• Open Access

Regulation of chronological aging in Schizosaccharomyces pombe by the protein kinases Pka1 and Sck2

Authors


Luis A. Rokeach, Pascal Chartrand and Gerardo Ferbeyre, Department of Biochemistry, Université de Montréal, Montréal, Québec, Canada HC3 3J7. Tel.: (514) 343–6324; fax: (514) 343 2210; e-mail: luis.rokeach@umontreal.ca, p.chartrand@umontreal.ca and g.ferbeyre@umontreal.ca

Summary

Budding yeast shows a progressive decline in viability after entering stationary phase, a phenomenon known as chronological aging. We show here that the fission yeast Schizosaccharomyces pombe also undergoes chronological aging and that the process is regulated by genes controlling two related nutrient signalling pathways. The first pathway includes the serine/threonine cAMP-activated protein kinase Pka1 and the second pathway comprises the serine/threonine kinase Sck2, a homologue of Saccharomyces cerevisiae SCH9. A double mutant for pka1 and sck2 displayed an additive effect on prolonging the fission yeast lifespan, suggesting that these genes regulate related but independent pathways. These long-lived mutants also accumulated less reactive oxygen species and had a delayed initiation of apoptosis compared with wild-type cells. We also found that strains carrying pka1 deletion but not those with sck2 deletion gained resistance to oxidative stress due to exposure to H2O2 or menadione. On the other hand, the additional increase in lifespan shown by the Δpka1Δsck2 double-mutant strain correlated with an increased resistance to both oxidative stress and heat shock. These results underscore the importance of nutrient signalling pathways and reactive oxygen species on organismal lifespan and establish S. pombe as a new model organism to study the molecular mechanisms underlying aging.

Introduction

The viability of individual yeast cells decreases with the number of cell generations (replicative aging) or with the time they spend in stationary phase (chronological aging). Replicative aging is linked to exponential growth and measures the number of daughters a single mother cell is able to generate before its death. Replicative aging in yeast has been associated with the accumulation of rDNA circles in the nucleolus of old cells (Sinclair & Guarente, 1997). Overexpression of the NAD+-dependent deacetylases Sir2 and Hst2 extends the replicative potential of yeast cells and reduces the accumulation of rDNA circles (Kaeberlein et al., 1999; Lamming et al., 2005). In addition, caloric restriction, a condition that extends lifespan in a variety of species including flies, worms and mice also extends replicative lifespan (Lin et al., 2000). Chronological aging relates to the time a yeast population can survive in stationary phase (Fabrizio et al., 2001). Most chronological lifespan studies in yeast were performed in postdiauxic stage, which is characterized by a high degree of cell metabolism and respiration. It has been proposed that the loss of viability observed in this stage is mainly due to ethanol production (Fabrizio et al., 2005), toxicity induced by reactive oxygen species (ROS), and loss of mitochondrial function (Longo & Fabrizio, 2002). More importantly, it has been established that death in postdiauxic cultures is not the result of simple starvation. First, individual yeast cells die at a high rate when their glycogen content is still high (Longo et al., 1997). Second, yeast cells survive better in water than in synthetic complete medium (SCM), and switching from SCM to water diminishes cell death (Longo et al., 1997). Third, cell-cycle mutants that arrest growth in the presence of nutrients also lose viability (Motizuki et al., 1995). Finally some genetic mutations extend lifespan without changing metabolic rates (Fabrizio & Longo, 2003).

Although the processes that limit lifespan during replicative and chronological aging may differ, it is known that passage through the stationary phase accelerates replicative aging (Ashrafi et al., 1999). Moreover, disabling mutations in the gene coding for the protein kinase SCH9 increases both replicative and chronological lifespan in yeast, suggesting a common link between the two processes (Kaeberlein et al., 2005).

Previous studies have found that deletion of the genes SCH9, CYR1 and RAS2 enhances chronological lifespan of the budding yeast (Fabrizio et al., 2001; Fabrizio & Longo, 2003). The encoded proteins mediate glucose signalling; this pathway stimulates cell proliferation and glycolysis, but decreases stress resistance (Longo, 2003). CYR1 is an adenylate cyclase (Matsumoto et al., 1983) and RAS2 a small GTPase (Powers et al., 1984), and both belong to the cAMP/PKA signal transduction pathway. Increased chronological lifespan due to deletion of these genes requires the stress response transcription factors Msn2 and Msn4, and the ROS-detoxifying enzyme Sod2 (Fabrizio et al., 2003). SCH9 encodes a serine/threonine protein kinase and is a high-copy suppressor of the growth defect caused by deletion of the PKA signalling pathway (Toda et al., 1988), suggesting that it controls a redundant or parallel signalling pathway to PKA. Sch9 pro-aging functions seem to be conserved since its Caenorhabditis elegans homologue, AKT, also decreases longevity acting downstream of the insulin-like receptor DAF2 (Kenyon et al., 1993; Hekimi & Guarente, 2003).

