| Edited by Hirokazu Inoue. Kunihiro Ohta: Corresponding author. E-mail: kohta@riken.jp. Note: The microarray data for wild type and mutants in this article have been deposited in NCBIs Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) and are accessible through GEO Series accession number GSE6620. Supplementary materials in this article are at http://wwwsoc. nii.ac.jp/gsj3/sup/82(1)Kugou/ |
Double-strand DNA break (DSB) repair is an essential process to repair meiotically induced DSBs at recombination hot spots and DNA damages caused by extra- and intracelullar factors such as ionizing radiation, chemical modification, abortive DNA replication, and active oxygen radicals. DSB repair is achieved by various recombination proteins belonging to the Rad50 epistasis group. Among them, MR complex: Rad50, Mre11, Xrs2/Nbs1 are involved in the initial steps of homologous recombination, such as introduction of DSBs and processing of DSB ends (reviewed in D’Amours and Jackson, 2002). In meiosis, homologous recombination is triggered by a transient meiosis-specific DSB formation catalyzed by the meiotic nuclease Spo11 and other coactivating proteins (Keeney and Neale, 2006), including Mre11. After meiotic DSB formation, DNA ends are processed to form 3’ overhangs, which provide highly recombinogenic intermediates. Other factors (i.e., Rad51, Rad52, Rad54, Rad55, Rad57, etc) are mainly required for later recombination processes including heteroduplex formation and strand exchange reactions.
Many of these recombination proteins are well conserved in eukaryotes. In some cases, these genes are responsible for human genetic diseases that exhibit chromosomal instability. For example, hypomorphic mutations in Mre11 induce the “ataxia telangiectasia-like disorder” (ATLD) in human, lymphoid cancer predisposition, and developmental defects in neuronal and immune system (Stewart et al., 1999). Most of those defects can be attributed to the loss or impairment of DSB repair or proper checkpoint response, but possibly as well as by other types of deficiencies coupled to cellular differentiation. However, little is known about the link between Mre11 function and cellular developmental processes.
To investigate the roles of Mre11 in gene regulation coupled to differentiation, we studied the transcriptional profiles during yeast sporulation in various mutants, such as mre11Δ, rad50Δ, and spo11-Y135F (containing a mutation in the catalytic domain of the Spo11 (Bergerat et al., 1997)) using DNA microarrays, and compared them to the profile in a wild-type strain. We found that the meiotic expression profile in mre11Δ was generally unaffected, but the activation of a specific class of meiotic genes was severely affected. Interestingly, many of them were related to spore wall biogenesis. We confirmed that the transcriptional deficiency in mre11Δ is due to neither absence of DSB formation nor uncontrolled activation of DNA damage checkpoint. These data suggest that Mre11 may be involved in the transcriptional regulation of genes for spore wall development in yeast meiosis.
For all experiments, we used yeast strains with the SK1 background and their genotypes were as follows: MJL1720 (MATa/MATα ura3/ura3 lys2/lys2 ho::LYS2/ho::LYS2 leu2/leu2 arg4-bgl/arg4-nsp cyh2-z/cyh2-z), YMD1599 (MATa/MATα ura3/ura3 lys2/lys2 ho::LYS2/ho::LYS2 leu2/leu2 arg4-bgl/arg4-nsp mre11Δ(acc-aor51) ::URA3/mre11Δ(acc-aor51)::URA3), YMD316 (MATa/MATα ura3/ura3 lys2/lys2 ho::LYS2/ho::LYS2 leu2/leu2 arg4-bgl/arg4-nsp cyh2-z/cyh2-z mre11D16A:: URA3/mre11D16A::URA3),YMD325(MATa/MATα ura3/ura3 lys2/lys2 ho::LYS2/ho::LYS2 leu2/leu2 arg4-bgl/arg4-nsp cyh2-z/cyh2-z mre11ΔC49::URA3/mre11ΔC49 ::URA3), XDU278 (MATa/MATα ura3/ura3 lys2/lys2 ho::LYS2/ho::LYS2 xrs2Δ::URA3/xrs2Δ::URA3), RKD101 (MATa/MATα lys2/lys2 ho::LYS2/ho::LYS2 ura3/ura3 rad50Δ::hisG/rad50Δ::hisG), ORD6167 (MATa/MATα trp1/trp1 his4/his4 leu2/leu2 ura3/ura3 spo11Δ::hisG-URA3::hisG/spo11Δ::hisG-URA3::hisG), YHS50 (MATa/MATα ura3/ura3 lys2/lys2 ho::LYS2/ho::LYS2 rad24Δ::URA3/rad24Δ::URA3 mre11Δ(acc-aor51)::URA3 /mre11Δ(acc-aor51)::URA3), RKD1312 (MATa/MATα ura3/ura3 leu2/leu2 his4/his4 trp1/trp1 arg4-bgl/arg4-nsp nuc1Δ::LEU2/nuc1Δ::LEU2 spo11Δ::hisG-URA3:: hisG/spo11Δ::hisG-URA3::hisG pCY28-spo11Y135F).
