digested and purified as well as meiotic samples. The appropriate amounts of either DNA sequence standard or molecular weight marker were then mixed with digested mitotic DNA and loading buffer. After heat denaturation for 5 min at 100°C, the DNA fragments were loaded and separated on a sequencing gel with 1⫻ TBE buffer at 1,700 V and 70 W for ⬃4 h. DNA fragments were electrotransferred to noncharged nylon membranes (GeneScreen) by using a TE 90–GeneSweep sequencing gel transfer unit (Hoefer Scientific Instruments) according to the supplied instructions. Probe preparation and hybridization. The preparation of strand-specific probes was performed as described previously (17) with some modifications. Briefly, DNA fragments used as templates for the preparation of 32P-labeled probes were amplified by PCR and gel purified. Strand-specific probes were prepared by multiple-cycle primer extension using an appropriate primer listed in Table 2. The reaction mixture in a final volume of 12.5 ␮l included 5 ng of template DNA fragment; 13.3 ␮M concentrations each of dGTP, dATP, and dTTP; 50 ␮Ci of [␣-32P] dCTP (6,000 Ci/mmol; GE Healthcare Bio-Sciences Corp.); 1⫻ Taq polymerase buffer; 2.5 mM MgCl2; and 0.625 U of Taq DNA polymerase. The reaction mix was incubated at 94°C for 4 min, and 30 cycles of primer extension (94°C for 15 s, 50°C for 30 s, and 72°C for 1 min) were performed. Prehybridization, hybridization, and washing were performed at 57°C. Image analysis. Parental and DSB bands were visualized by using a phosphorimager and quantified by using ImageQuant as described previously (25). Technically, the visualization of DSB signals mapped at nucleotide resolution requires the blots to be exposed to the phosphorimaging plate for a long time, with the consequence that the signal from the parental fragment (unbroken) is saturated. Therefore, the signals of the DSB peaks in the lane were quantified in two steps. Most accurately, the cleavage frequencies of individual sites (signal not saturated) relative to one another were estimated as the total of DSB signals per lane. Then, the parental (unbroken) signal was quantified, and the signal strength of each DSB band per lane on the Southern blots expressed as a ratio of the specific versus the total (i.e., DSB plus parental) signal per lane. The resulting values are represented by bars of various thickness in the experiments reported in Fig. 6 to 11. Thus, the final maps provide an accurate representation of the

VOL. 29, 2009


3503

TABLE 2. Primers used for sequence standards and probes Gene or locus tag

GAL2

YAL048c

YNL328c

GAL80

YCR048w

Primera

Sequenceb

Template for DNA sequence standardsc

HindF† HindR* AccF* AccR†

AGCTTCCTTCCGGGATGG GTCCAAGGCACATGGA TCCCATCTCAAGATGGGGAG CTACAAAACTTATCCTATCTCCAC

pGAL2pr#1

NheF† NheR* XbaF* XbaR†

CTAGCACAATGGAGTATGACGACTT GCAACATGGCTTTCCAA GGGAAACAGTACTATTTGATTTTG CTAGAAGCGAgACTAAGTAGGAG

pYAL048c-seq

MlyF† MlyR* NcoF* NcoR†

ACACAAGTTTTAGGTCGGC AGATGCTGGTTTGATGC GAACCATTTCACTAAAGAGG CATGGAGGATCATATCCCAATcAA

HaeF† HaeR* BglF* BglR†

CCTTCTTCCAATGCTAATCC GGGTATAAATATATATTACAAAGCGG TTCATAATAACCAAAAGCTCATAAC GATCTCTTGTTGTAGTCCATGAC

PCR product amplified by HaeF and BglR

BsuF† BsuR* XmnF* XmnR†

TCAGGACGTGACATTATTTCTTCTCTGATATATTTCC CGATAAAAAGAAGAGATGAGCAG ACCCTGGCAACATGC TCTTCGTCTTGCAACAAATCCTTAGTCTCCGTCATTC

–

pGAL2pr#1

pYAL048c-seq pYNL328c-seq pYNL328c-seq

PCR product amplified by HaeF and BglR

–

a

*, primer used for probe labeling; †, primer used for the preparation of sequence standard. b Lowercase letters indicate single-nucleotide polymorphisms in the SK1 strain background. c –, PCR product amplified by the following primers: GCACGGCTGCAGATTCAAAAGCTCCAACGC and CGTGGCGAAACCTCCATGGACTCCTGTGGC.

relative DSB frequencies at each phosphodiester bond (used here in all our conclusions), but the absolute frequencies are potentially overestimated. Estimation of DSB frequency at each specific sequence. For each of the 10 possible pairs of nucleotides and for each of the four possible nucleotides flanking an Spo11 cleavage site, we determined the number of times that sequence occurred within the genomic regions examined in the present study, namely, for pGAL2, the sequences shown in Fig. 6C excluding UAS-E and 15-nt sequences from UAS-E; for the YAL048c ORF, pYNL328c, and pGAL8, the sequences shown in Fig. 7B excluding the UAS and surrounding 15-nt sequences on both sides of UAS; for pYCR048, the region from positions ⫺296 to ⫺95 relative to the YCR048w translation start; for pGAL2_abcdE_csm1, the sequences 17 to 23 nt from UAS-E; for pGAL2_abcdE_csm2, the sequences 18 to 24 nt from UAS-E; and for pGAL2_abcdE_csm3, the sequences 17 to 26 nt from UAS-E. We counted the number of times a DSB could be detected there, irrespective of the relative intensities of DSBs at each cleavage position. The cleavage frequency was represented as the percentage of each possible site that was detectably cleaved (i.e., the percentage of the number cleaved/the number present in mapped regions). The observed cleavage frequency at each sequence was compared to the frequency expected if DSBs occurred at random (i.e., without regard to DNA sequence) within the mapped regions by using the Fisher exact test. The DSB intensities detected at each specific sequence were summated and divided by the number of times that sequence occurred within the mapped regions.

RESULTS Targeted DSBs are recombinogenic and induce COs efficiently. We previously reported that Gal4BD-Spo11 induces novel DSBs in the promoter region of the GAL2 gene, which contains UAS sites, and this DSB targeting strongly stimulates meiotic gene conversion (21). To test whether targeted DSBs also stimulate COs, the drug-resistant cassettes KanMX and HphMX were inserted upstream and downstream of the GAL2 gene to create appropriate flanking markers (Fig. 1A). Flanking markers did not create additional DSB sites within themselves or at the insertion junctions (data not shown). Tetrad analysis showed that the pADH1GAL4BD-SPO11