Saccharomyces cerevisiae has been shown to be a successful model to study aging. However, aging in budding yeast is influenced by the formation of ribosomal DNA circles, a process controlled by the histone deacetylase SIR2. These studies have generated a considerable interest in the possibility that SIR2 also increase longevity in mammals. Intriguingly, the increase in chronological lifespan in budding yeast due to mutations in PKA pathway and in Sch9, or caloric restriction can be largely extended in a strain deleted for SIR2 (Fabrizio et al., 2005). Hence, SIR2 mediates paradoxical pro-aging and anti-aging effects. We reasoned that studying aging in another species of yeast might help in identifying universal aging pathways. To develop such a model we have chosen the fission yeast Schizosaccharomyces pombe. Several cellular processes like cell division, cell-cycle regulation and signalling pathways in S. pombe are closer to mammalian cells than the same pathways in S. cerevisiae (Sipiczki et al., 2004). This striking similarity to mammalian cells constitutes a compelling reason to develop S. pombe as a model system to study aging. Like the budding yeast, the fission yeast also offers a convenient genetic system that makes it relatively straightforward to isolate mutants for a given phenotype. Replicative aging of S. pombe has already been measured by Barker & Walmsley (1999), but no long-lived mutant of this sort has been isolated as yet.

In budding yeast, mutations extending longevity commonly affect the glucose-sensing signalling pathway. In S. pombe, this pathway regulates growth through the activation of the cAMP-dependent protein kinase A (Hoffman & Winston, 1991; Maeda et al., 1994). Most of the genes of this pathway have been dubbed git (glucose insensitive transcription). The git2+ gene, also called cyr1+, codes for the adenylate cyclase (Hoffman & Winston, 1991); git6+/pka1+ encodes the catalytic subunit of PKA (Byrne & Hoffman, 1993); git8+/gpa2+ encodes a Gα subunit (Isshiki et al., 1992; Nocero et al., 1994) and git5+ encodes a Gβ subunit (Landry et al., 2000). The G proteins Git8 and Git5 transduce glucose signalling from a GPCR named Git3+ (Welton & Hoffman, 2000). There are two S. pombe homologues of the budding yeast SCH9 gene: sck1+ and sck2+ (Fujita & Yamamoto, 1998). Overexpression of sck1+ or sck2+ suppresses the defects associated with git mutants in the cAMP/PKA pathway of S. pombe (Jin et al., 1995; Fujita & Yamamoto, 1998). The deletion of both sck1+ and sck2+ has no obvious phenotype, but results in less efficient spore germination (Fujita & Yamamoto, 1998).

To establish S. pombe as a model organism to study aging, we studied the lifespan in stationary phase of wild-type S. pombe and mutant strains for pka1, sck1 and sck2. As in the budding yeast, we found that the cAMP/PKA pathway controls chronological aging in fission yeast. Of the two homologues of SCH9, we found that only sck2+ regulates aging, and the pathway in which it belongs seems to be independent of the cAMP/PKA pathway. Long-lived S. pombe mutants maintained their viability for longer times and accumulated less ROS and caspase activity compared to wild-type cells. Surprisingly, of Δpka1 and Δsck2, only pka1 deletion mutants were more resistant to oxidative stress compared to the wild type, while only the double mutant Δpka1Δsck2 displayed a significant increased resistance to heat stress. Hence, by reporting the first long-lived mutants of S. pombe, we introduce here this organism as a model for the study of aging.

Results

Chronological aging of Schizosaccharomyces pombe

As an initial step, we wished to characterize the chronological aging of S. pombe by analysing the survival of a wild-type strain in stationary phase. As shown previously in S. cerevisiae, the content of the growth medium greatly affects chronological lifespan (Fabrizio & Longo, 2003). We chose to carry out the experiment in SCM containing 2% glucose because in budding yeast this condition is not associated with growth during the stationary phase (Fabrizio & Longo, 2003). Chronological aging in this medium is not simply starvation because growth of a wild-type strain in minimum medium (MM) resulted in approximately threefold longer survival than in SCM (data not shown). We measured cell survival from the early stationary phase, 10 h after the cultures ceased to grow, until the time when 99.9% of the cells were dead. Survival was scored by counting colony-forming units (CFU) from aliquots taken at different time points that were then diluted and plated on yeast extract completed medium (YEC; Moreno et al. 1991). We established that after 4 ± 1 days, 50% of the wild-type S. pombe population grown in SCM was able to give rise to colonies on plates (Fig. 1). After 15 ± 2 days, 99.9% of this population was replicatively dead. To verify that mutations in the genetic markers present in the wild-type reference strain (leu1–32, ura4-D18 and ade6-M210) would not influence survival, a prototrophic strain without these markers was tested in this assay. It was found that the latter had the same lifespan as the auxotrophic population (Supplementary Fig. S1A,B), suggesting that these mutations have no effect on longevity. Furthermore, we occasionally noticed regrowth during stationary phase when using MM (data not shown) as it was described in low caloric conditions with baker yeast, possibly due to a better adapted subpopulation (Fabrizio & Longo, 2003). The use of SCM prevents the cells from entering such a state of regrowth.