The sequence 50–310 bp upstream of SPS4, containing SPS4 MSE, was amplified and flanked with SpeI site to the 5’ end and BamHI site to the 3’ end by PCR. The PCR was performed with ExpandTM High Fidelity PCR System (Boehringer Mannheim), using the forward primer (5’-GGACTAGTGCGGTCGTTAAGACACAAGGAC), and the reverse primer (5’-CGGGATCCTGTTCTTGTTGCTTGCCTCTC). The added restriction recognition sites are underlined. The PCR product was cloned into the pCR2.1 TOPO vector (Invitrogen). To construct pYES2.1-pSPS4-V5-lacZ, the SpeI-BamHI fragment of pCR2.1-pSPS4 was inserted into the SpeI-BamHI site of pYES2.1/V5-His/lacZ (Invitrogen). To construct pRS415-pSPS4-V5-lacZ, the NaeI-BamHI fragment of pYES2.1-pSPS4-V5-lacZ was replaced with the NaeI-BamHI fragment of pRS415 (Stratagene).
Transformation of the plasmids containing SPS4 promoter-V5-lacZ reporter gene into yeast was performed with a Fast Yeast Transformation Kit (GenoTechnology). Transformants carrying pYES2.1-pSPS4-V5-lacZ or pRS415-pSPS4-V5-lacZ were grown overnight at 30°C in 1 ml of SD-uracil or SD-leucine medium, respectively. 15 μl of culture was spotted onto SPM plate (1% potassium acetate, 0.1% yeast extract, 0.05% dextrose, 1/5 level of amino acids except for uracil or leucine, 2% agar) containing 40 μg/ml X-Gal, and incubated for 2 days at 30°C.
Cells were grown on a YPD plate for 2 days. The cells were inoculated into 10 ml of SPS presporulation medium with antifoam (Ohta et al., 1998), and grown overnight at 30°C. The presporulation culture was inoculated into 250 ml of SPS medium, and the cells were grown to a density of 4.0×107 cells/ml at 30°C. The cells were washed once with water, suspended in SPM (1% potassium acetate, 1/5 level of amino acids, and polypropylene glycol) to a density of 2.0×107 cells/ml, and incubated at 30°C for sporulation.
Total RNA was prepared from 4×109 cells using RNeasy Maxi Kit (QIAGEN), precipitated with ethanol, and redissolved with DEPC-treated water to at a concentration of 3 mg/ml. mRNA was isolated from 250 μg of the total RNA using Oligotex-dT30 (TaKaRa). Processes from cDNA synthesis to preparation of biotin labelled cRNA for microarray hybridization were performed as described in the Expression Analysis Technical Manual (Affymetrix) using the SuperScript Choice System for cDNA synthesis (Invitrogen), the BioArray High Yield RNA Transcript Labeling Kit (Enzo), and the RNeasy Mini Kit (QIAGEN). 200 μl of the hybridization mixture was applied to a S98 GeneChip (Affymetrix), hybridized for 16 hours at 45°C at 60 rpm in a Hybridization Oven 640 (Affymetrix). The array was washed and stained using EukGE-WS1v3 protocol in a Fluidics Station (Affymetrix), and scanned with an Hp GeneArray Scanner (Affymetrix). Calculation of signal intensities in each array, normalization, and comparison analysis were performed using Microarray Analysis Suite 5.0 (Affymetrix) at the default settings. Exploration of significant Gene Ontology (GO) terms were carried out with web-based GO Term Finder (Boyle et al., 2004) at Saccharomyces Genome Database (http://db.yeastgenome.org/cgi-bin/GO/goTermFinder). Hierarchical clustering, GO-Mapping, and visualization of expression were performed using Genesis version 1.7.0 (Institute for Genomics and Bioinformatics, Graz University of Technology) (Sturn et al., 2002). MySQL database for GO-Mapping was constructed with go_20060827-termdb-tables (the Gene Ontology Consortium). GO-Mapping file is based on gene_association.sgd (Saccharomyces Genome Database).
Northern blot analysis was performed as described previously (Hirota et al., 2001). Total RNA was prepared from 1×109 cells. A total of 2 μg RNA was loaded per lane, separated in a 1.2% agarose formaldehyde gel, and blotted onto a Biodyne Nylon 6,6-B membrane (PALL). Blots were hybridized overnight with randomly 32P-labeled probes (Amersham). DNA fragments used for the random priming labeling were amplified from genomic DNA by PCR, and cloned into the pCR2.1 TOPO vector (Invitrogen). After confirming sequences, each plasmid was digested with appropriate restriction enzymes. Primer pairs used for the PCR and restriction enzymes were as follows. The SPS4 probe was prepared by the PCR amplification using the primer pair (5’-CAAATTTGAATATAGTGAAGGTTACAAAACC and 5’-CTAGATGTGGTAAATCTCCTCCCTCAG), and further exercised by EcoRI (the fragment size is 1020 bp). For the CMD1 probe, 5’-ATGTCCTCCAATCTTACCGAAGAA and 5’-CAAAGCAGCGAATTGTTGAATG, EcoRI, 457 bp. For the DMC1 probe, 5’-TCTGTTACAGGAACTGAGATCGATAGTG and 5’-GTCACTTGAATCGGTAATACCTTTTTCAC, HincII and EcoRI, 772 bp. For the IME2 probe, 5’-CAAGCTTGCTAACGAAATCGATCAGATATGG and 5’-CAAGCTTGAATTGGATAGATGAGGCTCGAAC, HindIII, 1014 bp. For the TCM1 probe, 5’-ATGTCTCACAGAAAGTACGAAGCACC and 5’-CCTTCTTCAAAGTACCCATGAAAGC, EcoRI, 1174 bp. For the YCL048W probe, 5’-AACCGTATTACTAGGAAAAGTTGTTTATTCG and 5’-CAATAACATAGCAACAAGGCATGTACC, EcoRI, 1128 bp. The added restriction recognition sites are underlined.