strain exhibited a three- to fourfold higher rate of COs than the control strains with wild-type SPO11 or with wild-type SPO11 and expressing only Gal4BD (Fig. 1B). It demonstrates that DSBs targeted by Gal4BD-Spo11 are recombinogenic and also stimulate COs. Next, to obtain constructs that target DSBs with even higher frequency, we tested the recombinogenic activity when the meiosis-specific REC8 promoter (pREC8) drives the fusion protein. Interestingly, in comparison with the pADH1GAL4BD-SPO11 strain, the pREC8GAL4BD-SPO11 strain showed an ⬃3-fold higher level of COs (Fig. 1B) and a 2-fold higher level of DSB frequency at the GAL2 promoter (⬃12% with pADH1 [21], ⬃21% with pREC8 [see Fig. 2C]). The configuration of flanking markers (cis or trans, see Fig. 1A) did not significantly affect the frequency of crossing over (Fig. 1B). The genetic data were confirmed by physically detecting recombinant products generated by COs between the flanking markers in SPO11 and pREC8GAL4BD-SPO11 strains (Fig. 1C). As expected, pREC8GAL4BD-SPO11 showed a high level (20 to 25%) of recombinant products, whereas that of SPO11 was low (⬃5%) (Fig. 1D). Importantly, although the frequency of recombination became very high, we noted that the distribution of parental and recombinant products in tetrads (mostly tetratype) remained similar to the wild type. As we previously observed for enhanced gene conversion (21), this result indicates that recombination at the same site on both sister chromatids in the same meiosis is rare. Since the pREC8GAL4BDSPO11 strain introduced a higher frequency of targeted DSBs at the GAL2 promoter, we used the pREC8GAL4BD-SPO11 construct for further analysis of targeted DSBs. Hereafter, pREC8GAL4BD-SPO11 is indicated as GAL4BD-SPO11.

3504


FIG. 1. Genetic and physical assay of crossing over in the GAL2 region. (A) The positions of flanking markers are shown. Numbers represent the positions where the flanking marker is inserted. The first A of the GAL2 ORF is ⫹1. (B) Genetic assay for COs. Diploids bearing flanking markers KanMX and HphMX upstream and downstream of GAL2 in trans or in cis configurations were sporulated and dissected. According to the phenotype of drug resistance against G418 and hygromycin B, the numbers of PD, T, and NPD were determined to calculate the genetic distance in centimorgans as described in the text: 100(T ⫹ 6 NPD)/2(PD ⫹ T ⫹ NPD). ORD8706 (SPO11 trans), ORD8714 (SPO11 cis), ORD8723 (SPO11, pADH1GAL4BD trans), ORD8715 (pADH1GAL4BD-SPO11 trans), ORD8704 (pREC8GAL4BDSPO11 trans), ORD8713 (pREC8GAL4BD-SPO11 cis), ORD8797 (pREC8GAL4BD-SPO11 cis abcde), and ORD9226 (pREC8GAL4BDSPO11 cis ab2cd2e) were used (Table 1). (C) Schematic drawing of detection of recombinant products. Genomic DNA digestion by XbaI and NruI makes parental (P1 and P2) and recombinant (R1 and R2)

MOL. CELL. BIOL.

DSB targeting by Gal4BD-Spo11 requires specific binding of the Gal4BD moiety to the UAS. To test whether specific binding of the Gal4BD moiety to UAS sites provides recruitment of Spo11 to the site and consequently production of novel DSBs, we abolished the binding function of the Gal4BD moiety or the UAS sequences. Amino acid substitutions in Gal4, C14Y and L32P (Fig. 2A), decrease Gal4-binding activity to the UAS (13). We took advantage of these substitutions to specifically abolish the DNA-binding function of the Gal4BD moiety within Gal4BD-Spo11 and checked natural and targeted DSB formation (Fig. 2). Mutations within the Gal4BD moiety, which abolish its function of DNA binding, eliminated targeted DSB formation within the YCR048w ORF and at the GAL2 promoter (both containing Gal4 UAS sequences), whereas natural DSB formation in the YCR048w promoter was not affected (Fig. 2). We next mutated the UAS sites of GAL2 promoter to test their requirement for DSB targeting. The Gal4 UAS is composed of the 17-bp recognition site (CGGN11CCG). The palindromic CGG triplets at the ends are essential for Gal4 binding in vitro (16). For mutating the UAS, we performed nucleotide substitution in one of the two essential triplets. There are five sequences that correspond to this consensus in the GAL2 promoter, and we refer to them as UAS-A, -B, -C, -D, and -E (Fig. 3A). UAS-D and -E, overlap each other. No mutation of a single UAS showed a significant decrease (or increase) of the DSB frequency in comparison to the wild type (Fig. 3B, ABCDE versus aBCDE, AbCDE, ABcDE, ABCdE, and ABCDe). In contrast, quadruple mutations in which four out of five UAS are inactivated decreased the frequency to different magnitudes (Fig. 3B, Abcde, aBcde, abCde, abcDe, and abcdE), suggesting that Gal4BD-Spo11 has affinity for each UAS and gives rise to DSB targeting. This is consistent with the observation of Gal4-dependent and -independent footprints at most UAS sites (12). The Abcde mutation showed discrete DSB signals near the UAS-A (Fig. 3B), suggesting that Gal4BD-Spo11 bound to respective UASs compete with one another to generate DSB nearby; therefore, the Abcde mutation liberates UAS-A from this competition and generates more DSBs in the vicinity of UAS-A. A strain with quintuple mutations, which has no canonical UAS within the GAL2 promoter, still targeted DSBs (7.1%) (Fig. 3B, abcde). Since wildtype Spo11 introduced a low level of DSB at the GAL2 promoter with the mutation abcde (2.5% at 9 h in meiosis [data not shown]), the DSBs observed in the GAL4BD-SPO11 cells are likely to be Gal4BD dependent. These observations raised a possibility that the length of the sequence between essential triplets could be variable in vivo to create degenerate Gal4

fragments of different sizes. These fragments were detected by Southern blot using a DNA probe located in the GAL2 ORF. (D) Detection of recombinant fragments. Genomic DNA was prepared from SPO11 (ORD8706) and pREC8GAL4BD-SPO11 (ORD8704) diploids taken at the indicated times after transfer to sporulation medium, and parental and recombinant fragments were detected by Southern blot analysis. After digestion by XbaI and NruI, DNA fragments were separated in 0.6% agarose. Signal intensities of P1, P2, R1, and R2 shown in the panel were quantified, and the percentage of recombinant product was calculated: 100 (R1 ⫹ R2)/(P1 ⫹ P2 ⫹ R1 ⫹ R2).

VOL. 29, 2009

FIG. 2. Requirement for Gal4BD for DSB targeting. (A) The configurations of Gal4BD-Spo11 and Gal4BD105-Spo11 and the positions of mutations introduced in the Gal4BD moiety are shown. The DNAbinding function of the Gal4BD moiety within Gal4BD-Spo11 was specifically abolished by amino acid substitutions in Gal4, C14Y, and L32P, which are reported to decrease Gal4-binding activity to the UAS (13). (B) Targeted and natural DSB formation in YCR048w. DSBs in the rad50S strains were analyzed by Southern blot hybridization. The positions of ORFs (gray arrows), restriction sites, Gal4 UAS (horizontal bar), and probes are indicated at the right of each gel image. DSB signals were quantified and are represented as the percentage of signal intensity per lane. Genomic DNA from rad50S diploids, SPO11 (ORD8421), GAL4BD-SPO11 (ORD8050), gal4BDC14Y-SPO11 (ORD9216), and gal4BDL32P-SPO11 (ORD9217) was digested by AseI and separated on a 0.8% agarose gel. The blot was probed with a YCR048w internal fragment. Both mutations eliminated targeted DSB formation at the YCR048w ORF, whereas natural DSB formation in the YCR048w promoter was not affected. (C) Targeted DSB formation in GAL2 strains identical to strains in Fig. 2B. Genomic DNA was digested by BanII and separated on a 1.0% agarose gel. The blot was probed with a GAL2 internal fragment. Both mutations eliminated targeted DSB formation at the GAL2 promoter.