Figure 1.

Chronological lifespan of wild-type Schizosaccharomyces pombe. Survival of auxotrophic strain cultured in synthetic complete medium (SCM) 2% glucose as a carbon source and estimated by CFU (colony-forming units) counting. Y-axis is shown in logarithmic scale (see Supplementary Fig. S1B for nonlogarithmic scale).

While CFU counting provides information on the number of cells that can resume growth, such measurement does not reflect the viability of the cells in culture. To follow the viability of the yeast population, we used the fluorescent dye Phloxin B. This marker is passively taken up by cells and is actively excreted only by ‘metabolically active’, i.e. living cells (Supplementary Fig. S3). Thus, Phloxin B stains dead cells, which can be readily counted under the fluorescence microscope or by using flow cytometry (Fig. 2A,B). By using Phloxin B in wild-type yeast cultures, it was found that the proportion of metabolically dead cells was lower than the number of dead cells estimated from the CFU counting (Fig. 2C). Therefore, the lifespan measured using Phloxin B was considerably longer than the one measured by CFU (Table 1). For example, after 6 days in stationary phase, around 60% of the population was metabolically alive but only 2.3% of these cells were able to form colonies. Considering metabolic activity, 46 days were necessary to attain the threshold of 99.9% mortality, against 15 days for CFU counting (data not shown). These results show that CFU counting informs about the proportion of replicatively active cells in a population and underestimates the number of metabolically active cells, i.e. living cells estimated by Phloxin B staining.

Figure 2.

Differences in the estimation of cell survival followed by CFU (colony-forming units) counting or Phloxin B staining. (A) Cells from exponentially growing cultures, 6-day and 14-day-old cultures were stained with the fluorescent vital dye Phloxin B, and analysed by fluorescence microscopy. Nonstained cells under fluorescence microscope were considered as alive. (B) Percentage of wild-type living cells measured as the number of Phloxin B unstained cells (left peak) vs. stained cells (right peak) during stationary phase and quantified by flow cytometry. (C) Survival curves of an auxotrophic culture in 2% glucose measured by CFU counting or by vital staining with Phloxin B. The percentage of unstained cells was estimated by flow cytometry analysis.

Table 1.  Survival at day 10 estimated by CFU (colony-forming units) counting or phloxin B staining. Survival is given as the percentage of total living cells at day 0 (fold factor of survival of a given mutant compared to wild-type)
StrainCFU (%)Phloxin B (%)
WT 2.359
Δsck232(13.9)80(1.35)
Δpka131.5(13.7)89.5(1.5)
Δpka1Δsck267(29)94.7(1.6)

The cAMP/Pka1 pathway controls longevity in stationary phase

The cAMP/PKA1 pathway controls chronological lifespan in S. cerevisiae (Longo, 1999; Fabrizio et al., 2001). We advanced that mutations in genes regulating the cAMP/PKA pathway will increase longevity in S. pombe as well. The cAMP/Pka1 pathway of S. pombe is required for growth and exit from stationary phase (Jin et al., 1995; Fujita & Yamamoto, 1998). The unique PKA catalytic subunit known to date in S. pombe is coded by pka1+, and cells deleted for this gene are viable (Maeda et al., 1994). The Δpka1 cells were analysed for their ability to survive in stationary phase. The survival of Δpka1 estimated by CFU counting was found to be threefold longer than wild type (Fig. 3A). The Δpka1 mutant also stayed metabolically active for longer time compared to wild type. For example, after 6 days, 89% of the population of this mutant was still alive compared to 59% for wild-type culture (Fig. 3C and Table 1). Therefore, there seems to be a good correlation between the ability of yeast cells to stay metabolically active (Phloxin B positive) in stationary phase and their ability to resume cell proliferation after being replated on rich medium.

Figure 3.

Deletions of sck2+ and pka1+ increase survival in stationary phase. (A) Survival of WT, Δpka1, Δsck1, Δsck2 and Δsck1Δsck2 strains was evaluated by CFU (colony-forming units) counting. Y-axis is shown in logarithmic scale (see Supplementary Fig. S2A for nonlogarithmic scale). (B) Survival of WT, Δpka1,Δsck2 and Δpka1Δsck2 strains evaluated by CFU counting. Y-axis is shown in logarithmic scale (see Supplementary Fig. S2B for nonlogarithmic scale). (C) Percentage of dead cells measured by Phloxin B staining of WT, Δsck2,Δpka1 and Δpka1Δsck2 strains. Stained cells were considered as dead and were quantified by flow cytometry.

