Yeast replicative aging: a paradigm for defining conserved longevity interventions

Abstract

The finite replicative life span of budding yeast mother cells was demonstrated as early as 1959, but the idea that budding yeast could be used to model aging of multicellular eukaryotes did not enter the scientific mainstream until relatively recently. Despite continued skepticism by some, there are now abundant data that several interventions capable of extending yeast replicative life span have a similar effect in multicellular eukaryotes including nematode worms, fruit flies, and rodents. In particular, dietary restriction, mTOR signaling, and sirtuins are among the most studied longevity interventions in the field. Here, we describe key conserved longevity pathways in yeast and discuss relationships that may help explain how such broad conservation of aging processes could have evolved.

INTRODUCTION

The replicative life span (RLS) of a mother cell, which is defined as the number of daughters produced prior to irreversible cell cycle arrest (Mortimer & Johnston, 1959), is a characteristic phenotype of budding yeast (Saccharomyces cerevisiae) that arises due to the asymmetric nature of cell division (Steinkraus et al., 2008). When a yeast mother cell divides, the mother retains more volume than the daughter cell (Hartwell & Unger, 1977). In addition to cell size, a variety of factors are also differentially partitioned between the mother and daughter cells, apparently through both active and passive mechanisms (Henderson & Gottschling, 2008; McMurray & Thorner, 2009). It has been proposed that this differential partitioning has been selected for during evolution to retain damaging material in the mother cell while providing the highest level of fitness for the daughter cells (Erjavec et al., 2008; Kaeberlein, 2010). As a consequence of this asymmetry, the mother cell ‘ages’, while the daughter cells retain a youthful phenotype, including full RLS, even when born from old mother cells (Egilmez & Jazwinski, 1989; Kennedy et al., 1994).

Over the past two decades, much progress has been made at understanding the nature of the ‘senescence factors’ that accumulate with replicative age and limit mother cell life span. The first asymmetrically inherited form of molecular damage proposed to cause aging in yeast was the accumulation of extrachromosomal ribosomal DNA circles (ERCs), which are formed from homologous recombination between adjacent repeats within the ribosomal DNA (rDNA; Sinclair & Guarente, 1997). Since then, several additional types of molecular damage have been shown to also undergo asymmetric inheritance and are proposed to contribute to replicative senescence, including damaged nuclear pore complexes (Kaeberlein, 2008; Shcheprova et al., 2008), dysfunctional mitochondria (Lai et al., 2002; McFaline-Figueroa et al., 2011; Delaney et al., 2013b), cytosolic protein aggregates and oxidatively damaged proteins (Aguilaniu et al., 2003; Erjavec & Nystrom, 2007; Erjavec et al., 2007), and vacuolar dysfunction (Hughes & Gottschling, 2012). The extent to which these asymmetrically inherited aging factors from yeast contribute to aging in multicellular organisms remains an active area of investigation.

LONGEVITY DETERMINANTS CONSERVED BETWEEN YEAST AND MULTICELLULAR EUKARYOTES

Dietary restriction

The most studied intervention capable of slowing aging in multicellular eukaryotes is dietary restriction (DR), defined as a reduction in nutrient availability in the absence of malnutrition (Masoro, 2005). First shown to extend life span in rats in the 1930s (McCay et al., 1935), DR has since been found to promote longevity and healthy aging in a variety of other species, from yeast to rhesus monkeys (Kennedy et al., 2007; Omodei & Fontana, 2011). Recent studies have indicated that genetic background can have a large influence on the response to DR in both yeast and mice (Liao et al., 2010; Schleit et al., 2012, 2013), and the mechanisms underlying these effects, as well as the degree to which the downstream mechanisms of DR are conserved, are an active area of investigation. Inhibition of the mechanistic target of rapamycin (mTOR) by DR appears to be a shared mechanism for slowing aging in mammals, invertebrates, and yeast; however, the mTOR-dependent and mTOR-independent effects of DR can differ in a species- and tissue-specific manner (Stanfel et al., 2009; Kapahi et al., 2010). Among the effects of DR that are apparently conserved across species and have been associated with life span extension are enhanced mitochondrial function, increased autophagy, reduced mRNA translation, and increased resistance to different forms of stress (Kaeberlein, 2013).