UAS recognition site(s) by combination(s) of the remaining, more distant, intact half-site triplets. Therefore, we introduced additional mutations (Fig. 3A, b2 and d2) to eliminate all possible combinations. These mutations showed an insignificant (Fig. 3B, ab2cde, 4.4%) and significant (Fig. 3B, abcd2e, 2.6%,


3505

and ab2cd2e, 3.2%) decrease of DSB frequency compared to that of the quintuple mutation (abcde). Thus, this suggests the existence of functional but degenerate Gal4 UAS sites in which the distance between the conserved triplet can be longer than 11 nt. This may explain why some Gal4BD-Spo11 DSB targeted sites do not contain the canonical UAS (24). The effect of UAS mutations in the GAL2 promoter was further assessed by the genetic assay used above. Consistent with DSB frequency, quintuple (abcde) and septuple (ab2cd2e) mutations showed approximately two- and fivefold reduction in percent CO compared to the wild type (ABCDE), respectively (Fig. 1B). To further examine the dependency on UAS, we mutated the UAS in two recently discovered DSB target sites, YAL048c and YNL328c, which each have a single UAS in the ORF and promoter, respectively (24). Gal4BD-Spo11-specific DSB targeting was observed near the UAS site. The mutation of each UAS drastically decreased the frequency of DSB targeting at both loci (Fig. 4). Taken together, the data show that most of the targeted DSB formation requires a Gal4 UAS site. We conclude that the specific binding of the Gal4BD moiety to a UAS recruits the Spo11 moiety to target DSBs in the vicinity of the UAS. High-resolution mapping of targeted DSBs. Naturally, meiotic DSBs show multiple cleavage positions that are irregularly spaced within a hot spot (7, 17, 28). We sought to determine whether the DSB distribution within a hot spot is due to heterogeneous binding of Spo11 to DNA. For this, we restricted Spo11 binding to a single site (only one Gal4 UAS) and mapped the resulting DSBs to a nucleotide resolution. We chose to focus on the GAL2 promoter, which contains five canonical UASs, as well as some degenerate UASs. First, we mapped targeted DSBs in the wild-type GAL2 promoter (ABCDE) at modestly higher resolution using highconcentration agarose gels (1.7%). Hereafter, the DSB mapping using this high concentration of agarose gel or a small polyacrylamide gel (see below) are referred to as “high-resolution mapping.” In contrast, the DSB mapping using a sequencing gel is referred to as “nucleotide resolution mapping.” Six DSB sites distributed around each UAS were observed (Fig. 3C, numbered 1 to 6). The strongest DSB (number 5) was targeted between UAS-C and -D. In order to obtain a situation with a strong DSB coming from only one UAS, we mapped targeted DSBs in the series of UAS mutants used above (Fig. 3C and data not shown). When only UAS-D (abcDe) or -E (abcdE) was present, only one strong DSB remained close to UAS-D and -E (Fig. 3C) and, importantly, in these two mutants, the DSB pattern was different from the residual DSB pattern observed in the quintuple mutant (abcde). This suggests that the UAS sites, both canonical and degenerate, compete with one another for binding to Gal4BD-Spo11 and therefore compete for directing subsequent DSB formation. It is likely that UASs with strong affinity would be preferentially bound by Gal4BD-Spo11 and be able to eliminate the DSBs from lower-affinity, degenerate UASs. Since we used GAL4⫹ strains in the present study, the chromatin structure required for DSB targeting in the GAL2 promoter might be established by the wild-type Gal4 protein bound to UASs. However, this possibility is unlikely because the Gal4BD construct did not stimulate recombination compared to the wild type (Fig. 1B),

3506


MOL. CELL. BIOL.

FIG. 4. Targeted DSBs in YAL048c and YNL328c ORFs. (A) Targeted DSB formation in the vicinity of a UAS in the YAL048c ORF. Genomic DNA was prepared from SPO11 (ORD8421), GAL4BDSPO11 (ORD8050), and GAL4BD-SPO11 with the UAS mutation in the YAL048c ORF (ORD9225). The DNA was digested with PstI and separated on a 1.0% agarose gel. The mutation introduced in the UAS is indicated in lowercase under the gel image. (B) Targeted DSB formation in the vicinity of a UAS in the YNL328c promoter. Genomic DNA was prepared from SPO11 (ORD8421), GAL4BD-SPO11 (ORD8050), and GAL4BD-SPO11 strains with the UAS mutation in the YNL328c promoter (ORD9221). The DNA was digested with HincII and separated on a 1.0% agarose gel. The mutation introduced in the UAS is indicated in lowercase under the gel image.

nor did the strain with gal4 deleted reduce DSB targeting in the GAL2 promoter (data not shown). Therefore, the effects of UAS mutations on DSB targeting likely reflect direct effects on Gal4BD-Spo11 binding to UASs. We thus pursued the use of the abcDe and abcdE mutants for nucleotide resolution mapping. DSB targeting by Gal4BD-Spo11 is restricted to a small window of space near the UAS. To measure precisely the distance between DSB sites and the UAS, we first performed DSB 3⬘ end mapping at single-nucleotide resolution in the wild-type GAL2 promoter using AccI and HindIII digests and strand-specific probe A (see Fig. 6A, ABCDE; for the procedure, see Fig. 5 and Materials and Methods). Meiosis-specific DSBs were detected at the previously designated bands 2 to 5 FIG. 3. Requirement of Gal4 UAS for DSB targeting. (A) Positions of five Gal4 UAS sites in the GAL2 promoter—UAS-A, -B, -C, -D, and -E—and mutations at Gal4 UAS sites. The upper portion of the panel shows the position of each UAS. Numbers represent the position of the 5⬘ C in each UAS, when the first A of GAL2 ORF is ⫹1. UAS-D and -E are overlapping. The lower portion of the panel shows the mutations introduced in each UAS. One of two essential triplets (open rectangle) was mutated. In mutants a, c, and e, the triplet on the right was mutated, whereas the left triplet was mutated in mutants b and d. In mutants b2 and d2, both triplets were mutated. Each mutation indicated as lowercase letters creates a novel restriction enzyme site (a, BspHI; b, PvuI; c, EcoRI; d, SnaBI; e, XbaI; b2, NdeI; and d2, DraI). (B) Targeted DSBs frequency in a series of UAS mutants at the GAL2 promoter. DSBs in the rad50S strains were analyzed by Southern blot hybridization using a 0.8% agarose gel. The positions of ORFs (gray arrows), restriction sites, Gal4 UAS (horizontal bar), and probes are indicated at the right of gel image. DSB signals were quantified and represented as % of signal intensity per lane. Genomic DNA were prepared from GAL4BD-SPO11 with pGAL2_ABCDE (ORD8050), aBCDE (ORD8731), AbCDE (ORD8758), ABcDE (ORD8759), ABCdE (ORD8741), ABCDe (ORD8760), Abcde (ORD9200), aBcde (ORD9201), abCde (ORD9208), abcDe (ORD8799), abcdE

(ORD9207), abcde (ORD8796), ab2cde (ORD9209), abcd2e (ORD9210), ab2cd2e (ORD9220), and SPO11 with pGAL2_ABCDE (ORD8421) diploids taken at 6 h after transfer to the sporulation medium. The graph shows the average of quantified DSB frequency and the standard deviations of three independent experiments. (C) High-resolution mapping of targeted DSBs in the GAL2 promoter. The positions of ORFs (gray arrows), restriction sites, Gal4 UAS (black boxes), and probes are indicated on the right of the gel image. Genomic DNA was prepared from cells in meiosis taken at the indicated times after transfer to sporulation medium. DNA was digested with AseI and AlwNI and separated on a 1.7% agarose gel. The blot was probed with a DNA fragment corresponding to the sequence from positions ⫺244 to ⫹147 relative to the GAL2 translation start. Horizontal arrows indicate six targeted DSB sites around each UAS. Genomic DNA from pGAL2_abcde (ORD8796) was digested with AlwNI and either BspHI, PvuI, EcoRI, SnaBI, or XbaI, and the mixture was used as a molecular weight marker for each UAS (lane marked M).

