A cytosolic NAD-dependent deacetylase, Hst2p, can modulate nucleolar and telomeric silencing in yeast

Abstract

In budding yeast, the silent information regulator Sir2p is a nuclear NAD-dependent deacetylase that is essential for both telomeric and rDNA silencing. All eukaryotic species examined to date have multiple homologues of Sir two (HSTs), which share a highly conserved globular core domain. Here we report that yeast Hst2p and a mammalian Hst2p homologue, hSirT2p, are cytoplasmic in yeast and human cells, in contrast to yHst1p and ySir2p which are exclusively nuclear. Although yHst2p cannot restore silencing in a sir2 deletion, overexpression of yHst2p influences nuclear silencing events in a SIR2 strain, derepressing subtelomeric silencing while increasing repression in the rDNA. In contrast, a form of ySir2p carrying a point mutation in the conserved core domain disrupts both telomeric position effect (TPE) and rDNA repression at low expression levels. This argues that non-nuclear yHst2p can compete for a substrate or ligand specifically required for telomeric, and not rDNA repression.

Fig. 1. Comparative alignments and phylogenetic tree of Sir2 family members. (A) Diagrammatic representations of Sir2p homologues are aligned with respect to the evolutionarily conserved core domain (light grey boxes). The PRSS3 score is indicated for the pairwise comparison of the indicated core domain with that of the yeast Sir2p or Hst2p (labelled ScSir2 and ScHst2). The PRSS3 program (http://www.ch.embnet.org/software/PRSS) calculates the optimal score of a protein sequence alignment and evaluates the significance of this score. Scores with higher homology to ySir2p are indicated in grey and those closer to yHst2p are in italics. The percentage amino acid identity shared with the ySir2p core domain is calculated by the Gap program of GCG (with ran = 100, gap = 12, gap extension = 2), and is indicated for the yHst family and the Homo sapiens SirT2 protein. Unshaded boxes are regions of no significant identity (<20%) with ySir2p. Cross-hatched boxes indicate 50% identity in N-terminal extensions, and black boxes indicate 55% identity to ySir2p. The amino acid boundaries of the domains selected for comparison are indicated. Vertical black bars represent cysteine pairs of a putative zinc finger motif. We have used the nomenclature of Frye (1999) for human SirTuins 1–5, and named a previously unreported human cDNA, HsSirT6. The HsSirT2 is identical to the gene called hSir2L (Afshar and Murnane, 1999) and hSir2A (Sherman et al., 1999). (B) The phylogenetic unrooted tree of eukaryotic Sir2 homologues was generated using CLUSTALW (Higgins et al., 1996) and TREEVIEW (Page, 1996), which compare the core domain sequences of homologues identified in cDNA and unigene libraries from Escherichia coli, S.cerevisiae, K.lactis, Schizosaccharomyces pombe, C.albicans, Leishmania major, H.sapiens, Mus musculus, Caenorhabditis elegans and Drosophila melanogaster. The database accession No. for each gene listed is given in Materials and methods. The Sir2 subfamily is indicated in grey and the Hst2 subfamily in italics.

Fig. 2. yHst2p at high levels of overexpression is dominant-negative for TPE. The effects of ectopic expression of yeast Hst2p on silencing of the TelVR::ADE2 reporter gene (Singer and Gottschling, 1994) were determined after growth on glucose (Glu, low expression levels) and galactose (Gal, high expression levels). Isogenic sir2::HIS3 (GA427) and SIR2 (GA426) strains, carrying the TelVR::ADE2 reporter gene, were used in (A). (A) The plasmid indicated above each panel was introduced into the sir2::HIS3 strain GA427 labelled sir2Δ or in the SIR2 strain GA426, and were grown under limiting adenine conditions (see Materials and methods). pJG45, pJG45-ySir2 and pJG45-yHst2 express the indicated yeast protein fused at their N-termini to the B42-NLS-HA peptide. pGAL-yHst2 encodes an HA-Hst2p fusion. All fusions are expressed under the inducible GAL10 pomoter, and are expressed at low levels on glucose and induced at least 40-fold on galactose. Both on glucose and galactose media, sir2Δ colonies carrying the parental vectors (shown here only for pJG45) are white, indicating ADE2 expression, whereas Sir2⁺ colonies carrying the parental vectors alone (shown here only for pJG45) have a red/white sectored appearance like the untransformed strain, indicating metastable ADE2 silencing. (B) The effects of strong ectopic expression of yHst2p, with and without a fused SV40 NLS, on silencing of RDN1::mURA3/HIS3, was determined in isogenic sir2::kanMX4 (GA759) and SIR2 (GA758) strains. Strains were transformed with the plasmids indicated at the right of the graph and described above in (A). Galactose induces strong expression from the GAL10 promoter. Five-fold serial dilutions of each transformant were plated on both –TRP and –TRP–URA media with the indicated carbon source. URA3 expression is calculated as described (see Materials and methods). The sir2::kanMX4 strain (GA759) carrying either vector alone gives a value of ∼1 for this calculation (100% of the colonies express URA3), while the SIR2 strain (GA758) with either vector gives a value of ∼0.1 (10% of the cells express URA3). The mean and standard deviation were calculated from at least three independent transformants of each plasmid.

