This Article Abstract Full Text (PDF) Supplemental material Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Gómez, E. B. Articles by Forsburg, S. L. Search for Related Content PubMed PubMed Citation Articles by Gómez, E. B. Articles by Forsburg, S. L.

Schizosaccharomyces pombe mst2⁺ Encodes a MYST Family Histone Acetyltransferase That Negatively Regulates Telomere Silencing

Molecular and Cell Biology Laboratory,¹ Regulatory Biology Laboratory, The Salk Institute, La Jolla, California 92037,³ Molecular and Computational Biology Section, University of Southern California, 1050 Childs Way, Los Angeles, California 90089-2910,² Department of Molecular, Cellular, and Developmental Biology, University of Colorado, Boulder, Colorado 80309⁴

Received 22 February 2005/ Returned for modification 23 March 2005/ Accepted 21 July 2005

	ABSTRACT

Top Abstract Introduction Materials and Methods Results Discussion References

 
Histone acetylation and deacetylation are associated with transcriptional activity and the formation of constitutively silent heterochromatin. Increasingly, histone acetylation is also implicated in other chromosome transactions, including replication and segregation. We have cloned the only Schizosaccharomyces pombe MYST family histone acetyltransferase genes, mst1⁺ and mst2⁺. Mst1p, but not Mst2p, is essential for viability. Both proteins are localized to the nucleus and bound to chromatin throughout the cell cycle. mst2 genetically interacts with mutants that affect heterochromatin, cohesion, and telomere structure. Mst2p is a negative regulator of silencing at the telomere but does not affect silencing in the centromere or mating type region. We generated a census of proteins and histone modifications at wild-type telomeres. A histone acetylation gradient at the telomeres is lost in mst2 cells without affecting the distribution of Taz1p, Swi6p, Rad21p, or Sir2p. We propose that the increased telomeric silencing is caused by histone hypoacetylation and/or an increase in the ratio of methylated to acetylated histones. Although telomere length is normal, meiosis is aberrant in mst2 diploid homozygote mutants, suggesting that telomeric histone acetylation contributes to normal meiotic progression.

	INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

 
DNA is packaged into histone-containing nucleosomes and from there into higher order chromatin structures (45, 50). Not surprisingly, the enzymes that modify chromatin have diverse roles in the cell that affect multiple DNA-dependent events, including transcription, repair, and replication (8, 12, 32, 40, 43, 74). In particular, histones are subject to covalent modifications, including phosphorylation, methylation, ubiquitylation, sumoylation, and acetylation. These modifications create an epigenetic "histone code" that targets chromatin for transcriptional activation or repression, changes in genome stability, response to DNA damage, or differentiation (24, 32, 41, 76, 88).

Histone acetylation on lysine residues in the histone tails is the most extensively studied of these chromatin modifications. Acetylation on different lysines is achieved by a balance in activity of histone acetyltransferases (HATs) and histone deacetylases (27, 41, 75, 76). HAT enzymes can be divided into several broad groups based on their conserved sequence domains. Two large families are the GNAT group, including Gcn5/PCAF, and the Moz-Ybf2/Sas3-Sas2-Tip60 group (MYST) family, and there are several smaller groups, including the TAF1, CBP/p300, SRC-1, and HAT1 enzymes (14, 53).

The MYST family of HATs has been closely examined in the budding yeast Saccharomyces cerevisiae, which has three members: Esa1p, Sas2p, and Sas3p (2, 16, 21, 43, 74, 81). ScESA1 encodes an essential HAT that primarily modifies nucleosomal histone H4, and conditional mutants undergo RAD9-checkpoint-dependent arrest with 2C DNA content (16, 74). The protein is part of the NuA4 and piccolo complexes required for transcriptional activation as well as for double-strand break repair (2, 8, 11, 19). ESA1 interacts genetically with members of the nucleosome assembly complex SPN/FACT (26) and physically with DNA polymerase (25, 26, 43, 67, 91, 92). In addition to its HAT domain, it contains a chromodomain, a motif implicated in assembly of large complexes on specific regions of chromatin (12). In contrast, ScSas2p and ScSas3p lack chromodomains. ScSas2p is part of the SAS-I protein complex and interacts with chromatin assembly factors (56, 68). As well as affecting silencing, Scsas2 suppresses temperature-sensitive (ts) mutations in the origin recognition complex (ORC) subunits ORC2 and ORC5 (21). ScSas2p antagonizes the silencing factor ScSir2p, such that sas2 mutants lead to wider distribution of Sir2-mediated silencing at telomeres (49, 78). ScSas3p is a component of the NuA3 complex which also interacts physically with the SPN/FACT complex (43). Scsas3 is synthetically lethal with Scgcn5, a GNAT family HAT, suggesting the proteins overlap to acetylate histone H3 (38).

In humans, there are additional members of the MYST family. These include Tip60p (which binds to and is repressed by the human immunodeficiency virus type 1 Tat protein), the Mozp (monocytic leukemia zinc finger) protein, which is involved in translocations in acute myeloid leukemia, and the related Morfp factor (32, 85). The recently identified Hbo1 protein also falls into this class (13, 40). All these proteins contain the conserved MYST domain. In addition, Tip60p contains a chromodomain; this motif is also found in ScEsa1p, making these likely orthologues. Mozp and Morfp contain PHD and H15 domains, while Hbo1p contains a zinc finger. These additional domains are not found in ScSas2p and ScSas3p, suggesting there may be some functional divergence between the fungal and metazoan MYST proteins.

Regulated histone acetylation and deacetylation is closely associated with transcriptional activity and particularly in the formation of constitutively silent heterochromatin (59). The protein complexes and histone modifications between heterochromatic regions in budding and fission yeast are different (39, 57). Silencing in S. cerevisiae is mediated by Sir-containing complexes, whereas silencing in Schizosaccharomyces pombe is mediated primarily by Swi6p-containing complexes (39). Fission yeast heterochromatin contains hypermethylated histone H3 K9, whereas budding yeast histone H3 K9 is not methylated (6, 61, 77). Furthermore, a recent report using histone mutants showed that the formation of Swi6-dependent heterochromatin at centromeres is mediated mainly by modifications of the histone H3 tail, whereas Sir3-dependent heterochromatin in budding yeast relies on hypoacetylation of H4 tails and, in particular, of H4 K16 (57). Thus, although both fission and budding yeast require specific histone modification and binding of specialized proteins to assemble heterochromatin, there are substantial differences in the actual modifications and proteins involved (57).

