In Schizosaccharomyces pombe, three genes, sir2⁺, hst2⁺, and hst4⁺, encode members of the Sir2 family of conserved NAD⁺-dependent protein deacetylases. The S. pombe sir2⁺ gene encodes a nuclear protein that is not essential for viability or for resistance to treatment with UV or a microtubule-destabilizing agent. However, sir2⁺ is essential for full transcriptional silencing of centromeres, telomeres, and the cryptic mating-type loci. Chromatin immunoprecipitation results suggest that the Sir2 protein acts directly at these chromosomal regions. Enrichment of Sir2p at silenced regions does not require the HP1 homolog Swi6p; instead, Swi6-GFP localization to telomeres depends in part on Sir2p. The phenotype of sir2 swi6 double mutants supports a model whereby Sir2p functions prior to Swi6p at telomeres and the silent mating-type loci. However, Sir2p does not appear to be essential for the localization of Swi6p to centromeric foci. Cross-complementation experiments showed that the Saccharomyces cerevisiae SIR2 gene can function in place of S. pombe sir2⁺, suggesting overlapping deacetylation substrates in both species. These results also suggest that, despite differences in most of the other molecules required, the two distantly related yeast species share a mechanism for targeting Sir2p homologs to silent chromatin.

---

SACCHAROMYCES cerevisiae Sir2p is a member of a ubiquitous protein family characterized by a conserved NAD⁺-dependent protein deacetylation domain (reviewed in BUCK et al. 2004). The family members are regulatory proteins that control a diverse set of functions from transcriptional silencing (reviewed in RUSCHE et al. 2003) to aspects of skeletal muscle differentiation (FULCO et al. 2003). Consistent with these wide-ranging cellular roles, substrates of Sir2p family members include tubulin (NORTH et al. 2003), p53 (LUO et al. 2001; VAZIRI et al. 2001; LANGLEY et al. 2002), the archael chromatin protein Alba (BELL et al. 2002), an acyl-coenzyme A synthetase (STARAI et al. 2002), histones (reviewed in GOTTSCHLING 2000; SHORE 2000), a forkhead transcription factor (BRUNET et al. 2004; MOTTA et al. 2004), and the pol I subunit TAF(I)68 (MUTH et al. 2001). The NAD⁺-dependent activity of Sir2p homologs may be critical to their roles as sensors of cellular metabolic states (GUARENTE 2000; SMITH et al. 2000; SANDMEIER et al. 2002; BITTERMAN et al. 2003; DENU 2003). Although there has been a dramatic increase in the study of this family in recent years, no comprehensive understanding yet exists for the range of functions that the SIR2 gene family members perform in any one organism and the ways in which these functions vary from one species to another.

In S. cerevisiae, there are five members of the SIR2 gene family, SIR2 itself and the four HST (homolog of SIR two) genes. The S. cerevisiae SIR2 gene was first identified on the basis of its contribution to transcriptional silencing of the unexpressed copies of mating-type genes used as templates in mating-type switching (KLAR et al. 1979; RINE and HERSKOWITZ 1987). Further studies uncovered a role for Sir2p in silencing transcription of reporter genes inserted ectopically at telomeres (APARICIO et al. 1991). Models for silencing at telomeres and the silent mating-type loci in S. cerevisiae propose that DNA-binding proteins, including Rap1p, Abf1p, ORC, and Ku70/80, bind to DNA and recruit a complex containing Sir4p and Sir2p. The Sir2 protein then deacetylates histone substrates, perhaps most importantly lysine 16 of histone H4. This deacetylation event increases the affinity of Sir3p for histones and is thought to create and propagate a heterochromatin-like structure refractory to the production of stable transcripts (reviewed in GREWAL and MOAZED 2003; RUSCHE et al. 2003). The S. cerevisiae Sir2 protein also has additional, less well-understood functions. These include repressing origin activation (PASERO et al. 2002), transcription of reporter genes (BRYK et al. 1997; SMITH and BOEKE 1997), and homologous recombination (GOTTLIEB and ESPOSITO 1989) within the rDNA repeats, negatively regulating aging (reviewed in BITTERMAN et al. 2003; HEKIMI and GUARENTE 2003), and participating in meiotic checkpoint control (SAN-SEGUNDO and ROEDER 1999).

Like S. cerevisiae SIR2, the other S. cerevisiae and S. pombe SIR2 gene family members play a number of roles. The S. cerevisiae HST1 gene is closely related to SIR2 and can partially substitute for SIR2 in some silencing functions (BRACHMANN et al. 1995; DERBYSHIRE et al. 1996). HST1 represses expression of the middle-sporulation genes during vegetative growth (XIE et al. 1999), acts to repress de novo synthesis of NAD⁺ (BEDALOV et al. 2003), and represses expression of FLO10, a gene involved in invasive growth (HALME et al. 2004). The S. cerevisiae HST2 gene encodes a protein with robust in vitro deacetylase activity (LANDRY et al. 2000; SMITH et al. 2000) that also represses FLO10 expression (HALME et al. 2004) and affects silencing when overexpressed (PERROD et al. 2001). The HST3 and HST4 genes have overlapping cellular functions in S. cerevisiae (BRACHMANN et al. 1995), and their closest S. pombe homolog, hst4⁺, is a nuclear and nucleolar protein required for chromosomal integrity, chromosome stability, and centromeric and telomeric silencing (FREEMAN-COOK et al. 1999). In S. pombe, there are three SIR2-related genes: the hst4⁺ gene (FREEMAN-COOK et al. 1999); the hst2⁺ gene, which has not yet been well characterized (M. DERBYSHIRE, personal communication); and the sir2⁺ gene, which is closely related to S. cerevisiae SIR2 and HST1 (SHANKARANARAYANA et al. 2003).

In S. pombe, transcriptional silencing occurs at the centromeres, telomeres, and the silent mating-type loci mat2 and mat3 (reviewed in GREWAL 2000; HUANG 2002). Transcriptional silencing within these different regions requires distinct sets of proteins. Perhaps the best-understood proteins, which are required for silencing at all three regions, are the histone methyltransferase Clr4p and the histone binding protein Swi6p. Clr4p, a homolog of the Drosophila SU(VAR)3-9 protein, methylates histone H3 lysine 9 (REA et al. 2000; NAKAYAMA et al. 2001). Swi6p, a homolog of the Drosophila HP1 protein, colocalizes with silenced regions through self-association and through interactions with methylated histone H3 lysine 9 (reviewed in SINGH and GEORGATOS 2002; MAISON and ALMOUZNI 2004). A model for the mechanism of silencing in S. pombe proposes that histone H3 methylation by Clr4p increases the nucleosome binding affinity of Swi6p, thereby creating a repressed chromatin structure that can be further spread by Swi6p (NAKAYAMA et al. 2001; reviewed in GREWAL and MOAZED 2003).

Mutations in some of the genes required for silencing at the cryptic mating-type loci, such as clr4⁺, swi6⁺, and rik1⁺, cause phenotypes that reflect a role for heterochromatin in centromere function. However, mutations in other genes required for silencing at the cryptic mating-type loci, such as clr1⁺, clr2⁺, and clr3⁺, do not cause defects in centromere function (reviewed in GREWAL 2000; HUANG 2002). The role of these latter genes in centromeric function may be overlapping or less critical. Alterations of histone H3 residues lysine 9, lysine 14, or serine 10 result in loss of silencing at the otr1 domain of cen1, loss of Swi6p localization to telomeric and centromeric foci, and defects in centromere function (MELLONE et al. 2003). Thus, multiple residues in the histone H3 tail are critical for the creation of repressed chromatin. Silencing in S. pombe also requires RNA interference (RNAi) machinery (reviewed in BAILIS and FORSBURG 2002; MATZKE and MATZKE 2003), and it appears that transcripts from silenced regions facilitate the targeting of Clr4p to those silenced regions (reviewed in EKWALL 2004).

Silencing in S. cerevisiae is phenomenologically similar to silencing in S. pombe and position-effect variegation in Drosophila. Indeed, Drosophila sir2 mutants are suppressors of position-effect variegation (NEWMAN et al. 2002; ROSENBERG and PARKHURST 2002; ASTROM et al. 2003). However, aside from Sir2p, Rap1p (CHIKASHIGE and HIRAOKA 2001; KANOH and ISHIKAWA 2001; PARK et al. 2002), and histones (MOORE et al. 1979; MELLONE et al. 2003), surprisingly few of the silencing proteins found in fission yeast or in Drosophila appear to be involved in silencing in budding yeast (reviewed in GREWAL and ELGIN 2002; HUANG 2002; GREWAL and MOAZED 2003). Thus, we were motivated to study the S. pombe sir2⁺ gene because it provided a potential common link between the differing mechanisms of silencing in S. pombe and in S. cerevisiae. In addition, because heterochromatin in S. pombe shares many mechanistic features with heterochromatin in mammals, it provides a potential window into the role of SIR2 family members in mammalian silencing.

