Investigation of the Mechanism of Meiotic DNA Cleavage by VMA1-Derived Endonuclease Uncovers a Meiotic Alteration in Chromatin Structure around the Target Site

Strains.The yeast strains used in this study are listed in Table 1. All strains were derivatives of the SK1 strain, which enters meiosis in a highly synchronous manner (18). Strains were constructed by standard genetic crosses, transformation, and other genetic procedures. The VMA1-101 and VMA1-201 alleles have a complete deletion of the VMA1 intein-coding sequence (VMA1⁻). The VMA1, VMA1-202, and VMA1-207 alleles contain the VMA1 intein-coding sequence (8). VMA1-207 has a mutation in the VDE coding region that results in a loss of DNA-binding activity. The BamHI site, which is located 866 bp upstream of the VMA1 start codon, is replaced with a SalI site in the VMA1-202 and VMA1-207 allele, enabling us to distinguish the VMA1⁺ homing product from the VMA1⁺ donor in Southern analysis (9).

For the expression of VDE at the ADE3 or URA3 gene locus under the VMA1 promoter, cells were constructed as described previously (24, 26). For the expression of Flag-tagged VDE, the synthetic oligonucleotides 5′-CAT GGA CTA CAA GGA CGA TGA CGA TAA GGG TAC-3′ and 5′-CCT TAT CGT CAT CGT CCT TGT AGT C-3′ were annealed and cloned into the NcoI-KpnI site of plasmids harboring the VMA1 promoter-fused VDE coding region (26).

To construct the ADE2::VRS allele, the HindIII-digested fragment of pYO2411, which contains the VRS at the XbaI site near the ATG start codon of the ADE2 gene in pASZ10 (34), was transformed into the predeleted ADE2 gene locus. For the construction of the MCM4::VRS allele, the XbaI restriction site was inserted adjacent to the ATG start codon by a two-step PCR method (15) using primer pairs 5′-AGT CCT CGA GAC TGG GAG AAT CTT CGA GCC AAT GC-3′ and 5′-GTT GGA GAG CTA GAC TGT TGT CTA GAC ATT TTA AGT TCT TGA G-3′ and 5′-CTC AAG AAC TTA AAA TGT CTA GAC AAC AGT CTA GCT CTC CAA C-3′ and 5′-AGT CGC GGC CGC CCA GAA GCG CTC GCA GAA GAA CG-3′, and then a PCR product was cloned into pRS306 (33). The synthetic oligonucleotides 5′-CTA GCT ATG TCG GGT GCG GAG AAA GAG GTA ATG AAA-3′ and 5′-CTA GTT TCA TTA CCT CTT TCT CCG CAC CCG ACA TAG-3′ were annealed and cloned into the XbaI site to produce pRS306MCM4VRS. After SphI digestion and integration at the MCM4 gene locus, pop-out of the URA3 marker and a part of the gene were performed on 5-fluoroorotic acid plates, resulting in the generation of the MCM4::VRS allele.

Deletion mutations of each gene were constructed by the PCR-mediated gene disruption method (32) and confirmed by PCR.

Synchronous sporulation, physical analysis of intein-coding sequence homing, immunoblotting of VDE, endonuclease activity assay, and immunostaining of VDE.Synchronous sporulation, physical analysis of homing, and the preparation of protein samples and immunoblotting were carried out as described previously (9). For the endonuclease assays of VDE, cells expressing Flag-tagged VDE were sampled at various time points, frozen, and resuspended in lysis buffer (200 mM Tris-HCl [pH 8.0], 400 mM ammonium sulfate, 10 mM MgCl₂, 1 mM EDTA, 10% glycerol) containing protease inhibitors. After we vortexed the cell extracts with glass beads, they were clarified by sonication and centrifugation. Protein concentrations were estimated by a bicinchoninic acid assay (Pierce), and aliquots of cell extracts were incubated with anti-Flag-Sepharose resin (Sigma) overnight at 4°C, washed, and then eluted in elution buffer (25 mM Tris-HCl [pH 8.5], 100 mM KCl) with 150 μg/ml 3× Flag peptide (Sigma). Aliquots of the eluates were incubated with 0.5 μg of pBSΔVDE linearized with XmnI (19) in 25 mM Tris-HCl (pH 8.5), 100 mM KCl, and 2.5 mM MnCl₂ at 30°C. Nuclear staining and immunofluorescence microscopy were performed as described previously (24), with some modifications. Alexa Fluor 488-conjugated goat anti-rabbit immunoglobulin antibody was used as a second antibody.