During the analyses of meiotic phenotypes of mre11 mutant strains, we noticed that null and separation-of-function mutants of MRE11 exhibit deficiency in spore wall development. In the mre11Δ mutants, spore formation is impaired: 39% of spore formation compared to nearly 99% in wild type strains with the SK1 background, as previously reported (Ajimura et al., 1993; Johzuka and Ogawa, 1995; Furuse et al., 1998; Usui et al., 1998). Interestingly, in mre11 mutant cells, morphology and robustness of spore wall were severely affected (Fig. 1A). A couple of days after sporulation, the poorly formed spores from the mre11Δ cells often collapsed, whereas spores from wild type cells are very robust and stably present (Fig. 1A). Such deficiency in spore wall formation may not simply be due to a defect in early meiotic prophase processes such as meiotic DSB formation and meiotic chromosome segregation, since spo11Y135F, a catalytic tyrosine mutant of Spo11 (Bergerat et al., 1997), can efficiently generate rigid spores (Fig. 1A), although they are all inviable.
 View Details | Fig. 1. Effects of mre11 deletion on sporulation. (A) Meiotic cells of wild type, mre11Δ and spo11Y135F. The cells were collected at the indicated time after transfer into SPM and fixed with 3.8% formaldehyde solution, except for the samples in the lowest three panels (3 days + Zymolyase). The cells were collected at indicated time and treated with Zymolyase 100T (lowest panels). (B)–(D) Effects of mre11 deletion within the meiotic gene clusters defined by previous studies (Chu et al., 1998; Primig et al., 2000). Genes with significant changes in the transcript level during meiosis were selected. The 1669 genes for wild type and the 1597 genes for mre11Δ were classified within the clusters defined by Chu and Primig. The percentage of the genes belonging to Chu’s (Chu et al., 1998) (B) or Primig’s (SK1 MAT a/MATα, Primig et al., 2000) (C) clusters is indicated with filled (wild type) and open bar (mre11Δ). (D) Proportion of the genes in the Primig’s clusters of well-correlated genes in both SK1 and W303 (Primig et al., 2000). Note that the mre11 deletion resulted in specific effects on genes classified as Chu’s “middle” and Primig’s “cluster 5 and 5a”. We consider that the difference in Chu’s “late” and Primig’s “cluster 3 and 6a” may not be significant, since the number of genes in these clusters is relatively small. |
To further study the effects of mre11 mutations on spore wall development, we analyzed the genome-wide gene expression profile in mre11Δ cells, and compared it with the one in wild type cells. mRNA samples were purified from cells at 0 hour (t=0h), when preconditioning of cells to sporulation is finished (premeiosis), and 4 hours (t=4h) after transfer into the sporulation medium, when meiotic recombination events actively occur. The latter time point is critical for the expression of later sporulation genes, which are induced during and after meiotic recombination. Transcriptional profiles of mre11Δ and wild type cells were analyzed using Affymetrix Yeast Genome S98 Array that covers approximately 7000 ORFs. In this analysis, we focused on 6135 ORFs excluding some genes such as putative ORFs and Ty elements.
We compared all signals for mre11Δ and wild type at t=0h (premeiosis) with those at t=4h (prophase) using Affymetrix MAS 5.0 software. Considering detection p-values (p-value ≤ 0.04) and change p-values (change p-value ≤ 0.0025 or change p-value ≥ 0.9975), out of 6135 genes, we selected 3905 and 3729 genes as genes with significant transcripts for mre11Δ and wild type cells, respectively. We further set up threshold criteria for meiotically regulated genes at a 1.4 log2 fold (signal log2 ratio > 1.4 or <–1.4, when meiotic values are compared to premeiotic ones) for wild type and at a 1.3 log2 fold (signal log2 ratio >1.3 or <–1.3) for mre11Δ. Such threshold criteria has been used in previous genome wide transcriptional analysis in meiosis (Primig et al., 2000), in which approximately 1600 meiotically regulated genes were identified in the SK1 background strain. With these criteria, we selected 1669 genes (905 up-regulated, 764 down-regulated) for wild type and 1597 genes (924 up-regulated, 673 down-regulated) for mre11Δ as meiotically regulated genes.