This article presents an identification and characterization of the S. pombe sir2⁺ gene. While this work was in preparation, an independent study corroborating some of the findings reported here was published (SHANKARANARAYANA et al. 2003). We discovered that the sir2⁺ gene encodes a nuclear protein that silences expression from centromeres, telomeres, and the cryptic mating-type loci. Because the protein was enriched at centromeric and telomeric sites and at sites between the silent mating-type loci, S. pombe Sir2p is likely to act directly at the genomic regions that it silences. Genetic, biochemical, and localization studies place Sir2p before Swi6p in a genetic pathway that leads to silencing at telomeres and the silent mating-type loci, but suggests independent effects at the centromeres. Finally, we show that the S. cerevisiae SIR2 gene was able to functionally substitute for the distantly related S. pombe sir2⁺ gene in silencing at the cryptic mating-type locus mat2.

Yeast strains and culture conditions:

The genotypes of the strains used are listed in Table 1. Standard culture conditions and genetic methods were used (MORENO et al. 1991; FORSBURG 2003; http://www.pombe.net). S. pombe strains were grown in yeast extract plus supplements (YES) or Edinburgh minimal medium (EMM). Malt extract (ME) plates were used for sporulation. 5-Fluoroorotic acid (5-FOA) plates are EMM supplemented with 1 g/liter 5-FOA (Toronto Research Chemicals, North York, Ontario, Canada), and 1.5x 5-FOA plates are EMM supplemented with 1.5 g/liter 5-FOA. Thiabendazole (TBZ) plates are YES supplemented with 20 µg/ml TBZ in DMSO (Sigma, St. Louis). G418 plates are YES plates supplemented with 200 mg/liter G418. S. pombe cells were grown at 32° on plates and 30° in liquid culture unless otherwise noted.

View this table: In this window In a new window   TABLE 1 S. pombe strains used in this study

 

Identification and cloning of the sir2⁺ cDNA:

During a routine search of new S. pombe sequence (WOOD et al. 2002), a fragment of a gene with homology to S. cerevisiae SIR2 was identified on chromosome II. Primers 5'sir2-intron and 3'sir2 were used to amplify a 635-bp fragment from a cDNA library (a gift from F. LaCroute, Center de Génétique Moléculaire, Gif-sur-Yvette, France). The fragment was isolated, radioactively labeled, and used to probe a second cDNA library (a gift from L. Guarente, Massachusetts Institute of Technology, Cambridge, MA). Two clones, pLP1115 and pLP1116, which appeared to contain the same insert in opposite orientations, were isolated from the second library. The 1600-bp fragments from the library clones were subcloned as NotI fragments into Bluescript pKS+ (Stratagene, La Jolla, CA) in both orientations and verified by reamplification of the inserts using the 5'sir2 and 3'sir2 primers.

PCR primers (5'–3') used for identification and cloning were as follows: 5'sir2 (GGAGAATCAAACTTCCAC) anneals just downstream of the predicted splice site of the second intron; 5'sir2-intron (CAATGTCATGGGTCTTTTG) bridges the predicted 64-bp intron with the 5'-end annealing upstream of the 5' splice site and the 3'-end annealing downstream of the 3' splice site; and 3'sir2 (ATATACTGCAGTTAGCGGCCGCTACAACTACCGTTTGTTGTGC) anneals upstream of the predicted stop codon and contains a nonannealing tail that encodes a NotI restriction enzyme site directly upstream of the sir2⁺ stop codon.

The sir2⁺ cDNA was sequenced to reveal a 1425-bp ORF encoding a 475-amino-acid predicted protein (Figure 1A). The predicted protein sequence (Figure 1A) differed in two places from the predicted sequence (CAB38511 at NCBI. First, the initial in-frame ATG in the cDNA was used as the presumed translational start site instead of the second in-frame ATG, which added an additional nine amino acids at the N terminus. Second, comparison of the cDNA sequence with the S. pombe genomic sequence (WOOD et al. 2002) led to identification of the positions of the five sir2⁺ introns. We observed that the splice acceptor site for the first intron differed from the acceptor site predicted solely from the genomic sequence. Consequently, at amino acid 62, 5 amino acids were omitted from the protein sequence relative to the predicted sequence at NCBI (Figure 1A). Thus, from comparison of our cDNA sequence to the genomic sequence, it appears that either the sir2⁺ transcript is alternatively spliced or, more likely, the splice junction predicted from the genomic sequence is incorrect.

View larger version (113K): In this window In a new window Download PPT slide   FIGURE 1.— (A) Protein sequence alignment of the S. pombe and S. cerevisiae members of the SIR2 gene family. Red-shaded residues indicate amino acids that are identical in five or more of the eight proteins aligned and yellow-shaded residues indicate amino acids that are similar in five or more of the eight proteins aligned. Blue-shaded residues indicate amino acids that are identical in S. pombe Sir2p and either S. cerevisiae Sir2p or S. cerevisiae Hst1p, or both. indicates the relative positions of the introns within the coding sequences of the SIR2 RNA. (B) Phylogenetic relationship of the Sir2 and Hst proteins of S. pombe and S. cerevisiae. For comparison, the closest human homolog of S. pombe Sir2p is also shown. The conserved core regions are shaded; percentage identity is to the S. pombe Sir2p core.

 
The protein sequence alignment and phylogenetic tree for the Sir2 family members were generated using Clustal X (ftp://ftp-igbmc.u-strasbg.fr/pub/ClustalX/), MacBoxShade (http://www.isrec.isb-sib.ch/ftp-server/boxshade/MacBoxshade/), and Phyllip (http://evolution.genetics.washington.edu/phylip.html) on the Macintosh. Residues specific to S. pombe Sir2p, S. cerevisiae Sir2p, and S. cerevisiae Hst1p were identified by manual inspection.

Construction of sir2::kanMX6 allele:

A complete start-to-stop null mutant was created using a PCR-based method (BAHLER et al. 1998). Briefly, primers psir2F and psir2R were used to amplify a kanMX6 cassette with 5'- and 3'-ends identical to sir2⁺. The 1.6-kb product was gel purified and used to transform S. pombe cells. Transformed cells were plated onto YES plates and grown at 30° for 16 hr. The cells were then replica plated to YES + G418 plates and grown at 30° for 3 days. Large colonies were streaked to YES + G418 plates and primers kanmx and psir2ko were used to confirm the potential knockout transformants by colony PCR (see below).

Primers (5'–3') used for null mutant construction were as follows: psir2F (CCTTTCTCAAACGAAGCTTTAATTTATACTTCCTTAAAAATAATTGATTAGTCATATTTTCATGTTATACAAACTCTGTCCGGATCCCCGGGTTAATTAA) contains a nonannealing 5'-tail homologous to sir2⁺ upstream of its start codon and a 3'-end that anneals to the 5'-end of the kanMX6 gene; psir2R (AATCTGGCTCTCATACTCGATATATGTATATAATATATTTGCTTCTACATTCAACTTCGATAGATCGATCAGTCCTAAAAGAATTCGAGCTCGTTTAAAC) contains a nonannealing 5'-tail homologous to sir2⁺ downstream of its stop codon and a 3'-end that anneals to the 3'-end of the kanMX6 gene; psir2ko (CACAGTCCTTGGTATTG) anneals downstream of the stop codon; and kanmx (GCTAGGATACAGTTCTCACATCACATCCG) anneals within the kanMX6 gene.

Construction of sir2-myc:

Thirteen copies of the sequences encoding the human c-myc epitope were added to the C terminus of sir2⁺ using a PCR strategy (BAHLER et al. 1998). Primers 5'-tag and psir2R were used to amplify a cassette containing the 13 myc tags and kanMX6 with 5'- and 3'-ends homologous to the C terminus of sir2⁺, which directed insertion at the sir2⁺ locus. The 2.4-kb product was gel purified and used to transform S. pombe cells. Transformants containing the myc-tagged sir2⁺ were selected as described above for the creation of the sir2 strains and confirmed by colony PCR (see below). The sir2-myc allele is functional by the criteria that it was expressed and was able to substitute for sir2⁺ in silencing the ura4⁺ reporter at mat3 (data not shown).