Fig. 2. yHst2p at high levels of overexpression is dominant-negative for TPE. The effects of ectopic expression of yeast Hst2p on silencing of the TelVR::ADE2 reporter gene (Singer and Gottschling, 1994) were determined after growth on glucose (Glu, low expression levels) and galactose (Gal, high expression levels). Isogenic sir2::HIS3 (GA427) and SIR2 (GA426) strains, carrying the TelVR::ADE2 reporter gene, were used in (A). (A) The plasmid indicated above each panel was introduced into the sir2::HIS3 strain GA427 labelled sir2Δ or in the SIR2 strain GA426, and were grown under limiting adenine conditions (see Materials and methods). pJG45, pJG45-ySir2 and pJG45-yHst2 express the indicated yeast protein fused at their N-termini to the B42-NLS-HA peptide. pGAL-yHst2 encodes an HA-Hst2p fusion. All fusions are expressed under the inducible GAL10 pomoter, and are expressed at low levels on glucose and induced at least 40-fold on galactose. Both on glucose and galactose media, sir2Δ colonies carrying the parental vectors (shown here only for pJG45) are white, indicating ADE2 expression, whereas Sir2⁺ colonies carrying the parental vectors alone (shown here only for pJG45) have a red/white sectored appearance like the untransformed strain, indicating metastable ADE2 silencing. (B) The effects of strong ectopic expression of yHst2p, with and without a fused SV40 NLS, on silencing of RDN1::mURA3/HIS3, was determined in isogenic sir2::kanMX4 (GA759) and SIR2 (GA758) strains. Strains were transformed with the plasmids indicated at the right of the graph and described above in (A). Galactose induces strong expression from the GAL10 promoter. Five-fold serial dilutions of each transformant were plated on both –TRP and –TRP–URA media with the indicated carbon source. URA3 expression is calculated as described (see Materials and methods). The sir2::kanMX4 strain (GA759) carrying either vector alone gives a value of ∼1 for this calculation (100% of the colonies express URA3), while the SIR2 strain (GA758) with either vector gives a value of ∼0.1 (10% of the cells express URA3). The mean and standard deviation were calculated from at least three independent transformants of each plasmid.

Fig. 3. yHst1p is enriched in the non-nucleolar nucleoplasm and yHst2p is cytoplasmic. The indicated proteins were localized by indirect immunofluorescence on fixed yeast cells as described in Materials and methods. In all cases, the nucleolar marker Nop1p was localized with anti-Nop1 (rabbit antiserum or mouse monoclonal antibody, as appropriate; Gotta et al., 1997) and a Cy5-coupled secondary antibody. This is shown in the first panel of each row and is red in the merged images. In the first row, localization of ectopically expressed ySir2p (α-Sir2, green in merged image) was examined in a diploid sir2::HIS3 strain (GA194) after transformation with pADH-ySir2. The inset shows the localization of ySir2p to the telomeric foci and the nucleolus, when the cells have been washed, after fixation, in 1% Triton–0.02% SDS to improve accessibility (see Gotta et al., 1997). yHst1-Myc is detected by the monoclonal 9E10 (α-Myc) in the haploid strain GA1154 (SIR2) and the isogenic sir2::HIS3 strain GA1155 (inset). yHst2-Myc was examined in GA1276 (SIR2) and the isogenic sir2::HIS3 strain GA1229 (inset). Both fusions are genomic and under their endogenous promoters. The NLS-containing HA-tagged yHst2p expressed from pJG45-yHst2 was examined in transformants of the diploid wild-type strain GA225 expressing either low or high levels of the protein after 4 h of galactose induction, as indicated. The localization of the HA-tagged yHst2 fusion protein, which is encoded by pGAL-yHst2 and lacks a detectable NLS, was examined in transformants of GA225 under conditions of low and high expression, as indicated. The merge is shown in colour, with Nop1p in red and ySir2p, c-Myc or HA epitopes in green. Coincidence of the two signals is yellow. Bar = 2 µm.

