Valter D. Longo, Division of Biogerontology Andrus Gerontology Center, University of Southern California, 3715 McClintock Avenue, Los Angeles, CA 90089-0191, USA. E-mail: email@example.com
Simple model systems have played an important role in the discovery of fundamental mechanisms of aging. Studies in yeast, worms and fruit flies have resulted in the identification of proteins and signalling pathways that regulate stress resistance and longevity. New findings indicate that these pathways may have evolved to prevent damage and postpone aging during periods of starvation and may be conserved from yeast to mammals. We will review the yeast S. cerevisiae model system with emphasis on the chronological life span as a model system to study aging and the regulation of stress resistance in eukaryotes.
The classic approach to the study of aging in yeast is based on the measurement of replicative life span. Mother cells of the yeast S. cerevisiae reproduce asymmetrically by originating buds (daughter cells). Daughter cells are smaller than mothers and can be easily recognized and removed by micromanipulation after budding occurs. Replicative (budding) life span is defined as the total number of daughter cells generated by a mother cell (Mortimer, 1959). Replicative senescence can be caused by the accumulation of extrachromosomal ribosomal DNA circles (ERCs) in old mother cells (Sinclair et al., 1997). ERCs are self-replicating units produced in the nucleolus by rDNA homologous recombination, which segregate asymmetrically to the mother cell during cell division. The silent information regulator proteins Sir2, Sir3, Sir4, which control chromatin silencing at different DNA sites, can regulate ERCs production and aging (Kaeberlein et al., 1999). In particular, the NAD-dependent histone deacetylase Sir2 inhibits rDNA recombination and ERCs accumulation (Kaeberlein et al., 1999). Increasing the dosage of Sir2 extends the replicative life span whereas the deletion of the SIR2 gene decreases replicative longevity (Kaeberlein et al., 1999). Notably, overexpression of the homologue of SIR2 extends longevity in C. elegans, suggesting that this conserved gene may affect both replicative senescence and chronological aging in other eukaryotes (Tissenbaum & Guarente, 2001).
Other genes that regulate replicative aging have been identified. In particular, RAS1 and RAS2 show opposite roles in replicative aging. The deletion of RAS1 extends, whereas the deletion of RAS2 shortens replicative longevity (Sun et al., 1994). As described later in this review, Ras may play opposite roles in the regulation of replicative and chronological longevity since the deletion of RAS2 extends the chronological life span. The LAG1 and LAG2 genes are also implicated in the control of replicative life span. The deletion of each of these genes separately decreases life span by 40% and 50%, respectively. Conversely, their overexpression increases the number of buds generated by yeast mother cells (D’Mello N et al., 1994). The mechanisms behind Lag1/Lag2 life span extension are still unknown. However, Lag1 has been recently implicated in ceramide synthesis and a possible role of this function in regulation of aging has been proposed (Guillas et al., 2001). Replicative aging has been linked to the retrograde response, which is activated by an intracellular signalling pathway from the mitochondrion to the nucleus and leads to the transcription of several genes encoding for metabolic enzymes. The retrograde response is usually induced in yeast strains that have lost part of the mitochondrial DNA and consequently do not have fully functional mitochondria (petite strains). These strains show a significant life span extension, which depends upon the activation of the retrograde response. In fact, the deletion of the RTG2 gene, which is a downstream mediator of the retrograde response, abolishes the life span extension effect (Kirchman et al., 1999).
Calorie restriction (CR), an intervention that lengthens life span in several organisms including mammals, also extends yeast replicative life span (Lin et al., 2000). This effect is mediated by Sir2 and requires increased respiration rates. In fact, calorie restriction failed to extend the life span of a petite strain lacking cytochrome c1 (Lin et al., 2002). Intriguingly, the elimination of the electron transport chain appears to have opposite effects on replicative longevity since in some strains it also promotes replicative life span extension (Kirchman et al., 1999).