Chromatin preparation, digestion by MNase, and hybridization.The preparation of crude chromatin fractions from yeast cells and the treatment of chromatin with micrococcal nuclease (MNase) were performed as described previously (27, 28), with the minor modification that spheroplasts be prepared by incubation with a higher concentration (0.1 mg/ml) of Zymolyase 100T (Seikagakukogyo Co. Ltd) for 5 min on ice, due to the recently improved solubility and activity of the commercially available Zymolyase. DNA was digested with the indicated restriction endonucleases, separated in agarose gels, and transferred to Biodyne B membranes (Pall) by alkaline transfer. Radioactive probes were prepared by random priming of DNA fragments purified by agarose gel electrophoresis from appropriate PCR products. The membranes were hybridized according to the method of Church and Gilbert (5).

ChIP assay.Chromatin immunoprecipitation (ChIP) was performed basically as described previously (36). Aliquots of cells taken from a synchronous sporulation culture were fixed in 1% formaldehyde for 35 min at room temperature. Immunoprecipitations were carried out with affinity-purified anti-VDE antibody (24) bound to protein G-coated Dynabeads (Dynal). After treatment with RNase A and proteinase K, samples were purified using a QIAquick PCR purification kit (QIAGEN). Multiplex PCR amplification by AmpliTaq Gold (Perkin-Elmer) was performed with primer pairs, the sequences of which are described elsewhere (10). PCR products were separated on 2% agarose gels containing ethidium bromide and analyzed with FLA3000 (Fujifilm).

VDE is constitutively present and exhibits endonuclease activity in both mitotic and meiotic cell extracts.The simplest explanation for the meiosis specificity of DSB formation by VDE is that DNA cleavage is controlled by the amount or activity of VDE. However, as reported previously (9, 24), VDE is expressed abundantly in mitotic cells as well as in meiotic cells. We consequently examined the endonuclease activity of VDE during mitosis and meiosis. Cells expressing Flag-VDE were removed and harvested at various time points from synchronously sporulating cultures. The Flag-VDE was then recovered with anti-Flag antibody and incubated with linear double-stranded DNA containing the VRS. Immunoprecipitated VDE obtained from mitotic and meiotic cells exhibited equivalent levels of endonuclease activity (Fig. 1A). Therefore, we conclude that VDE is constitutively expressed and active during both mitosis and meiosis, ruling out the possibility that VDE-mediated DNA cleavage is regulated by quantitative or qualitative differences in VDE during meiosis.

• Open in new tab
• Download powerpoint

FIG. 1.

Examination of endonuclease activity and effects of nuclear localization or meiotic progression on meiosis-specific homing. (A) Endonuclease activity of immunoprecipitated VDE. Cells expressing Flag-tagged VDE were subjected to synchronous sporulation and sampled at various time points in SPM. Using an anti-Flag antibody, VDE was immunoprecipitated and assayed by cleavage of the 3.3-kb linearized DNA containing the VRS that produces 2.2- and 1.1-kb fragments after digestion at the VRS. Substrates were incubated for 0, 15, 30, and 60 min at 30°C. (B) Physical detection of homing by Southern blotting. VMA1-207/VMA1⁻ heterozygous diploid cells ectopically expressing no VDE, wild-type VDE, or NLS-fused VDE were subjected to synchronous sporulation. Since the VMA1-207 allele contains an intein-coding sequence encoding inactive VDE (R90A), only VDE that is ectopically expressed can produce DSBs and the cleaved recipient is repaired using a VMA1-207 donor as a template, resulting in the production of a VMA1⁺ homing product. DNA was isolated from cells at the indicated times after incubation in SPM and subjected to Southern analysis after BamHI digestion. The BamHI site polymorphism makes it possible to distinguish VMA1⁺ homing products from VMA1-207 donors. (C) Physical analysis of homing in mutants deficient in progression of meiotic program. Southern analysis was performed on the VMA1⁺/VMA1⁻ heterozygous diploid cells in which meiotic entry, DNA replication, or nuclear divisions were blocked by the introduction of each mutation. (D) VDE localization in meiotic mutants at 4 h in SPM. The upper panels show localization of VDE, and the lower panels show the nuclei. DAPI, 4′,6′-diamidino-2-phenylindole.