In the yeast replicative aging paradigm, DR is typically accomplished by reducing the glucose concentration of the medium from 2% to 0.5% or lower (Lin et al., 2000). Growing cells on the nonfermentable carbon source glycerol has also been described as an alternative method of DR (Kirchman & Botta, 2007; Delaney et al., 2013a). In addition to changing the amount of glucose or the carbon source altogether, restriction of amino acids has also been reported to extend RLS (Jiang et al., 2000). This method has not been studied in detail, however. Several genetic models of DR have also been described in which mutations in nutrient-responsive signaling pathways lead to reduced activity of nutrient-responsive kinases including protein kinase A (PKA), mTOR, and Sch9 (Kennedy et al., 2007; Longo et al., 2012).

In yeast, the reduction in glucose from DR is associated with several physiological changes, including a metabolic shift away from fermentation toward mitochondrial respiration, as indicated by the induction of many genes involved in mitochondrial respiration, including several electron transport chain components (Lin et al., 2002; Barros et al., 2004). Although mammalian cells do not generally utilize fermentation under aerobic conditions, the induction of mitochondrial respiration in response to DR does appear to be conserved, at least for certain tissues (Hempenstall et al., 2012). Overexpression of HAP4, a transcription factor that activates respiratory gene expression, can also increase yeast RLS (Lin et al., 2002). In respiratory-deficient yeast that have lost their mitochondrial DNA (mtDNA), some strain backgrounds have an increased RLS (Kirchman et al., 1999; Woo & Poyton, 2009), while in other strains there is a decrease or no effect compared with wild type (Kirchman et al., 1999; Kaeberlein et al., 2005b). Importantly, DR can still increase RLS, even in respiratory-deficient yeast lacking mitochondrial DNA (Kaeberlein et al., 2005b). These data suggest that increasing respiration may be sufficient to increase RLS, but is not necessary for the prolongevity effects of DR in yeast. Whether the life span extension associated with increased respiration is mechanistically similar to life span extension from DR is currently unclear.

Reactive oxygen species (ROS) generated from mitochondrial respiration have been proposed as the primary driver of aging in humans and other species, a theory known as the free radical theory of aging (Harman, 1956, 2006). Superoxide levels have been measured to increase with replicative age in yeast, suggesting that free radicals could contribute to replicative aging (Lam et al., 2011). During exponential growth, DR increases mitochondrial respiration and elevates the production of mitochondrial hydrogen peroxide and superoxide levels, while antioxidant capacity also increases in a corresponding manner (Weinberger et al., 2010; Sharma et al., 2011). DR has also been reported to decrease release of hydrogen peroxide (Barros et al., 2004). In terms of sensitivity to exogenous oxidative stress, DR has been reported to not enhance resistance or perhaps even slightly sensitize cells to hydrogen peroxide or paraquat in one study (Lin et al., 2002), but other studies have found that DR increases resistance to hydrogen peroxide (Molin et al., 2011).

Along with enhanced mitochondrial respiration, inhibition of mTOR by DR results in up-regulation of autophagy (macroautophagy) in yeast as well in multicellular eukaryotes (Johnson et al., 2013). Autophagy is a cellular recycling process that can remove aged or damaged cellular components, and evidence indicates that autophagic function declines with age in several organisms (Cuervo, 2008; Rajawat et al., 2009), although it is unclear whether this is also true in yeast. Many yeast autophagy-related genes are dispensable for RLS extension by DR (Tang et al., 2008). This differs from the case in Caenorhabditis elegans, where RNAi knockdown of autophagy genes can prevent life span increases from both DR and mTOR inhibition (Hansen et al., 2008). Autophagy is required for DR-induced extension of chronological life span (stationary- phase survival) in yeast (Alvers et al., 2009a, b), perhaps suggesting that yeast chronological aging may more closely resemble aging in C. elegans in this context. This interpretation is complicated, however, by the fact that complete loss of autophagy is detrimental in both systems (lethal in C. elegans), and activation of autophagy independent of mTOR inhibition has not yet been shown to be sufficient to extend life span in either system. Thus, a definitive answer to the question of the important of autophagy in aging has yet to be determined.