We have developed a different system to measure yeast longevity based on the chronological longevity of a population of non-dividing yeast (chronological life span) (Longo et al., 1996; Fabrizio et al., 2001; Longo & Fabrizio, 2002). The chronological life span is monitored either during a hypometabolic phase (stationary phase) or during a high-metabolic state (post-diauxuc phase). Studies of the high-metabolism life span have been crucial in establishing a link between stress resistance and longevity and have provided evidence for the conservation of portions of the pathways that regulate life span in phylogenetically distant eukaryotes (Fabrizio et al., 2001).
Although the relationship between the mechanisms that regulate the replicative and chronological life span are poorly understood, several studies suggest that these two survival paradigms are related: (1) passages through stationary phase decrease the replicative life span of yeast (Ashrafi et al., 1999). The longer the yeast are incubated in stationary phase, the shorter their replication potential once they are switched to fresh medium. (2) Mutations that decrease the activity of the PKA pathway increase both types of life span (Longo, 1997; Lin et al., 2000; Fabrizio et al., 2001). (3) The deletion of SOD1, which encodes for cytoplasmic superoxide dismutase, or of both SOD1 and SOD2 (mitochondrial superoxide dismutase) dramatically reduces yeast chronological life span and decreases replicative life span (Laun et al., 2001; Longo et al., 1996).
Extensive studies on the role of the PKA pathway in both aging paradigms showed that decreasing its activity extends both types of life span. However, whereas the effect of the cAMP/PKA pathway on chronological life span requires Msn2/Msn4 and Sod2 and is associated with increased stress resistance, its effect on replicative life span appears to be independent from stress resistance and dependent on the expression of the Sir2 deacetylase (Lin et al., 2000). Therefore, the chronological and replicative life spans appear to be regulated by overlapping but distinct mechanisms.
Chronological life span: survival in the post-diauxic and stationary phases
Yeast and other micro-organisms have evolved to survive under adverse conditions commonly encountered in the wild, such as starvation. In fact, most micro-organisms are estimated to survive in a low-metabolism stationary phase under nutrient-depleted conditions (Werner-Washburne et al., 1996). In the wild, yeast organisms are likely to exit stationary phase only during the rare periods when all the nutrients required for growth become available. For this reason we perform most of our experiments in either a medium containing a limited amount of nutrients [synthetic dextrose complete (SDC)] or in water. Wild type yeast incubated in SDC medium survive for 6–7 days at high metabolic rates and then begin to die rapidly. When yeast grown in SDC are switched to water between days 1 and 5, metabolic rates decrease and survival is extended (see below). However, since long-lived mutants isolated by incubation in SDC also live longer when incubated in water, we believe that analogous pathways and mechanisms regulate survival in both paradigms. Yeast grown and incubated in the nutrient-rich YPD medium also survive for months in a low-metabolism stationary phase. However, it is not clear whether YPD medium allows some growth to occur during the supposedly ‘stationary’ phase (see survival in water/YPD).
To understand how yeast age and identify conserved pathways that regulate longevity in many eukaryotes, it is important to reproduce conditions similar to those under which these pathways have evolved. Although the chronological life span paradigm may appear to be a starvation phase that does not resemble the life span of higher eukaryotes, non-dividing yeast are not starving but are slowly utilizing the nutrients stored intracellularly at the end of the growth phase (see below). An environment that lacks nutrients may not be common for certain mammals but it is very common for micro-organisms. In some circumstances even mammals have learned how to respond to long periods of starvation. For example, black bears and Turkish hamsters alternate between a high- and a low-metabolism hibernation phase, in which stored nutrients are utilized to survive. The longer the period Turkish hamsters spend under hibernation the longer the life span (Lyman et al., 1981). Similarly, yeast forced to enter the low-metabolism stationary phase by incubation in water survive longer than yeast grown and incubated in SDC medium, which maintain high metabolic rates.
In addition to the high-metabolism post-diauxic life span (SDC), and the low-metabolism stationary phase, under particularly severe starvation conditions, diploid S. cerevisiae can form haploid spores that may survive for years in a dormant state. Yeast spores may be the equivalent of the worm dauer larva, which also live much longer than adult worms (Riddle, 1988; Guarente, 2001). The food supply determines whether worms grow and become metabolically active adults or exit development at the L2 larva stage to enter the low-respiration dauer larva stage. Although most yeast diploid organisms enter and remain in stationary phase and only a minority of diploid organisms form spores (Codon et al., 1995), all of our life span studies are performed using haploid strains which behave similarly to diploid cells under most conditions but do not sporulate.