Meiosis-specific DNA cleavage by VDE is not simply due to nuclear import.VDE is reported to be excluded from the nucleus in vegetatively growing cells, and its nuclear localization is induced in response to nutrient limitation (24). It has also been revealed that the attachment of a nuclear localization signal (NLS) to VDE causes the accumulation of VDE in the nucleus and induces homing events to some extent even during mitosis (24). To determine the precise contribution of the nuclear localization of VDE to meiosis-specific DSB, we conducted Southern analyses to examine the effects of NLS addition to VDE on homing. We prepared a heterozygous VMA1-207/VMA1⁻ diploid as a tester strain. VMA1-207 is an allele containing the intein-coding sequence, but its VDE has an R90A mutation that abolishes the DNA binding and cleavage activity of VDE (14, 26). As shown in Fig. 1B, we confirmed that neither cleavage nor homing products could be detected in the tester strain. VDE or NLS-fused VDE was expressed under the VMA1 promoter in this tester strain. The expression of wild-type VDE resulted in DSB formation at the VRS during meiosis (Fig. 1B). Homing products are faintly observed in premeiotic cells resulting from mitotic leak of homing, as reported elsewhere (9, 24), but greatly accumulate in the later meiotic stages (Fig. 1B). When the NLS-fused VDE was expressed, the kinetics of the VDE-mediated DSB formation and appearance of the homing products were generally similar to those observed for normal VDE expression (Fig. 1B). These results suggest that meiotic induction of DSB formation and homing are not necessarily consequences of meiotic nuclear localization of VDE and implicate the existence of other factors that regulate the meiosis-specific actions of VDE.

Entry into the meiotic program is necessary for DSB formation by VDE.In yeast, meiosis is induced in diploid cells under certain conditions of nutrient limitation. After the induction of meiosis, premeiotic DNA replication, two successive nuclear divisions, and sporulation take place in sequential order. Previously, it was shown that full activation of VDE-mediated DSB formation requires premeiotic DNA replication since blocking premeiotic DNA replication reduces and delays DNA cleavage by VDE (9).

To examine detailed requirements of meiotic progression for VDE-mediated DSBs, we introduced a mutation in genes of key regulators of meiotic events and observed homing by Southern analysis. In the ime1Δ and ime2Δ mutants, which are defective in the activation of early events of meiosis, including entry into the meiotic program and the induction of premeiotic DNA replication (6, 7, 37), VDE-mediated DSBs and homing products were not observed (Fig. 1C). In contrast to the case of the clb5Δ clb6Δ double mutant, in which the progression of premeiotic DNA replication is prevented (35), VDE-mediated DSBs were hardly observed at later time points in these mutants (Fig. 1C). These results suggest that entry into the meiotic program is indispensable for the activation of VDE-mediated DSB formation. As shown in Fig. 1D, VDE was localized both in the cytoplasm and in the nucleus in each mutant after 4 h of incubation in sporulation medium (SPM). Thus, it is suggested that a nutritional limitation only, which is sufficient for nuclear entry of VDE, is insufficient for the induction of homing, consistent with the results for NLS-fused VDE (Fig. 1B). On the contrary, the ndt80Δ mutant, which leads to a failure to induce the middle sporulation genes and a subsequent arrest before nuclear divisions (4, 37), exhibited an induction of DSBs and intein-coding sequence homing (Fig. 1C), indicating that VDE-mediated DSB formation is an event which occurs before nuclear divisions. Thus, these results lead us to hypothesize that the nuclear-localized VDE, which was induced by nutritional limitation, is subjected to further regulation that inhibits VDE-mediated DSB formation until meiotic entry or that activates DSB formation during meiotic progression.

