Accelerated Aging and Nucleolar Fragmentation in Yeast sgs1 Mutants

See allHide authors and affiliations

Science  29 Aug 1997:
Vol. 277, Issue 5330, pp. 1313-1316
DOI: 10.1126/science.277.5330.1313


The SGS1 gene of yeast encodes a DNA helicase with homology to the human WRN gene. Mutations in WRNresult in Werner's syndrome, a disease with symptoms resembling premature aging. Mutation of SGS1 is shown to cause premature aging in yeast mother cells on the basis of a shortened life-span and the aging-induced phenotypes of sterility and redistribution of the Sir3 silencing protein from telomeres to the nucleolus. Further, in old sgs1 cells the nucleolus is enlarged and fragmentedchanges that also occur in old wild-type cells. These findings suggest a conserved mechanism of cellular aging that may be related to nucleolar structure.

The SGS1 gene ofSaccharomyces cerevisiae is a member of the RecQ helicase family that includes human BLM (mutations in which cause Bloom's syndrome) (1), human RECQL(2), and WRN (3). Patients with Werner's syndrome contain two mutant alleles of WRN and display many symptoms of old age including graying and loss of hair, osteoporosis, cataracts, atherosclerosis, loss of skin elasticity, and a propensity for certain cancers (4). Cells isolated from patients with Werner's syndrome divide approximately half as many times in culture as those from normal individuals (4).

Yeast cells lacking topoisomerase III activity (top3) are unable to unwind negatively supercoiled DNA efficiently and thus grow extremely slowly. Mutations in SGS1 were first identified by their ability to suppress the slow-growth phenotype of top3strains (5); the Sgs1 protein was subsequently shown to interact physically with both topoisomerases II and III (5-7). SGS1 is required for the fidelity of chromosome segregation and the suppression of recombination at the ribosomal DNA (rDNA) array and other loci (5-7).

Cell division in S. cerevisiae is asymmetric, giving rise to a large mother and a small daughter cell. Mother cells undergo, on average, a fixed number of cell divisions and show characteristic changes with age, such as cell enlargement (8) and sterility (9). Age-related sterility is due to a loss of transcriptional silencing at the mating type loci HMR aand HMLα, and the resulting simultaneous expression of both a and α information (10). Old yeast cells also exhibit a loss of telomeric silencing (11). In young cells, silencing at HM loci and telomeres is mediated by the Sir2/3/4 protein complex (12, 13). In addition, Sir2p silences marker genes inserted at rDNA (14) and suppresses recombination between rDNA repeats (15).

The genes SIR2, SIR3, and SIR4, as well as UTH4, are determinants of life-span in yeast (16). Null mutations in these genes shorten life-span (17), whereas overexpression of UTH4 extends life-span. Life-span extension by the Sir complex is associated with a redistribution of these factors from telomeres (and HM loci) to the nucleolus in old cells—a process requiring UTH4(16, 18).

We investigated the role, if any, of SGS1 in yeast longevity. Life-span analysis of the wild-type S. cerevisiaeW303-1A strain and an otherwise isogenic sgs1 strain was carried out by micromanipulation until each mother ceased to divide. The average life-span of sgs1 cells (9.5 divisions) was about 40% that of the wild-type strain (24.5 divisions) (Fig.1A). The maximum life-span of thesgs1 strain was similarly reduced (18 versus 40 divisions).

Figure 1

Mutant sgs1 cells have a short life-span. (A and B) Mortality curves of wild-type (W303-1A MAT a ade2-1 can1-100 his3-11,15 leu2-3,112 trp1-1 ura3-1) and isogenic mutant derivatives sgs1::HIS3,ada1::hisG, and hap5::hisGwere determined by the removal of daughters from virgin mothers by micromanipulation until senescence of the mother cell as determined by a cessation of cell division. Average life-spans (and sample sizes) were as follows: wild type, 24.5 (39); sgs1, 9.5 (30);ada1, 3.3 (23); and hap5, 16.6 (28).

To determine whether the sgs1 cells actually died from aging, rather than general sickness, we examined the age-specific phenotype of sterility as scored by their ability to respond to mating pheromone. In the sgs1 mutant, young cells that had completed less than 50% of their life-span were almost always fertile (Fig. 2A). However, over 60% of cells in the last one-fifth of their life-span were sterile. This pattern, previously observed for wild-type cells, was due to a loss of transcriptional silencing at HMLα (10). Deletion of HMLα reduced the frequency of sterility in theMAT a sgs1 strain to almost zero (Fig. 2B).