Survival in SDC: post-diauxic phase
Most of our chronological longevity studies are performed by monitoring survival in the high-metabolism post-diauxic phase (SDC medium) (Fig. 1A). The SDC studies are started by diluting overnight cultures to an initial density of 1–2 106 cells mL−1 (OD600 of 0.1–0.2) in 10–50 mL of synthetic complete medium containing 2% glucose (SDC) as well as a four-fold excess of the supplements Trp, Leu, Ura and His. The SDC medium contains glucose, yeast nitrogen base, agar, ammonium sulphate (nitrogen source), sodium phosphate, vitamins, metals and salts. Yeast cultures are incubated at 30 °C in flasks with a volume/medium ratio of 5 : 1, shaking at 220 r.p.m. After approximately 10 h of exponential growth, the glucose concentration in the medium reaches very low levels and yeast switch from a fermentation- to a respiration-based metabolism. After this switch, called ‘diauxic shift’, yeast catabolize the ethanol accumulated during the fermentative phase and obtain most of the energy from mitochondrial oxidative phosphorylation. When yeast organisms are incubated in SDC, the diauxic shift is followed by a post-diauxic phase, in which growth continues slowly until approximately 48 h, and then stops. In the post-diauxic phase metabolic rates remain high until day 5–6 (day 0 = dilution day) (Fig. 1B). The final density reached at day 3 varies from strain to strain and is usually between 7 and 15 (OD600). The mean survival of wild type strains depends on their genetic background and ranges from 6–7 days (DBY746/SP1) to 15–20 days (S288C/BY4700). A diauxic-shift-like switch from fermentation to respiration may also occur after reducing the glucose concentration in the medium from 2 to 0.5%. This form of calorie restriction increases respiration rates and causes an extension of the yeast replicative life span (Lin et al., 2002).
In a standard post-diauxic experiment we monitor survival by measuring the ability of an individual yeast cell/organism to form a colony (colony forming units or CFUs) within 3 days of plating onto YPD plates. CFUs are normally monitored until at least 99.9% of the population dies. The number of CFUs at day 3 is considered to be the initial survival (100% survival) and is used to determine the age-dependent mortality. Day 3 was selected considering that in our wild type strains DBY746 and SP1 the population density does not normally increase after day 3, suggesting that the great majority of the cells have stopped dividing. Other methods have been used in our laboratory to confirm that the loss of CFUs correlates with death (see Viability loss section).
The ‘post-diauxic phase’ differs for organisms incubated in SDC and those incubated in rich YPD medium (Werner-Washburne et al., 1996). In fact, incubation in YPD promotes a 6–7 day post-diauxic phase characterized by slow growth and low respiration, followed by entry into a non-dividing hypometabolic stationary phase in which organisms are highly resistant to multiple stresses and survive for up to 3 months. By contrast, incubation in SDC triggers the entry into an alternative post-diauxic phase in which only minimal growth occurs after 48 h and metabolic rates remain high until the population begins to die.
Survival in water/YPD: stationary phase
Yeast incubated in YPD or water survive much longer than yeast grown and maintained in SDC medium (Fig. 1A). In fact, the mean life span of strains DBY746 and SP1 in water is approximately three times longer than in SDC (15–20 days). Our experiments are usually repeated in water to simulate an alternative environment that may be commonly encountered by yeast in the wild. Instead, we do not monitor survival in YPD since this rich medium may promote growth after the culture reaches the maximum density. After more than 99% of wild type DBY746 and SP1 yeast incubated in SDC die, in about 30% of the studies a better-adapted subpopulation is able to grow back by utilizing the nutrients released by dead cells (unpublished results). A similar phenomenon called ‘gasping’ is observed for populations of bacteria (Zambrano & Kolter, 1996). By contrast, incubation of yeast in YPD appears to promote major increases in viability before the majority of the population has died, suggesting that some growth may occur when viability is high. Growth would create a mixed population containing both young and old organisms, which would invalidate the survival studies.