Deletion of sck2 increases longevity in stationary phase

To investigate in more detail chronological aging in S. pombe, we focused on the two homologues of the budding yeast SCH9 gene, sck1+ and sck2+. The Δsck1 and Δsck2 deletion mutants were grown until stationary phase and their chronological longevities were analysed. The sck1 knockout had the same survival as the wild-type reference strain SP14000 (Fig. 3A). However, the sck2 knockout exhibited a lifespan threefold longer than the wild-type strain while the double mutant Δsck1Δsck2 displayed the same survival as the Δsck2 strain (Fig. 3A), indicating that sck1+ has no pro-aging effect. Phloxin B staining of Δsck2 was followed during stationary phase and showed that dead cells appeared later in this mutant (Fig. 3C). Neither Δsck2 nor Δsck1Δsck2 exhibit a significant delay in their generation doubling time (Fujita & Yamamoto, 1998) as observed in long-lived SCH9 mutant strains of S. cerevisiae (Toda et al., 1988). Therefore, the increased longevity of sck2 mutants in S. pombe is not a consequence of slow growth. Altogether, these results suggest that sck2+ is the functional homologue of SCH9 since they both have a similar effect on the control of aging.

We found that both pka1+ and sck2+ control longevity. Next we wanted to investigate whether they act in the same pathway. In S. cerevisiae, the PKA1 and SCH9 kinases control separate but partially redundant signal transduction pathways (Roosen et al., 2005). Indeed, overexpression of one of these kinases compensates for the loss of function of the other (Toda et al., 1988). A recent study has shown that they have common target genes, which explains the redundancy of their function (Roosen et al., 2005). Interestingly, the double deletion Δpka1Δsck2 showed an increase in lifespan compared with the single mutants Δpka1 and Δsck2 (Fig. 3B). This result suggests that the deletion of both cAMP/Pka1 and Sck2 pathways has an additive effect on longevity.

Cell-cycle arrest of wild-type and mutant strains during stationary phase

The longer the cells spend time in stationary phase, the lower chance they have to reverse their cell-cycle arrest and resume cell division when the nutritive environment becomes favourable. We observed that deletion of sck2 or pka1 extends chronological lifespan enabling these mutants to prolong the period where re-entrance into the cell cycle is possible. Therefore, we investigated if these mutants may have a different cell-cycle arrest. Most S. pombe cells in logarithmic phase show a typical G2 (2 N) DNA content. Furthermore, it has been previously reported that in stationary phase, fission yeast stops its cell cycle either in G1 (N) or in G2 (2 N) and that the survival of cells arrested in G1 is the same as that of cells arrested in G2 (Costello et al., 1986). Using flow cytometry, we found that wild-type and mutants displayed no difference in their cell-cycle profile, neither in exponential phase nor in stationary phase (Fig. 4). Interestingly, after 6 days the Δpka1 mutant displayed an increased number of polyploid cells (4 N) similar to wild-type cells during log phase, whereas Δsck2 displayed a much lower peak. This probably represents binuclear cells that have completed S phase in each nucleus but have not finished cytokinesis. Together, our data suggest that long-term survival in stationary phase does not depend on this kind of cell-cycle arrest (G1 or G2 arrest) since long-lived mutants have cell-cycle profiles identical to wild-type.

Figure 4.

Wild-type, Δsck2,Δpka1 and Δpka1Δsck2 strains in log phase and stationary phase show no major differences in their cell-cycle arrest. Wild-type and mutant strains were fixed with ethanol, and stained with propidium iodide and quantified by flow cytometry. The N, 2 N, and 4 N peaks are indicated.

Lower accumulation of ROS and decreased apoptosis in Δpka1, Δsck2, Δpka1Δsck2 long-lived mutants

One of the main causes of cellular damage are the reactive oxygen species (ROS) produced by cellular metabolism. ROS accumulate proportionally with age and mutations. Therefore, increasing the antioxidant capacity of the cells lengthens the mean lifespan in several species (Finkel & Holbrook, 2000). In order to better understand the cause for the longer lifespan of pka1 and sck2 deletions mutants, the accumulation of ROS was investigated in these strains during stationary phase using dihydrorhodamine 123 (DHR123), a free-radical fluorescent probe (Fig. 5A). The results confirmed a low concentration of ROS in long-lived Δpka1 and Δsck2 mutants compared with wild-type cells (Fig. 5B), while the double deletion of pka1 and sck2 displayed an even lower accumulation of ROS during stationary phase.

Figure 5.

Strains Δpka1,Δsck2,Δpka1Δsck2 show slower accumulation of reactive oxygen species (ROS) during stationary phase as compared to wild type. (A) Wild-type and Δpka1Δsck2 strains were incubated with the ROS probe DHR123 and cells were observed under fluorescence microscopy. (B) Percentages of ROS accumulating cells are displayed for wild-type, Δsck2,Δpka1 and Δpka1Δsck2 strains. At least 400 cells per count were examined and the average of DHR123-positive cells from two cultures was calculated.