Protein kinase A

The first nutrient response pathway to be implicated in life span extension from DR in yeast was the cAMP-dependent protein kinase A (PKA). The cyclic AMP (cAMP) pathway is stimulated by the presence of glucose. After glucose addition, the G-protein-coupled receptor Gpr1 along with the Ga subunit Gpa2 activates adenylate cyclase (Cyr1), which catalyzes the conversion of AMP to cAMP (Kraakman et al., 1999; Rolland et al., 2000). cAMP activates protein kinase A (PKA), which consists of three partially redundant subunits, Tpk1, Tpk2, and Tpk3 (Toda et al., 1987). PKA plays an important role in the regulation of cell growth, metabolism, and resistance to stress. Deletion of GRP1, GPA2, or HXK2 (hexokinase that phosphorylates glucose) can extend the RLS of yeast (Lin et al., 2000). Yeast contains two small Ras GTPase proteins, Ras1 and Ras2, which are capable of stimulating cAMP production. The guanine nucleotide exchange protein Cdc25 exchanges GDP for GTP to activate Ras proteins. Mutation of Cdc25 can also extend life span, and this effect is independent of the Msn2/4 stress-responsive transcription factors, which the PKA pathway negatively regulates (Lin et al., 2000). Deletion of RAS1 or overexpression of RAS2 can also extend replicative life span (Sun et al., 1994). Exogenous cAMP addition, PKA activation by a disruption of the negative regulator Bcy1, and overexpression of CYR1 all decrease RLS (Sun et al., 1994). Given the role of Gpr1 and Gpa2 in glucose sensing, it was hypothesized that mutations in these genes and others that reduce PKA signaling may genetically mimic DR (Lin et al., 2000). This model is consistent with subsequent observations that deletion of GPA2 or GPR1 fails to extend life span additively with DR and further extends the life span of cells lacking both SIR2 and FOB1 (discussed further below; Kaeberlein et al., 2004). Taken together, these results demonstrate that the cAMP/ PKA pathway is important for the regulation of replicative life span and support the idea that reduced signaling through this pathway modulates the response to DR.

mTOR AND SCH9

Budding yeast has played a central role in the discovery of mTOR as well as a role as a key modulator of aging. Studies involving mTOR signaling began with the use of rapamycin, a metabolite produced by the bacterium Streptomyces hygroscopicus, which can inhibit cell cycle in both yeast and mammalian systems (Vezina et al., 1975; Heitman et al., 1991). Selection for rapamycin resistance in S. cerevisiae identified mutations in the then novel TOR1 and TOR2 genes (Heitman et al., 1991; Cafferkey et al., 1993; Helliwell et al., 1994). Yeast Tor proteins are PIK-related serine/threonine protein kinases that are c. 80% similar at the amino acid level and regulate cell growth and metabolism in response to nutrient availability (Cafferkey et al., 1994; Keith & Schreiber, 1995). Tor1 functions within the multisubunit mTOR complex 1 (TORC1), while Tor2 functions within both TORC1 and mTOR complex 2 (TORC2). In higher eukaryotes, there is only one mTOR protein, which functions within both of the conserved mTOR complexes. Both complexes have a variety of functions: TORC2 is involved in regulating actin polarization, lipid metabolism, and (in yeast) cell wall integrity; TORC1 modulates mRNA translation, carbon and amino acid metabolism, autophagy, and stress responses. Sch9 is the yeast ortholog of S6 kinase, a substrate of mTOR that has diverse functions in multicellular eukaryotes, including the regulation of global mRNA translation, body size, and insulin-like signaling (Urban et al., 2007; Stanfel et al., 2009). In addition to extending life span, deletion of SCH9 in yeast results in substantial defects in ribosome biogenesis, mRNA translation, doubling time, and cell size (Toda et al., 1988; Jorgensen et al., 2002, 2004).

The role of Tor1 in yeast replicative aging was first identified from an unbiased screen of 564 haploid gene deletion strains, where loss of TOR1 as well as other TOR pathway-related genes (SCH9, URE2, ROM2, YBR238C, RPL31A, RPL6B) was found to increase RLS (Kaeberlein et al., 2005a). Similar to mutations in the PKA pathway, DR did not further extend life span in tor1Δ or sch9Δ mutants, and both of these mutants extended life span independently of Sir2 (Kaeberlein et al., 2005c). Genetic inhibition of mTOR signaling also extends life span in worms (Vellai et al., 2003; Jia et al., 2004), flies (Kapahi et al., 2004), and mice (Selman et al., 2009; Lamming et al., 2012), and in each of these organisms, evidence places the mTOR pathway downstream of DR. In addition, pharmacological inhibition of TOR signaling by rapamycin has been shown to increase life span in yeast (Powers et al., 2006; Medvedik et al., 2007), nematodes (Robida-Stubbs et al., 2012), fruit flies (Bjedov et al., 2010), and mice (Harrison et al., 2009; Anisimov et al., 2011; Miller et al., 2011). Recently, studies have correlated mTOR signaling and expression of mTOR pathway genes with longevity in people (Passtoors et al., 2013). Collectively, these data suggest that mTOR functions as a highly conserved modulator of growth and longevity in organisms from yeast to humans.