VOL. 29, 2009

FIG. 5. The procedure of nucleotide level resolution mapping of 5⬘ and 3⬘ DSB ends. Genomic DNA from meiotic cells was digested with appropriate restriction enzymes. After separation on a sequencing gel and blotting, 3⬘ DSB ends were detected with a strand-specific DNA probe. For the detection of 5⬘ ends, 3⬘ ends were filled in with a polymerase to mimic the size of 5⬘ ends.

from the high-resolution mapping (Fig. 3C). The GAL2 promoter mutants abcDe and abcdE showed a cluster of DSB signals at the band 5 position, which were observed as a single band in the high-resolution mapping (Fig. 3C and 6A). It is likely that Gal4BD-Spo11 bound to UAS-D and/or UAS-E prefers to cleave DNA at this position. In addition, the mutant abcDe retained somewhat weaker signals at the position of band 4 (Fig. 6A). Next, DSBs at the band 5 position were mapped to detect both 5⬘ and 3⬘ DSB ends with an AccI digest and probe A (Fig. 6B, left panel). Signals from 5⬘ DSB ends are shifted to a higher molecular weight compared to those of the 3⬘ DSB ends. DSBs were also mapped with HindIII digestion and probe B on the other side of DSBs (Fig. 6B, right panel). As previously reported for the wild-type Spo11, our data confirm that the DSB ends produced by Gal4BD-Spo11 contain a 2-nt 5⬘ overhang (Fig. 6C) (17). However, some 3⬘ DSB ends, which we detected with a single probe, were not identical to the 5⬘ ends. For example, the mutant abcdE showed three discrete signals by the 3⬘ end detection, whereas only one signal was exhibited by the 5⬘ end detection (Fig. 6B). The 3⬘ end may be filled in either in vivo or during the preparation of genomic DNA. Therefore, to accurately map DSBs on both strands, we performed single-nucleotide resolution mapping by using two different restriction enzyme digestions and probes on both sides of the DSBs. We also mapped the targeted DSBs in the quadruple mutants abcDe and abcdE (Fig. 6B and C). DSBs were located nt 17 to 24 and nt 15 to 22 ([nt 17-24/15-22 nt] 5⬘ end and 3⬘ end, respectively) and nt 44-45/42-43 from UAS-D in the abcDe mutant. The abcdE mutant showed a cluster of DSBs at nt 18-24/16-22 from UAS-E and a strong DSB at nt 22/20. Taking these data together, we conclude that Gal4BD-


3507

Spo11 has a prominent tendency to cleave ⬃20 nt from its Gal4BD-driven binding site. To assess whether this ⬃20-nt distance is a general phenomenon, targeted DSBs were mapped at three other sites, YAL048c, YNL328c, and the GAL80 locus, which all have a single Gal4 UAS. DSBs were located on both sides of the UAS in all three loci. Again, the distances between major DSB sites and the UASs were ⬃20 nt (Fig. 7A and B). When we aligned all of the targeted DSBs mapped at singlenucleotide resolution according to the distance from their UAS (Fig. 7C), we found that most of the DSBs were concentrated within a zone located 17 to 25 nt from UAS (for 5⬘ ends) that appears to define a “DSB targeting window.” Two notable exceptions are the DSBs located 44 to 45 nt from UAS-D in the GAL2 promoter with the quadruple mutation abcDe and the DSBs located 29 to 32 nt from the YNL328c UAS. These outlying DSBs might be the result of degenerate UAS, as suggested above (Fig. 3B and C). These results demonstrate that Gal4BD-Spo11 binding to a single site restricts the distribution of DSBs to a narrow window near the binding site. This pattern is in marked contrast to the natural DSBs induced by Spo11, which are spread over a wider region within a hot spots. Our findings thus raise the possibility that the natural multiplicity of DSBs reflects the heterogeneous binding of Spo11 within hot spots (see also the Discussion). The structure of the Gal4BD fusion partner does not solely determine the position of the DSB targeting window. We next focused on the molecular mechanism that establishes the DSB targeting window. A simple model could be that the structure of the Gal4BD moiety determines the distance between the cleavage site and the UAS. Gal4 binds to DNA as a homodimer, including a DNA recognition element (amino acid residues 7 to 40) and a dimerization domain (amino acid residues 50 to 94) (Fig. 2A) (11). The cocrystal structure of the Gal4BD dimer (residues 1 to 100) and DNA containing a UAS shows that the two globular DNA recognition elements of a dimer make base-pair-specific contacts to the CGG triplets at the ends of the UAS, and the dimerization domain forms an intertwined helical bundle to contact the phosphate backbone within the inner 11 bp (11). Although the structure of the region after the dimerization domain has not yet been solved, it is known that a Gal4BD construct that is shorter than the one we used in the present study (105 instead of 147 amino acid residues) has full function for dimerization and DNA binding (3). To test whether the length of the region linking the dimerization domain and Spo11 determines the range in which Spo11 binds and cuts DNA, we examined DSB targeting with a strain expressing Gal4BD105-Spo11 (Fig. 2A), which lacks 42 amino acid residues after the dimerization domain. In contrast to our prediction, Gal4BD105-Spo11 did not move the DSBs closer to the UAS but, instead, narrowed the DSB targeting window. Both GAL4BD-SPO11 and GAL4BD105-SPO11 showed nearly identical DSB distributions in the GAL2 promoter with UAS mutations abcdE (pGAL2_abcdE), except that a DSB site at nt 18/16 from UAS-E disappeared in GAL4BD105-SPO11 cells (Fig. 8B and C). In the GAL80 promoter, the ORF-distal DSB at nt 18/16 also disappeared, and ORF-proximal DSBs at nt 18/16 and nt 25/23 were decreased in GAL4BD105-SPO11, while other

3508


MOL. CELL. BIOL.