Recent studies have reported that unicellular organisms like the budding and fission yeasts can undergo an apoptosis-like phenomenon triggered in response to stresses such as low concentrations of acetic acid, hydrogen peroxide, high concentration of mating pheromone, nutrients and DNA-damage conditions (Burhans et al., 2003; Zhang, 2003). In higher eukaryotes, apoptosis is mediated by a family of proteases known as caspases and S. pombe possesses a similar enzyme, the metacaspase Pca1p (Low et al., 2005). Moreover, typical apoptotic markers such as activated caspase are observed in old yeast cells accumulating ROS (Madeo et al., 2002). Therefore, long-lived mutants that display lower intracellular levels of ROS are expected to enter later in apoptosis. We tested this hypothesis with the pka1 and sck2 mutants and measured the number of cells with induced caspase activity by using a fluorescent probe. With this assay, we found that the percentage of cells exhibiting caspase activation increased much more rapidly for the wild-type strain than for the long-lived mutants Δsck2, Δpka1 and Δpka1Δsck2 (Fig. 6A,B). Furthermore, the less these strains accumulated ROS (Fig. 5B), the less they triggered the apoptosis-like program. Curiously, the number of wild-type and Δsck2 cells displaying caspase activity decreased after 20 days. Since viability is low by this time, a caspase-independent mechanism of cell death may be acting late in stationary phase. Alternatively, cells well advanced in the execution of apoptosis may not display caspase activity anymore, and hence will be counted as a negative.

Figure 6.

Caspase induction during aging of Δsck2, Δpka1 and Δpka1Δsck2 strains is delayed compared to wild type. (A) Wild-type and Δpka1Δsck2 strains were stained with an apoptotic marker (FITC-VAD-fmk) during exponential and stationary phase, and observed by fluorescence microscopy. (B) Quantification of caspase-positive cells was done by flow cytometry as described in the Experimental procedures. The average measurements of two independent cultures is presented.

Longevity correlates with resistance to heat-shock and oxidative stress with Δpka1 mutant but not with Δsck2 mutant in S. pombe

Previous studies showed that most chronological long-lived mutants of S. cerevisiae show high oxidative stress or heat resistance (Fabrizio & Longo, 2003). In agreement, long-living pka1 and sck2 mutants of S. pombe, but not sck1 mutants, showed low levels of ROS. In order to find out whether pka1 and sck2 mutants control lifespan in S. pombe by increasing stress resistance, we performed several stress resistance assays. To carry out these tests, wild-type or mutant yeasts were grown to stationary phase and 3 days later the cells were submitted to hydrogen peroxide, menadione or heat shock. The survival of these stationary phase cells after treatment was measured by drop tests on solid rich medium. We noticed that wild-type S. pombe is resistant to high concentrations of hydrogen peroxide and menadione (Fig. 7). In fact, a 45-min treatment in 1 m H2O2 was necessary to kill a significant proportion of cells, whereas only 30 min in 100 mm H2O2 was sufficient to kill wild-type S. cerevisiae (Fabrizio et al., 2001). Likewise, a 0.75 m menadione treatment for 3 h had to be applied to our strains, while only 20 µm for 1 h of the same superoxide/H2O2 generating agent was enough to reduce S. cerevisiae viability (Fabrizio et al., 2001). For Δsck1, Δsck2, Δsck1Δsck2 and Δpka1Δsck2 mutant strains we did not score any difference in the sensitivity to H2O2 during stationary phase when compared to wild type (Fig. 7A). We did observe a moderate increase in resistance to H2O2 in the pka1 mutant strain as compared to the reference strain. This trend was clearly enhanced when we tested the sensitivity to another pro-oxidant, menadione. Figure 7A shows that Δpka1 and Δpka1Δsck2 mutant strains grew better after treatment in menadione compared with wild-type.

Figure 7.

Stress resistance of long-lived mutants in stationary phase. All strains were treated by short exposure to high concentration of oxidative stress: H2O2 (1 M) and menadione (0.75 M) (A) or heat shock (B). Dilutions were then spotted onto rich media to measure survival.

The relationship between resistance to heat shock and longevity seems to be more complex. After 3 days in stationary phase, Δpka1, Δsck2, Δsck1Δsck2 and Δpka1Δsck2 mutants displayed a moderate resistance compared to wild-type and the Δsck1 mutant strain (Fig. 7B). Surprisingly, the double mutant Δpka1Δsck2 showed a much higher resistance to heat shock in comparison to wild type.

Discussion

We have shown that S. pombe cells progressively lose their viability after entering the stationary phase, a phenomenon known as chronological aging in S. cerevisiae (Fabrizio & Longo, 2003). We found that chronological aging in S. pombe was accelerated by glucose signalling because cells bearing mutations in genes controlling this pathway such as pka1 and sck2 lived longer. Since glucose also accelerates aging in S. cerevisiae, our results indicate a conserved mechanism of lifespan reduction operating in yeast during the stationary phase.

Studying chronological aging in S. pombe is not simply an extension of similar efforts in S. cerevisiae. The physiology of S. pombe is considered closer to higher eukaryotes and our results started to reveal differences in chronological aging between fission and budding yeast. For example, although the cAMP/PKA pathway controls longevity in both species, no mutants of the PKA holoenzyme (composed of the subunits Tpk1, Tpk2 and Tpk3, and the regulatory subunit Bcy1p) have been identified to increase lifespan in budding yeast. We show here, for the first time, that a deletion of pka1, coding for the PKA catalytic subunit in S. pombe, increases longevity. Furthermore, although mutations in a single gene coding for the protein kinase SCH9 extends lifespan in the budding yeast, S. pombe possesses two SCH9 homologues, sck1+ and sck2+ (Fujita & Yamamoto, 1998). We discovered that of the two, sck2+ is the longevity regulator in S. pombe. Since Sck1p and Sck2p have similar structures, studying the differences in activators and effectors of these two proteins should shed considerable light on the pathways controlling aging in yeast.