Primers (5'–3') used for construction of the myc-tagged strains were as follows: 5'-tag (GCTTTCGAAACGGATTTGGATATAAAATTTGAGGAGCCCAGCACCTATCATATCACGAGCACAACAAACGGTAGTTGTCGGATCCCCGGGTTAATTAA) has a nonannealing 5'-end that is homologous to sir2⁺ directly upstream of the stop codon and a 3'-end that anneals to the 5'-end of the 13 myc-kanMX6 cassette; psir2R, psir2ko, and kanmx are described above.

Colony PCR:

Single colonies were picked, resuspended in 30 µl water, and boiled for 5 min. A total of 5 µl of this cell suspension was used with 15 µl master mix: 2 µl reaction buffer, 1 µl 50 mM MgCl, 2 µl 10 mM dNTP stock (2.5 mM each dNTP), 3.8 µl water, 3 µl forward primer (stock concentration 5 pmol/µl), 3 µl reverse primer (stock concentration 5 pmol/µl), and 0.2 µl Taq polymerase (5 units/µl). Thirty PCR cycles were as follows: 94° 1 min, 45° 1 min, 72° 1 min.

UV and TBZ sensitivity assays:

Liquid cultures were grown to saturation, diluted fivefold in a 96-well microtiter plate, and spotted onto YES plates or YES plates containing 20 µg/ml TBZ in DMSO using a pin replicator. For the UV assay, the YES plates were treated with 80 µJ/m² UV irradiation using a Stratalinker (Stratagene, La Jolla, CA) immediately after plating and again after 12 hr. Control plates were left untreated. The plates were placed in a dark box to prevent photoreactivation and incubated at room temperature (20°).

Silencing assays:

Liquid cultures were grown to saturation, diluted fivefold in a 96-well microtiter plate, and spotted onto YES, EMM complete, supplemented EMM, and 5-FOA plates using a pin replicator. Plates were incubated until full growth was achieved (3–10 days). Plates containing hst4 mutant strains were grown for extended periods to ensure that any differences seen were not simply due to the slow growth of the hst4 mutant strains.

Sporulation determinations:

Freshly growing strains were patched to ME plates and incubated at 25° for 7 days. Cells from patches were fixed and DAPI stained. Images of the cells were captured under differential interference contrast and fluorescence microscopy and scored for numbers of sporulated and unsporulated cells.

Sir2-myc immunofluorescence analysis:

Immunofluorescence analysis was performed as previously described (GOMEZ et al. 2002) with some modifications. After spheroplasting with 2 mg/ml of zymolyase 20T (Seikagaku, Rockeville, MD), cells were resuspended in 2 ml PEMS buffer (100 mM PIPES, pH 6.9, 1 mM EGTA, 1 mM MgSO₄, 1.2 M sorbitol) + protease inhibitors (P-8215, Sigma). Then, a 1/10 volume of PEMS + 0.25% Triton X-100 was added, followed by a 10-min incubation at room temperature. Three percent BSA in PEMS was used as blocking buffer. Rabbit anti-myc antibody (sc-789, Santa Cruz) was used at a 1:125 dilution. After three 1-ml washes with blocking buffer, cells were incubated with Alexa Fluor 488 chicken anti-rabbit antibody (A-21441, Molecular Probes, Eugene, OR) at a 1:500 dilution for 1 hr at room temperature in the dark. Cells were washed twice with blocking buffer and once with 1 ml PEMBAL buffer (100 mM PIPES, pH 6.9, 1 mM EGTA, 1 mM MgSO₄, 1% BSA, 0.1% NaN₃, 100 mM L-lysine monohydrochloride). Cells were mounted and stained with 4',6-diamidino-2-phenylindole (DAPI). Microscopy was performed with a Leitz Laborlux S microscope. Images were captured with a Hamamatsu (Bridgewater, NJ) digital camera using Improvision Openlab software.

Chromatin immunoprecipitation:

Chromatin immunoprecipitation (ChIP) was performed as described previously (EKWALL and PARTRIDGE 1999) with some modifications. Strains were grown to A₅₉₅ = 1 in YES and formaldehyde fixed. Protein lysates were sonicated in 3 ml final volume and quantified. Rabbit anti-myc (sc-789, Santa Cruz) or monoclonal anti-HA (MMS-101R, Covance) antibodies were incubated overnight with 2 mg of protein lysate. A 1:1 mix of protein A/protein G Sepharose 4 Fast Flow (17-0974-01/17-0618-01, Amersham Biosciences) was blocked with 0.1% BSA (A-7638, Sigma) and 0.3 mg/ml salmon testes DNA (D-7656, Sigma) and used to precipitate the antibodies. The immunoprecipitated DNA was resuspended in 50 µl of resuspension buffer (10 mM Tris-HCl pH 8, 0.1 mM EDTA, and 0.05% SDS) and the input DNA in 50 µl of resuspension buffer + 3 µg of RNase followed by a 1-hr incubation at 37°. Immunoprecipitated DNA was analyzed by real-time quantitative PCR using SybrGreen (Applied Biosystems, Foster City, CA) as a marker for DNA amplification on an ABI Prism 7900 apparatus (Applied Biosystems) with 40 cycles of three-step amplification. Values obtained with the untagged or no-antibody immunoprecipitations were subtracted from the corresponding values. Input DNA was used to obtain a standard curve to estimate the amount of DNA in each sample. Data were graphed as the percentage of immunoprecipitated DNA with respect to the total input of DNA used in each immunoprecipitation.

Primers used for the quantitative PCR were as follows: Q-cnt F (GTATTAGTGGTCGGTTTTCTTTTTGTT); Q-cnt R (CGGCGAAATGCTTCAGACAT); Q-imr F (CCTTTACTGGAAAATTGTCGATATTACTAC); Q-imr R (CGTTGCAATTATAAGAAACTAAGCTACTCA); Q-dg F (CGGTCTTTGCAGGACTCTTGA); Q-dg R (CCACCACAATTTAACCCGATTAG); Q-dh F (CTAATCATAAGAGGCAATGGGAATG); Q-dh R (ATTCATGTCGTAGATGTGACGTCA); Q-MAT2 F (TAGTATTCTGTCGAAATTATCGAAAGCTA); Q-MAT2 R (GAAACTGAAGCAGGGAAAAATGTAG); Q-NSU70 F (ACCGTATTTCATTTCTATTTCTTTATTCAA); Q-NSU70 R (GGGAATTTAGGAAGTGCGGTAA); Q-701 F2 (TACTTCGTTATCCAGCTCACCATG); Q-701 R2 (GGGATGAAAAATTTGAAGTTGACTC); Q-702 F (GTTACTCGCCTGCCTCTACCAT); Q-702 R (GTCAGTCAAGTTAATGAGTCATGAAGAA); Q-562 F (ACCCACGAACCCTCTCATCTT); Q-562 R (GGTCACTGGGTCAACAGGTTAAA); Q-E11 F (CAAGTTGTGGTCGGCCTTG); Q-E11 R (CCCGTACGCTTATCTACTTGTTATAAATG); Q-E12 F (AATCGGGAAGTTGCGTCTCTAA); Q-E12 R (TGGACCGCATTCAAAGATGTAC); Q-E13 F (GGTGCAAAGCAGGCAGAGA); Q-E13 R (CATAAAGATGGTACTTCAAAATATGCCTAT).

Localization of Swi6-GFP and Pcp1-GFP:

Cells were grown in EMM complete medium at 32°, harvested by centrifugation, and attached to poly-L-lysine-coated coverslips for live observation of GFP fluorescence. For DAPI staining, cells were incubated briefly in 0.5 µg/ml of DAPI in PBS for 60 sec and washed in PBS, prior to placement on the coverslips. Mounting medium was made from a stock of 10 mg/ml p-phenylenediamine in 1 M Tris and diluted 1:9 in 100% glycerol. The distribution of numbers of Swi6-GFP spots was determined by counting 400 cells for each strain.