VDE binding to the VRS occurs during meiosis.To test whether VDE can bind to its target DNA before digestion, we examined the in vivo association of VDE to the VRS by a ChIP assay. Cells homozygous for VMA1⁻ expressing VDE were cross-linked with formaldehyde at various time points to generate protein-DNA cross-links. After fragmentation of the chromatin, VDE was immunoprecipitated with an anti-VDE antibody, and the precipitated DNA was purified. Purified DNAs were analyzed by PCR using primer sets for the sequence around the VRS (VRS3 and VRS5). Compared to the signals in the control input samples or the signals by the primer set for telomeric sequences, the anti-VDE-precipitated fraction exhibited a significant enrichment of the sequence detected by the VRS3 and VRS5 primer sets (Fig. 2A and B). These enrichments were not detected in cells expressing VDE^R90A, which is defective in DNA binding, or in the ime1Δ, ime2Δ, or clb5Δ clb6Δ mutants, whereas the ndt80Δ mutant showed meiotic enrichment of the signals for the sequence near the VRS (Fig. 2C). These results indicate that VDE binding to the VRS is closely correlated with DSB formation by VDE and that VDE can be detected at the target site only when DSB formation takes place. Taken together, it is implied that meiosis-specific DNA cleavage by VDE is regulated by an accessibility of VDE to the target site.

• Open in new tab
• Download powerpoint

FIG. 2.

Detection of VDE binding to the VRS. (A) Diploid cells homozygous for VMA1⁻- expressing VDE or DNA-binding-deficient VDE (R90A) were sampled at the indicated times during meiosis. Cells were treated with formaldhyde, and then the cell extract was prepared, sonicated to shear the chromatin, and immunoprecipitated with anti-VDE antibody. DNA from input samples and in the immunoprecipitates (IP) was amplified by multiplex PCR with primers for the 3′ region adjacent to the VRS (VRS3), the 5′ region adjacent to the VRS (VRS5), and the telomeric region of chromosome VI (TEL). (B) The ratios of the VRS signals to the TEL signal were normalized to those of the input samples at each time point. Error bars denote the standard deviations among three independent experiments. (C) A ChIP assay with anti-VDE antibody was performed for the indicated mutants. The signals obtained for the 4.5-h samples are shown as in panel B.

Chromatin configuration changes around the VRS during meiosis.The results of the ChIP assay led us to explore the determinant of the accessibility of VDE to the VRS. We examined the possibility of cleavage regulation by the substrate, namely, the chromatin structure around the VRS, because regulations through chromatin structure are involved in many DNA metabolic processes, including DNA replication, transcription, and recombination. We prepared chromatin fractions from diploid strains homozygous for VMA1⁻ at various time points in meiotic culture. The isolated chromatin was digested with various concentrations of MNase. We further digested the MNase-digested DNA with BamHI and SphI. The DNA fragments were then analyzed by Southern hybridization using a short probe for the sequence adjacent to the BamHI site (indirect end-labeling). Figure 3A shows the MNase-sensitive sites in chromatin of the VMA1⁻ locus in cells at various time points exhibiting a ladder-like pattern of sensitive sites in the open reading frame of VMA1⁻. The VRS was mapped at a position adjacent to the sensitive site on the 5′ side. We detected a prominent MNase-sensitive site appearing immediately at the 5′ site adjacent to the VRS after 3 and 4.5 h of meiotic culture, whereas MNase sensitivity at this site was very faint in MNase-digested chromatin from mitotic and early meiotic cells (Fig. 3A, arrowheads). Importantly, this MNase-sensitive site was not detected in naked DNA partially digested with MNase (Fig. 3A), indicating that the appearance of this site reflects changes in local chromatin structure around the VRS during meiosis. In this experiment, most of the cells completed premeiotic DNA replication by 3 h and started the first meiotic division at 6 h (Fig. 3B and C), indicating that chromatin alteration near the VRS occurs around the period of VDE-mediated DSB formation.

• Open in new tab
• Download powerpoint

FIG. 3.