Much of the effort on understanding how reduced mTOR signaling increases RLS has focused on its regulation of mRNA translation. This is, in part, because multiple ribosomal protein and translation initiation factor gene deletions have been found to increase RLS (Chiocchetti et al., 2007; Steffen et al., 2008, 2012). Of these, at least six have orthologs in C. elegans that similarly modulate worm life span (Table 1), in addition to Tor1 and Sch9 themselves. Tor1 has also been shown to bind directly to the 35S rRNA gene promoter and stimulate Pol I-mediated synthesis of 35S rRNA gene (Li et al., 2006), but whether this function is important for aging remains to be determined. As discussed above, although regulation of autophagy by mTOR is clearly important for its effects on longevity in other organisms (Johnson et al., 2013), autophagy appears to be dispensable for extension of RLS in yeast.

Table 1.

Conserved longevity modifiers

Yeast gene	Worms	Flies	Mice	References
TOR1	let-363	dTOR	mTOR	Vellai et al. (2003), Kapahi et al. (2004), Kaeberlein et al. (2005a, c), Powers et al. (2006) and Lamming et al. (2012)
SCH9	rsks-1	dS6K	S6K1	Fabrizio et al. (2004), Kapahi et al. (2004), Hansen et al. (2007), Pan et al. (2007) and Selman et al. (2009)
SIR2	sir-2.1*	dSir2*		Kaeberlein et al. (1999), Tissenbaum & Guarente (2001), Rogina & Helfand (2004) and Burnett et al. (2011)
RPD3		Rpd3		Kim et al. (1999), Jiang et al. (2002) and Rogina et al. (2002)
DBP3	B0511.6			Curran & Ruvkun (2007) and Smith et al. (2008)
PMR1	eat-6			Lakowski & Hekimi (1998) and Smith et al. (2008)
YGR130C	erm-1			Curran & Ruvkun (2007) and Smith et al. (2008)
IDH1, IDH2	F43G9.1			Hamilton et al. (2005) and Smith et al. (2008)
TIF4631	ifg-1			Curran & Ruvkun (2007), Pan et al. (2007) and Smith et al. (2008)
TIF1, TIF2	inf-1			Curran & Ruvkun (2007) and Smith et al. (2008)
PKH2	pdk-1			Paradis et al. (1999) and Smith et al. (2008)
TIS11	pos-1			Curran & Ruvkun (2007) and Smith et al. (2008)
YPT6	rab-10			Hansen et al. (2005) and Smith et al. (2008)
RPL19A	rpl-19			Hansen et al. (2007) and Smith et al. (2008)
RPL6B	rpl-6			Hansen et al. (2007) and Smith et al. (2008)
RPL9A	rpl-9			Hansen et al. (2007) and Smith et al. (2008)
SAM1	sams-3			Curran & Ruvkun (2007) and Smith et al. (2008)
HSE1	sem-5			Curran & Ruvkun (2007) and Smith et al. (2008)
AFG3	spg-7			Curran & Ruvkun (2007), Smith et al. (2008) and Delaney et al. (2013a)
SPT4	spt-4			Hamilton et al. (2005) and Smith et al. (2008)
ALG12	T27F7.3			Curran & Ruvkun (2007) and Smith et al. (2008)
INP51, INP53	unc-26			Lakowski & Hekimi (1998) and Smith et al. (2008)
ADH1	W09H1.5			Hamilton et al. (2005) and Smith et al. (2008)
SIS2	Y46H3C.6			Hamilton et al. (2005) and Smith et al. (2008)
COX4	cco-1			Dillin et al. (2002) and Miceli et al. (2011)
Intervention	Worms	Flies	Rodents	References
Dietary restriction	X	X	X	McCay et al. (1935), Klass (1977), Lakowski & Hekimi (1998), Lin et al. (2000), Mair et al. (2003) and Kaeberlein et al. (2006)
Rapamycin	X	X	X	Medvedik et al. (2007), Harrison et al. (2009), Bjedov et al. (2010), Miller et al. (2011), Robida-Stubbs et al. (2012) and Wilkinson et al. (2012)
Resveratrol*	X*	X*		Howitz et al. (2003), Wood et al. (2004), Kaeberlein et al. (2005d) and Bass et al. (2007)
Spermidine	X	X		Eisenberg et al. (2009)
Heat shock	X	X		Lithgow et al. (1995), Khazaeli et al. (1997) and Shama et al. (1998)

^*Refers to data that have been questioned in subsequent peer-reviewed papers.