Cross-complementation experiments:

Transformations of sir2⁺ and sir2 strains were performed using a lithium acetate protocol (SCHIESTL and GIETZ 1989). Plasmids used for the complementation assays were the pREP-1 vector (MAUNDRELL 1990) and pLP1051, which contains the S. cerevisiae SIR2 gene inserted behind the nmt⁺ thiamine repressible promoter of pREP1. Liquid cultures were grown to saturation in selective medium with a thiamine concentration of 0.5 mM, serially diluted fivefold, and plated onto EMM –leucine, EMM –leu –uracil, and EMM –leu 5-FOA plates, with thiamine concentrations of 0, 0.05, 0.5, 1, and 15 mM. Images were captured after 4 days of growth.

Identification of the closest S. pombe homolog of S. cerevisiae SIR2:

A segment of a gene with homology to S. cerevisiae SIR2 was amplified from an S. pombe cDNA library and used as a probe to isolate two cDNA clones. The cDNA sequence of one of the clones was compared to the S. pombe genomic sequence (WOOD et al. 2002) and was found to encode a predicted 475-amino-acid protein with five introns (Figure 1A). Comparison of the predicted protein sequence with the Sir2p and Hstp family members in S. cerevisiae and S. pombe (Figure 1, A and B) revealed that this cDNA corresponded to S. pombe sir2⁺ (see below).

The Sir2p sequence deduced from cDNA sequencing differed from the Sir2p sequence deduced from genomic sequencing (WOOD et al. 2002) in that it starts nine amino acids earlier and has one different splice site junction (see MATERIALS AND METHODS for details). In the 5' leader region of the sir2⁺ cDNA, an out-of-frame ORF starts 19 bases upstream from the presumed initiating ATG and stops a few nucleotides before a second in-frame ATG. This complicated leader structure and the large number of introns may serve to regulate splicing or translation of the sir2⁺ message or to keep translational levels low.

The predicted sir2⁺ protein sequence was compared to the other Sir2p family members in S. pombe and in S. cerevisiae (Figure 1A). The S. pombe Sir2 protein is most closely related to the S. cerevisiae Sir2 and Hst1 proteins (Figure 1B). Like all Sir2p family members, it shares a conserved "core" domain, which has NAD⁺-dependent protein deacetylase activity in all family members tested (reviewed in BUCK et al. 2004). In addition, sequences scattered within and amino-proximal to the core domain are uniquely shared by S. pombe Sir2p, S. cerevisiae Sir2p, and Hst1p. Alterations of several of these conserved sequences in the S. cerevisiae Sir2 protein cause a loss of silencing, and in some cases, the loss of silencing is specific to a particular silenced genomic region (COCKELL et al. 2000; CUPERUS et al. 2000; ARMSTRONG et al. 2002; GARCIA and PILLUS 2002).

S. pombe sir2 null mutants have distinct phenotypes from S. pombe hst4 null mutants:

To begin a characterization of the S. pombe sir2⁺ gene, a sir2 null allele was created by a complete deletion of the sir2⁺ open reading frame and replacement with the kanMX6 gene. Strains containing the sir2 allele were tested to see if they shared similar phenotypes with strains lacking Hst4p, the only other characterized S. pombe Sir2p family member (FREEMAN-COOK et al. 1999). Strains with an hst4 mutation grow slowly, have elevated rates of chromosome loss, and are sensitive to UV and TBZ, a microtubule-destabilizing agent (FREEMAN-COOK et al. 1999). By contrast, sir2 strains did not have overt growth defects and showed only subtle increases in chromosome loss compared to wild-type strains (data not shown). In addition, the sir2 strains were not sensitive to UV or TBZ, and the sensitivity of hst4 strains to these treatments was not exacerbated in sir2 hst4 double-mutant strains (Figure 2). Thus, it seems that the S. pombe sir2⁺ gene has some nonoverlapping functions distinct from those of hst4⁺.

View larger version (24K): In this window In a new window Download PPT slide   FIGURE 2.— The sir2 mutant is not sensitive to UV or TBZ. Liquid cultures of the strains FY1648 (wild type), LPY4451 (sir2), LPY3278 (hst4), and LPY4464 (sir2 hst4) were grown in rich medium to saturation and then serially diluted fivefold and spotted to YES plates for the growth control (left) and UV treatment (center) or YES plates with 20 µg/ml TBZ, a microtubule-destabilizing drug (right).

 

The sir2 mutant has region-specific defects in centromeric silencing:

The SIR2 gene is required for position effects in budding yeast (reviewed in RUSCHE et al. 2003), and we investigated whether sir2⁺ participates in the position effect at centromeres, telomeres, and the silent mating-type loci in fission yeast. Silencing of the S. pombe ura4⁺ reporter 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 (BOEKE et al. 1984; ALLSHIRE et al. 1994). Silencing of the S. pombe his3⁺ reporter can be detected in a similar manner by monitoring growth on medium lacking histidine.

To determine if the sir2⁺ gene product silences at centromeres, a series of sir2 strains (Table 1) carrying the ura4⁺ reporter inserted into different centromeric regions (Figure 3; ALLSHIRE et al. 1994, 1995) was tested for growth on medium lacking uracil and on medium containing 5-FOA. For strains with the reporter located within the central core domains of cen1 or cen3 (TM1 or TM3, respectively), the growth of the sir2 and wild-type strains was similar on all plates, including the 5-FOA plates. This indicates that sir2 strains are not inherently sensitive to 5-FOA and that sir2 strains do not have obvious defects in silencing the central core domain. However, sir2 strains did have modest defects in silencing of the inner (imr) and outer (otr) repeats of cen1 (Figure 3), as evidenced by increased growth on –uracil plates and decreased growth on 5-FOA plates. Interestingly, these phenotypes differ from hst4 mutant phenotypes, which include modest silencing defects at the otr and imr repeats of cen1, but striking silencing defects at the central domain of cen3 (TM3), especially at room temperature (FREEMAN-COOK et al. 1999). S. pombe centromeres have distinct domains (reviewed in SULLIVAN et al. 2001), and it appears that whereas both sir2⁺ and hst4⁺ function to silence the flanking repeat arrays, hst4⁺ functions additionally to silence within the central core.

View larger version (58K): In this window In a new window Download PPT slide   FIGURE 3.— The sir2 mutant has distinct defects in centromeric silencing. Strains with a sir2 mutation were tested for their ability to silence the ura4+ reporter gene at the central domain of chromosome I (TM1; A) or III (TM3; B), the imr repeat of chromosome I (C), and the otr repeat of chromosome I (D). A map of chromosome I, with positions of the ura4+ reporters indicated, is at the top (not to scale). The letter B at the top indicates the relative position of the TM3 reporter (actually on chromosome III). Liquid cultures of the strains were grown to saturation in rich medium, serially diluted fivefold, and spotted onto plates. After incubation for 4 days, plates were photographed. The strains used to test silencing at (A) TM1 were FY336 (wild type) and LPY4477 (sir2); at (B) TM3, LPY2398 (wild type) and LPY4472 (sir2); at (C) imr, FY498 (wild type) and LPY4475 (sir2); and at (D) otr, FY971 (wild type) and LPY4468 (sir2).

 

S. pombe Sir2p promotes telomeric silencing:

S. pombe centromeres are more complex and heterochromatic than those of budding yeast (reviewed in SULLIVAN et al. 2001; BJERLING and EKWALL 2002). To determine if S. pombe Sir2p represses transcription of regions that are heterochromatic in both yeasts, silencing at telomeres and the silent mating-type loci was tested in sir2 strains. Telomeric silencing was tested using two different telomeric reporter constructs, one in which the ura4⁺ reporter was located at the telomere of a minichromosome (NIMMO et al. 1994) and one in which the his3⁺ reporter was located adjacent to a natural telomere (NIMMO et al. 1998). In both cases, silencing of the telomeric reporter was partially lost in sir2 mutants (Figure 4, A and B). Thus, the S. pombe Sir2 protein, like the S. cerevisiae Sir2 protein, is required for complete telomeric silencing. In addition, the demonstration that the S. pombe Sir2 protein repressed expression of both the his3⁺ and the ura4⁺ reporter indicates that Sir2p-mediated transcriptional repression, like silencing in general, is gene nonspecific.