FIG. 6. Single-nucleotide resolution mapping of targeted DSBs of the wild type and UAS mutants in the GAL2 promoter. Strains with the rad50S mutation were used. The positions of ORFs (gray arrows), restriction sites, Gal4 UAS (black boxes), and probes are indicated beside each gel image. Lanes marked GC and AT are sequencing standards containing G and C standards and A and T standards, respectively. The strains used here are identical to those used in Fig. 3. (A) The 3⬘ DSB ends were mapped as described in Fig. 5 using a strand-specific probe adjacent to AccI (probe A). The positions of each UAS are indicated as open boxes. The numbers of bands correspond to those of Fig. 3C. (B) The 5⬘ and 3⬘ ends of DSBs were analyzed as described in Fig. 5 using strand-specific probe adjacent to AccI (probe A in the left panel) and HindIII (probe B in the right panel). (C) Location of DSBs in the GAL2 promoter region. The numbers represent the distance from either UAS-D or -E. The thickness of the bars provides representation of the signal strength of each DSB band on the Southern blots. The range of the signal intensity (% DSB/lane) is indicated in the box (see Materials and Methods). Since the 3⬘ DSB end may be filled in either in vivo or during the preparation of genomic DNA (i.e., in panel B the mutant abcdE showed discrete three signals by the 3⬘ end detection; instead, only one signal was exhibited by the 5⬘ end detection), 5⬘ DSB ends mapping using two different restriction enzyme digestions and probes on both side of the DSBs were used to show accurate mapping DSBs without this ambiguity.

VOL. 29, 2009


3509

FIG. 7. Nucleotide resolution mapping of targeted DSBs in the YAL048c ORF, the YNL328c ORF, and GAL80. (A) The mapping of targeted DSBs in the YAL048c ORF, the YNL328c promoter and the GAL80 promoter were performed as described in Fig. 6. Genomic DNA was prepared from GAL4BD-SPO11 (ORD8050). (B) Location of targeted DSBs mapped in panel A. Other features are as described for Fig. 6C. (C) Targeted DSBs mapped in Fig. 6 and panel B are shown. Numbers correspond to the distance from the UAS. -L and -R indicate DSBs mapped on the left and the right ends of the UAS in panel B.

DSBs were unchanged (Fig. 9B and C). In contrast, all of the natural DSB sites in the YCR048w promoter were identical between the GAL4BD-SPO11 and GAL4BD105-SPO11 strains (Fig. 10B and C), excluding the possibility that the 42-aminoacid deletion within the Gal4BD moiety affects the DNAbinding function of Spo11 itself. Thus, the Gal4BD moiety restricts DSB formation within the DSB targeting window and the size of window is slightly affected by the deletion of a part of Gal4BD moiety. However, the structure of the Gal4BD moiety by itself is not sufficient to account for the establishment of the DSB targeting window. Possible explanations are offered below. The DNA-binding function of Spo11 contributes to the selection of cleavage sites in the DSB targeting window. Gal4BDSpo11 somehow restricted DSB formation within the DSB targeting windows, but the distribution of DSBs inside the window was different in each DSB targeting site (Fig. 7C). Since amino acid substitutions in the putative DNA-binding

surface of Spo11 drastically alter the local distribution of DSBs within a natural DSB hot spot, Spo11 itself has been proposed to contribute to the selection of DSB formation sites (8). To determine whether DSB distributions within the DSB targeting window are dependent on Spo11’s function to select the sites of cleavage, we introduced amino acid substitutions into the Spo11 moiety of our Gal4BD-Spo11 construct. We picked four mutations, spo11R131K, spo11E235A, spo11F260R, and spo11D290A, which generate unusual DSB patterns at site I of the his4::LEU2 hot spot (8). Our Gal4BDSpo11 mutants also showed different DSB patterns from Spo11 in the natural DSB formation site, YCR048w (Fig. 10). Gal4BD-Spo11R131K, Gal4BD-Spo11E235A, and Gal4BDSpo11D290A introduced DSBs at fewer sites. Gal4BDSpo11F260R introduced DSBs at as many sites as Spo11 but drastically changed the DSB pattern (Fig. 10). The effect of mutations in Gal4BD-Spo11 were similar and in good agreement with mutations in Spo11 at his4::LEU2 hot spot (8) and

3510


MOL. CELL. BIOL.

FIG. 8. Effects of shortening the Gal4BD moiety and mutations in the Spo11 moiety on DSBs targeted at pGAL2_abcdE. Targeted DSBs were mapped in GAL4BD-SPO11 (ORD9207), shortened Gal4BD moiety GAL4BD105-SPO11 (ORD9270), and GAL4BD-spo11 mutants (R131K, ORD9271; E235A, ORD9274; R260R, ORD9272; and D290A, ORD9273) at pGAL2_abcdE. (A) High-resolution mapping of DSBs targeted by modified Gal4BD-Spo11 in pGAL2_abcdE (acrylamide gel). The positions of the probe, restriction sites, and ORF are indicated on the right of the gel image. (B) Single-nucleotide resolution mapping of DSBs targeted by modified Gal4BD-Spo11 in pGAL2_abcdE. Open arrowheads indicate DSB positions which observed in GAL4BD-SPO11 but disappeared in GAL4BD105-SPO11. Filled circles indicate DSB positions specifically generated in GAL4BD-spo11F260R. (C) Location of targeted DSBs observed in panel B. Other features are as described for Fig. 6C.

at the YCR048w hot spot (S. Keeney, unpublished data), respectively. We also examined targeted DSB distributions of each GAL4BD-spo11 mutant at pGAL2_abcdE and pGAL80. Spo11 R131K, E235A, and D290A mutations decreased the frequency of targeted DSBs but did not seem to alter their position when looked at by high-resolution mapping (Fig. 8A and

9A). Nucleotide resolution mapping confirmed that the amount was decreased, but distribution was only slightly affected, if at all (Fig. 8 and 9). In contrast, Gal4BDSpo11F260R changed the DSB intensity at several positions and created new DSB positions: nt 25/23 from UAS-E at pGAL2_abcdE and nt 24/22 and nt 25/23 from the UAS at pGAL80. In summary, our results from two targeted DSB loci

VOL. 29, 2009


3511

FIG. 9. Effects of shortening the Gal4BD moiety and of mutations in the Spo11 moiety on DSBs targeted at the GAL80 promoter region. Targeted DSBs were mapped in SPO11 (ORD8421), GAL4BD-SPO11 (ORD8050), GAL4BD105-SPO11 (ORD9228), and GAL4BD-spo11 mutants (R131K, ORD9260; E235A, ORD9262; R260R, ORD9258; and D290A, ORD9240) in the GAL80 promoter region. (A) High-resolution mapping of DSBs targeted by modified Gal4BD-Spo11 in the GAL80 promoter region (acrylamide gel). The position of the probe, restriction sites, UAS, and ORF are indicated on the right of the gel image. (B) Single-nucleotide resolution mapping of DSBs targeted by modified Gal4BD-Spo11 in the GAL80 promoter region. Open arrowheads indicate DSB positions which were observed in GAL4BD-SPO11 but disappeared in GAL4BD105-SPO11. Open circles indicate DSB positions which were observed in GAL4BD-SPO11 but disappeared in GAL4BD-spo11F260R. Filled circles indicate DSB positions specifically generated in GAL4BD-spo11F260R. (C) Location of targeted DSBs observed in panel B. Other features are as described for Fig. 6C.

3512


MOL. CELL. BIOL.