Recent studies also link histone acetylation with DNA replication proteins. The human MYST family member Hbo1p interacts with replication proteins Mcm2p and Orc1p (13, 40). Increased histone acetyltransferase activity correlates with the onset of DNA replication in rat liver cells (89), and histone acetylation facilitates replication elongation in vitro (1). Overexpression of a member of the Gcn5-containing HAT complex partially suppresses the ts phenotypes of an mcm5 mutant (20). Unexpectedly, some budding yeast mcm5 mutants also have defects in telomeric silencing (20), which may implicate replication proteins more directly with silencing.

Based on the differences in heterochromatin in fission and budding yeasts and our long-standing interest in replication, we looked for MYST family HATs in the fission yeast genome. We identified two, which we call Mst1p and Mst2p. Mst1p is an essential protein structurally similar to ScEsa1p and Tip60p. Mst2p is nonessential and clearly related to both ScSas2p and ScSas3p. In this report, we further analyze the phenotypes associated with mst2 and find that loss of Mst2p leads to an increase in telomere silencing. This effect appears to be independent of other S. pombe proteins that influence telomere structure and silencing, including Swi6p, Sir2p, Taz1p, and Rad21p. Instead, our data suggest that the ratio of acetylated to unacetylated histones changes the boundary of silencing at the telomere. The mst2 mutants suffer a disordered meiosis that might also reflect loss of normal telomere functions. This study suggests a distinct function for SpMst2p in the regulation of telomere function.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

 
Strains and manipulations. Strains used in this study are listed in Table 1. Strains were grown and maintained on yeast extract plus supplements (YES) or Edinburgh minimal medium (EMM) with appropriate supplements, using standard techniques (60). Matings were performed on synthetic sporulation agar (SPAS) (35) plates for 2 to 3 days at 25°C. Temperature-sensitive cells were grown at 25°C, and non-ts cells were grown at 32°C. Transformations were carried out by electroporation (48). Double mutants were constructed by standard tetrad analysis or random spore analysis. G418 plates were YES supplemented with 100 µg/ml G418 (Sigma).

Strains FY1812 and FY1818 expressing Mst1-HA and Mst2-HA, respectively, from their own locus and endogenous promoter were constructed as follows. Plasmids pEBG72 (nmt-mst1-HA) and pEBG73 (nmt-mst2-HA) were constructed by PCR amplifying and subcloning mst1⁺ and mst2⁺ genes into the triple-hemagglutinin (HA)-tagging nmt expression vector pSLF172 (28). Next, plasmids pEBG72 and pEBG73 were digested with AccI/SmaI and HincIII, respectively, and the fragments containing the 3' end of the tagged genes were subcloned into pJK148 (47), creating pEBG80 (mst1) and pEBG90 (mst2). Following linearization with BclI (mst1) and StuI (mst2), pEBG80 and pEBG90 were integrated into the mst1 or mst2 locus, respectively, creating a partial tandem duplication with the full-length copy expressing the tagged protein and an unexpressed gene fragment. Expression of the tagged proteins was confirmed by Western blotting (unpublished data).

The protein sequence alignment and phylogenetic tree for the MYST family members were generated using Clustal X (ftp://ftp-igbmc.u-strasbg.fr/pub/ClustalX/) and Phylip (http://evolution.genetics.washington.edu/phylip.html). BLAST 2 sequence alignment was used to calculate the percentage identities, similarities, and gaps between the MYST domains of Mst2p and the other MYST family members (84).

In situ chromatin binding assay. An in situ chromatin binding assay was performed as described previously (31). Cells were mounted and stained with 4',6-diamidino-2-phenylindole (DAPI). Microscopy was performed with a Leica DMR microscope. Images were captured with a Hamamatsu digital camera using Improvision Openlab software and assembled in Canvas 8.0 (Deneba).

UV, TBZ, HU, and MMS sensitivity assays. Cells were grown to an A₅₉₅ of 1, serially diluted fivefold, and spotted onto YES plates or YES plates containing 10 µg/ml thiabendazole (TBZ), 5 mM hydroxyurea (HU), or 0.005%, 0.007%, and 0.01% methyl methanesulfonate (MMS) and incubated for 3 to 5 days at 25°C or 32°C. For UV irradiation assays, 1,000 cells were plated in duplicate onto YES and immediately treated with the indicated doses of UV irradiation using a Stratalinker (Stratagene). Control plates were left untreated. Plates were incubated at 32°C for 2 to 3 days. Colonies were counted, and percentage survival was obtained relative to the number of colonies grown on the untreated control plates.

Silencing assays. To analyze silencing using the ade6⁺ marker, mutant strains were crossed to wild-type cells that had the full-length ade6⁺ gene inserted in the different heterochromatic regions and a truncated ade6 gene at the ade6 endogenous locus (ade6N/N). The presence of full-length ade6⁺ and ade6N/N was verified by PCR using the following primers: ade6-F, 5'-CCAGGAAAGTGTTGAAAAAG, and ade6-R, 5'-CTTCAAACTGAGAAGTTGGG. Wild-type strains FY2328, FY2996, and FY2997, mst2 strains FY2364 and FY2365, sir2 mst2 strain FY2700, sir2 strain FY2702, taz1 strain FY2367, mst2 taz1 strains FY3298 and FY3299, swi6 strain FY3293, and mst2 swi6 strain FY3295 were used for telomeric silencing assays. For TM(cnt) centromeric region analysis, wild-type FY3036 and mst2 FY3039 strains were used, and for the outer repeat region (otr), wild-type FY2219 and mst2 FY2412 strains were used. For the mating type region, wild-type FY3001 and mst2 FY3003 strains were used. To analyze colony color, cells were streaked on EMM plates lacking leucine and uracil and containing a low concentration of adenine (low adenine; 7.5 µg/ml) and were incubated for 3 days at 32°C. To analyze telomeric silencing using the ura4⁺ reporter, mst2::kanMX6 mutants were crossed to cells carrying minichromosome Ch16-M23 with the ura4⁺ marker inserted immediately after the telomeric repeats, and for the analysis of the "native" telomere the mutants were crossed to cells with ura4⁺ inserted 300 bp from the left telomere of chromosome 1, tel1L::ura4⁺. For centromeric inner (imr) region silencing analysis, mst2::kanMX6 mutants were crossed to cells with ura4⁺ inserted in the imr region of centromere 1. Wild-type FY3043 and mst2 FY3044 were used for telomeric silencing assays using the ura4⁺ reporter. Wild-type FY1589 and mst2 FY3046 were used for inner centromeric silencing (imr::ura4⁺). Cells were grown to an A₅₉₅ of 1, serially diluted fivefold, spotted onto YES, EMM-ura, or EMM supplemented with 1 mg/ml 5-fluoroorotic acid (5-FOA) and incubated for 3 to 4 days at 32°C.