View larger version (66K): In this window In a new window Download PPT slide   FIGURE 4.— The Sir2 protein silences telomeric regions and the cryptic mating-type loci. (A) A sir2 mutation causes defects in silencing the ura4+ reporter gene at the minichromosome telomere Ch16ade6-M216m23::ura4+-TEL. Silencing assays were performed as described in Figure 3, except that growth and silencing on 5-FOA were tested on media lacking adenine to select for maintenance of the minichromosome. The strains used were LPY3536 (wild type) and LPY4573 (sir2). (B) The sir2 mutant has defects in silencing at a his3+-marked telomere. Silencing was evaluated as described in Figure 3, except that strains were tested for silencing on minimal medium lacking histidine. Strains shown were FY1863 (wild type) and LPY4470 (sir2). (C) Strains with a sir2 mutation have strong defects in silencing the ura4+ reporter gene at the mat3 locus. Silencing assays were performed as described in Figure 3, except that 5-FOA plates were 1.5x, and plates were incubated at 30° for 3 days. The h90 mat3-M(EcoRV)::ura4+ strains tested were FY597 (wild type), LPY6112 (sir2), and LPY6114 (sir2). The control strain FY340 (Ura+), which is phenotypically Ura+, has a null mutation at the ura4 locus and an expressed copy of the ura4+ reporter inserted randomly in the genome (ALLSHIRE et al. 1994).

 

The S. pombe Sir2 protein silences the cryptic mating-type loci:

To determine if the S. pombe Sir2 protein was required for transcriptional repression at the silent mating-type loci, strains containing the ura4⁺ reporter inserted near the silent mat3 locus (THON and KLAR 1992) were tested for expression (Figure 4C). Wild-type strains grew poorly on medium lacking uracil and grew well on medium containing 5-FOA, reflecting strong silencing of the reporter. In striking contrast, sir2 strains grew well on medium lacking uracil and grew poorly, if at all, on medium containing 5-FOA. Growth of the sir2 strains was similar to a control strain (ALLSHIRE et al. 1994; labeled Ura⁺) in which the ura4⁺ reporter was randomly inserted and expressed from a euchromatic region. Thus, Sir2p is required for silencing at the mat3 locus. Sir2p was also required to silence the ura4⁺ reporter and endogenous genes at mat2 (see below). This suggests that the role of Sir2 proteins in silencing the cryptic mating-type loci has been conserved in S. cerevisiae and S. pombe despite a lack of conservation of many of the other proteins required for silencing in either yeast species (HUANG 2002).

The sir2⁺ and swi6⁺ genes function in concert to silence the cryptic mating-type loci:

A number of proteins, including Swi6p, silence the cryptic mating-type loci in S. pombe. To further characterize the role of Sir2p in silencing within this region and its relationship to Swi6p, silencing of sir2, swi6, and sir2 swi6 double mutants was compared in mat1-Msmt-0 mat2-P(XbaI)::ura4⁺ strains, which are unable to switch mating type and contain the ura4⁺ reporter centromere distal to mat2-P (THON et al. 1994). The sir2 strains showed strong silencing defects, especially on the –uracil plate, but were not completely derepressed (Figure 5A), indicating that some residual silencing occurred in the absence of sir2⁺. In swi6-115 strains, the extent of derepression was more variable and sometimes exceeded that of the sir2 strains. The double-mutant (sir2 swi6-115) strains looked similar to and sometimes slightly more derepressed than the swi6 strains. The strains were also tested for their ability to undergo "haploid meiosis," which is a consequence of derepression of the P genes at mat2 in this genetic background (KELLY et al. 1988). None were sporulation proficient (THON et al. 1994; data not shown), indicating that the P genes were not completely derepressed. These results are consistent with a model in which Sir2p and Swi6p work in a single genetic pathway leading to silencing at the cryptic mating-type loci, but that one or more additional, independent pathways also contribute to silencing.

View larger version (80K): In this window In a new window Download PPT slide   FIGURE 5.— The sir2 and swi6 mutants have similar silencing defects at the mat2 locus. (A) Silencing of the mat2-P(XbaI)::ura4+ reporter in sir2, swi6, and sir2 swi6 double-mutant strains. Silencing assays were performed as described in Figure 3, except that 5-FOA plates were 1.5x, and plates were incubated at 30° for 3 days. The mat1-Msmt-0 mat2-P (XbaI)::ura4+ strains tested were LPY6264 (wild type), LPY6258 and LPY6259 (sir2), LPY8771 and 8772 (swi6-115), and LPY8768 and LPY8767 (sir2 swi6-115). The Ura+ control strain FY340 (Ura+) has a null mutation at the ura4 locus and an expressed copy of the ura4+ reporter inserted randomly in the genome (ALLSHIRE et al. 1994). (B) Silencing of the mat2-P(XbaI)::ura4+ reporter in combination with the cis-acting (BglII-BssHII) deletion mutation in sir2, swi6, and sir2 swi6 strains. Silencing assays were performed as described in Figure 3, except that plates were incubated at 30° for 3 days. The mat1-Msmt-0 (BglII-BssHII)mat2-P(XbaI)::ura4+ strains tested were LPY6277 (wild type), LPY6270 (sir2), LPY8781 (swi6-115), and LPY8774 (sir2 swi6-115).

 
To gain more insight into the relationship between sir2⁺ and swi6⁺, sir2, swi6-115, and sir2 swi6-115 strains were tested for silencing in a sensitized mat1-Msmt-0 (BglII-BssHII)mat2-P(XbaI)::ura4⁺ background. In this background, the reporter is in the same position, but a deletion of cis-acting sequences centromere proximal to mat2 causes increased silencing defects in combination with trans-acting silencing mutations like swi6 (THON et al. 1994). The sir2 strain was completely derepressed and looked like the swi6-115 strain (Figure 5B), indicating that Sir2p and Swi6p, with the cis-acting site, are needed for full repression of the mat2 locus. The sensitized strains were also tested for their ability to undergo haploid meiosis (Table 2). As expected, wild-type strains did not sporulate, even with the cis-acting deletion, but the sir2, swi6-115, and sir2 swi6-115 strains were able to sporulate. Interestingly, sir2 strains sporulated significantly less well than the swi6 and sir2 swi6 strains, which did not differ significantly. Therefore, by two different assays, swi6 strains were slightly more derepressed than sir2 strains and had levels of derepression similar to the double-mutant strains. These results strengthen the model that Sir2p and Swi6p function in the same genetic pathway leading to silencing at the silent mating-type loci and that this pathway is distinct from the pathway utilizing the cis-acting sequences near mat2.

View this table: In this window In a new window   TABLE 2 Relative contributions of sir2+ and swi6+ to silencing

 

Sir2-myc is a nuclear protein throughout the cell cycle:

Because Sir2p is required for silencing at centromeres, telomeres, and the silent mating-type loci, we tested whether its cellular localization was consistent with it acting directly at those regions. The Sir2 protein was epitope tagged at the carboxyl terminus with 13 copies of the c-myc epitope and expressed under the control of the sir2⁺ promoter. Using immunofluorescence analysis, anti-myc antibodies detected a specific nuclear signal in strains expressing Sir2-myc (Figure 6). The signal was found throughout the DAPI-staining region and appeared more concentrated in some areas of the nucleus than in others, but was not found in discrete foci. Sir2-myc was visualized in the nucleus in mononuclear (G2) and binuclear (M, G1, and S-phase) cells, suggesting that it is present throughout the cell cycle.

View larger version (59K): In this window In a new window Download PPT slide   FIGURE 6.— Sir2-myc is a nuclear protein throughout the cell cycle. A strain expressing Sir2-myc (FY2773) and a wild-type untagged strain (WT; FY259) were grown to A595 = 1 in YES, harvested, and prepared for immunofluorescence analysis. Sir2-myc was visualized using anti-myc antibodies and a secondary antibody conjugated to Alexa Flour 488 (myc). DNA was visualized by DAPI staining (DAPI). Bar, 10 µm.

 
When other proteins that function in telomere maintenance and in silencing, such as Taz1p and Swi6p, are epitope tagged and visualized microscopically, they appear concentrated in a few foci or dots within the nucleus. These foci correspond to telomeres (both proteins), centromeres (Swi6p), and the silent mating-type regions (Swi6p; EKWALL et al. 1995; COOPER et al. 1997). Experiments using the technique of fluorescence recovery after photobleaching suggest that the interactions of Swi6p with chromatin are highly dynamic and that transient histone binding by Swi6p underlies heterochromatin formation in fission yeast (CHEUTIN et al. 2004). Although immunofluorescence analysis does not distinguish between soluble nuclear or chromatin-bound proteins, the relatively uniform Sir2-myc localization suggests that Sir2p is not restricted to centromeres, telomeres, and the silent mating-type region but rather may be distributed to more sites within the nucleus.