Chromatin configuration at the VMA1⁻ locus during meiosis. (A) Chromatin was prepared in diploid cells homozygous for VMA1⁻ at various times of incubation in SPM, treated with 0, 5, 10, or 20 U/ml of MNase, and redigested with BamHI and SphI. Naked DNA was also treated with 0.2, 0.4, or 0.8 U/ml of MNase. MNase-sensitive sites were detected by indirect end-labeling using a probe for the sequence adjacent to the BamHI site. The vertical gray arrow indicates the position of the coding region of the VMA1⁻ locus. The arrowheads show MNase-sensitive sites near the VRS that became prominent during meiosis. Numbers indicate the positions in nucleotides relative to the VDE cutting site that is shown by the horizontal arrow. Meiotic progression was monitored by DNA content, which was analyzed by a fluorescence-activated cell sorter (B), and by nuclear division, which was analyzed by microscopy of 4′,6′-diamidino-2-phenylindole-stained cells (C).

We next examined the effect of VDE on chromatin structure. We analyzed chromatin configuration in a diploid strain homozygous for VMA1⁻ but expressing VDE under the VMA1 promoter at the URA3 gene locus. At 0 h, the patterns of the MNase-sensitive sites were very similar to those observed in the strain lacking VDE (Fig. 3A and 4A). Unfortunately, robust VDE-mediated DSBs made it impossible to detect the chromatin structure around the VRS after 4 h of incubation in SPM (Fig. 4A). Thus, we prepared strains expressing VDE^D326V, which is defective in endonuclease activity, and VDE^R90A, which is defective in DNA binding (11, 26). In each case, we observed chromatin alteration around the VRS during meiosis as in the wild-type cells (Fig. 4B), suggesting that the chromatin change is not affected by the expression or target binding of VDE.

• Open in new tab
• Download powerpoint

FIG. 4.

Chromatin alteration during meiosis occurs independently of VDE expression and is dependent on meiotic progression. Results of chromatin analysis of VMA1⁻ homozygous cells expressing wild-type VDE (A), endonuclease-defective VDE (D326V) (B), and VDE defective in DNA binding (R90A) (B) are shown. Samples were treated with 0, 5, or 10 U/ml of MNase. Indirect end-labeling was performed, and MNase-sensitive sites were analyzed as described in the legend to Fig. 3. The arrowheads indicate the chromatin alteration site. (C) Chromatin structure of the haploid cell in SPM. Chromatin was prepared in VMA1⁻ haploid cells sampled at the indicated times of incubation in SPM and treated with 0, 5, or 10 U/ml of MNase. The sir2Δ mutation permits haploid cells to enter the meiotic program. (C) Chromatin structure around the VRS in mutant cells defective in meiotic progression. Chromatin was prepared from VMA1⁻ homozygous diploid cells after 1.5 and 4.5 h of incubation in SPM and subjected to indirect end-labeling.

Chromatin alteration around the VRS takes place in correlation with DNA binding and cleavage by VDE.It is possible that the observed chromatin alteration occurs in response to metabolic changes caused by starvation, as in the case of nuclear entry of VDE. We consequently examined whether haploid cells undergo chromatin changes around the VRS during culture in SPM. As shown in Fig. 4C, the MNase-sensitive site near the VRS remained very faint even after 4 h of incubation in SPM, indicating that chromatin transition does not take place in starved haploid cells in SPM. In addition, we introduced a sir2Δ mutation into the haploid strain, which resulted in the derepression of the normally silent mating type loci HMR and HML, leading to the coexpression of a and α mating types and permitting haploid meiosis. After 4 h of incubation in SPM, the sir2Δ haploid cells exhibited the same chromatin structure as that observed in diploid cells (Fig. 4C). These data suggest that the chromatin alteration around the VRS is coupled with meiotic induction, not with metabolic changes. To confirm this notion, we further examined the dependence of the chromatin alteration on a set of meiotic events. In the ime1Δ and ime2Δ mutants, chromatin alteration around the VRS was barely detected even after 4.5 h of meiotic culture (Fig. 4D), indicating that the chromatin changes are dependent on entry into meiotic programs. The clb5Δ clb6Δ double mutation also severely affected the chromatin alteration around the VRS. In contrast, the ndt80Δ mutant exhibited normal chromatin alteration (Fig. 4D). Thus, the dependency of chromatin alteration on meiotic events is very similar to the case of DNA binding and cleavage by VDE.