Telomere length analysis. Strains FY1421 (taz1), FY1204 (rhp51), FY2056 (swi6), FY1889 (mst2), and FY258 (wild type) were grown for 50 generations, and genomic DNA was isolated. Fifteen micrograms of DNA was digested with EcoRI, separated on a 1% agarose gel, and analyzed by Southern blotting using as a probe the G-specific oligonucleotide 5'-GGGTTACAGGTTACAGGTTACA-3' (17).

RNA purification and reverse transcriptase PCR (RT-PCR) analysis. Cells were grown to an A₅₉₅ of 0.5 to 0.7. Total RNA was extracted with the RNeasy kit (Invitrogen) following the manufacturer's instructions. RNA was quantified, and its integrity was checked by agarose gel. First, cDNA strand was prepared with equal concentrations of total RNA using the Superscript First Strand Synthesis System for RT-PCR kit (Invitrogen), following the manufacturer's instructions, and was used in PCR amplifications. The expression of the full-length ade6⁺ (tel::ade6⁺) and the truncated ade6 (ade6) gene were quantified by real-time quantitative PCR using SybrGreen (Applied Biosystems) as a marker for DNA amplification on an ABI Prism 7900HT apparatus (Applied Biosystems). Primers used specifically amplify full-length ade6⁺ or the truncated mini-ade6 (Table 2) and were designed with Primer Express software (Applied Biosystems). Quantified cDNA was used to obtain a standard curve to estimate the amount of DNA in each sample. Results are plotted as fold expression relative to the wild type after normalization to 18S rRNA values.

Meiosis time course. Wild-type strains FY253 and FY258 and mst2 strains FY1889 and FY1890 were mated on SPAS plates for 12 h at 25°C and streaked onto plates lacking adenine to select for diploids. Colonies were streaked on YES and replica plated to YES plus phloxin B. Dark pink colonies (diploid cells) were isolated and streaked on YES. Eight independent mst2 diploids and four independent wild-type diploids were patched on SPAS plates at 25°C to induce meiosis and ethanol fixed after 15 h. Cells were stained with DAPI, and 200 asci per diploid were analyzed. Asci with more than four DAPI spots or unequal DAPI-stained bodies were considered "abnormal asci." Normal asci were those with four equal DAPI-stained bodies. Cells were visualized with a Leica DMR microscope, and images were captured with a Hamamatsu digital camera and Improvision Openlab software (Improvision).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

 
S. pombe has two distinct MYST family HAT proteins. A series of BLAST searches using different MYST domain sequences revealed that S. pombe has two open reading frames, SPAC637.12c and SPAC17G8.13c, with high similarity to MYST family histone acetyltransferases. We named these two putative acetyltransferases Mst1p and Mst2p, respectively. The Mst2p sequence was first identified, but not molecularly characterized, by Reifsnyder et al. (70). Figure 1A shows a phylogenetic tree and a schematic representation of different MYST family acetyltransferases comparing S. pombe Mst1p and Mst2p with other MYST domain proteins. mst1⁺ codes for a 463-amino-acid protein, 70% similar to S. cerevisiae Esa1p. Like Esa1p and Tip60p, Mst1p has a chromodomain motif, which is implicated in chromatin association (12, 44). The mst2⁺ gene encodes a 407-amino-acid protein with homology restricted to the MYST domain and amino acid similarity ranging between 50 to 70% (Fig. 1A). Mst2p does not contain a chromodomain or other motifs outside of the MYST/HAT domain, similar to S. cerevisiae Sas2p and Sas3p.

View larger version (26K): [in this window] [in a new window]   FIG. 1. (A) Phylogenetic tree and schematic alignment showing the relationship between putative MYST family acetyltransferases, S. pombe (Sp) Mst1p (GenBank accession number AL034583) and Mst2p (GenBank accession number Z69795), human (h) Hbo1p, Tip60p, Mozp, and Morfp, and S. cerevisiae (Sc) Esa1p, Sas2p, and Sas3p. Zinc fingers (zf), chromodomains (chromo), and H1 and PHD domains are indicated. Percentage of identity, similarity, and gaps with respect to the MYST domain of SpMst2p are shown below the MYST domains of each protein (identity/similarity/gap). Numbers above the Mst2p MYST domain indicate amino acids (aa). (B) SpMst1p is essential for cell survival. One copy of the mst1+ gene was disrupted in diploid cells with ura4+. Ura+ transformants were induced to undergo meiosis, and spores were analyzed by tetrad dissection on YES plates. After 3 to 5 days at 25°C, colonies were replica plated to minus uracil plates (–ura) and incubated for 3 to 5 days at room temperature.

Spmst2

Mst1p and Mst2p are nuclear and chromatin bound throughout the cell cycle. To localize Mst1p and Mst2p we tagged the endogenous genes with simian virus 5 and HA epitopes and performed in situ chromatin binding assays (31, 46). Using this technique, we can visualize the localization of proteins in the cell and their association with chromatin. When permeabilized cells are treated with 1% Triton X-100 before fixation for immunofluorescence, proteins not tightly bound to chromatin are released from the nucleus. An example of this method can be visualized with the Mcm2-HA and Orp1-HA control proteins (Fig. 2A). SpMcm2 is nuclear and binds chromatin at the end of M phase, staying bound through G₁ and S (binucleate cells in S. pombe); SpOrp1p is nuclear and binds chromatin throughout the cell cycle. When cells are not treated with the detergent, Mcm2-HA and Orp1-HA are visualized in the nuclei of mononucleate (cells in G₂ phase) and binucleate cells (cells in M/G₁/S phase); when treated with 1% Triton X-100, Mcm2-HA is only in binucleates, but Orp1-HA is in both mono- and binucleate cells (Fig. 2A). Both Mst1p and Mst2p behaved like Orp1-HA: they were localized in the nucleus and bound to chromatin throughout the whole cell cycle. Under these experimental conditions we saw no evidence for region-specific localization. The remainder of this report will address the function of nonessential mst2⁺, and further characterization of mst1⁺ will be described in a separate report.