VOL. 29, 2009


indicate that, with or without mutation in the Spo11 moiety, DSB targeting is restricted to a defined window. However, inside the window, Spo11 itself contributes to the selection of specific cleavage positions. Spo11 has a subtle sequence preference for cleavage. These observations raise the question of how the Spo11 moiety chooses cleavage sites within the DSB targeting window. Because nucleotide resolution mapping has thus far revealed no sequence consensus, Spo11 is thought to have no nucleotide sequence specificity for cleavage site selection. However, direct introduction of mutations at cleavage sites has not yet been performed. Therefore, we took advantage of the highly specific targeted DSB at nt 22/20 from UAS-E in pGAL2_abcdE (Fig. 6B and 8B). We introduced nucleotide substitution(s) either at the 3⬘ end (cleavage site mutation 1 [csm1]), the 5⬘ end (csm2), or both ends (csm3) of the cleavage site (Fig. 11A). Nucleotides occurring at low frequency near DSB ends of natural DSB hot spots (the ARG4 promoter and the YCR048w promoter) (17) were introduced. These nucleotide substitution(s) affected the position of the DSB (Fig. 11B and C). A C-to-G substitution at the 3⬘ end (csm1) eliminated the original DSB at nt 22/20 from UAS-E and split it into two new DSBs flanking the original position. On the other hand, a T-to-G substitution at the 5⬘ end (csm2) did not affect the original DSB, but it did generate a new DSB position at an adjacent nucleotide. Multiple substitutions at both the 5⬘ and 3⬘ ends (csm3) abolished the original DSB position and created two new positions. These results indicate that Spo11 has a subtle sequence preference and is less likely to cleave at a position where there is a G residue 5⬘ of the scissile bond (i.e., to leave a G as the 3⬘-most residue at a DSB) (Fig. 11A). If so, Spo11 would be unlikely to introduce a DSB at the sequence 5⬘-G/NNC-3⬘ (the shill indicates the cleavage position), which we did not find among the 67 DSB sites we mapped in the present study (see discussion). However, since some DSBs with a G nucleotide at the 3⬘ end are observed, this sequence preference of Spo11 might not be absolute. DISCUSSION In this study, we showed that Spo11 fused with Gal4BD introduced DSBs restricted to a DSB targeting window nearby the Gal4 UAS and that inside the DSB targeting window, the Spo11 moiety of the fusion protein chooses preferred sequences to cleave and introduces multiple DSBs. It is clear that Gal4BD does not affect Spo11 function when the fusion protein introduces natural DSBs, since the DSB distribution of Gal4BD-Spo11 showed a pattern almost identical to that of Spo11 at the YCR048w promoter (Fig. 10C). In the targeted DSB sites, each mutation within the Spo11 moiety affected DSB frequency and/or cleavage position (Fig. 8 and 9). These mutations are located in putative DNA binding surface of

3513

Spo11 protein, and the effect of each mutation showed somewhat similar trends in both natural and targeted DSB sites. Taken together with the full requirement of all known DSB proteins (21), we favor the idea that Gal4BD-Spo11 behaves like Spo11 even when its DSB introduction is restricted to a DSB targeting window, and the DNA-binding feature of Spo11 protein dictates the cleavage position. If these are the case, the fact that Gal4BD-Spo11 restricts DSB formation within the DSB targeting window would suggest that the multiplicity of the natural DSB distribution is caused by the heterogeneous binding of Spo11 to DNA within a hot spot based on a weak DNA sequence specificity. However, because of the neomorphic nature of Gal4BD-Spo11, it is difficult to completely dissociate characteristics of Spo11 from effects of Gal4BD. Since other proteins than Spo11 are required for DSB formation, it remains possible that the local multiplicity of DSBs is caused by heterogeneous binding of another DSB protein and that Spo11 and/or DSB protein complex normally slide along the DNA. In any case, it is worth noting that this feature of meiotic DSB formation may contribute to diversify recombination initiation sites in meiosis with the evolutionary benefit of locally increasing the functional robustness of the active region. It avoids a situation in which a single mutation(s) in a recombination hot spot would be likely to eliminate all DSB formation and, therefore, might contribute to keeping a recombination hot spot hot. In addition, analysis of our DSB targeting system suggests a subtle sequence preference of Spo11 and provides a method to precisely approach this issue. Molecular architecture of DSBs targeting by Gal4BD-Spo11 and establishment of the DSB targeting window. The Gal4BD moiety, presumably binding to a UAS via its DNA recognition domain, may sterically prevent Spo11 itself from binding to the UAS, so that Spo11 binding instead takes place next to the UAS occupied. A shortening of the protein region linking the dimerization domain and Spo11 narrowed the size of the window but did not slide the window closer to the UAS (Fig. 8 and 9). The disappearance of the UAS-distal DSBs in this case is relatively easy to interpret: the shortening of the tether may simply limit Spo11’s access to DNA at more distant sites. However, shortening of the tether also eliminated DSBs proximal to the UAS. This result led us to posit that some proteins could exist in the region between the edge of a UAS and the DSB sites. It could be Spo11 itself and/or other required DSB proteins. A Spo11 monomer may occupy ⬃10 bp on DNA as predicted from comparison to the related TopoVIA structure (14). In addition, since nine other DSB proteins are absolutely required for DSB targeting, these factors and a Spo11 moiety are likely to form a complex. Spo11, and in particular the catalytic tyrosine residue of Spo11, needs access to DNA beyond this complex and, therefore, the limitation of a short tethering region could make access proximal to the UAS sites

FIG. 10. Effects of shortening the Gal4BD moiety and mutations in the Spo11 moiety on formation of natural DSBs in the YCR048w promoter region. Strains are the same as in Fig. 9. (A) High-resolution mapping of natural DSBs by Spo11 and modified Gal4BD-Spo11 in the YCR048w promoter region (acrylamide gel). The position of the probe, restriction sites and ORF are indicated on the right of the gel image. (B) Singlenucleotide resolution mapping of natural DSBs by Spo11 and modified Gal4BD-Spo11 in the YCR048c promoter region. (C) Location of natural DSBs observed in panel B. (D) Overview of natural DSB positions shown in Fig. 10C. Numbers correspond to positions relative to the translation start of YCR048w. Other features are as described for Fig. 6C.

3514


MOL. CELL. BIOL.

FIG. 11. Effect of mutations at a DSB site. (A) Mutations at a frequent DSB targeting site in pGAL2_abcdE. Cleavage site mutations (csm1, csm2, and csm3) are introduced either at the 3⬘ DSB end, the 5⬘ DSB end, or both ends. (B) Targeted DSBs of GAL4BD-SPO11 at pGAL2_abcdE with or without cleavage site mutations (wild type [WT], ORD9207; csm1, ORD9233; csm2, ORD9234; csm3, ORD9235). (C) Location of targeted DSBs observed in panel B. Mutations are indicated in lowercase with a gray background. Other features are as described for Fig. 6C. (D) Biased sequence context around cleavage sites. For each of the 14 pairs of nucleotides flanking a Spo11 cleavage site, we determined the number of times that sequence occurred within the genomic regions and counted the number of times a DSB could be detected there, irrespective of the relative intensity of DSBs at each cleavage position. Black bars indicate the percentage of each possible site that was cleaved; numbers next to the bars show the frequencies (number cleaved/number present in mapped regions). The observed cleavage frequency at each sequence was compared to the frequency expected if DSBs occurred at random (i.e., without regard to the DNA sequence) within the mapped regions (*, P ⬍ 0.05; and **, P ⬍ 0.01 [Fisher exact test]). Gray bars indicate the percentage of mapped DSBs that occurred within the indicated sequences, normalized for the estimated intensity of DSBs (see Material and Methods for details).