For life span studies in water, yeast are grown and incubated for 3 days in SDC, and are then washed with sterile distilled water and re-suspended in sterile water. Viability is monitored by measuring CFUs every 2 days. The cells are washed three times with water every 2 days to remove all the nutrients released by dead yeast. Incubation in water and the removal of nutrients released by dead organisms minimizes the chance of growth during long-term survival in stationary phase. Notably, the life span in the post-diauxic and stationary phases (water) appear to be regulated by similar mechanisms since the ras2Δ and sch9Δ mutants survive longer than wild type yeast in both SDC and water. Incubation in water decreases metabolic rates and may increase life span simply by promoting a slower rate of senescence.
Oxidative damage and longevity in yeast
Micro-organisms have played a key role in the characterization of anti-oxidant enzymes and identification of the sources and targets of reactive oxygen species. Like mammalian cells, S. cerevisiae expresses a cytosolic CuZn superoxide dismutase (Sod1) and catalase (Ctt1), as well as a mitochondrial MnSOD (Sod2). By monitoring the chronological life span of yeast mutants lacking specific genes it was possible to understand how different anti-oxidant enzymes affect cell damage and aging. Cytoplasmic and mitochondrial superoxide dismutases, but not catalase and metallothionein, are required for long-term survival (Longo et al., 1996) (Fig. 2). Sod2 is required under both low and normal oxygen conditions, whereas cytoplasmic Sod1 is mainly required under normal aeration (Longo et al., 1996). The expression of human SOD1 in yeast sod1 null mutants completely reverses the survival defects, suggesting that the function of this enzyme is conserved from yeast to mammals. In fact, these results are consistent with studies performed using mice lacking sod1. Although sod1 knockout mice showed few abnormalities, cultured fibroblasts obtained from these mice were 80 times more sensitive to the superoxide generator paraquat than wild type cells and grew poorly in air (Huang et al., 1997), indicating that, as reported for yeast, Sod1 is only required under a high concentration of oxygen or superoxide.
Unlike most experimental organisms, yeast have the ability to grow either by respiration utilizing non-fermentable carbon sources (ethanol, lactate), or by fermentation utilizing glucose (low respiration). We have exploited this feature to determine whether mitochondrial damage precedes death. We have called the percentage of live yeast able to utilize non-fermentable carbon sources and respiration to grow the Index of Respiratory Competence (IRC). Using the IRC, it was possible to define the sequence of events that lead to the death of mutants lacking mitochondrial SOD (sod2) and to explore the role of mitochondrial damage in the death of wild type yeast in stationary phase. In both sod2 mutants and wild type yeast, death was preceded by a decrease in IRC suggesting that loss of mitochondrial function preceded death (data not shown). The characterization of yeast sod2 null mutants also resulted in the identification of mitochondrial aconitase and succinate dehydrogenase, both 4Fe−4S cluster binding proteins, as the primary targets of mitochondrial superoxide (Longo et al., 1999). These results are consistent with studies in sod2 knockout mice in which the activity of mitochondrial aconitase was reduced by more than 67% in the heart and brain, and succinate dehydrogenase was reduced by more than 65% in the heart and skeletal muscle (Melov et al., 1998). As expected, in sod2 mice the cells most severely affected by the increased concentration of mitochondrial superoxide were post-mitotic. The similarities between non-dividing yeast and mice lacking either SOD1 or SOD2 suggest that the chronological life span of yeast is a valuable model system for mechanisms of oxidative damage in mammalian cells, particularly post-mitotic cells.