View larger version (87K): [in this window] [in a new window]   FIG. 2. (A) SpMst1p and SpMst2p are nuclear- and chromatin-bound throughout the cell cycle. The wild type (FY258) and strains expressing Mst1-HA (FY1812), Mst2-HA (FY1818), Mcm2-HA (FY798), and Orp1-HA (FY793) were permeabilized and washed with the indicated concentrations of Triton X-100. DNA was visualized by DAPI staining (DNA) and HA-tagged proteins using the HA antibody and a secondary antibody conjugated to Cy3 (HA). Bar, 10 µm. (B to E) mst2 cells are sensitive to HU (A) and slightly sensitive to MMS (D) and UV irradiation (E). For HU, TBZ, and MMS sensitivity analyses, cells were grown in liquid cultures to an A595 of 1, serially diluted fivefold, spotted on the indicated plates, and incubated at 25°C for 3 to 5 days. Equal numbers of cells were spotted on each plate. To analyze UV sensitivity, 1,000 cells were plated on YES, irradiated with 50, 100, 150, 200, or 300 J/m2, and incubated at 23°C for 3 days in the dark. Cells were counted and the percentage survival with respect to the number of colonies on the untreated plates was plotted versus the UV dose. Strains used include the wild type (FY258), rad3 (FY1107), hsk1ts (FY2104), mst2 (FY1889), rad21ts (FY1159), mst2 rad21ts (FY2032), swi6 (FY2056), taz1 (FY1421), and mst2 taz1 (FY2330 and FY2331). WT, wild type.

mst2 genetically interacts with mutants that affect heterochromatin, cohesion, and telomere structure.

rad3

View this table: [in this window] [in a new window]   TABLE 3. mst2 genetic interactionsa

mst2

taz1

mst2

Since Taz1p, Swi6p, and proteins involved in cohesion all affect telomere function, we investigated the effect of Mst2p at the telomeres. We compared this to its effect at the centromere and mating type locus, two other well-characterized heterochromatic domains in fission yeast.

Mst2p is a negative regulator of silencing at telomeres. Silencing in fission yeast has largely been studied in regions of heterochromatin at the telomeres, the centromeres, and the mating type loci (34). Although there are some factors specific for each region, they all share a common pathway that involves sequential deacetylation and methylation of histones, which allows binding of the Swi6/HP1 heterochromatin protein and association of cohesin proteins (6, 7, 61, 66). Swi6p and other components of this pathway are conserved in fission yeast and metazoans but are not found in budding yeast.

We investigated the role of Mst2p in heterochromatin silencing by analyzing the effect of mst2 mutation in the expression of a reporter gene when inserted at a telomere, mating type, or centromeric regions cnt (TM), imr, and dg. For telomeric silencing we used a strain that has the ade6⁺ gene inserted immediately proximal to the telomeric repeat of the minichromosome Ch16-M23 right telomere (62). This minichromosome is a 350-kb derivative of the original 530-kb Ch16-M23 minichromosome (64) and lacks telomere-associated sequence (TAS) elements. The expression of this gene is subjected to position effect variegation, and cells plated on low adenine can exhibit red-pink-white-sectored colonies (Fig. 3A; see also Fig. S1 in the supplemental material) (3). Cells lacking mst2 and having ade6⁺ at the Ch16-M23 telomere were streaked on low-adenine plates lacking leucine to select for the minichromosome and were analyzed. mst2 cells show a deep red colony color, indicating that Mst2p negatively regulates telomeric silencing (Fig. 3A; see also Fig. S1 in the supplemental material). The presence of the ade6⁺ gene and ade6N/N was confirmed by PCR (Fig. 3B). Similar results were observed for independent isolates, and the phenotype was complemented by plasmid-borne mst2-HA (data not shown). To analyze the expression of the ade6⁺ gene at the telomere in wild-type and mst2 cells, we purified total RNA and used random hexamers to prepare the cDNA strand. Two different amounts of cDNA were used in lineal PCR amplifications, and after 26 cycles only the truncated ade6 transcript (ade6) could be amplified in wild-type and mst2 cells (Fig. 3B). As a control we used a strain lacking the sir2 gene (see below) and carrying Ch16::LEU2⁺ tel::ade6⁺, where the expression of the full-length ade6⁺ gene at the telomere is derepressed. After 34 cycles, the expression of the full-length ade6⁺ gene at the telomere of wild-type cells was detected, while no amplification was observed in mst2 mutants (Fig. 3B). We quantified the expression of both the full-length ade6⁺ gene at the minichromosome telomere and the truncated ade6 at its own locus by using real-time quantitative PCR. We designed primers that would only amplify the full-length ade6⁺ gene or the truncated one (Table 2) and used them in real-time quantitative PCR experiments (Fig. 3C). Data represent fold expression relative to the wild type after normalization to 18S rRNA values. As shown in Fig. 3C, in wild-type and mst2 cells the truncated ade6 gene is similarly expressed, while the expression of the full-length ade6⁺ is approximately 25 times higher in wild-type cells. These results indicate that in mst2 cells the expression of the ade6⁺ gene at the telomere is specifically repressed while no effect is seen in the expression of the truncated ade6 gene located in ade6's own locus.

View larger version (40K): [in this window] [in a new window]   FIG. 3. Mst2p negatively regulates telomeric silencing of ade6+. (A) Silencing assay using the ade6+ gene at the telomere of minichromosome Ch16. Wild-type (FY2997), mst2 (FY2364), taz1 (FY2367), mst2 taz1 (FY3298 and FY3299), sir2 (FY2702), mst2 sir2 (FY2700), swi6 (FY3293), and mst2 swi6 (FY3295) strains carrying minichromosome Ch16-M23 tel::ade6+ and the ade6N/N minigene were streaked on low-adenine plates lacking leucine and incubated at 32°C for 3 days. (B and C) The expression of the full-length ade6+ gene at the minichromosome Ch16 telomere is highly repressed. Wild-type (FY2997), mst2 (FY2365), and sir2 (FY2702) cells were grown to an A595 of 0.5 to 0.7. Genomic DNA and total RNA were extracted and quantified. The cDNA strand was prepared with equal concentrations of total RNA and used in PCR amplifications (B). To ensure PCR linearity, different concentrations of cDNA were used. Primers employed amplify both the full-length ade6+ at the Ch16 telomere (tel::ade6+) and the truncated ade6 (ade6) gene at its own locus. (C) Real-time quantitative PCR analysis. Expression of the full-length ade6+ (tel::ade6+) and the truncated ade6 (ade6) gene were quantified in wild-type (FY2997) and mst2 (FY2365) cells using primers that specifically amplify full-length ade6+ or the truncated mini-ade6. Results are plotted as fold expression relative to the wild type after normalization to 18S rRNA values. WT, wild type.

32 ade6N/N

taz1

It was recently shown that the S. pombe sir2⁺ gene is required for normal telomeric silencing (29, 72). Data from budding yeast suggest that ScSas2p antagonizes ScSir2p, and the increased silencing observed in some budding yeast sas2 strains reflects wider distribution of ScSir2p (49, 78). We analyzed the role of SpSir2p in the increased telomeric silencing observed in mst2 mutants. As predicted, cells lacking both sir2 and mst2 exhibit a white colony color when streaked in low-adenine plates lacking leucine, indicating that the increased silencing in mst2 cells remains Sir2p dependent (Fig. 3A).