Fig. 4. Telomeric foci remain intact at the nuclear periphery despite yHst2p-mediated disruption of silencing. Upper panels: ySir4p-Myc was localized in SIR2 cells (GA1275) transformed with pJG45-ySir2 after 4 h induction on galactose, using mouse anti-Myc antibodies (red signal in the merge). Nuclear localization of the highly overexpressed HA-Sir2p is demonstrated by immunostaining with anti-HA in the first panel (visualized in green in the merge). ySir4p-Myc is partially delocalized as compared with the punctate pattern observed in cells that do not overexpress ySir2p (GA1275, inset A). Lower panels: ySir4-Myc was localized in the same cells transformed with pGAL-yHst2 and induced for 4 h on galactose. The cytoplasmically localized HA-yHst2p is visualized in green; anti-Myc (ySir4p-Myc) is in red in the merged image. The inset B shows control cells that do not overexpress yHst2p. Bar = 2 µm.

Fig. 5. hSirT2 is cytoplasmic in human cells. (A) Whole-cell extracts of a human embryonic kidney cell line (Phoenix cells, labelled Φ, see Materials and methods) and of the same cells transfected with pCMV-hSirT2-Myc6His (labelled ΦTRX) were analysed by western blotting using anti-His, anti-Myc, column-purified anti-hSirT2 (see Materials and methods) and the same anti-hSirT2 mixed with an excess of a bacterial extract expressing MBP–hSirT2, to demonstrate the specificity of the anti-hSirT2 antibody. Each lane was loaded with ∼40 µg of total protein. Molecular weight markers (kDa) are indicated on the left of the blot. (B) Protein samples are as in (A), with the addition of a total cell extract from HeLa cells. A 40 µg aliquot of protein was analysed in each lane by western blotting using affinity-purified anti-hSirT2 and the same purified antibody mixed with bacterial extract expressing MBP, to identify endogenous hSirT2p. Molecular weight markers (kDa) are indicated on the left of the blot. (C) hSirT2-Myc6His and endogenous hSirT2p are cytoplasmic in Phoenix cells. Row a: Phoenix cells transfected with pCMV-hSirT2-Myc6His were stained with anti-pore (detected by Cy5-conjugated secondary antibodies, red in the merge) and anti-Myc (detected by DTAF-conjugated secondary antibodies, green in the merge). Row b: to test the specificity of the anti-Myc, non-transfected Phoenix cells were stained with anti-pore and anti-Myc. Row c: Phoenix cells transfected with pCMV-hSirT2-Myc6His were stained with anti-pore (red in the merge) and column-purified anti-hSirT2 (green in the merge). Row d: as c, except that the column-purified anti-hSirT2 antibodies were pre-incubated with an excess of extract from bacteria overexpressing MBP–hSirT2p. The immune complexes were removed by centrifugation prior to staining the fixed cells. Row e: non-transfected Phoenix cells were stained with anti-pore (red in the merge) and column-purified anti-hSirT2 (green in the merge). To detect the low level signal of endogenous hSirT2, the laser intensity (488 nm) was increased 40-fold over the scanning conditions used in a–d. Row f: the specificity of column-purified anti-hSirT2 is demonstrated by pre-mixing an excess of bacterial extract expressing MBP–hSirT2 with the purified antibody, prior to staining non-transfected Phoenix cells with anti-pore (red in the merge) and the depleted anti-hSirT2 (green in the merge). Bar = 15 µm.