Previous genome wide transcriptional analyses by Chu et al. and Primig et al. revealed that transcription levels of about 1000 and 1600 were significantly changed during meiosis in SK1 background strain, respectively (Chu et al., 1998; Primig et al., 2000). These genes were grouped into 8 clusters (metabolic, early I, early II, early middle, middle, mid late, late and repression (Chu et al., 1998)) and 7 clusters (cluster 1~7 (Primig et al., 2000)) based on the expression patterns. In this study, we referred to the first 8 clusters as Chu’s cluster and the late 7 clusters as Primig’s cluster.
To compare our data with the previous genome wide transcriptional analyses, we grouped our meiotically regulated genes according to the criteria for Chu’s and Primig’s clusters. In summary, most of the clusters in mre11Δ and wild type cells show no significant differences from the previous analysis on wild type cells (SK1 background) by Chu et al. and Primig et al. (Chu et al., 1998; Primig et al., 2000), although some of the middle genes, especially those belonging to “Chu’s middle gene cluster” or “Primig’s cluster 5” that contain many known meiosis-specific genes, exhibited distinct difference in levels of transcripts (Fig. 1B–D).
DNA microarray analysis revealed that approximately 70% of genes (1177/1669 for wild type, 1177/1597 for mre11Δ) were meiotically regulated in mre11Δ and wild type cells, as shown in the Venn diagrams are shown in Fig. 2A. The intersection of both sets (659 up-regulated and 518 down-regulated genes) includes many key meiotic genes required for transcriptional regulation (e.g., IME1, IME2, IME4, NDT80), meiotic recombination (e.g., SPO11, DMC1, MSH5), SC formation (e.g., HOP1, RED1, ZIP1), cohesion (e.g., REC8, SCC2, SPO13), and spore formation (e.g., SPO71, SPS100). These indicate that the majority of meiotic genes are similarly regulated in both mre11Δ and wild type cells.
 View Details | Fig. 2. Cluster analyses. (A) Venn diagrams summarizing meiotically regulated genes in wild type cells (1669 genes) and those in mre11Δ cells (1597 genes). The Venn diagram on the right shows the cluster number. The one on the left shows the number of genes in each cluster. Upper regions (up) indicate “genes up-regulated during meiosis”, and lower ones (down) indicate “genes down-regulated during meiosis”. (B) Venn diagrams showing the genes influenced in mre11Δ at meiosis 4 hours. The genes within the clusters shown in (A) were further selected by a criterion for the absolute signal log2 ratio (wild type vs. mre11Δ > 1.4). (C) Number of meiotic genes and total genes in each cluster. Black portions (meiosis) indicate the genes belonging to the meiotic genes classified in either Chu’s or Primig’s analyses (SK1 MAT a/MATα). (D) Classification of the genes in each “cluster diff.” with respect to the categories defined by Chu et al. (Chu et al., 1998). (E)–(G) Proportion of the classified genes in (D) into cluster 1 and cluster diff-1 (E), cluster 2 and cluster diff-2 (F), and cluster 3 and cluster diff-3 (G). Explanatory notes are indicated in (D). (H) Classification of the genes in each “cluster diff.” with respect to the categories defined by Primig et al. (SK1 MAT a/MATα). (I)–(K) Proportion of the classified genes in (H) into cluster 1 and cluster diff-1 (I), cluster 2 and cluster diff-2 (J), and cluster 3 and cluster diff-3 (K). Explanatory notes are indicated in (H). |
To find out the meiotically regulated genes, of which expression is influenced by the mre11Δ mutation, we have set additional criteria: the absolute signal log2 ratios of wild type versus mre11Δ > 1.4. 256 meiotically regulated genes were selected by these criteria and categorized into six subsets (Fig. 2B). We referred to each subset of the two Venn diagram as cluster 1 to 6 (Fig. 2A and Supplementary Table 1) or cluster diff-1 to 6 (Fig. 2B and Supplementary Table 2).
Cluster 1 (meiotically up-regulated in wild type but not in mre11Δ) has 246 genes, out of which 62 genes were significantly up-regulated by the deletion of MRE11 (Fig. 2B, cluster diff-1). We confirmed that 122 out of 246 genes (50%) in cluster 1 and 48 out of 62 genes (77%) in cluster diff-1 are found in the group of meiotically induced genes previously reported (Chu et al., 1998; Primig et al., 2000), but the other type of genes are not classified into that group (Fig. 2C). When we assigned the cluster diff-1 genes to the “meiotic gene clusters” previously identified by Chu et al. (Chu et al., 1998) and Primig et al. (Primig et al., 2000), it turned out that many of genes in cluster diff-1 belongs to the “middle gene” class as defined by Chu et al. (Fig. 2D, “Chu’s middle gene cluster”; Fig. 2H “Primig’s cluster 5”). The proportion of the cluster diff-1 genes classified into the “middle gene” is higher than those in cluster 1 (“Primig’s cluster 5”, 80% versus 46%; “Chu’s middle gene cluster”, 89% versus 65%; see Fig. 2E and I).