Studies of the mechanism by which reduced mRNA translation can extend RLS have focused on regulation of the transcription factor Gcn4. Gcn4 is a transcriptional activating factor involved in general amino acid control (GAAC) and has hundreds of target genes, including enzymes required for amino acid biosynthesis and transport (Hinnebusch, 2005). Gcn4 expression is normally kept low due to the presence of inhibitory upstream open reading frames (uORFs) in the 53 untranslated region (UTR) of the GCN4 mRNA; however, under certain conditions where translation is reduced, Gcn4 expression is increased resulting in an activation of Gcn4 target genes. Most notably for RLS, this occurs in yeast when there is a deficiency of ribosomal large (60S) subunits, but not of small (40S) subunits (Steffen et al., 2008). Extension of RLS in several long-lived large subunit ribosomal protein deletion mutants is fully dependent on Gcn4, consistent with the model that activation of Gcn4 plays a causal role in the enhanced longevity. Unlike the ribosomal protein deletion mutants, however, Gcn4 is only partially required for RLS extension from DR, deletion of TOR1, or deletion of SCH9. Thus, additional longevity-promoting factors must also contribute in these cases.

While inhibition of mTOR signaling decreases cytoplasmic translation, mitochondrial translation of mitochondrially encoded subunits of the electron transport chain increases, and there is a corresponding increase in mitochondria respiration (Bonawitz et al., 2007). This mTOR-mediated regulation of respiration is important for chronological life span extension (Bonawitz et al., 2007), but it remains unknown whether it also contributes to RLS extension. Reduced translation from mTOR inhibition has been previously shown to suppress the replicative age-associated mitochondrial degeneration caused by mutation of the adenine nucleotide translocase Aac2p or deletion of mitochondrial prohibitins (Wang et al., 2008), and one recent study proposes that an imbalance between cytoplasmic and nuclear translation can slow aging in nematodes and mice (Houtkooper et al., 2013).

Sirtuins

Sirtuins (sir-two-ins) are a conserved family of NAD-dependent protein deacetylases named after their founding member, the yeast histone deacetylase Sir2 (Finkel et al., 2009). The first evidence that sirtuins modulate age-related phenotypes came from studies of yeast replicative aging, where it was found that deletion of Sir2 shortens RLS, while overexpression of Sir2 increases RLS (Kennedy et al., 1997; Kaeberlein et al., 1999). Subsequent studies reported life span extension from overexpression of Sir2 orthologs in both worms (Tissenbaum & Guarente, 2001) and flies (Rogina & Helfand, 2004), establishing Sir2 activation as the first genetic intervention capable of promoting longevity in each of these three common nonmammalian model systems.

Since these initial reports, the role of sirtuins in aging has become a complicated issue. Many studies have established important functions for the mammalian Sir2 ortholog, SirT1, in a variety of age-related disease models (Donmez & Guarente, 2010; Guarente, 2011), but, thus far, overexpression of SirT1 has not yielded life span extension in a mammal. Recently, another sirtuin lacking a direct yeast ortholog, SirT6, was reported to extend life span when overexpressed in male, but not in female mice (Kanfi et al., 2012). At the same time as evidence for a role for sirtuins in mammalian aging has increased, however, the case in invertebrate models has become questionable (Delaney et al., 2011b). Multiple laboratories have reported an inability to reproduce the initial reports of life span extension in both worms and flies (Burnett et al., 2011), although other reports have suggested modest effects on life span from sirtuins overexpression in these organisms (Rizki et al., 2011; Viswanathan & Guarente, 2011). Importantly, the initial result that overexpression of Sir2 results in robust extension of RLS has been confirmed in additional strain backgrounds and extended through genetic interaction studies (Kaeberlein et al., 2004, Stumpferl et al., 2012).