The Sir2 protein binds to centromeres, the silent mating-type loci, and telomeric regions:

Sir2p functioned to silence discrete genomic regions, yet appeared by immunofluorescence to generally pervade the nucleus without distinct focal localization. If Sir2p is acting directly to silence, we would expect it to have physical association with the regions it affects. To test if the Sir2 protein physically localizes to these chromosomal regions, strains expressing the Sir2-myc-tagged protein were used in ChIP analysis.

Sir2-myc-tagged and untagged control strains were formaldehyde fixed and lysates immunoprecipitated with anti-myc antisera. The immunoprecipitated DNA was amplified by quantitative real-time PCR. After subtracting background amplification from the untagged strain, the amount of enrichment of immunoprecipitated sequences was quantified relative to input DNA. Compared to a site 6.2 kb from telomeric repeats, the Sir2-myc protein was enriched to varying degrees at centromeric sites, at a site within the K-region between mat2 and mat3 (mat), and at telomere-proximal sites (Figure 7A). Taken together with the silencing data, these results suggest that the Sir2 protein is directly required at these sites of enrichment to create a heterochromatic state. The Sir2p enrichment observed was relatively modest and similar to that seen for Clr3p at mat3-M::ura4⁺ (BJERLING and EKWALL 2002). This might indicate that the association of Sir2p with these silenced regions, although critical, is transient, narrowly localized, of low abundance, or masked from the antibodies and is consistent with the lack of detection of discrete foci in immunofluorescence (see above).

View larger version (30K): In this window In a new window Download PPT slide   FIGURE 7.— Sir2-myc binds to telomeric, centromeric, and silent mating-type regions, independently of Swi6p. Untagged wild-type (FY259 or FY264) or tagged sir2-myc (FY2773), sir2-myc swi6 (FY2745), and taz1-HA (TN327) strains were prepared for ChIP as described in MATERIALS AND METHODS. Sir2-myc was immunoprecipitated with anti-myc antibodies, Swi6p was immunoprecipitated with anti-Swi6 antibodies, and Taz1-HA was immunoprecipitated with anti-HA antibodies. Real-time quantitative PCR was carried out with primers corresponding to cnt, central domain; imr, inner region; dg and dh, outer repeats region; mat, cryptic mating-type region; and telomeric regions 0.38, 2.3, 6.2, 9, 10.6, 12.2, and 14.2 kb from the telomeric repeats. Untagged values for Sir2-Myc and Taz1-HA were subtracted from the corresponding tagged values. For Swi6, the no antibody values were subtracted. The percentage of DNA precipitated relative to the amount of input DNA was plotted for each locus. Data are the mean plus standard deviation from two independent experiments. (A) Sir2-myc was immuno-precipitated from a wild type (wt) or swi6 mutant background. (B) The pattern of Swi6p enrichment differed from Sir2-myc. (C) Sir2-myc, Swi6p, and Taz1p have distinct patterns of abundance at TAS. A cartoon of the telomere including position of primers used in ChIP analysis is depicted along the x-axis.

 
SHANKARANARAYANA et al. (2003) recently showed that Swi6p localization to silenced regions depends on sir2⁺. To test the converse, if Sir2-myc association with silent chromatin depends on the presence of Swi6 protein, binding was analyzed in a swi6 strain. Neither the degree nor the distribution pattern of Sir2-myc binding was significantly affected in a strain lacking Swi6p (Figure 7A). However, a slight increase in the dg region is observed in cells lacking Swi6p. Swi6p is highly enriched in this region (Figure 7B), suggesting that Sir2-Myc might be spreading toward the outer repeats in swi6 cells.

Sir2-myc appeared to show different degrees of association with the various centromeric regions, with greatest relative abundance at the central core region (cnt) and least relative abundance at the dg and dh outer repeats. To compare the Sir2-myc centromeric distribution pattern to the distribution pattern of Swi6p binding, ChIP analysis was performed using anti-Swi6p antisera (Figure 7B). Swi6p binding was relatively more abundant at the dg and dh outer repeats and less abundant at cnt and imr, as was previously seen for Swi6p binding to the ura4⁺ reporter at centromeric sites (PARTRIDGE et al. 2000). Thus, the two proteins showed distinctive patterns of enrichment at the different centromeric domains.

We further explored the relative distribution of Sir2-myc, Swi6p, and the telomere-associated protein Taz1p by performing ChIP analysis using primer sets that amplified sites along telomere-associated sequences (TAS; Figure 7C). Sir2-myc enrichment was greatest at the most telomere-proximal site, a distance of 0.38 kb from the telomeric repeats, was diminished 2.6 kb from the telomeric repeats, and appeared to plateau at a relatively low level 6.2 kb from the telomeric repeats (Figure 7C, top). In contrast, the patterns of accumulation of Taz1-HA and Swi6p differed from Sir2-myc and from each other. Levels of Taz1-HA enrichment were highest 0.38 kb from the telomeric repeats and then appeared to drop to low levels of enrichment 2.6 kb from the telomeric repeats before terminating at a region between 9 and 11 kb from the telomere (Figure 7C, middle). Qualitatively similar results were seen previously for Taz1p binding (SADAIE et al. 2003). Swi6p binding was least enriched 0.38 kb from the telomeric repeats, was more enriched 2.6 kb from the telomeric repeats, and appeared to plateau at its highest levels starting 6.2 kb from the telomeric repeats (Figure 7C, bottom). Swi6-GFP localization to the telomere is independent of Taz1p (SADAIE et al. 2003), but the pattern of occupancy defined by Swi6p and Sir2p may reflect a spreading mechanism for Swi6p binding.

The Sir2 protein is required for Swi6-GFP localization:

In whole cells, a Swi6-GFP fusion protein forms two to six foci in the nucleus that correspond to centromeres, telomeres, and the silent mating-type loci (EKWALL et al. 1995). Assembly of Swi6-GFP in heterochromatin regions depends upon the Clr4p methyltransferase, so that in cells lacking Clr4p, Swi6-GFP appears diffusely nuclear (EKWALL et al. 1996). We observed that most sir2 cells typically have a single Swi6-GFP spot, in contrast to wild-type or clr4 mutants (Figure 8, left and bottom). Thus, Sir2p is necessary for localization of Swi6p at most, but not all, of the foci within the nucleus.

View larger version (25K): In this window In a new window Download PPT slide   FIGURE 8.— Sir2p promotes localization of Swi6-GFP to telomeres. (Top, left) The number of Swi6-GFP foci is decreased in sir2 mutants. Paired images of wild type (strain FY1586) or sir2 (strain FY2680) cells that expressed Swi6-GFP using DAPI staining to visualize DNA (A and B) or epifluorescence to visualize GFP (A' and B'). (Top, right) The Swi6-GFP foci remaining in sir2 strains colocalize with the spindle-pole body. C and D depict GFP fluorescence in cells that expressed Pcp1-GFP in a wild-type (FY2803) or sir2 (FY2864) strain background; E and F depict GFP fluorescence in cells that expressed both Pcp1-GFP and Swi6-GFP in a wild-type (FY2846) or sir2 (FY2847) strain background. (Bottom) Histogram showing the percentage of cells with zero, one, or two or more Swi6-GFP foci in wild-type (WT; FY1586), hst4 (FY2737), sir2 (FY2680), sir2 hst4 (FY2738), and clr4 (FY2735) strains.

 
Although sir2 mutants have silencing defects at telomeres and at centromeres, they do not show the chromosome segregation defects (see above) that are typical of swi6 or clr4 mutants with attenuated centromere function (ALLSHIRE et al. 1995; BERNARD et al. 2001; NONAKA et al. 2002). This led us to speculate that the single spot remaining in sir2 strains might be functional Swi6 protein remaining at the centromere, thus ensuring normal chromosome segregation. This possibility was investigated by simultaneously observing the localization of Swi6p and the spindle pole body, which associates with the centromeres in fission yeast (FUNABIKI et al. 1993).

Pcp1p, a homolog of Spc110p, is a component of the spindle pole body (FLORY et al. 2002). Cells expressing Pcp1-GFP alone should have a single GFP spot. Indeed, a single spot was observed in wild-type or sir2 control cells expressing Pcp1-GFP alone (Figure 8, C and D). In wild-type cells expressing both Swi6-GFP and Pcp1-GFP, two or more spots were observed (Figure 8E), consistent with the number of spots reflecting the sum of the Swi6-GFP and Pcp1-GFP centromeric spot and the Swi6-GFP telomeric spots. For sir2 strains, we predicted that if Swi6-GFP were associating with centromeres, then a single spot would be observed in cells coexpressing Swi6-GFP and Pcp1-GFP, whereas if Swi6-GFP were associated with telomeres, then there would be more than one spot in cells coexpressing Swi6-GFP and Pcp1-GFP. We observed a single spot in sir2 cells, consistent with the pronounced Swi6p signal being at the centromere (Figure 8F). Thus, Sir2p is likely to be required for localization of Swi6p to the telomeres, but not for its localization to centromeric regions.