S. pombe heterochromatin silencing is mediated primarily by Swi6p-containing complexes (39). As shown in Fig. 3A, swi6 cells as well as the double mutant mst2 swi6 show a white colony color when streaked on low-adenine plates lacking leucine, indicating that Swi6p is also needed for the increased silencing observed in mst2 cells.

We further investigated the role of Mst2p in telomeric silencing by analyzing the effect of a mst2 mutation (mst2::kanMX6) on the expression of a ura4⁺ reporter gene inserted immediately proximal to the telomeric repeat of minichromosome Ch16-M23 (62). This construct also lacks TAS elements. Silencing of ura4⁺ can be detected by decreased growth on medium lacking uracil or by increased growth on the toxic substrate 5-FOA; loss of silencing is manifested as increased growth on medium lacking uracil and decreased growth on 5-FOA (10). As shown in Fig. 4A, wild-type cells (Ch16 tel::ura4⁺) can grow on plates lacking uracil and adenine and do not grow on 5-FOA, indicating that in this construct the ura4⁺ gene located at the telomere is actively transcribed. Four different isolates were analyzed for mst2 (mst2::kanMX6) (Fig. 4A). All isolates can grow as wild types on plates lacking uracil and adenine, suggesting that cells mutant for mst2 express the ura4⁺ gene inserted in the minichromosome telomere to levels similar to that of the wild type. Interestingly, all isolates can grow, to different extents, on 5-FOA, indicating that some cells are repressing the expression of the ura4⁺ gene, consistent with variegated expression. These results show that mst2 deletion causes a slight increase in telomeric silencing when using the ura4⁺ marker.

View larger version (69K): [in this window] [in a new window]   FIG. 4. Mst2p negatively regulates telomeric silencing of ura4+. (A) Silencing assay using the ura4+ gene at the telomere of minichromosome Ch16. The wild type (FY3284) and different isolates of mst2 (1, FY3289; 2, FY3290; 3, FY3291; and 4, FY3292) carrying minichromosome Ch16-M23 tel::ura4+ and the ura4-DS/E minigene were grown to an A595 of 1 on EMM-adenine, serially diluted fivefold, spotted onto EMM-adenine, EMM-adenine-uracil, or EMM-adenine plus low uracil, supplemented with 1 mg/ml 5-fluoroorotic acid (5-FOA) plates, and incubated for 3 to 4 days at 32°C. For growth control on the different plates, uracil minus (h– ura4-D18 leu1-32) and uracil plus (h– leu1-32) strains were used. (B) Silencing assay using the ura4+ gene inserted at a native telomere (tel1::ura4+) or the centromeric imr (imr::ura4+) region. For the telomere assay, wild-type (FY3043) and mst2 (FY3044) strains were used; for the centromere imr region analysis, wild-type (FY1589) and mst2 (FY3046) strains were used. Cells were grown to an A595 of 1, serially diluted fivefold, spotted onto YES, EMM-uracil, or EMM plus low uracil, supplemented with 1 mg/ml 5-fluoroorotic acid (5-FOA) plates, and incubated for 3 to 4 days at 32°C. For growth control on the different plates, the uracil minus strain FY258 was used. (C) PCR amplifications to analyze the expression of the ura4+ (tel::ura4+) and the truncated ura4 minigene (ura4-DS/E). The wild type (FY3284) and mst2 isolate 1 (1-mst2 Ch16 tel::ura4+; FY3289) carrying minichromosome Ch16-M23 tel::ura4+ and wild type FY3043 and mst2 FY3044 with the ura4+ marker inserted at a native telomere were grown to an A595 of 0.5 to 0.7. Genomic DNA and total RNA were extracted and quantified. The cDNA strand was prepared from equal concentrations of total RNA and used in PCR amplifications. To ensure PCR linearity, different concentrations of cDNA were used. Primers employed amplify both full-length ura4+ (tel::ura4+) and truncated ura4 (ura4-DS/E). (D) mst2 cells have normal telomere length. Cells were grown in liquid cultures for 50 generations. Genomic DNA was isolated, digested with EcoRI, separated on a 1% agarose gel, and analyzed by Southern blotting with an -32P-labeled telomere repeat DNA probe. WT, wild type.

To extend our analysis of the role of Mst2p in silencing, we also analyzed the effect of mst2 in the expression of the ade6⁺ gene at other heterochromatic regions: the mating type and the central domain, cnt, and the outer repeat, dg, of the centromere. No difference in colony color was observed between wild-type and mst2 cells when streaked on low-adenine plates (data not shown). Similarly, when using the ura4⁺ gene inserted in the centromeric inner (imr) region, there was no difference in growth on plates lacking uracil or with 5-FOA between wild-type and mst2 cells (Fig. 4B). Similar results were observed for independent isolates (data not shown). Taken together, these results indicate that Mst2p is a negative regulator of telomeric silencing and does not affect heterochromatin at the centromere or the mating type.

Mst2p does not affect telomere length. Silencing at telomeres is affected by general heterochromatin factors, such as Swi6p (4), as well as specific telomere factors that influence telomere length, such as Taz1p (17). Because mst2 cells have increased silencing at telomeres, we investigated telomere length in these mutants. We compared the telomere lengths of mst2 to telomeres in wild-type, taz1, and swi6 cells. As a control for slight changes in telomere length we used rhp51 mutant cells. Figure 4D shows that mst2 telomere length is similar to that of the wild type and swi6 (23), indicating that Mst2p does not have a role in telomere length control. As previously shown, taz1 has very long telomeres (17). We observe that rhp51 has slightly longer telomeres than the wild type; this is in contrast to a previous report (90) but has been seen by others (S. G. Pasion, personal communication) and could reflect some uncharacterized strain differences. These results show that the effect of Mst2p on telomere silencing is independent of telomere length control and, presumably, independent of Taz1p and suggests that mst2 and taz1 have synthetic growth defects because they independently affect separate aspects of telomere function.

Mst2p does not affect the heterochromatic localization and distribution of Swi6p, Rad21p, Sir2p, and Taz1p. Our silencing experiments indicate that Mst2p is a negative regulator of telomeric silencing and requires Sir2p and Swi6p for these effects. However, we found that mst2 affected expression of reporters in different telomere constructs quite differently. In an independent approach, we therefore examined the effect of Mst2p on normal wild-type telomeric chromatin unperturbed by any reporter constructs and created a map of histone modifications and localization and distribution of telomeric proteins by ChIP and real-time quantitative PCR (Fig. 5A). This provides the first detailed census of proteins and histone modifications in the context of a normal, fully wild-type telomere. We compared this to the effect of Mst2p on centromeric and mating type heterochromatin regions (Fig. 5B).