We further analyzed these genes using “Gene Ontology (GO) Term Finder” (Boyle et al., 2004). In the analyses, we set the threshold of significant p-value at ≤1.0×10–5. In cluster 1, a group of genes for “reproductive cellular physiological process” scored a particularly significant value (p-value 1.23×10–11) (Fig. 3B and Supplementary Table 3). More importantly, genes for “process of spore wall assembly”, which belong to the gene group for “reproductive cellular physiological process”, exhibited the highest score (p-value 1.3×10–9 in cluster 1, and p-value 9.33×10–16 in cluster diff-1; Fig. 4C, Supplementary Table 3 and 4).
 View Details | Fig. 3. Process annotation and expression profiles of each cluster. Genesis 1.7.0 was used to analyze and visualize the expression patterns (left, wild type; right, mre11Δ) alongside GO process category. Cluster names defined in Fig. 2 (A) and (B) are indicated on the left side of each expression pattern. Gene lists of each cluster were analyzed by GO Term Finder. GO processes are shown on the right side of each expression pattern. Light grey boxes indicate the process annotation of each gene. Red boxes indicate genes with particularly significant changes within the cluster. (A) Scale of signal log2 ratio from low (yellow: expression reduced during meiosis;) to high (purple: expression induced during meiosis). (B) Up-regulated genes at 4 hours after transfer into SPM. (C) Down-regulated genes at 4 hours after transfer into SPM. |
 View Details | Fig. 4. Expression profiles of subcategories in annotated processes. GO processes and cluster number are shown on the left sides of the expression pattern. Subcategories of the GO process are shown on the right side of the expression pattern. (A) Scale of signal log2 ratio from low (yellow) to high (purple). (B) Chromosome segregation and cell cycle. Genes noted (1) (YOR313C, YOL091W, YHL007C) are classified into both chromosome segregation and cell cycle and reproductive process. (C) Genes for reproductive process. (D) Genes for metabolism and transport. |
Cluster 2 and cluster diff-2 consist of 659 genes and 42 genes, respectively (Fig. 2A and B). Genes in cluster 2 were meiotically induced in both wild type and mre11Δ. This cluster includes many genes required for meiosis and spore formation: for example, genes for meiotic recombination (p-value 8.03×10–10), synapsis (p-value 2.22×10–6), M phase of meiotic cell cycle (p-value 1.4×10–20), chromosome segregation (p-value 5.22×10–9), and spore wall assembly (p-value 4.17×10–6) (Fig. 3B, Fig. 4B, and Supplementary Table 5). Therefore, in general, the deletion of MRE11 does not influence the meiotic transcription of most meiotic genes. However, the absolute values of the transcripts for the cluster diff-2 genes (a subset of genes in the cluster 2 genes) further decreased by the deletion of MRE11. As the cluster diff-1 genes, many of the cluster diff-2 genes (79%, 33/42) belong to the previously assigned group of “Chu’s middle genes” or “Primig’s cluster 5” (Fig. 2D, F, H, and J), which include some genes involved in the “process of spore wall assembly” (Fig. 3B, Fig. 4C and Supplementary Table 6).
The cluster 3 includes gene that are significantly induced in mre11Δ but not in wild type cells (Fig. 2A). They are involved in metabolic functions (e.g., nitrogen compound metabolism, catabolism, phosphorous metabolism, and regulation of cellular metabolism), signal transduction, response to stimulus, and protein modification, although their score was not so significant (p-value 0.00021 – 0.06861) except allantoin metabolism (p-value 7.78×10–5) (Fig. 3B, Fig. 4D, Supplementary Table 7 and 8). Unlike cluster diff-1 and cluster diff-2, proportion of middle genes was only modest (Fig. 2G and K).
Most of the meiotically repressed genes (categorized into “cluster 4–6”) are related to metabolism, cytokinesis, ribosomes, and mitochondria (Fig. 3C and Supplementary Table 9–16). Cluster 4 contains 246 genes involved in ribosome biogenesis (p-value 4.23×10–12), amine metabolism (p-value 8.9×10–6), cytokinesis (p-value 6.59×10–5), and steroid metabolism (p-value 3.85×10–5) (Fig. 2A, Fig. 3C, and Supplementary Table 9). Cluster diff-4 contains genes involved in cytokinesis (pvalue 6.59×10–5) (Supplementary Table 10). Cluster 5 consists of 518 genes which were annotated as components of ribosomes (p-value 2.16×10–106) and mitochondria (p-value for mitochondrial ribosome: 1.42×10–45, p-value for mitochondrial envelope: 2.8×10–10) (see Fig. 2A and Supplementary Table 13). Some of the cluster 5 genes are involved in translation (p-value 4.92×10–71) including tRNA metabolism (p-value 3.04×10–5) and ribosomes (Fig. 3C and Supplementary Table 11). Cluster 6 contains components of mitochondrial parts (p-value 4.44×10–8) and ribosome (p-value 7.71×10–7) (Supplementary Table 16). In cluster diff-6, no GO term reached a score high enough to be selected (Supplementary Table 15).