Fig. 6. The ysir2^P394L mutant form is non-functional and disrupts TPE in wild-type cells. (A) The effects of ectopic expression of yeast Sir2p mutants and human SirT2 on silencing of the TelVR::ADE2 reporter gene (Singer and Gottschling, 1994) was determined after growth on glucose (Glu, low expression) and galactose (Gal, high expression). Isogenic sir2::HIS3 (GA427) and Sir2⁺ (GA426) strains were transformed with the plasmids indicated above each panel and were grown under limiting adenine conditions such that colonies accumulate red pigment when ADE2 is repressed. The plasmids pJG45-ySir2 and pJG45-ysir2^P394L express the indicated yeast protein fused at their N-termini to the B42-NLS-HA peptide. pRD-hSirT2 and pRD-hsirT2^P182L encode HA epitope-tagged hSirT2p and hsirT2^P182Lp. All are under GAL10 control. Although these hSirT2 constructs lack their first 33 amino acids, we obtained identical results by expressing a full-length hSirT2 clone fused to a C-terminal Myc epitope under ADH control (data not shown and see below). (B) The effects of wild-type and mutated yeast Sir2p (plasmids pJG45-ysir2^P394L and pJG45-ysir2^LGG) on silencing of the TelVIIL::URA3 reporter were determined in a Sir2⁺ strain carrying the TelVR::ADE2 marker (GA503). GA503 was transformed with the plasmids indicated on the right: pJG45-ysir2^P394L, pJG45-ysir2^LGG, pJG45-ySir2 and the vector alone. Ten-fold serial dilutions starting with 5 × 10⁵ cells of each transformant were plated on glucose plates lacking tryptophan (–TRP), tryptophan and uracil (–TRP–URA), and on galactose plates with the same selectivity (growth on –TRP/Gal was identical to that observed on –TRP, data not shown). Growth on –URA indicates loss of TPE on glucose or galactose in the presence of the mutated forms of ySir2p, while low level expression of wild-type ySir2p improves TPE. On –TRP plates, the reddish colour indicates subtelomeric repression of the TelVR::ADE2 marker, whereas white indicates full derepression. (C) The effect of ectopic expression of wild-type and mutants of ySir2p and hSirT2p on silencing of a RDN1::mURA3 reporter was determined in isogenic sir2::kanMX4 (GA759) and Sir2⁺ (GA758) strains, transformed with the plasmids indicated on the right of the graph. Plasmids pJG45, pJG45-ySir2, pJG45-ysir2^LGG and pJG45-ysir2^P394L are described in (B), and are all galactose inducible. Five-fold serial dilutions of each transformant were plated on both –TRP and –TRP–URA plates on the indicated carbon source. The fraction of colonies in which the URA3 gene is expressed is described in Materials and methods. For sir2::kanMX4 strain (GA759), transformation with either vector alone gives a value of ∼1 (100% of colonies express URA3) while the Sir2⁺ strain (GA758) with either vector alone gives ∼0.1 (∼10% of cells express URA3). Standard deviations and means were calculated from at least three independent transformants of each plasmid. (D) Epitope-tagged proteins expressed in GA426 after transformation with the plasmid indicated above each lane were detected by western blots of whole-cell extracts of each transformant. For all constructs, except those labelled p2µ, proteins were extracted after growth on galactose/raffinose. Approximately 30 µg of protein extract was loaded per lane. Equal loading was checked by blotting with a constitutively expressed protein, p55^RNase H (Karwan et al., 1990).

Fig. 7. Model for yHst2p effects on silencing in yeast. A model is shown to account for the dominant-negative effects of cytoplasmic yHst2p on silencing in yeast. yHst2p is cytoplasmic at both low and high levels of expression, and its absence has no effect on TPE or rDNA silencing. At high levels of expression, cytoplasmic yHst2p affects TPE and rDNA silencing much like highly overexpressed ySir2p: rDNA silencing improves while TPE is disrupted. We propose that yHst2p sequesters or modifies a ligand of ySir2p that is essential for TPE, and which shuttles between the nucleoplasm and cytoplasm. By sequestering or modifying this unknown ligand, yHst2p may disrupt TPE, releasing a pool of ySir2p that can relocalize to the nucleolus and improves rDNA silencing. The limiting ySir2p/yHst2p ligand must not be necessary for rDNA repression. In this model, fluctuations in levels of cytoplasmic ySir2-like proteins, as well as changes in the amount of this ligand are predicted to influence Sir protein function at telomeres and at the rDNA.