View larger version (54K): [in this window] [in a new window]   FIG. 5. Mst2p positively regulates telomeric histone acetylation. ChIP and real-time quantitative PCR were carried out to analyze the heterochromatic localization and distribution of Swi6p, Rad21-HA, and Taz1-HA proteins as well as modified histones in wild-type and mst2 cells. For acetylated histones and Swi6p ChIPs, wild-type h– FY2997 and h+ FY2996 as well as and mst2 h– FY2365 and h+ FY2364 strains were used. For H3 K9-methylated histones, the h– strains were analyzed. For Rad21-HA ChIPs, wild-type h– FY3049 and h+ FY3048 as well as mst2 h– FY2903 and h+ FY2902 strains were used. For Taz1-HA, wild-type h– FY2784 and h90 FY2392 as well as mst2 h– FY2787 and h90 FY2786 strains were used. For Taz1-HA and Rad21-HA, the untagged control immunoprecipitation values were subtracted from the tagged ones; for Swi6p and modified histones, the minus-antibody values were subtracted. Data for each strain was obtained from two to four independent immunoprecipitations. Mean values ± standard deviations are plotted. The percentage of DNA precipitated relative to the amount of input DNA is plotted for each amplicon. Note that the scale of the y axis is different from ChIP to ChIP. (A) Telomere protein occupancy and histone modifications analysis. Real-time quantitative PCR was carried out with primers corresponding to telomeric regions 0.38, 2.3, 6.2, 7.1, 9, 10.6, 12.2, 14.2, 15.6, 16.9, 17.6, 18.2, and 18.9 kb from the telomeric repeats. For Swi6p, data for h– and h+ cells are shown. As a nontelomeric region, the fbp1+ promoter region was analyzed. A scheme of the telomere including positions of primers used in ChIP analysis and described putative coding sequences in cosmids c212 in chromosome 1 and pT2R1 in chromosome 2 is depicted along the x axis. Arrows indicate gene orientation. (B) Centromeric and mating type region analyses. Real-time quantitative PCR was carried out with primers corresponding to the centromeric regions cnt, imr, dg, and dh and cryptic mating-type region. IP, immunoprecipitation; wt, wild type.

We examined association of the telomeric proteins Swi6p, Rad21p, Sir2p, and Taz1p, which are known to be involved in telomeric silencing; we already showed that mutations in these show genetic interactions with mst2 (Table 3 and data not shown). We observed no difference between wild-type and mst2 cells in the localization and distribution of Swi6p, Rad21p, Taz1p (Fig. 5A), and Sir2p (data not shown) (29) in any of the heterochromatic regions analyzed. Thus, within the resolution of this experiment, Mst2p does not influence the localization of known heterochromatin or telomere proteins, and its effects on silencing are likely to be mediated by other means or at a subtle level beneath our detection threshold.

Interestingly, we found Swi6p distribution in telomeres changed with the mating type, although the total amount of Swi6p in the cells was similar (see Fig. S2 in the supplemental material). h^– strains had 3-fold more Swi6p in the telomeric region located 6.2 kb from the telomeric repeats, with a modest but still detectable difference observed in the 7.1-kb region (Fig. 5A). Thus, in h^– strains Swi6p levels are low 0.38 kb from the telomeric repeats, slightly increase towards the 2.6-kb region, and show an abrupt increase at 6.2 kb before dropping back to levels similar to those in the 2.6-kb region (Fig. 5A). Similar to h^– strains, h⁺ cells are enriched for Swi6p 0.38 kb from the telomere end and slightly increased 2.6 kb from the repeats, but at 6.2 and 7.1 kb the levels only slightly increased (Fig. 5A). Swi6p localization and distribution have recently been reported for h⁺ cells (29). In light of this unexpected result, we analyzed silencing in h^– and h⁺ strains using Ch16-M23 tel::ade6⁺. No differences were observed in colony color after streaking cells onto low-adenine plates (see Fig. S1 in the supplemental material), which suggests that this difference in the levels of Swi6p does not affect silencing.

Rad21p levels are low proximal to the telomere end but start increasing 9 kb away from the repeats and reach a maximum level 12.2 kb from the telomere, after which they plateau (Fig. 5A). Interestingly, although it was shown that Rad21p cohesin binding to heterochromatin is Swi6p dependent (7), these proteins do not have the same telomeric distribution; furthermore, there are no differences between h⁺ and h^– strains. However, at centromere I and the mating type region, Swi6p and Rad21p show similar distribution patterns, as expected (Fig. 5B) (7, 69).

Taz1p is highly enriched close to the telomeric repeats (0.38 kb), drops in abundance in the region between 2.6 kb and 9 kb, and is almost undetected 10.6 kb away from the telomeric repeats (Fig. 5A). Similar results were recently published by Sadaie et al. and Freeman-Cook et al. (29, 71). We found no difference in h^– and h⁹⁰ strains. We also found no difference in Sir2p heterochromatin localization and distribution in mst2 compared to what we observed in the wild type (29 and data not shown).

Mst2p positively regulates telomeric histone acetylation. Since Mst2p had no strong effect on distribution of telomeric heterochromatin proteins, we next analyzed histone modifications in heterochromatic regions. We used antibodies to acetylated histone H4 (H4Ac), acetylated lysine 9 histone H3 (H3K9-Ac), and methylated lysine 9 histone H3 (H3K9-Me) in ChIP experiments, followed by real-time quantitative PCR analysis. For H4 and H3 K9 acetylation we analyzed the same heterochromatic regions as described above for h^– and h⁺ strains; no differences between mating types were observed with either antibody. For H3 K9 methylation we only analyzed telomeric heterochromatin in h^– strains. When using this antibody no differences were observed between the wild type and mst2. Histone H3 K9 methylation increases with the distance from the telomeric repeats, following a similar pattern as the one observed for Rad21p (Fig. 5A).

Our analysis using H4Ac and H3K9-Ac antibodies showed that in wild-type cells, histone acetylation increases steadily with the distance from the telomeric ends with a pronounced transition to higher levels occurring between 15.6 and 16.9 kb. In mst2 mutants, histone acetylation also increases with the distance from the telomeric repeats, reaching a maximum in the 10.6- to 12.2-kb region (Fig. 5A). However, the sharp transition to higher levels is not seen, and in the region beyond 16.9 kb, levels are reduced by 2.5-fold (Fig. 5A). Thus, Mst2p affects the acetylation state of this region. We examined the sequence in this transition region and found it contains a pseudogene (SPAc212.09c) with two frameshifts (93) (Fig. 5A). In contrast, no difference in histone acetylation was detected between wild-type and mst2 cells in the centromeric and mating type regions (Fig. 5B). This suggests that region-specific differences distinguish the telomeres from other heterochromatic loci.