Global expression studies have shown that sck1+ and sck2+ are expressed in S. pombe (Chen et al., 2003). Surprisingly, their expressions are regulated in opposite ways under different stress conditions such as oxidative stress, heat, osmolarity stress or cadmium treatment. In these conditions sck1+ is up-regulated whereas sck2+ is down-regulated (Chen et al., 2003). Taken together, this expression data and our results suggest, that under certain stress conditions, S. pombe has evolved pathways to reduce Sck2 expression but in stationary phase there is still sufficient Sck2 activity to reduce viability.

Since the cAMP/Pka1 pathway controls chronological lifespan in S. pombe, what could be the downstream targets of this pathway? One possible target is Rst2p, which is known to be controlled by Pka1p and inhibited by the presence of glucose (Higuchi et al., 2002). This transcription factor has been shown to bind the stress response element STREP that resembles the S. cerevisiae stress response element STRE (Marchler et al., 1993). The expression of genes with a STREP motif in their promoter region is enhanced after glucose starvation (Hoffman & Winston, 1991). Our study showed that mutants without active cAMP/Pka1 pathway accumulated less ROS, suggesting that enzymes involved in ROS generation and/or degradation may be under the control of the cAMP/Pka1 pathway. Since this mutant is more resistant to hydrogen peroxide and menadione, we currently favour the idea that they act by suppressing ROS accumulation. Yet, no antioxidant protein gene is known to have a STREP motif and to be regulated by Rst2/Pka1. In S. pombe antioxidant genes are under the control of the Atf1 transcription factor (Shiozaki & Russel, 1996; Wilkinson et al., 1996) and the Pap1 transcription factor (Toone et al., 1998). While a direct connection between Pka1p and Pap1p is not clear, a link between the cAMP/Pka1 and Wis1/Spc1/Aft1 pathways has already been suggested (Stettler et al., 1996). Moreover, nutritional limitation is known to activate the kinase Spc1, an element of the Wis1 MAP kinase cascade that is upstream of Atf1 (Shiozaki et al., 1997). Our results support the possibility of a crossed regulation between these two pathways.

While both pka1+ and sck2+ affected S. pombe longevity separately, the double deletion mutant Δpka1Δsck2 showed an additive effect promoting a better survival than the two single deletion mutants Δpka1 and Δsck2. This additive effect was not only visible at the level of cell viability, but also on ROS accumulation, caspase activation and surprisingly on resistance to heat shock. These results suggest that either both factors act on the same gene targets and have an additive effect, or they act on different genes with complementary function. Indeed, Roosen and collaborators (2005) have shown in S. cerevisiae that PKA and Sch9p act through separate signalling cascades, and affect synergistically or oppositely common target genes. A similar pattern can be envisaged in S. pombe since sck2+ is a high-copy suppressor of pka1 and both affect longevity, suggesting partially overlapping functions.

While the increased lifespan of Δpka1 and Δsck2 mutant strains correlated with low accumulation of ROS in stationary phase, strains carrying pka1 deletion but not those with sck2 deletion gained resistance to oxidative stress such as H2O2 and menadione (Fig. 7). On the other hand, the additional increase in lifespan shown by the Δpka1Δsck2 double mutant strain correlated with an increased resistance to both oxidative stress and heat shock. Although several models may explain these correlations, the results are consistent with the idea that Sck2 controls a pathway of ROS production but not resistance to ROS. In contrast, Pka1p may control the pathways of ROS resistance or elimination as described in S. cerevisiae (Fabrizio et al., 2003). Combined, the effects of these two deletions may increase lifespan by both decreasing ROS production and increasing ROS resistance. Of note, Perez-Campo and colleagues (1998) have reported that longevity among different species negatively correlates with the amount of ROS produced by the mitochondria. Our results are consistent with that possibility and suggest that studying chronological aging in S. pombe may help to identify genes controlling ROS production.

We also observed that a large proportion of metabolically active (Phloxin B positive) cells in stationary phase were unable to resume growth when plated on complete solid medium. It is possible that these cells may have accumulated too many oxidative damages on their DNA and proteins and will ultimately undergo apoptosis. Alternatively, these cells are already dead but cannot take up or retain Phloxin B. We exclude the latter possibility, because when we killed the cells using heat or sodium azide, all of them were stained by Phloxin B (Supplementary Fig. S3).