Figure 2

Mutant sgs1 cells age prematurely. (A to D) Cells of various ages were scored for their ability to undergo cell cycle arrest and schmooing in response to the yeast mating pheromone, α-factor. After 4 hours of α-factor challenge, cells were transferred to fresh medium to complete their life-span. Life-span is not affected by deletion ofHML (10). Number of cells in each data set for each age group is shown above bar. Cells that failed to respond or divide after α-factor challenge were excluded from the data set. Disruption plasmids pCW9-1, pDM212 (19), and pADA1KO (20) were used to create hmlαΔ::LEU2,hap5::hisG, and ada1::hisGstrains, respectively.

Is sterility in yeast an aging-specific phenotype? We addressed this by examining two short-lived strains that were mutant for the transcriptional activator HAP5 (19) or the coactivator ADA1 (20). Although the two strains have significantly shortened life-spans (Fig. 1B), neither thehap5 nor the ada1 strains became sterile at a high frequency as they grew older (Fig. 2, C and D). Thus, sterility is an aging-specific phenotype that occurs prematurely in sgs1mutants.

Sgs1p was visualized by indirect immunofluorescence of fixed yeast cells (16). In >90% of cells, Sgs1p staining (green) was concentrated on one side of the nucleus (blue) (Fig.3, A and B). Costaining cells for yeast Nop1p (red), an abundant nucleolar protein (21,22), demonstrated that the concentration of Sgs1p coincided with the nucleolus (Fig. 3C).

Figure 3

Nucleolar localization of Sgs1p and nucleolar fragmentation in old cells. Escherichia coli recombinant 6histidine–tagged Sgs1p (amino acids 1071 to 1447) was affinity-purified with Ni-NTA agarose (28). Immunoglobulin Y (IgY), which recognized 0.1 ng of Sgs1p, was affinity-purified from chicken serum with the use of recombinant Sgs1p. (A) Young wild-type cells stained for DNA with 4′,6′-diamidino-2-phenylindole (DAPI) (blue) and yeast fibrillarin with a monoclonal antibody to Nop1p (22) and Cy3-conjugated goat antibody to mouse (anti-mouse) IgG (red). (B) Sgs1p stained with sheep fluorescein isothiocyanate (FITC)–conjugated IgG against chicken anti-Sgs1p (green). (C) Merged image of Sgs1p and Nop1p (overlap is yellow). sgs1 mutant strains treated identically (insets A to C) showed only background staining. (D to J) Cells were stained with anti-Nop1p (red) and DAPI (blue) (18): (D) young sgs1cells; (E) old sgs1 cells [average bud scar count (ABSC) of magnetic sort ± SD: 9 ± 2]; (F) old wild-type cells (ABSC = 26 ± 3); (G) old sgs1 sir3::URA3 cells (ABSC = 8 ± 2); (H) oldada1 cells (ABSC = 2 ± 1); (I) starved wild-type cells; and (J) cells overexpressing Sgs1p from an inducible GAL-promoter (29). SIR3 was disrupted with pDM42. Starved cells were obtained by growing first in yeast extract–bactopeptone–glucose to stationary phase, then in 2.4 g/liter yeast nitrogen base without amino acids for 12 hours.

We then examined whether nucleolar changes occur during the life-span of sgs1 cells. As observed for young wild-type cells (21), the nucleoli of young sgs1 cells were well-defined structures occupying about 20% of the nucleus (Fig. 3D). In about 50% of old sgs1 cells, the nucleolus was fragmented into several bodies that occupied a large fraction of the nucleus (Fig. 3E). In the remaining cells, a substantial enlargement of the structure was evident, which may be a prelude to fragmentation. In young wild-type cells expressing SGS1 at high levels, a similar pattern of nucleolar fragmentation was observed (Fig. 3J), indicating that abundant Sgs1p may mimic the null phenotype by unbalancing the stoichiometry of an Sgs1p complex.

To determine whether fragmentation of the nucleolus was due to a scaling of a normal aging phenotype or was a consequence of thesgs1 mutation per se, we examined the nucleoli of old wild-type cells. A substantial fraction (about 45%) of these had enlarged or fragmented nucleoli, similar to those in thesgs1 cells (Fig. 3F). Fragmentation of the nucleolus in either young or old ada1 cells was not observed (Fig. 3H).