The mechanistic basis for life span extension from Sir2 overexpression is thought to be mediated primarily through its role in promoting stability within the ribosomal DNA (rDNA) locus. The rDNA is prone to recombination between adjacent repeats, which can lead to excision of self-replicating extrachromosomal rDNA circles (ERCs). ERCs accumulate in the mother cell nucleus with age and are proposed to represent one limiting factor for RLS (Sinclair & Guarente, 1997). Deletion of SIR2 results in increased recombination within the rDNA and increased levels of ERCs (Kaeberlein et al., 1999). The role of ERCs as a direct cause of aging remains uncertain, however, and additional studies have suggested that Sir2 also modulates rDNA stability and aging independently of ERCs through the regulation of transcription from a noncoding bidirectional promoter within the rDNA spacer (Kobayashi & Ganley, 2005). There is also some evidence that additional functions of Sir2, such as regulating acetylation of histones near telomeres, modulation of protein aggregation, and promoting asymmetric inheritance of oxidatively damaged proteins, could contribute to its longevity-promoting effects in yeast (Aguilaniu et al., 2003; Erjavec & Nystrom, 2007; Dang et al., 2009; Knorre et al., 2010; Cohen et al., 2012; Sampaio-Marques et al., 2012; Salvi et al., 2013).

AMP Kinase

AMP-activated protein kinase (AMPK, yeast homolog Snf1) is a serine/threonine kinase that is activated under low energy conditions (e.g. a high AMP to ATP ratio) and controls multiple biochemical pathways. Activation of AMPK has been shown to extend life span in C. elegans (Apfeld et al., 2004) and has been proposed to modulate age-related processes through a variety of effectors, including mTOR and sirtuins, in mammals (Salminen & Kaarniranta, 2012). In yeast, loss of Sip2, a negative regulator of Snf1, results in increased Snf1 activity, decreased replicative life span, and signs of accelerated replicative aging (Ashrafi et al., 2000; Lin et al., 2003). Acetylation of Sip2 facilitates impairment of Snf1 activity, and Sip2 is increasingly deacylated with replicative age, leading to increased Snf1 activation (Lu et al., 2011). Mutations in Sip2 preventing acetylation result in a shortened RLS, while Sip2 acetylation mimetic mutants are long-lived. Similarly, deletion of a Sip2 acetyl transferase, NuA4, decreases RLS, and deletion of Rpd3, a Sip2 deacetylase, increases RLS (Kim et al., 1999). DR extended the shortened life span of sip2 null and Sip2 nonacetylatable mutants, but did not further extend a Sip2 acetylation mimetic mutant. This suggests that Sip2 is not necessary for DR-mediated life span extension, but both of these factors may function within the same longevity pathway. Similar to DR, loss of SCH9 could extend the shortened life span of sip2 and Sip2 nonacetylatable mutants, and Sch9 was found to be a substrate of Snf1 (Lu et al., 2011).

CONSERVATION OF AGING: EVOLUTIONARY CONSIDERATIONS

Although there is no longer any question about the existence of both genetic and environmental determinants that similarly modulate yeast replicative aging and aging in multicellular eukaryotes (Table 1), the question of why such determinants exist is an interesting and important one. Indeed, arguments have been put forward that are critical of standard approaches to using yeast as a model for aging in higher eukaryotes (Gershon & Gershon, 2000; Bilinski et al., 2012). On the surface, replicative aging of yeast mother cells would seem to be unrelated to chronological aging in an organism such as C. elegans, which, with the exception of the germ line, exists as a postmitotic adult organism. Even further removed would be mammals, which exist as a combination of mitotic and postmitotic cells and tissues with complex organ systems and systemic signals that yeast lack altogether.

Why then is maximal cell replicative capacity in yeast mechanistically similar to, or at least regulated by, the same factors as, adult longevity in multicellular animals? One potential explanation is the link between reproduction and aging. In multicellular eukaryotes, most (but not all) long-lived mutants also have reduced fecundity. This has been proposed to result from a trade-off between resources spent toward reproduction vs. resources spent toward maintenance of the soma, referred to as the disposable soma theory of aging (Kirkwood, 1977). Although counterexamples to this theory have now been established where reproduction can be uncoupled from longevity (Partridge et al., 2005; Flatt, 2011), it is generally the case that long-lived mutants show enhanced resistance to different forms of stress (Zhou et al., 2011). This relationship may suggest that pathways that modulate longevity are actually selected for based on their roles modulating growth vs. stress resistance. When conditions are favorable, then growth and reproduction are favored, but when conditions are poor (suboptimal nutrition, temperature, oxygen, osmolality, etc.), increased stress resistance and reduced fecundity are favored. In this model, life span extension is a coincidental byproduct of enhanced resistance to a temporarily unfavorable environment.