It has been proposed that yeast cells commit suicidal apoptosis in stationary phase to keep a small fraction of cells alive and able to resume growth when nutrients become available (Fabrizio et al., 2004; Herker et al., 2004). According to this interpretation, the pro-aging functions of Pka1 and SCH9/Sck2 were selected for the sake of the population. Alternatively, these proteins may be important for logarithmic growth while their effects on reducing fitness in stationary phase have not been neutralized by natural selection. The latter argument may be valid if yeasts living in stationary phase do not significantly contribute to subsequent generations. Although S. pombe and S. cerevisiae diverged several million years ago, pka1+ and SCH9/sck2+ have a pro-aging effect in both species. This conservation is consistent with the idea of a selective value for aging and death in stationary phase. It is intriguing that there are mechanisms in S. pombe to down-regulate sck2+ expression during certain stresses but these mechanisms do not seem to fully operate during the stationary phase. Again, this supports the idea that aging in stationary phase has a selectable value in yeast. Decoding the molecular components of the conserved aging pathway in S. pombe may shed some light on the aging process of higher eukaryotes, including humans. Proof that this is indeed the case is well illustrated by our results showing that mutations in pka1+ and sck2+ extend lifespan in S. pombe, and reduce the accumulation of ROS and the activation of caspase in these cells. Both ROS and caspase-mediated cell death have been implicated in the functional declines associated with aging in mammals (Balaban et al., 2005) and ROS production negatively correlates with species lifespan (Perez-Campo et al., 1998). Hence, we propose to add S. pombe to the list of model organisms for the study of aging.

Experimental procedures

Strains and media

Schizosaccharomyces pombe strains used in this study are listed in Table 2. All disruption strains were obtained by mating JX770 with a wild-type strain and different combinations of disruptions were isolated on selective media, except for sck1 disruption in SP14073, which is detailed below. Cultures were grown in synthetic complete medium (SCM) and MM (Edinburgh minimal medium; Moreno et al., 1991) with 2% glucose. SCM is supplemented with 111 mg L−1 of amino acids and a fourfold excess of the supplements that strains do not synthesize (Ade, Ura, Leu and His). Solid rich medium (YEC), was made of YE 5 g L−1 supplemented with 111 mg L−1 of adenine, uracil, leucine, histidine. Liquid cultures were grown at 30 °C with shaking at 250 rotations per minute.

Table 2.  Genotypes of strains used in this study
StrainGenotypeOrigin
  1. WT, wild type.

SP14078 (prototrophic WT)hC. Hoffman
SP14000 (auxotrophic WT)h ade6-M210 ura4-D18 leu1–32This study
SP14073h ade6-M210 leu1–32 ura4-D18 sck1::neoRThis study
SP14086h ade6-M210 leu1–32 ura4-D18 his7–366 sck2::LEU2This study
SP14094h ade6-M210 leu1–32 ura4-D18 his7–366 sck1::neoR sck2::LEU2This study
CHP453h leu1–32 ura4 -D18 his7–366 pka1::ura4+C. Hoffman
SP14089h ade6-M216 leu1–32 ura4-D18 his7–366 pka1::ura4+ sck2::LEU2This study
JX770h ade6-M216 leu1–32 ura4-D18 his7–366 pka1::ura4+ sck1::his7+ sck2::LEU2M. Yamamoto

Disruption of sck1+

The two flanking regions of sck1+ were cloned into the pDRIVE vector (QIAGEN, Mississauga, Canada) after their amplification by PCR from plasmid pBC1 (Jin et al., 1995) and using the pair of primers: 5′FLSCK1 for (5′-CGGGTACCCTCTGTTTCGATACTCC-3′), 5′FLS CK1rev (5′-GTAACACCAGACGAGGA-3′); and 3′FLSCK1-sense (5′-GCGCTCGAGTAAGCTTGCTCCAATC-3′), 3′FLSCK1-antisense (5′-TGCTCTAGATCATGCAAAC AACAGG-3′). Then pDRIVE-5′FLsck1 (bearing the 5′ flanking region) was digested with KpnI and SphI, and pDRIVE-3′FLsck1 (bearing the 3′ flanking region) with XhoI and XbaI. First, 5′FLsck1 was ligated to pDRIVE-NeoR linearised by KpnI and SphI digestion. Subsequently, 3′FLsck1 was ligated to pDRIVE-5′FLsck1-neoR cut with XhoI and XbaI to obtain pDRIVE-sck1::neoR. The deletion cassette was removed from pDRIVE-sck1:neoR by KpnI and XbaI double digestion, and used to transform SP14000 strain using a PEG/Li protocol (Elble, 1992). Transformants were selected on MM supplemented with adenine, uracil, leucine and G418. The knockout of sck1+ was validated by PCR and Southern blot analysis.

Survival in stationary phase

Yeast strains were streaked on MM and grown for 5 days at 30 °C. From these plates we inoculated a preculture on SMC and grew it until it reached OD595 3–4. Then, this preculture served to start a second preculture that was cultured until OD595 5–6. This latter preculture served to inoculate a 200 mL culture at initial OD595 0.2–0.4. This culture was grown until the end of exponential phase when OD595 stopped increasing and reached the maximum density (Table 3). Ten hours after this point, the culture was separated into three independent 50-mL cultures in 250-mL flasks (keeping a medium/volume ratio of 1 : 5 as described in Fabrizio & Longo, 2003). Next, a first aliquot was diluted and plated on complete yeast extract solid medium. After 10 days, the total number of colonies was counted, with this number representing 100% survival and day 0 of the curve (see Supplementary Tables S1 and S2). The following measurements were done every 3–5 days. The two mutant strains, CHP453 and SP14089, have a slightly slower growth rate than the wild type as described by Fujita & Yamamoto (1998). To compensate for these differences, we adjusted the initial quantity of cells used to start cultures with these strains.