Enlargement and fragmentation of the nucleolus may have been a response to an age-related reduction in protein synthesis. We therefore examined the nucleoli of starved young cells and found them to be ∼25% the size of wild-type nucleoli and with a more rounded appearance (Fig.3I). Hence, starvation is not a signal for nucleolar enlargement.

Compelling evidence that sgs1 cells age prematurely would be the redistribution of Sir proteins to the nucleolus after only nine generations. In young sgs1 cells, Sir3p staining showed characteristic telomere spots (18, 23) (Fig.4B). However, in old sgs1cells, Sir3p staining was predominantly nucleolar. In old cells with fragmented nucleoli, Sir3p staining coincided with most or all nucleolar bodies (Fig. 4F).

Figure 4

Sir3p localized to nucleolar fragments of old sgs1 cells. Old sgs1cells were obtained by magnetic sorting (ABSC ± SD = 9 ± 2) and stained as in Fig. 3. Nop1p is shown in red (A, C, D, F); DNA is in blue (A and D); Sir3p was visualized with FITC-conjugated goat anti-rabbit IgG (green) (B, C, E, F). Overlap of Sir3p and Nop1p staining is shown in yellow. (A to C) Young sgs1 cells. (D to F) Oldsgs1 cells. (G to H) Twelve-generation old wild-type and sir3 cells stained for DNA (blue) and Nop1p (red). (G) Wild-type cells. (H) sir3mutant cells. Proportion of cells with small, large, or fragmented nucleoli (percent): wild type (82, 12, 6); sir3 (43, 38, 19).

Nucleolar enlargement and fragmentation could either represent a cause of aging or be a response to aging that promotes longevity. If fragmentation of the nucleolus is a longevity response in old cells, it might be caused by redistribution of the Sir complex to the nucleolus. Thus fragmentation would not occur in a sir null mutant. In old sgs1 sir3 cells nucleolar fragmentation occurred at the same high level as in the sgs1 single mutant (Fig. 3G), showing that the Sir complex does not cause fragmentation.

It was still possible that nucleolar enlargement and fragmentation were part of a longevity response, but did not require Sir protein redistribution. To test this possibility, we grewSGS1 wild-type strains, which were either SIR3(wild type) or sir3, for 12 divisions and sorted them for old cells. At this age, many of the sir3 cells are approaching senescence because of their shorter life-span. Nop1p staining showed that a substantial fraction of the nucleoli in thesir3 mutant were either enlarged or fragmented (38% and 19%, respectively), a proportion much higher than in the age-matched wild type (12% and 6%, respectively) (Fig. 4, G and H). This finding suggests that the redistribution of the Sir complex to the nucleolus represses nucleolar fragmentation and promotes longevity.

Our data show that deletion of SGS1 causes premature aging in yeast on the basis of three phenotypes: (i) the average life-span ofsgs1 cells is about 40% of wild type; (ii) sgs1cells prematurely assume the aging-associated sterility, whereas mutants for other yeast genes either do not display a shorter life-span or do not exhibit the age-specific phenotype of sterility; and (iii) the Sir protein silencing complex redistributes from telomeres to the nucleolus in old sgs1 cells, as observed in old wild-type cells.

We observed a profound enlargement and fragmentation of the nucleolus in old sgs1 and wild-type cells. These fragments may result from a breakdown of nucleolar architecture or result from the inherent instability of the rDNA itself. We favor the latter hypothesis. Further, the nucleolar fragments were often arrayed in an orderly series around the periphery of the nucleolus, perhaps reflecting a mechanism that promotes assymetric segregation of the fragments to mother cells.

We cannot rule out the possibility that nucleolar enlargement and fragmentation are simply markers of aging. However, we do not favor this view because (i) an independent study has indicated a central role for the nucleolus in determining life-span (16), and (ii) the localization of Sgs1p in the nucleolus suggests that a nucleolar defect resulting from its absence may be a direct cause of premature aging.

The data herein indicate that nucleolar enlargement and fragmentation represent a cause of yeast aging, and that these changes are delayed by the redistribution of Sir proteins to the nucleolus. The similar effect of the related SGS1 and WRN genes on yeast and human aging along with age-associated changes in rDNA content reported for several mammalian species (24, 25) suggest that a common mechanism may underlie aging in eukaryotes.


Stay Connected to Science

Navigate This Article