Homozygous mst2 diploid cells have meiotic defects. Silencing and ChIP experiments suggest that cells lacking mst2 have some telomere abnormalities. Telomeres are crucial during meiosis at several levels. Telomere-led movement, or "horse-tailing," is necessary for correct chromosome pairing (94), and many telomere mutants disrupt normal meiotic progression (for examples see references 15, 18, and 63). Since many mst2 phenotypes presented above are subtle, we reasoned that meiosis might sensitize the cells to any mst2-associated telomeric defects. Wild-type and mst2/mst2 diploids were induced to undergo meiosis, and after 15 h they were ethanol fixed and DAPI stained. As shown in Fig. 6, mst2/mst2 diploids have 28% abnormal asci. The majority of the abnormal asci had more than four DAPI spots and unequal DAPI-stained bodies, suggesting that they have mis-segregated or fragmented their chromosomes during meiosis. Wild type/wild type was indistinguishable from mst2/wild-type heterozygous diploids, with only 6% abnormal asci, so there is no haploinsufficiency (Fig. 6 and data not shown). When intact spores from tetrads were dissected or released by random spore analysis, their viability was not significantly reduced, suggesting that the abnormal segregants were not packaged into recognizable spores. Analysis of recombination frequency in the his4-lys4 interval shows no obvious difference between wild-type and mst2 diploids (data not shown).

View larger version (59K): [in this window] [in a new window]   FIG. 6. mst2 homozygote diploids have meiosis defects. Wild-type and mst2 homozygote diploids were isolated by mating h– and h+ haploid cells and ade6-M210/ade6-M216 complementation. Eight independent mst2 diploids and four independent wild-type diploids were induced to undergo meiosis and were ethanol fixed after 15 h. Cells were DAPI stained, and 200 asci per diploid were analyzed. Asci with more than four DAPI spots or unequal DAPI-stained bodies were considered "abnormal asci," as indicated by the dotted line. Normal asci were those with four equal DAPI-stained bodies, as shown by the solid line. Percentage of abnormal asci in wild-type and mst2 homozygote meiosis is plotted.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

 
In this work we have cloned the genes encoding both S. pombe MYST family histone acetyltransferases, mst1⁺ and mst2⁺. Both proteins are constitutively chromatin bound but appear to be broadly distributed and not confined to narrow regions. Interestingly, fission yeast has only two MYST family HATs, while S. cerevisiae has three. In S. cerevisiae, ScEsa1p is the catalytic subunit of the NuA4 and piccolo complexes, capable of acetylating histone H4 and H2A (2, 11, 16, 74). ScSas2p was recently shown to regulate the acetylation of histone H4 K16, preventing Sir3p binding and thus preventing further spreading of telomeric heterochromatin (49, 78). ScSas3p is the catalytic subunit of the NuA3 complex and, together with other HATs, including the Gcn5-containing complexes SAGA and ADA, is responsible for acetylation of histone H3 (33, 43, 51). It is significantly larger than Sas2p, so these genes do not appear to be simple duplicates.

Like its closest homolog, ScEsa1p, SpMst1p is essential for viability, while SpMst2p is not. Further evidence that SpMst1p is the orthologue of ScEsa1p comes from a recent paper characterizing fission yeast Alp5p (ScArp4p) showing that it can interact with SpMst1-myc protein in vivo and is required for global acetylation of histone H4 (58). It will be interesting to determine whether Mst1p is the responsible HAT for this histone H4 acetylation activity.

Which protein is the Mst2p functional homolog in S. cerevisiae? The budding yeast has two proteins with broadly similar structures. Our studies suggest that Mst2p has a role in telomere silencing similar to that of ScSas2p, but we showed that Mst2p affects levels of both histone H4 and H3 K9 acetylation, suggesting that it could be a histone H3 HAT like ScSas3p. However, we found that immunoprecipitated Mst2-HA protein has no HAT activity in in vitro assays (data not shown). Although some reports suggest that neither Sas2p nor the Sas2p-containing complex have HAT activity (2, 16, 43, 49, 56, 68, 74), recent studies showed that the SAS-I complex (Sas2p, Sas4p, and Sas5p) is capable of acetylating free histone H4 K16 and H3 K14 as well as nucleosomal histone H4 (73, 79). These data are consistent with a model that Mst2p is a bona fide histone acetyltransferase, although the identity of its target residues remains unproven.

Significantly, however, there are differences between SpMst2p and ScSas2p. Cells lacking Spmst2 are sensitive to HU, MMS, and UV irradiation (Fig. 2B, D, and E), but Scsas2 mutants are not (21, 68). sas2 suppresses some temperature-sensitive orc mutants (21), but mst2 does not (Table 3). Furthermore, we show that mst2 cells are not TBZ sensitive per se, but when combined with mutant genes that affect chromosome segregation that are TBZ sensitive, such as rad21ts, swi6, eso1ts, and mis4ts, their TBZ sensitivity is enhanced (Table 3). Thus, Mst2p may have a different spectrum of functions from ScSas2p, although they may overlap.

Cells lacking mst2 show sensitivity to agents including HU and MMS which cause replication fork stalling or pausing during S phase. This effect is synthetic with defects in other proteins, including Rad21p (cohesin), and suggests that Mst2p may play a broader role in genome stability, particularly during S phase.

Most strikingly, we find that Mst2p is a negative regulator of telomeric silencing. Both ade6⁺ and ura4⁺ markers are repressed when inserted close to a telomere. This effect was less striking when using constructs with the ura4⁺ marker at a telomere, suggesting that the use of different promoters and reporter genes can reveal or mask the effect of a mutation. Interestingly, as shown by the RT-PCR analysis, the ura4⁺ gene at the telomere in both constructs (Ch16 and "native telomere") is highly expressed compared to the mini-ura4 (Fig. 4C). In contrast, the ade6⁺ gene at the telomere of Ch16-M23 in wild-type cells is highly repressed compared to the mini-ade6 (Fig. 3B), suggesting that the role of Mst2p as a negative regulator of telomere silencing is more evident when using markers with weaker or attenuated promoters. In contrast, we did not observe silencing defects in the centromere or mating type region (Fig. 5B).