Interestingly, the distribution of Swi6-GFP foci in hst4 strains was not significantly different from wild type, nor was the distribution of Swi6-GFP foci in sir2 strains significantly different from sir2 hst4 strains (Figure 8, bottom). This suggests that Hst4p, despite participating in centromeric and telomeric silencing, plays little role in localization of Swi6p to telomeres or centromeres. These results thereby provide further distinction between the activities of Sir2p and Hst4p.

S. cerevisiae SIR2 can substitute for S. pombe sir2⁺ in silencing:

Because the S. cerevisiae Sir2 protein appears to have some of the same general cellular functions as the S. pombe Sir2 protein, we performed experiments testing the ability of the budding yeast Sir2p to replace the fission yeast Sir2p. S. pombe sir2 (BglII-BssHII)mat2-P(XbaI)::ura4⁺ and wild-type (BglII-BssHII)mat2-P(XbaI)::ura4⁺ strains were transformed with a plasmid containing the S. cerevisiae SIR2 gene under control of the thiamine-repressible S. pombe nmt⁺ promoter or with a vector control. Transformants were tested for growth and silencing using dilution series on media with varying concentrations of thiamine.

High levels of expression of SIR2 (no thiamine) interfered with the growth of both wild-type and sir2 strains (Figure 9; compare growth on –leucine plates). Similar toxic effects are seen when SIR2 is overexpressed in budding yeast (HOLMES et al. 1997). At more moderate levels of expression (0.5 mM thiamine), SIR2 restored silencing to sir2 strains, resulting in abundant growth on 5-FOA plates and modestly decreased growth on –leucine –uracil plates. Thus, the S. cerevisiae SIR2 gene was able to functionally substitute for the S. pombe sir2⁺ gene in silencing. Interestingly, moderate levels of expression of SIR2 also subtly increased silencing in wild-type strains, demonstrating that sir2⁺ may be limiting for silencing under these conditions.

View larger version (46K): In this window In a new window Download PPT slide   FIGURE 9.— Cross-species complementation of S. pombe sir2 by S. cerevisiae SIR2. Wild-type (LPY6276) and sir2 (LPY6272) S. pombe strains were transformed with a vector control (pREP-1) or a plasmid in which S. cerevisiae SIR2 was expressed under the control of the nmt+ promoter (pLP1051). The transformants were tested for growth (–leucine plates) and silencing of the (BglII-BssHII)mat2-P(XbaI)::ura4+ reporter construct (–leucine –uracil and –leucine 5-FOA plates) using a fivefold dilution series. Silencing and growth are shown under conditions in which SIR2 was highly (no thiamine) or moderately (0.5 µM thiamine) expressed.

 

To learn cellular functions of the S. pombe gene most closely related to S. cerevisiae SIR2, we cloned the sir2⁺ cDNA and initiated a characterization of the sir2⁺ gene. Like its budding yeast counterpart, the fission yeast sir2⁺ gene plays a major role in transcriptional silencing at telomeres, the silent mating-type loci, and centromeres (Figures 3–5; Table 2; SHANKARANARAYANA et al. 2003).

The results reported in this article, in the context of previous results, help clarify the mechanistic role of S. pombe Sir2p in silencing. ChIP experiments established that Sir2p is more abundant at centromeres, telomeres, and the silent mating-type loci than at more telomere distal sites (Figure 7), suggesting strongly that Sir2p functions directly at those silenced regions. Recombinant Sir2p specifically deacetylates histone H3 lysine 9 in vitro and loss of Sir2p causes increased acetylation of histone H3 lysine 9 at telomeres and the silent mating-type loci in vivo, implying that acetylated histone H3 lysine 9 is an in vivo substrate of Sir2p (SHANKARANARAYANA et al. 2003). Histone H3 lysine 9 is critical for silencing, because alteration of this residue results in loss of silencing at cen1 and loss of Swi6-GFP localization (MELLONE et al. 2003). When sir2⁺ (SHANKARANARAYANA et al. 2003) and clr4⁺ (NAKAYAMA et al. 2001) are mutated, methylation of histone H3 lysine 9 is lost at the silent mating-type loci. In addition, Sir2p (Figure 8; SHANKARANARAYANA et al. 2003) and Clr4p (EKWALL et al. 1996; NAKAYAMA et al. 2001) are required for Swi6p localization to telomeres and the silent mating-type loci. Thus, in heterochromatic regions, Sir2p is likely to contribute to silencing by deacetylation of histone H3 lysine 9, which is subsequently methylated by Clr4p and thereby bound with increased affinity by Swi6p.

The importance of Sir2p for silencing varies with the heterochromatic region. We observed that Sir2p appears to be most important at the cryptic mating-type loci and at telomeres. At centromeres, we found that the contributions of Sir2p are less critical and depend on the site of reporter gene insertion. This result is in contrast to that of SHANKARANARYANA et al. (2003), who observed somewhat stronger defects in centromeric silencing. At present, the nature of the differences in our results is not clear, but may reflect local differences in positions of the reporters or in media conditions of the assays.

Localization of Swi6p to centromeric foci did not require Sir2p, as viewed microscopically (Figure 8), but did require Sir2p as assessed by ChIP (SHANKARANARAYANA et al. 2003). Because sir2 null strains do not show the sensitivity to TBZ and greatly elevated rates of chromosome loss that would be expected if Swi6p were completely delocalized from the centromere (Figure 2; data not shown), it seems likely that ChIP provides a quantitatively sensitive assay to detect loss of Swi6p, but that sufficient amounts of Swi6p remain to allow efficient chromosome segregation.

Sir2p appears to share many similarities with the Clr3 histone deacetylase. Clr3p, which is thought to specifically deacetylate histone H3 lysine 14 (BJERLING and EKWALL 2002), is required to varying degrees for silencing at the cryptic mating-type loci, telomeres, and centromeres (EKWALL and RUUSALA 1994; THON et al. 1994; ALLSHIRE et al. 1995; GREWAL et al. 1998; BJERLING and EKWALL 2002). Histone H3 lysine 14 is a critical residue for centromeric silencing and for Swi6-GFP localization (MELLONE et al. 2003). For both Sir2p and Clr3p, alteration of the presumed target histone residue causes more severe centromeric defects than mutation of either sir2⁺ or clr3⁺ alone. Thus, it seems possible that Sir2p and Clr3p work additively in centromeric silencing. Because a lack of Sir2p causes increased acetylation of histone H3 lysine 14 and a lack of Clr3p causes increased acetylation of histone H3 lysine 9 (NAKAYAMA et al. 2001; SHANKARANARAYANA et al. 2003), the modification state of one residue directly affects the other. In mammalian cells, different mechanistic models propose that interplay between modifications of histone H3 tail residues modulate HP1 binding (FISCHLE et al. 2003; MATEESCU et al. 2004). Considering this, studies of the relationship between different histone H3 tail histone modifications and modifying enzymes will be essential for a more complete understanding of silencing.

A significant focus of Sir2p function centers on histone H3 lysine 9. Although S. pombe Sir2p also deacetylates histone H4 lysine 16 in vitro (SHANKARANARAYANA et al. 2003), and histone H4 K16 is important for silencing in budding yeast (reviewed in RUSCHE et al. 2003), alteration of histone H4 lysine 16 does not cause detectable silencing defects at otr1 in fission yeast (MELLONE et al. 2003). It is possible that histone H4 lysine 16 is important for silencing but that alteration of the lysine 16 residue causes phenotypes that mimic the phenotype of sir2 mutants, with strong silencing defects at the silent mating-type loci but not at otr1.

The isolation of the sir2⁺ gene allowed us to compare the functions of S. pombe Sir2p to Hst4p. Although both proteins help to silence at telomeres and centromeres (Figures 3 and 4; FREEMAN-COOK et al. 1999), many of their roles seem to differ. The conclusion that the two proteins have nonredundant functions is further supported from double-mutant analyses. The sir2 hst4 mutants do not have exacerbated UV and TBZ sensitivity compared to hst4 mutants (Figure 2), nor do they have more severe Swi6-GFP localization defects compared to sir2 mutants (Figure 8). Identification of the in vivo substrate(s) of Hst4p should help clarify the basis for the distinct roles of the two proteins.