To consider whether this relationship applies to the yeast replicative aging paradigm, it is important to define what is meant by fecundity in this system. There is a common misconception that RLS and fecundity are equivalent: that is, both refer to the number of daughter cells produced by the mother cell (the definition of RLS). This incorrect perception may arise, in part, due to the use of the terms ‘mother’ and ‘daughter’ when referring to the asymmetric cell division. With respect to the force of natural selection, however, RLS is not a measure of fecundity. Instead, growth rate (doubling time) is the appropriate measure of fecundity in this system. This can be easily understood when one considers the relative selective advantages associated with increased RLS vs. increased growth rate (Fig. 1). For a typical wild-type cell that has an RLS of 25 generations, the frequency of a cell achieving that age in a logarithmically growing population is roughly 1 in 33.6 million (2²⁵) cells. The selective advantage associated with each one-generation increase in RLS is negligible at this point, because only a vanishingly small proportion of the total population reaches the age necessary to contribute to the advantage. In comparison, the selective advantage from enhanced growth rate is the driving determinant of which genetic material is passed on to the future generations. A mutant with even a relatively small growth advantage will rapidly outcompete a wild-type cell (assuming all other parameters are held constant) because every cell in the population displays the advantageous phenotype.

Figure 1. Theoretical growth competition assays.

(a) Wild-type (WT) yeast (doubling time 90 min, replicative life span 25 generations) and a fast-growing strain (20% increase in doubling time). (b) Wild-type and a long-lived strain (20% increase in mean replicative life span). (c) Long-lived and fast-growing strains. For generation of these curves, it was assumed that all cells in the population replicatively senesce upon reaching the mean replicative life span.

Several prior studies have noted that many of the pathways that modulate yeast aging also modulate stress resistance (Longo & Fabrizio, 2002). For example, mTOR, PKA, and Sch9 each negatively regulate stress-responsive transcription factors such as Msn2, Msn4, and Rim15 (Longo & Finch, 2003; Sadeh et al., 2011; Longo et al., 2012). Under conditions where RLS is extended, signaling through these pathways is reduced, and stress resistance is promoted. Unexpectedly, however, it does not appear to be the case that these transcription factors always directly promote increased RLS (Fabrizio et al., 2004). Instead, additional downstream processes, such as regulation of mRNA translation through differential translation of specific mRNAs, induction of autophagy, and preservation of mitochondrial function, appear to be the primary mechanisms by which these pathways modulate replicative aging (Longo et al., 2012). Interestingly, these same downstream processes also modulate stress resistance. For example, many long-lived ribosomal protein mutants also show enhanced resistance to tunicamycin, a potent inducer of ER stress (Steffen et al., 2012).

Two studies recently attempted to quantitatively assess the fitness and stress response profiles of long-lived yeast mutants. In the first report, fitness was measured for 32 long-lived mutants by direct competition through co-culture with wild-type cells in a 3-week experimental paradigm consisting of repeated periods of growth and quiescence (Delaney et al., 2011a). Under these conditions, a significant fitness defect was detected in more than 60% of the long-lived strains, with most of these long-lived strains also showing a decreased growth rate. In the second study, resistance to four types of stress (heat shock, oxidative, DNA damage, and ER stress) was measured for 46 long-lived mutants (Delaney et al., 2013a). Although no significant correlation was observed between RLS and resistance to a particular stressor, a majority of long-lived mutants showed enhanced resistance to at least one, and often more than one, form of stress.

CONCLUSION

The utility of the yeast replicative aging paradigm in aging research has been demonstrated by studies that have identified longevity pathways in this system, which were subsequently determined to be conserved longevity pathways in higher eukaryotes (Fig. 2). This coupled with the other numerous well-established advantages of yeast (fast growth, low cost, ease of environmental and genetic manipulation, etc.) provides a clear incentive for the continued use of this system in research on the biology of aging. It is now clear that yeast replicative aging shares many features with aging in other systems at the level of sensing and responding to critical environmental parameters such as nutrient availability. It will be of interest to determine whether the downstream responses to these pathways that account for delayed aging are also shared.

Figure 2. Longevity pathways.

Conserved factors modulating life span from yeast to higher eukaryotes are illustrated in blue.