Taken together, these data suggest that the meiotic transcription of a subclass of “middle genes” in cluster 1 and cluster 2, many of which are related to spore wall biogenesis, is specifically and significantly dependent upon Mre11 functions.
Many of the middle genes (~70%) have the “middle sporulation element (MSE)” in their promoter regions (Chu et al., 1998). One of the key meiotic transcription factors, Ndt80, binds to MSE and induces many middle genes (Xu et al., 1995; Chu and Herskowitz, 1998). Most notably, the vast majority of the cluster diff-1 and diff-2 genes with ≥ 2 signal log2 ratio (86% and 80%, respectively) contain MSE in their promoter regions (Fig. 5A).
 View Details | Fig. 5. Promoter activity of pSPS4 harboring MSE. (A) Number of genes (≥ 2 signal log2 ratio) containing at least one MSE in the promoter. Cluster diff-2 genes can be further categorized into two subclusters, cluster diff-2-down (expression levels in wt > mre11Δ) and cluster diff-2-up (expression levels in mre11Δ > wt). The selection of MSE-containing genes among the genes influenced by mre11Δ deletion is based on previous criteria (Chu et al., 1998). The percentage of MSE containing genes is shown in the top of each bar. (B) Reporter gene assay using a pSPS4-lacZ construct. Each diploid mutant carrying the pSPS4-lacZ construct was inoculated on SD-leucine or sporulation medium plates containing X-Gal. Blue-color development was monitored two days after the inoculation. (C) Northern analysis of SPS4 and DMC1 transcripts during meiosis in mre11Δ rad24Δ, mre11Δ, and wild type strains. Ethidium bromide staining of ribosomal RNA is shown in the lowest panel as loading control. (D) Northern analysis of SPS4 and YCL048W transcripts during meiosis in wild type, rad50Δ, spo11Δ, and spo11Y135F diploids. Ethidium bromide staining of ribosomal RNA is shown in the lowest panel as loading control. (E) Northern analysis of SPS4 and YCL048W transcripts during meiosis in wild type, mre11Δ, and mre11ΔC49 diploids. Ethidium bromide staining of ribosomal RNA is shown in the lowest panel as loading control. (F) Northern analysis of SPS4, IME2 and CMD1 (loading control) transcripts during meiosis in wild type and xrs2Δ diploids. |
To examine whether Mre11 is required for the meiotic activation of MSE-containing promoter, we performed an artificial reporter assay. We have constructed a low copy plasmid containing a CEN sequence and the lacZ reporter gene, the upstream region of which has MSE-containing SPS4 promoter (pSPS4) (Hepworth et al., 1995). As indicated in Fig. 4B and C, the meiotic transcription of SPS4 was severely affected by the deletion of MRE11 (signal log2 ratio was 4.6 in wild type, but –0.6 in mre11Δ, representing 39-fold reduction in transcripts). The promoter activity was measured by detecting β-galactosidase activity in wild type and mre11Δ cells on plates containing X-gal. β-galactosidase activity in wild type cells was clearly detected at 48 hours after the inoculation on sporulation plates, but not detected at all on SD-leu plates. In contrast, β-galactosidase activity in mre11Δ cells could be detected in neither plate at the same point (Fig. 5B). This suggests that Mre11 is involved in the meiotic transcriptional activation via pSPS4 rather than increasing the stability of the SPS4 transcripts or modulating the mRNA processing/degradation.
DNA damage or recombination checkpoint can arrest meiotic cell cycle at pachytene, and thereby can prevent meiotic induction of middle genes mainly by down-regulation of the key transcription regulator Ndt80 (Chu and Herskowitz, 1998; Hepworth et al., 1998; Tung et al., 2000; Pak and Segall, 2002). The recombination checkpoint pathway is activated by an accumulation of unrepaired DSB and abnormal synapsis, and depends on Rad24, Rad17, Mec1, Mec3, and Ddc1 proteins (Bailis and Roeder, 2000; Hong and Roeder, 2002; Lydall et al., 1996). Since DSB formation does not occur in the absence of Mre11 (Johzuka and Ogawa, 1995), it is unlikely that the transcriptional defect in mre11Δ is simply due to an uncontrolled activation of the recombination checkpoint pathway.
To confirm this point, we examined the transcription at SPS4 in an mre11Δ rad24Δ double mutant at 6 hours of meiosis by northern analysis. In wild type cells, induction of SPS4 occurs at around 4 hours of meiosis. On the other hand, in both mre11Δ and rad24Δ mre11Δ mutants, the timing of induction is similarly delayed at least by one hour, and the marked reduction in SPS4 transcripts was also similarly detected, whereas the transcription of DMC1 is activated robustly in all strains (Fig. 5C). Therefore, we conclude that the transcriptional defect in mre11Δ is not a caused by the activation of meiotic recombination checkpoint.