The overexpression of CuZnSOD has been convincingly shown to increase the life span of fruit flies. (Orr & Sohal, 1994; Parkes et al., 1998; Sun & Tower, 1999.) To test whether the overexpression of anti-oxidant enzymes could also extend the life span of yeast, strains expressing several fold higher concentrations of these antioxidant enzymes were generated. The overexpression of SOD1 and SOD2 together increased chronological survival by 30% whereas the overexpression of each SOD alone had a significant but more modest effect (Fabrizio et al., 2003). These results are in agreement with the age-dependent reversible inactivation of the mitochondrial enzyme aconitase, which is caused by superoxide (Fabrizio et al., 2001).
Yeast has served as a powerful model system in understanding the function and mechanism of action of mammalian proteins. The anti-apoptotic human Bcl-2 protein was overexpressed in yeast sod mutants and wild type yeast to investigate its mechanism of action and to understand whether yeast may have components of a programmed cell death pathway. Human Bcl-2 partially reversed several defects of yeast lacking superoxide dismutases, consistent with the demonstrated anti-oxidant function of Bcl-2 in mammalian cells. Bcl-2 overexpression increased long-term viability and growth in 100% oxygen of sod1 and sod1sod2 mutants (Longo et al., 1997). Bcl-2 also increased survival in wild type cells, raising the possibility that portions of an apoptotic pathway are present in yeast. This hypothesis was recently supported by yeast studies in which the expression of the human pro-apoptotic protein Bax causes cell death, which can be reversed by coexpression of Bcl-2 (Matsuyama et al., 1998; Shaham et al., 1998). Furthermore, a caspase-related protease and several features of apoptosis were recently identified in yeast (Madeo et al., 1999; Madeo et al., 2002).
Viability loss and starvation in the post-diauxic and stationary phases
Yeast viability in most of our experiments is measured by counting the number of colonies generated after plating an aliquot of the culture onto YPD plates. In theory, each viable cell should divide and form a colony (CFU). We tested whether the loss of CFU correlates with the loss of viability in the post-reproductive phase. We measured the concentration of proteins released into the medium by dead and damaged wild type DBY746-plasmid controls and by SOD1SOD2 double overexpressors. The increase in protein concentration in the medium of both strains began 48 h after the loss of CFUs and continued for 7 days. In agreement with the CFU measurements, the increase in protein concentration in the medium of SOD1SOD2 overexpressors was delayed by 2 days compared with controls (Fabrizio et al., 2003). The release of protein was lower in SOD1SOD2 overexpressors than in wild type controls throughout the study. To characterize further the chronological life span, we also performed confocal microscopy and two-photon confocal microscopy in aging organisms. Three- to 7-day-old DBY746 yeast were stained with FUN-1, which stains intravacuolar structures only in metabolically active organisms with an intact plasma membrane. Yeast were also stained with Calcofluor, a marker of fungal cell wall. FUN-1 staining of DBY746 and sch9Δ mutants suggests that the loss of membrane integrity and metabolic activity correlates with the decrease in CFUs (Fabrizio et al., 2003). At day 7 the majority of sch9Δ mutants were marked with fluorescent intravacuolar structures whereas the majority of wild type organisms produced a diffuse green cytoplasmic fluorescence (unpublished results). Taken together, these results suggest that during the chronological life span, the ability to reproduce and form a colony correlates with death. However, further studies are necessary, considering that it may be difficult to distinguish between a damaged organism surviving at very low metabolic rates and a dead organism.