Table 3.  Maximum optical density (OD) reached by different strains in this study. The measurement has been done at the first point of survival curve representing 100% in the conditions described in the Experimental procedures
StrainNameMaximum OD*
  • *

    (n = 3). WT, wild type.

SP14000WT10.0 ± 0.3
SP14073Δsck1 9.6 ± 0.7
SP14086Δsck210.7 ± 0.5
SP14094Δsck1Δsck2 9.9 ± 0.3
CHP453Δpka110.6 ± 0.6
SP14089Δpka1Δsck211.7 ± 0.6

Cell survival was also followed using the Phloxin B fluorescent dye. Around 1.4 × 107 cells were collected and 5 mg L−1 of Phloxin B were added to the medium. After 2 h incubation at 30 °C with shaking, the cells were washed twice in 1× PBS pH 7.4 and used for fluorescence microscopy or flow cytometry analyses. Controls with exponentially growing cells, and cells killed by heat (65 °C for 1 h) or NaN3 (0.1% for 2 h), showed that Phloxin B stains only dead cells (Supplementary Fig. S3). For all microscopic analysis, we used a fluorescence inverted microscope Nikon TE2000U (Nikon, Mississauga, Canada). Images were acquired using a motion-picture camera CCD CoolSnapFX 12 (Photometrics, Tucson, AZ, USA) bit and treated with UIC Metamorph software (Molecular Devices Corporation, Downington, PA, USA).

In vivo staining of caspase activity

Cultures were harvested in exponential phase or stationary phase. Samples of 107 cells were washed once in 1 mL 1× PBS pH 7.4 and resuspended in 150 µL of 1× PBS containing 10 µm FITC-VAD-fmk (CaspACE, Promega, Madison, WI, USA). After incubation for 20 min at 30 °C, cells were washed in 1 mL 1× PBS, resuspended in 100 µL 1× PBS and analysed by FACS or under the fluorescence microscope. The percentage of apoptotic cells was determined using flow cytometry.

Flow cytometry analysis

Cells were stained with Phloxin B or FITC-VAD-fmk as described above. Flow cytometry analysis was performed using FACS Calibur (Becton Dickinson Biosciences, Mississauga, Canada) on 10 000 cells. Emission from the argon LASER was at 488 nm; emission settings were 515–545 nm (filter FL1) or 560–600 nm (filter FL2), respectively, for FITC-VAD-fmk and Phloxin B staining. The percentage of positive stained cells was determined as the population of fluorescent cells with a higher fluorescent intensity than an unstained negative control. Two independent cultures were analysed.

Cultures were treated with propidium iodide in order to stain DNA and determine the quantity of DNA in cells. Cells were fixed and stained following the protocol of Sazer & Sherwood (1990), with 2 h incubation at 37 °C with RNAse A. Samples containing 10 000 cells were analysed by flow cytometry using FL2-A detection line.

In vivo staining of ROS accumulation

Samples of 100 µL (1–2 × 107 cells) of culture were collected and incubated with 30 µm dihydrorhodamine123 for 30 min at 30 °C. Cells were applied on a poly lysine pretreated slide, washed by 1× PBS pH 7.4 and observed under fluorescence microscopy. Background fluorescence was manually removed using the UIC Metamorph software and at least 400 cells of two independent cultures were counted to determine the percentage of positively stained cells.

Stress resistance in stationary phase

Cells were cultured for 3 days in stationary phase as described above. After washing with water, the cell concentration was adjusted to OD595 of 0.5. Next, we added H2O2 (1 M final concentration) and incubated for 1 h at 30 °C, or menadione (0.75 M final concentration) and incubated for 3 h at 30 °C. Cells were pelleted and washed two times with 1 mL of water. Serial dilutions were spotted onto YE rich medium and incubated for 4 days at 30 °C. For heat treatment, equal amounts of cells (OD595 of 0.5) were heat shocked for 45 min at 51 °C. Serial dilutions were dropped onto YE rich medium and incubated for 5 days at 30 °C.

Acknowledgments

We thank Drs Charlie Hoffman and Masayuki Yamamoto for providing plasmids and strains. We also wish to thank Dr Charlie Hoffman and the members of the Rokeach lab for helpful discussions. We thank Serge Sénéchal for assistance with the FACS analyses. This work was funded by the Canadian Institutes for Health Research (CIHR). G. F. is supported by a fellowship from CIHR, and P. C. by a fellowship from the Fond de Recherche sur la Nature et les Technologies du Québec (NATEQ).

Ancillary