difficult. Thus, the distance between targeted DSB sites and the UAS might represent a space that is required for Spo11 itself or a Spo11 containing complex to assemble on DNA. If DNA binding by Gal4BD-Spo11 restricts the orientation of the DSB protein complex on DNA, it would be interesting to study the

proposed asymmetric feature of meiotic DSB formation (e.g., endonucleolytic release of Spo11-oligo DNA complex) (19). In addition, it is interesting that among DSB sites mapped within the DSB targeting window, relatively strong DSBs were observed at nt 22/20 from the UAS (Fig. 7C). In addition,

VOL. 29, 2009

FIG. 12. Molecular architecture of Gal4BD-Spo11. Gal4BD-Spo11 targets DSBs ⬃20 nt from UAS. DSB targeting by Gal4BD-Spo11 defines a DSB targeting window, nt 18-25/16-23 from the UAS. The schematic drawing represents Gal4BD-Spo11 on B-form DNA (10.5 bp/turn). In the drawing, the Gal4BD moiety (green; sphere, DNA recognition domain; cylinder, dimerization domain; line, neck tethering Gal4BD and Spo11) bound to Gal4 UAS (gray nucleotides indicate the essential triplets of the UAS) and Spo11 moiety (red and gray) introducing DSB nt 22/20 from UAS, which was frequently observed in the present study, are represented. Dots on DNA represent the DSB targeting window. At three turns away from the center nucleotide (red) of the UAS, a catalytic tyrosine residue (yellow triangle) of one Spo11 moiety faces the same side of the DNA helix with the dimerization domain of the Gal4BD dimer.

mutations at the cleavage site of pGAL2_abcdE affected the position of a DSB, but major DSBs were still positioned around nt 22/20 from UAS-E. This implies that, inside the DSB targeting window, the position around nt 22/20 from the UAS may be preferred when Gal4BD-Spo11 targets a DSB. In Fig. 12, we show a model in which Gal4BD-Spo11 introduces a DSB at nt 22/20 from the edge of the UAS on B-form DNA. This model is based on a cocrystal structure of Gal4BD with DNA (11) and Spo11 structure as inferred from the structure of TopoVI A subunit (14). In this conformation, the dimerization domain of the Gal4BD moiety and a catalytic tyrosine residue of one of the two Spo11 moieties, which cleaves 22 nt from the UAS, face the same side of the DNA helix. The limitation of the linker length may limit some conformations in which the dimerization domain and the tyrosine residue are positioned at opposite sides of the DNA helix. These structural constraints may provide the positional preference for DSB targeting seen with Gal4BD-Spo11, while the shorter linker form of Gal4BD105-Spo11 may not allow the Gal4BD moiety and Spo11 moiety to reach the same site(s). Thus, the DSB targeting window may be established under steric, distance (the length of the linker extension), and rotational constraints. Gal4BD-Spo11 targeted DSBs on both sides of the UAS at YAL048c, YNL328c, and GAL80, albeit with different spatial patterns and overall frequencies (Fig. 7). At the pGAL2_ abcdE, in contrast, we detected the targeted DSBs only on one side (Fig. 6 and 8). The structure of UAS and the binding of Gal4 to UAS are symmetrical. Therefore, it is likely that there are local features (perhaps the chromatin structure) that affect local accessibility of the Spo11 moiety to DNA and thus lead to asymmetric DSB targeting. DSB site choice by Spo11 and DSB proteins. By singlenucleotide resolution mapping of meiotic DSBs with Spo11 and a series of Gal4BD-Spo11 mutants, we provide further details of Spo11’s cleavage position choice. At a natural DSB hot spot in the YCR048w promoter, GAL4BD-SPO11 and GAL4BD105-SPO11 lost several DSB positions that are ob-


3515

served in SPO11 cells, but at most DSB positions both versions of Gal4BD (147 and 105) had no effect on the natural DSB distribution. These results indicate that Gal4BD-Spo11 is close to, but not fully, wild type in terms of cleavage position choice (Fig. 10). The distributions of natural DSBs were further affected by Spo11 mutations (Fig. 10). Among the mutations tested, three amino acid substitutions (R131K, E235A, and D290A) showed similar patterns of non-UAS-targeted DSB distributions in which one cleavage position became prominent and the others disappeared or decreased (Fig. 10A). In contrast, the non-UAS-targeted DSB distribution of GAL4BDspo11F260R cells seemed to be completely different in comparison to SPO11, GAL4BD-SPO11, and GAL4BD105-SPO11 (Fig. 10A). Similar patterns of the DSB distribution at YCR048w are observed with another version of Spo11 (Spo11HA3His6) harboring the same amino acid substitutions (Keeney, unpublished). By looking closely at these DSB patterns, we noticed that the GAL4BD-spo11 R131K, E235A, and D290A mutants defined two distinct types of DSB sites: DSB sites whose signal decreased or disappeared and sites whose signal was maintained or even increased (Fig. 10C). Unexpectedly, GAL4BDspo11F260R did not create new DSB positions, relative to either SPO11 or GAL4BD-SPO11, but the relative strengths of each signal were altered. Thus, the overall distribution of DSBs within the hot spot changes substantially in this mutant. Since, three separate SPO11 mutations—R131K, E235A, and D290A—showed a similar phenotype, it may be that these mutations cause Spo11 to lose or reduce a common function, presumably DNA binding. In contrast, the mutation F260R might increase this function (Keeney, unpublished). If so, what is the nature of the DSB position that remained prominent in the mutants R131K, E235A, and D290A? It might be a place in which the DSB protein complex prefers to assemble. Thus, there might be two types of cleavage position choice: type 1 defines positions preferred by Spo11 itself, while type 2 defines positions dictated primarily by factors other than Spo11. Since Spo11 appears to be heterogeneous in choosing positions to bind, the distribution of type 1 DSBs would show multiplicity within a hot spot. If Spo11’s DNA-binding function is perturbed (e.g., by mutations in the DNA-binding surface), then type 2 DSB sites would predominate, thus exhibiting less multiplicity. Finally, we showed that Spo11 itself has weak sequence preference. Our results suggest that Spo11 is unlikely to cleave DNA when the future 3⬘ DSB end is a G nucleotide. If this is the case, Spo11 would be unlikely to introduce a DSB at the sequence 5⬘-G/NNC-3⬘ (where the shill indicates the cleavage position). In fact, we did not find any G/NNC sequences among 67 DSB sites that we mapped in this study, whereas there are a total 14 GNNC sequences in the DSB targeting window of targeted DSB sites (pGAL2, the YAL048c ORF, pYNL328c, and pGAL80) and in pYCR048w (positions ⫺296 to ⫺95 relative to the YCR048w translation start) (Fig. 11D). In contrast, we found 11 DSBs out of 14 CNNG sequences in the mapped region (Fig. 11D), suggesting that Spo11 prefers to cleave C/NNG but not G/NNC. A similar sequence bias was also found in an independent study in which DSBs were mapped genome-wide by direct sequencing of oligonucleotides that are covalently attached to Spo11 (Keeney, unpublished).