We previously reported two separation-of-function mutations in MRE11 (Furuse et al., 1998). An mre11 mutant with a truncation of a C-terminal domain (mre11ΔC49) has a specific defect in the meiotic functions of Mre11 such as meiotic DSB formation. The C-terminal domain of Mre11 has a double strand DNA binding activity (Furuse et al., 1998; Usui et al., 1998). A substitution of N-terminal Asp16 to Ala (Mre11D16A), which is located in conserved phosphoesterase motif, causes mitotic and meiotic DSB repair defect (Furuse et al., 1998). In mre11D16A, meiotic recombination checkpoint is activated by an accumulation of unrepaired DSB, thus it is likely that the activation of middle genes is impaired. Using these two mutants, we conducted the pSPS4-lacZ reporter assay as described above. We found that the pSPS4 promoter was totally inactive in both mutants at 48 hours of meiosis (Fig. 5B), suggesting that the N- and C-terminal domains of Mre11 are involved in the meiotic control of pSPS4-driven transcription, possibly by different mechanisms.
We next investigated whether other components of MRX/MRN complex and Spo11 are involved in the transcriptional regulation of pSPS4 by northern analyses. In spo11Δ, spo11Y135F, and rad50Δ cells, SPS4 (and another MSE-driven gene, YCL048W: signal log2 ratio of 3.7 in wild type versus 1.6 in mre11Δ) transcript strikingly increased at 4 hours of meiosis as observed in wild type cell (Fig. 5D), while mre11Δ and mre11ΔC49 are severely defective in the activation of these genes (Fig. 5E). In xrs2Δ, there seems to be slight delay in the transcription activation at SPS4, but the final level at 6 hours of meiosis is comparable to the wild type level (Fig. 5F). These results suggest that Mre11 (and Xrs2 partially) is involved in the meiotic activation of SPS4 transcription.
We further investigated the expression profiles in spo11Y135F and rad50Δ by microarray analysis. In these mutants, induction of meiotic genes was generally unaffected (Fig. 6A and B). Nevertheless, induction of some meiotic genes (e.g., genes for chromosome segregation, spindle organization, regulation of cell cycle, and many genes required for meiosis) was even promoted in spo11Y135F and rad50Δ (Fig. 6B–E). Transcription of the genes involved in meiotic cell cycle (e.g., genes for meiotic recombination and synapsis) was reduced to some extent in spo11Y135F (Fig. 6C). These seem to be reasonable, since some DSB-defective mutants, such as spo11Δ, spo11Y135F, rad50Δ, rec102Δ, rec104Δ, rec114Δ, enter into the first meiotic division earlier than wild type (Galbraith et al., 1997; Jiao et al., 1999; Kee and Keeney, 2002; Malone et al., 2004). In addition, in rec102Δ and rec104Δ cells, meiotic transcription progresses earlier than wild type (Malone et al., 2004).
 View Details | Fig. 6. Comparison of induction and reduction levels in wild type, mre11Δ, rad50Δ, and spo11Y135F at meiosis 4 hours. (A) Visual presentation of the expression profiles of every genes influenced by the mre11Δ deletion. (B) Visual presentation of the expression of genes induced by expressing Ndt80 ectopically in mitotic cells (Chu et al., 1998). Genes containing at least one MSE sequence in the promoter region are indicated with light purple boxes on the right side of the expression pattern. (C) Expression profiles of genes classified as related to “cell cycle”. Gene categories (meiotic cell cycle, meiosis I, and mitotic cell cycle) are shown on the right side of the expression pattern (indicated with light purple boxes). (D) Expression profiles of genes classified as related to “chromosome segregation and spindle”. Gene categories (chromosome segregation and spindle) are shown on the right side of the expression pattern (indicated with light purple boxes). (E) Expression profiles of genes classified as related to “sporulation”. Gene category (spore wall) is shown on the right side of the expression pattern (indicated with light purple boxes). |
In summary, these results indicate that Mre11 is involved in the activation of a subgroup of meiotic middle genes, many of which are controlled by MSE.
Some of the DSB-deficient mutants, such as spo11Δ and spo11Y135F, can produce inviable but morphologically normal spores, whereas mre11Δ cells form fragile and abnormally shaped spores. This indicates that Mre11, in addition to DSB formation, is involved in some process of spore development. Concomitantly, the transcription of a subclass of middle genes, comprising genes required for meiotic division and spore wall development, is greatly impaired in mre11Δ mutant cells, whereas it appears promoted to some extent in spo11Y135F mutants. In addition, we confirmed that the spore formation deficiency is not due to a cell cycle arrest by the meiotic recombination checkpoint. Therefore, we speculate that the deficiency of mre11Δ in spore formation may be the consequence of a perturbed meiotic transcriptional program.