The VRS element does not necessarily cause chromatin alteration.To test the sequence dependency of the chromatin alteration around the VRS, we next studied chromatin configuration in a VMA1⁺ homozygous strain. In the VMA1⁺ allele, insertion of the intein-coding sequence interrupts the VRS element and VDE is no longer able to cleave the sequence. In this strain, the site where meiotic MNase sensitivity was observed in the VMA1⁻ strain showed high MNase sensitivity even at 0 h, and little change in the pattern of sensitive sites was detected after 4 h of incubation (Fig. 5A). This was not due to the presence of VDE in the VMA1⁺ cells, since the chromatin structure at 0 h was not affected regardless of the presence or absence of VDE (Fig. 4A). Thus, intein-coding sequence insertion at the VRS affects chromatin configuration around the VRS during mitosis, implying that a complete VRS element is required for proper chromatin changes to occur around the VRS.

• Open in new tab
• Download powerpoint

FIG. 5.

(A) Chromatin structure around the VRS in diploid cells homozygous for VMA1⁻ or VMA1⁺. In the VMA1⁺ allele, the VRS is interrupted by the insertion of an intein-coding sequence in the middle of the sequence. Samples were treated with 0, 5, or 10 U/ml of MNase. (B) Chromatin configuration at the ADE2 gene locus. Mitotic (0 h) and meiotic (4 h) chromatins were prepared from diploid cells homozygous for ADE2 or VRS-inserted ADE2 (designated ADE2::VRS), treated with 0, 5, 10, or 20 U/ml of MNase, and redigested with AseI. MNase-sensitive sites were detected by indirect end-labeling with a probe for the sequence adjacent to the AseI site. The vertical gray arrow indicates the position of the coding region for the ADE2 locus. The arrowheads show MNase-sensitive sites that became prominent in ADE2::VRS during meiosis. The right panel shows the positions of the VDE cutting site (horizontal arrow) and the chromatin alteration site (arrowhead). Numbers indicate the positions in nucleotides relative to the VDE cutting site. (C) Chromatin configuration at the MCM4 gene locus. Mitotic (0 h) and meiotic (4 h) chromatins were prepared from diploid cells homozygous for MCM4 or VRS-inserted MCM4 (designated MCM4::VRS) and then analyzed by end-labeling after treatment with 0, 5, or 10 U/ml of MNase. (D) Physical detection of the VDE-mediated DSB by Southern blotting. VMA1⁺ diploid cells homozygous for ADE2::VRS or MCM4::VRS were subjected to synchronous sporulation. DNA was isolated from cells at the indicated times after incubation in SPM and subjected to Southern analysis.

It has been shown that insertion of the VRS is sufficient to introduce a meiotic DSB by VDE and a recombination event at the site (25, 26). If the observed meiotic chromatin alteration plays a critical role in the regulation of meiosis-specific action of VDE, the insertion of the VRS at an ectopic site causes a chromatin alteration at the site. To study whether the chromatin alteration is specific to its original chromosomal position, we analyzed chromatin structure around the VRS that is inserted at a different chromosomal location. We constructed a diploid strain in which the VRS is inserted in frame near the ATG start codon of the ADE2 gene or the MCM4 gene and investigated premeiotic and meiotic chromatin structures around the gene loci with or without VRS. In each case, meiotic DSBs were observed when VDE was expressed from VMA1⁺ (Fig. 5D). In the VRS-embedded ADE2 allele, a prominent MNase-sensitive site appeared during meiosis, while the sensitive site was ambiguous at 0 h of meiosis (Fig. 5B, arrowheads). No such meiotic chromatin alteration was observed in the wild-type ADE2 allele, in which a distinct MNase-sensitive band was detected at the corresponding site at 0 h (Fig. 5B, arrowheads). Although meiotic alteration in chromatin was observed, the MNase-sensitive site locates relatively apart from the VRS insertion site. In the case of VRS integration into the MCM4 locus, meiotic change in chromatin was not significantly observed around the VRS, while the VRS insertion brought some alterations in the chromatin configuration along the gene locus. Thus, VRS does not necessarily determine the chromatin structure around itself. Therefore, it is suggested that chromatin structure per se is not the sole mechanism for meiosis-specific homing of the intein-coding sequence and that another factor(s) exists and regulates the meiosis-specific DNA binding and cleavage of VDE.