Although the Sir2 proteins from S. pombe and S. cerevisiae are very similar, the molecular mechanisms of silencing between the two organisms previously appeared quite distinct. Therefore, in some respects it was particularly surprising that the S. cerevisiae SIR2 gene was able to substitute for S. pombe sir2⁺ in silencing at mat2. This cross-species complementation reveals a fundamentally conserved aspect of silencing in fission yeast and budding yeast that extends beyond the need for a particular enzymatic activity to deacetylate histones. Indeed, the S. cerevisiae Sir2 protein is likely to be targeted to S. pombe silenced regions, even though S. pombe lacks the Sir3p and Sir4p proteins that promote targeting in S. cerevisiae. S. pombe does have a Rap1p homolog, which is required for telomeric silencing and is recruited to telomeres by Taz1p (CHIKASHIGE and HIRAOKA 2001; KANOH and ISHIKAWA 2001; PARK et al. 2002). However, S. pombe Rap1p is not localized to the silent mating-type loci by ChIP (KANOH and ISHIKAWA 2001) and is not required for silencing at mat3 (PARK et al. 2002). Consequently, it seems unlikely that Rap1p recruits S. cerevisiae Sir2p to the silent mating-type loci. Alternatively, S. cerevisiae Sir2p expressed in S. pombe might be able to utilize RNAi-mediated targeting (HALL et al. 2002; VERDEL et al. 2004). This notion is particularly intriguing considering that RNAi regulatory mechanisms have not been identified in S. cerevisiae. If true, perhaps RNAi was once a part of the S. cerevisiae repertoire. Examination of silencing mediated by S. cerevisiae SIR2 in rap1 or RNAi mutants will address these mechanistic questions.

The data presented here focus on the silencing functions of the fission yeast Sir2 protein, which may well have additional cellular roles. The relatively uniform nuclear staining pattern of Sir2-myc (Figure 6) differs from the pattern of Sir2p localization in S. cerevisiae, in which Sir2p staining is enriched within telomeric foci and the nucleolus (GOTTA et al. 1997). This difference in distribution may indicate that S. pombe Sir2p does not function exclusively at heterochromatic-like regions such as centromeres, telomeres, the silent mating-type loci, and the rDNA. Perhaps fission yeast Sir2p also performs the cellular functions of Hst1p, such as helping to modulate cellular NAD⁺ levels (BEDALOV et al. 2003) or repressing vegetative expression of a subset of genes required for meiosis (XIE et al. 1999). Future studies will establish if S. pombe Sir2p has functions found in mammals, such as modulation of the activity of transcription factors (BRUNET 2004; MOTTA et al. 2004), or additional functions found within S. cerevisiae, including repressing homologous recombination within the rDNA (GOTTLIEB and ESPOSITO, 1989) and checkpoint control in meiosis (SAN-SEGUNDO and ROEDER, 1999). Although many questions remain about the scope of its functions, it is clear that sir2⁺ plays an important conserved role in silencing the centromeres, telomeres, and silent mating-type loci.

We thank Trisha Davis for the Pcp1-GFP strain, Junko Kanoh for the Taz1-HA strain, and Leonard Guarente and Francois LaCroute for cDNA libraries. We also thank Joanna Lowell for the construction of the nmt⁺-SIR2 plasmid and Joaquin M. Espinosa for help with the real-time PCR experiments. We are grateful to Sandi Jacobson, Jeanne Wilson, and Russell Darst for their incisive comments on the manuscript. This work was initiated while L.F.-C. was a Howard Hughes Medical Institute predoctoral fellow at the Department of Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, Colorado. This work was supported by grants from the National Institutes of Health to L.P. (GM54778) and S.L.F. (GM59321).

¹ Present address: Department of Genetics, Yale University School of Medicine, New Haven, CT 06510.

² Present address: Molecular and Computational Biology Section, University of Southern California, Los Angeles, CA 90089-1340.

ALLSHIRE, R. C., J. P. JAVERZAT, N. J. REDHEAD and G. CRANSTON, 1994 Position effect variegation at fission yeast centromeres. Cell 76: 157–169.[CrossRef][Medline]

ALLSHIRE, R. C., E. R. NIMMO, K. EKWALL, J. P. JAVERZAT and G. CRANSTON, 1995 Mutations derepressing silent centromeric domains in fission yeast disrupt chromosome segregation. Genes Dev. 9: 218–233.[Abstract/Free Full Text]

APARICIO, O. M., B. L. BILLINGTON and D. E. GOTTSCHLING, 1991 Modifiers of position effect are shared between telomeric and silent mating-type loci in S. cerevisiae. Cell 66: 1279–1287.[CrossRef][Medline]

ARMSTRONG, C. M., M. KAEBERLEIN, S. I. IMAI and L. GUARENTE, 2002 Mutations in Saccharomyces cerevisiae gene SIR2 can have differential effects on in vivo silencing phenotypes and in vitro histone deacetylation activity. Mol. Biol. Cell 13: 1427–1438.[Abstract/Free Full Text]

ASTROM, S. U., T. W. CLINE and J. RINE, 2003 The Drosophila melanogaster sir2⁺ gene is nonessential and has only minor effects on position-effect variegation. Genetics 163: 931–937.[Abstract/Free Full Text]

BAHLER, J., J. Q. WU, M. S. LONGTINE, N. G. SHAH, A. MCKENZIE, III et al., 1998 Heterologous modules for efficient and versatile PCR-based gene targeting in Schizosaccharomyces pombe. Yeast 14: 943–951.[CrossRef][Medline]

BAILIS, J. M., and S. L. FORSBURG, 2002 RNAi hushes heterochromatin. Genome Biol. 3: REVIEWS1035.

BEDALOV, A., M. HIRAO, J. POSAKONY, M. NELSON and J. A. SIMON, 2003 NAD+-dependent deacetylase Hst1p controls biosynthesis and cellular NAD+ levels in Saccharomyces cerevisiae. Mol. Cell. Biol. 23: 7044–7054.[Abstract/Free Full Text]

BELL, S. D., C. H. BOTTING, B. N. WARDLEWORTH, S. P. JACKSON and M. F. WHITE, 2002 The interaction of Alba, a conserved archaeal chromatin protein, with Sir2 and its regulation by acetylation. Science 296: 148–151.[Abstract/Free Full Text]

BERNARD, P., J. F. MAURE, J. F. PARTRIDGE, S. GENIER, J. P. JAVERZAT et al., 2001 Requirement of heterochromatin for cohesion at centromeres. Science 294: 2539–2542.[Abstract/Free Full Text]

BITTERMAN, K. J., O. MEDVEDIK and D. A. SINCLAIR, 2003 Longevity regulation in Saccharomyces cerevisiae: linking metabolism, genome stability, and heterochromatin. Microbiol. Mol. Biol. Rev. 67: 376–399.[Abstract/Free Full Text]

BJERLING, P., and K. EKWALL, 2002 Centromere domain organization and histone modifications. Braz. J. Med. Biol. Res. 35: 499–507.[Medline]

BOEKE, J. D., F. LACROUTE and G. R. FINK, 1984 A positive selection for mutants lacking orotidine-5'-phosphate decarboxylase activity in yeast: 5-fluoro-orotic acid resistance. Mol. Gen. Genet. 197: 345–346.[CrossRef][Medline]

BRACHMANN, C. B., J. M. SHERMAN, S. E. DEVINE, E. E. CAMERON, L. PILLUS et al., 1995 The SIR2 gene family, conserved from bacteria to humans, functions in silencing, cell cycle progression, and chromosome stability. Genes Dev. 9: 2888–2902.[Abstract/Free Full Text]

BRUNET, A., L. B. SWEENEY, J. F. STURGILL, K. F. CHUA, P. L. GREER et al., 2004 Stress-dependent regulation of FOXO transcription factors by the SIRT1 deacetylase. Science 303: 2011–2015.[Abstract/Free Full Text]

BRYK, M., M. BANERJEE, M. MURPHY, K. E. KNUDSEN, D. J. GARFINKEL et al., 1997 Transcriptional silencing of Ty1 elements in the RDN1 locus of yeast. Genes Dev. 11: 255–269.[Abstract/Free Full Text]

BUCK, S. W., C. M