It is intriguing that most of the genes affected in mre11Δ contain one or several MSE in their promoter. By a reporter assay using the SPS4 promoter (pSPS4)-lacZ construct, we revealed that Mre11 is required for the meiotic transcription to be activated via the ~260 bp pSPS4 sequence harboring MSE. The pSPS4 activity has been shown to depend on the MSE cis-acting element and the Ndt80 trans-acting factor (Chu and Herskowitz, 1998; Hepworth et al., 1995). Thus, Mre11 is possibly directly or indirectly involved in the transcriptional regulation mediated by the MSE-Ndt80 interaction. To further study this point, we examined the relationship between Mre11-dependent and Ndt80-dependent genes. It has been reported that approximately 200 genes are induced in mitosis at over three-fold the basal transcription level when Ndt80 is ectopically expressed (Chu et al., 1998). Among them, we found that about 150 genes were meiotically induced in wild type, mre11Δ, rad50Δ, or spo11Y135F (Fig. 6B). However, the meiotic activation of Ndt80-dependent middle genes was impaired in mre11Δ. Thus, we assume that Mre11 participates in the gene regulation of a subclass, but not of all, of the Ndt80-dependent genes with MSE-containing promoter.
Effects of Mre11 deletion on transcriptional induction of NDT80 were only limited (signal log2 ratio was 3 in wild type versus 1.5 in mre11Δ). Thus, we suggest two possible models for the transcriptional regulation by Mre11. In the first model, Mre11 may promote chromatin binding of Ndt80 to MSE. We previously reported that the Micrococcal nuclease (MNase) sensitivity at DSB hotspots increases prior to DSB formation (Ohta et al., 1994), depending on the presence of the C-terminus domain of Mre11 (Furuse et al., 1998; Ohta et al., 1998). This requirement is very similar to what we observed about the transcriptional effects in the mre11 mutants. Mre11 may be involved in the modulation of the local chromatin configuration at promoter regions (Ohta et al., 1998) as well as at DSB sites (Robert et al., 2006), thereby facilitating the chromatin binding of Ndt80.
The second possible mechanism is that Mre11 is involved in the phosphorylation of Ndt80. Tung et al, reported that the phosphorylation of Ndt80 is required for the up-regulation of its activity (Tung et al., 2000). The Ndt80 phosphorylation is also known to depend on Ime2 (Sopko et al., 2002). It would be therefore interesting to examine the Ime2-mediated Ndt80 phosphorylation in mre11Δ during meiosis in future experiments.
Meiotic events are highly regulated through a sequential activation of premeiotic DNA replication, massive rearrangements in chromosomal architectures (including cohesion, monopolar centromeric alignment, and synapsis formation), and meiotic recombination. As demonstrated in previous publications, some of the early prophase events (e.g., premeiotic DNA replication) are tightly linked to meiotic recombinational events such as DSB formation (Borde et al., 2000; Debrauwere et al., 2001). These relationships are sometimes under the regulation by checkpoint pathways. For instance, when meiotic recombination repair is impeded, meiotic recombination checkpoint arrests the subsequent nuclear division and spore formation (Bishop et al., 1992). In addition to such standard recombination checkpoint, Malone et al. demonstrated that some proteins involved in DSB formation (such as Rec102, Rec104, Mer2/Rec107, Rec114, Rad50, and Spo11) are required for a proper progression of the first meiotic division and the transcriptional activation of middle genes (Galbraith et al., 1997; Jiao et al., 1999; Malone et al., 2004). Indeed, the mutants deficient for the above factors enter into the first meiotic division earlier than wild type, and the transcriptional activation of middle genes in rec102Δ and rec104Δ occur earlier than in wild type. In contrast, the mre11Δ mutant exhibits a specific reduction of transcripts for a subclass of middle genes. Therefore, it is likely that Mre11 not only functions as a “DNA damage sensor” (Stracker et al., 2004), but also might be involved in another type of checkpoint pathway that links meiotic recombination to the transcriptional regulation of middle genes. In this regard, DSB hotspots, which are present in nucleosome-free accessible chromatin regions (mostly in transcriptional promoters containing intergenic intervals), are presumed to function as “controlling centers” that ensure the integrity of the progression into the initial events of meiotic recombination and generate signals for the entry into the later successive phases of spore development. This notion is consistent with the previous observation that Mre11 binds to those DSB sites in meiotic prophase (Borde et al., 2004).
Such relationship between the functions of Mre11 and the regulation of developmental genes might also play a crucial role in somatic cells (Bundock and Hooykaas, 2002; Lussier et al., 1997; Yang et al., 2006). Indeed, it has been reported that hypomorphic mutations in Mre11 cause the “ataxia telangiectasia-like disorder” (ATLD) in human, which exhibits developmental defects in neuronal and immune system (Stewart et al., 1999). Thus, further investigation on the mechanisms of the Mre11-dependent gene regulation in yeast sporulation might provide some important clues to understand the reasons for the developmental defects observed in diseases related to MRN complex defects and chromosomal instability.
The authors thank Y. Ichikawa, R. Nakazawa and Y. Hosono for DNA sequencing (Bioarchitect Research Group, RIKEN) and experimental helps. We are also grateful to Y. Katou and S. Mori (Tokyo Inst. of Tech.), T. Yamada, W. Lin and other laboratory members in RIKEN for helpful discussions. This work was supported by grants for basic research from the Bio-oriented Technology Research Advancement Institution (to K. Ohta and T. Shibata) and grants-in-aid for scientific research on priority areas from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.
|
Index