3516


Thus, the sequence preference of the Spo11 cleavage position that emerges from the Gal4BD-Spo11 system in the present study is likely to be a general feature of Spo11 DSB formation. ACKNOWLEDGMENTS We thank all members of our laboratory, in particular V. Borde, N. Robine, and N. Uematsu for strains, reagents, and helpful discussions. We especially thank S. Keeney and all members of his laboratory, especially L. Kauppi and R. Kniewel, as well as F. Cole for support to H.M. during preparation of the manuscript. This study was supported by grants from the Association pour la Recherche contre le Cancer and the Agence Nationale de la Recherche (grant Blanc06-3-150811). H.M. was supported by postdoctoral fellowships from the Ministe`re de l’Education Nationale, de la Recherche, et de la Technologie and from the Institut Curie.

MOL. CELL. BIOL.

12.

13. 14. 15. 16. 17.

REFERENCES 1. Arora, C., K. Kee, S. Maleki, and S. Keeney. 2004. Antiviral protein Ski8 is a direct partner of Spo11 in meiotic DNA break formation, independent of its cytoplasmic role in RNA metabolism. Mol. Cell 13:549–559. 2. Baudat, F., and A. Nicolas. 1997. Clustering of meiotic double-strand breaks on yeast chromosome III. Proc. Natl. Acad. Sci. USA 94:5213–5218. 3. Beretta, G. L., M. Binaschi, E. Zagni, L. Capuani, and G. Capranico. 1999. Tethering a type IB topoisomerase to a DNA site by enzyme fusion to a heterologous site-selective DNA-binding protein domain. Cancer Res. 59: 3689–3697. 4. Blat, Y., R. U. Protacio, N. Hunter, and N. Kleckner. 2002. Physical and functional interactions among basic chromosome organizational features govern early steps of meiotic chiasma formation. Cell 111:791–802. 5. Blitzblau, H. G., G. W. Bell, J. Rodriguez, S. P. Bell, and A. Hochwagen. 2007. Mapping of meiotic single-stranded DNA reveals double-strandedbreak hotspots near centromeres and telomeres. Curr. Biol. 17:2003–2012. 6. Buhler, C., V. Borde, and M. Lichten. 2007. Mapping meiotic single-strand DNA reveals a new landscape of DNA double-strand breaks in Saccharomyces cerevisiae. PLoS Biol. 5:e324. 7. de Massy, B., V. Rocco, and A. Nicolas. 1995. The nucleotide mapping of DNA double-strand breaks at the CYS3 initiation site of meiotic recombination in Saccharomyces cerevisiae. EMBO J. 14:4589–4598. 8. Diaz, R. L., A. D. Alcid, J. M. Berger, and S. Keeney. 2002. Identification of residues in yeast Spo11p critical for meiotic DNA double-strand break formation. Mol. Cell. Biol. 22:1106–1115. 9. Gerton, J. L., J. DeRisi, R. Shroff, M. Lichten, P. O. Brown, and T. D. Petes. 2000. Inaugural article: global mapping of meiotic recombination hotspots and coldspots in the yeast Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 97:11383–11390. 10. Goldstein, A. L., and J. H. McCusker. 1999. Three new dominant drug resistance cassettes for gene disruption in Saccharomyces cerevisiae. Yeast 15:1541–1553. 11. Hong, M., M. X. Fitzgerald, S. Harper, C. Luo, D. W. Speicher, and R.

18. 19. 20. 21. 22. 23. 24. 25.

26. 27. 28. 29.

Marmorstein. 2008. Structural basis for dimerization in DNA recognition by Gal4. Structure 16:1019–1026. Huibregtse, J. M., P. D. Good, G. T. Marczynski, J. A. Jaehning, and D. R. Engelke. 1993. Gal4 protein binding is required but not sufficient for derepression and induction of GAL2 expression. J. Biol. Chem. 268:22219– 22222. Johnston, M., and J. Dover. 1987. Mutations that inactivate a yeast transcriptional regulatory protein cluster in an evolutionarily conserved DNA binding domain. Proc. Natl. Acad. Sci. USA 84:2401–2405. Keeney, S. 2007. Spo11 and the formation of DNA double-strand breaks in meiosis, p. 81–123. In D. H. Lankenau (ed.), Recombination and meiosis, vol. 2/2008. Springer-Verlag, Heidelberg, Germany. Keeney, S., C. N. Giroux, and N. Kleckner. 1997. Meiosis-specific DNA double-strand breaks are catalyzed by Spo11, a member of a widely conserved protein family. Cell 88:375–384. Liang, S. D., R. Marmorstein, S. C. Harrison, and M. Ptashine. 1996. DNA sequence preferences of GAL4 and PPR1: how a subset of Zn2 Cys6 binuclear cluster proteins recognizes DNA. Mol. Cell. Biol. 16:3773–3780. Liu, J., T. C. Wu, and M. Lichten. 1995. The location and structure of double-strand DNA breaks induced during yeast meiosis: evidence for a covalently linked DNA-protein intermediate. EMBO J. 14:4599–4608. Mancera, E., R. Bourgon, A. Brozzi, W. Huber, and L. M. Steinmetz. 2008. High-resolution mapping of meiotic crossovers and non-crossovers in yeast. Nature 454:479–485. Neale, M. J., J. Pan, and S. Keeney. 2005. Endonucleolytic processing of covalent protein-linked DNA double-strand breaks. Nature 436:1053–1057. Ohta, K., T. Shibata, and A. Nicolas. 1994. Changes in chromatin structure at recombination initiation sites during yeast meiosis. EMBO J. 13:5754– 5763. Pecina, A., K. N. Smith, C. Mezard, H. Murakami, K. Ohta, and A. Nicolas. 2002. Targeted stimulation of meiotic recombination. Cell 111:173–184. Petes, T. D. 2001. Meiotic recombination hot spots and cold spots. Nat. Rev. Genet. 2:360–369. Prieler, S., A. Penkner, V. Borde, and F. Klein. 2005. The control of Spo11’s interaction with meiotic recombination hotspots. Genes Dev. 19:255–269. Robine, N., N. Uematsu, F. Amiot, X. Gidrol, E. Barillot, A. Nicolas, and V. Borde. 2007. Genome-wide redistribution of meiotic double-strand breaks in Saccharomyces cerevisiae. Mol. Cell. Biol. 27:1868–1880. Vedel, M., and A. Nicolas. 1999. CYS3, a hotspot of meiotic recombination in Saccharomyces cerevisiae: effects of heterozygosity and mismatch repair functions on gene conversion and recombination intermediates. Genetics 151:1245–1259. Wach, A., A. Brachat, R. Pohlmann, and P. Philippsen. 1994. New heterologous modules for classical or PCR-based gene disruptions in Saccharomyces cerevisiae. Yeast 10:1793–1808. Wu, T.-C., and M. Lichten. 1994. Meiosis-induced double-strand break sites determined by yeast chromatin structure. Science 263:515–518. Xu, L., and N. Kleckner. 1995. Sequence nonspecific double-strand breaks and interhomolog interactions prior to double-strand break formation at a meiotic recombination hot spot in yeast. EMBO J. 14:5115–5128. Zenvirth, D., T. Arbel, A. Sherman, M. Goldway, S. Klein, and G. Simchen. 1992. Multiple sites for double-strand breaks in whole meiotic chromosomes of Saccharomyces cerevisiae. EMBO J. 11:3441–3447.