In S. cerevisiae, Sas2p and Sas3p can function as either activators or repressors of transcription, depending on the gene context. Mutation of sas2 or sas3 restores silencing to a derepressed HMR locus but enhances silencing defects of the HML locus in a sir1 background (21, 70). Loss of sas2 decreases silencing of the URA3 gene when the reporter is inserted near the telomere of the left arm of chromosome VII but increased silencing of this same marker when inserted in the rRNA genes (56, 70). No telomeric functions have been detected for Sas3p (70). Recent reports suggest that ScSas2p can negatively regulate telomeric silencing (49, 78). These reports demonstrate that ScSas2p and ScSir2p have opposing effects on the acetylation of H4 K16 and are needed to maintain the boundary at telomeric heterochromatin (49, 78). However, we did not observe a change in SpSir2p distribution in a mst2 mutant (29) (data not shown), although Sir2p is required for silencing in a mst2 strain (Fig. 3A). Thus, Mst2p does not appear to regulate Sir2p directly.

We carried out a detailed analysis of the protein composition and histone modifications of normal fission yeast telomeres I and II in wild-type and mst2 cells. We analyzed endogenous telomeres without integrated reporter genes, since we wanted to understand the role of Mst2p on native telomeres and to avoid confusion with different partial telomeric constructs that change the wild-type context. We generated a telomeric map for the distribution of Taz1p, Sir2p, Swi6p, and Rad21p and defined histone acetylation and H3 K9 methylation patterns using ChIP and real-time PCR (Fig. 5).

As shown previously, Taz1p and Sir2p are highly enriched close to the distal end of the telomere end and decrease in abundance towards the centromere-proximal side (29). We found that Swi6p abundance at specific telomeric regions changes with the mating type; our results suggest that h^– cells have a surplus of Swi6p that localizes to a specific region within the telomere. Interestingly, no differences in telomeric silencing are observed between h⁺ and h^– cells, although these could occur dynamically and not be apparent in our assay.

Fission yeast centromeres and mating type regions have hypermethylated histone H3 K9, and this modification is needed for Swi6p localization (65). In contrast to what is seen at the centromere, the distribution patterns of H3 K9-Me and Swi6p at the telomeres differ from one another. Significantly, in the 6.2-kb region of h^– strains, high levels of Swi6p do not correlate with hypermethylated histone H3. These results suggest that at the telomere either low levels of methylated histone H3 K9 are sufficient for Swi6p localization or that the mechanism for Swi6p binding to chromatin is different compared to that of the centromere and mating type regions. Telomeric distribution of Taz1p, Sir2p, Swi6p, and Rad21p as well as histone H3 K9 methylation were all independent of Mst2p, although histone H4 and histone H3 K9 acetylation were reduced in the region 16.9 kb from the telomere end (Fig. 5A).

The increased telomeric silencing we observe in mst2 may reflect the reduction in histone acetylation or the increased ratio of methylated versus acetylated histones. These data also show that in mst2 cells the increase in telomeric silencing is independent of Swi6p, suggesting that other mechanisms must be contributing to silencing at telomeres. A recent study characterizing SpArp6p showed that this protein is involved in telomeric silencing and binds telomeres in an Swi6p-independent manner; moreover, Swi6p localization was not affected in arp6 cells. The authors suggest that there is an Arp6p-mediated repression mechanism that works side by side with the Swi6p silencing machinery at telomeres (87). It will be interesting to test if Mst2p affects Arp6p localization and activity at telomeres.

Most intriguing, our work suggests that there may be specific telomere boundary elements in S. pombe. Mst2p affects the histone acetylation of a localized area, starting 16.9 kb from the telomeric repeat and extending at least 2 kb towards the centromere. This increase in histone acetylation could be acting as a barrier for the spreading of telomeric heterochromatin. A study analyzing the Gallus gallus ß-globin locus demonstrated that the region immediately surrounding a 16-kb block of heterochromatin had constitutively high levels of histone H3 K9 acetylation, revealing that heterochromatin boundaries can be marked by peaks of acetylated histones (52). ScSas2p and acetylation on histone H4 K16 are proposed to generate a boundary for the spreading of telomeric heterochromatin (49, 78). However, we were unable to immunoprecipitate acetylated histone H4 lysine 16 or ChIP Mst2-HA to the native or the minichromosome Ch16 telomere to test this model (data not shown). Similarly, S. cerevisiae Sas2p was never localized to telomeres, but the evidence highly suggests that Sas2p is the responsible HAT for the H4 K16 acetylation.

Interestingly, the subtelomeric region analyzed in this work is present in two different cosmids, c212 and pT2R1, that have been anchored to chromosomes I and II, respectively, indicating that the high identity between TAS elements of chromosomes I and II goes far beyond 20 kb from the telomeric ends.

Although the changes we saw in wild-type telomeres in mst2 were modest, the hypoacetylation of the telomeres and the synthetic growth defect with taz1 implicate Mst2p in normal telomere function. Further, we observed that homozygous mst2 diploids have an aberrant meiosis (Fig. 6). taz1 mutants also have severely disrupted meiosis (18, 63), while swi6 mutants suffer only very modest defects (86). These data suggest that telomeric acetylation is important for some aspect of fission yeast meiosis. We tested if recombination was affected in this mutant and found that wild-type and mst2/mst2 diploids have the same recombination frequency in the his4-lys4 interval (data not shown). Because normal telomeres are required for pairing, this function may be disrupted in mst2 cells. A second alternative is that meiotic chromosome segregation is abnormal, and a third possibility is that this reflects changes in gene expression. It has been shown that during nitrogen starvation (which induces meiosis), clusters of genes located close to telomeres are up-regulated (55), raising the possibility that disruption of these silent regions may affect a meiotic transcriptional program. Thus, the aberrant meiosis in mst2/mst2 diploids could be a consequence of downregulation of important subtelomeric meiosis-induced genes in addition to or instead of effects on telomere structure per se. Further examination of telomere structure and expression will be required to determine how this conserved protein affects chromosome function.

	ACKNOWLEDGMENTS

 
We thank Karl Ekwall, Robin Allshire, and Junko Kanoh for strains, Sally Pasion for discussion of unpublished results, and Lorraine Pillus, Angel Tabancay, and Will Dolan for helpful comments on the manuscript. We are grateful to Melissa Baker for assistance with the Q-PCR. We thank Ji-Ping Yuan for assistance with meiotic recombination assays. We thank Lorraine Pillus, Beverly Emerson, Dick McIntosh, and Paula Grissom for their hospitality and help to E.B.G. during the course of this work.

This study was supported by NIH grant R01 GM59321 to S.L.F.

	FOOTNOTES

 
* Corresponding author. Mailing address: Molecular & Computational Biology Section, University of Southern California, 1050 Childs Way, Los Angeles, CA 90089-2910. Phone: (213) 740-7341. Fax: (213) 740-8631. E-mail: forsburg{at}usc.edu.

Supplemental material for this article may be found at http://mcb.asm.org/.

	REFERENCES

Top Abstract Introduction