Since most of the extracellular nutrients are internalized during the post-diauxic and stationary phases, it is often believed that stationary phase yeast starve to death. However, the published data and our own results suggest that yeast organisms are not dying by starvation. During the long-term survival in stationary phase, yeast produce energy mainly by catabolizing the glycogen accumulated in the late exponential growth. Using strain S288C, Lillie & Pringle (1980) showed that the reserve carbohydrates glycogen and trehalose are available after an incubation of 90 days in rich YPD medium (stationary phase). Our studies of the high-metabolism post-diauxic phase also suggest that death is not caused by a depletion of reserve nutrients. Glycogen levels adjusted for population density are not significantly different between day 3 and day 7 (SDC medium), when only 30% of the organisms are viable (unpublished results). Thus, it would be surprising if yeast were maintaining high concentrations of the major reserve carbohydrate but were dying of starvation. The mean life span of strains DBY746 and SP1 is more than 15 days in water but less than 7 days in SDC medium. An extended survival in the complete absence of nutrients (water) is also not consistent with death by starvation. Furthermore, the switch from expired medium to water during the high death phase halts cell death, suggesting that dying organisms have not depleted reserve nutrients (unpublished results). Finally, overexpression of both SOD1 and SOD2 extends survival by 30% but does not affect metabolic rates. Thus, it is difficult to reconcile this role of increased protection against oxidative stress in extending survival with death by starvation, as increased investment in maintenance would require additional energy. Although these results suggest that starvation does not play a major role in the chronological life span of yeast, further studies are required to rule out this possibility. We are also currently testing the effect of the lack of amino acid biosynthetic enzymes and of the depletion of amino acids on the survival of various strains. Our preliminary results suggest that chronological life span (SDC) is not increased by adding to the expired medium the amino acids that strains SP1 and DBY746 are unable to biosynthesize (data not shown).
The identification of pathways that regulate stress resistance and longevity
The deletion of RAS2 increases stress resistance and doubles the chronological life span of yeast (Longo, 1997; Fabrizio et al., 2003). The identification of Ras mutations that increase Sod activity and longevity resulted in the description of the first eukaryotic longevity pathway (Longo, 1997). These results and the concurrent identification of genes that regulate longevity in worms and flies led to the hypothesis that in organisms ranging from yeast to humans, longevity is regulated by a similar signal transduction pathway that modulates Sod activity through the activation of stress resistance transcription factors (Longo, 1997, 1999). The role of ras2 mutations in extending longevity by activating stress resistance transcription factors Msn2/Msn4 and SODs (Longo, 1997) was recently conclusively demonstrated by showing that Msn2/Msn4 and Sod2 are required for longevity extension in ras2 mutants (Fabrizio et al., 2003). As described in the next section, the yeast Ras2 and Sch9 pathways share remarkable similarities with the worm insulin/IGF-I-like pathway and with the mammalian insulin/IGF-I pathway. Notably, Ras functions in the mammalian insulin/IGF-I signalling pathway but has not been implicated in the worm longevity pathway.
Our strategy to isolate long-lived mutants combined the selection of stress-resistant mutants with transposon mutagenesis (Ross-Macdonald et al., 1999). Because of the link between stress resistance and survival in higher eukaryotes, after generating a population of transposon-mutagenized yeast, we performed a preselection for mutants that were (a) thermotolerant (52 °C, 1 h) or (b) resistant to oxidative stress (1 mm paraquat, 9–10 days). Four thousand thermotolerant and 40 oxidative stress resistant mutants were isolated and their mean post-diauxic survival was measured and compared with that of the wild type strain. Mutants with life span significantly longer than wild type were identified and their mutated alleles were cloned as described (Ross-Macdonald et al., 1999). From the mutants isolated, the ones that carried transposon insertions in the CYR1 gene and in the promoter region of the SCH9 gene were the two longest lived (Fabrizio et al., 2001). These were also the only two mutants isolated independently in the thermotolerance and oxidative stress selection. These two genes encode for adenylate cyclase and for the Sch9 serine/threonine protein-kinase, respectively.
Cyr1 and Sch9 function in parallel signalling pathways (Ras/adenylate cyclase and Sch9 pathways) are both involved in mediating the glucose/nutrients-dependent response including the stimulation of growth and glycolysis and the down-regulation of stress resistance, glycogen accumulation and gluconeogenesis. Since the transposon insertion in both CYR1 and SCH9 genes appears to reduce the activity of the corresponding proteins, we conclude that the down-regulation of the glucose/nutrients signalling pathways extends yeast chronological life span (Fabrizio et al., 2001).
The deletion of RAS2, which encodes for an upstream regulator of Cyr1, doubles the yeast life span (Longo, 1997; Fabrizio et al., 2003). To verify that the effect of Sch9 and Ras2 on the life span regulation is not dependent on the yeast background we deleted SCH9 and RAS2 in another strain (SP1) and measured the chronological life span of the corresponding strains. The life span extension of the deletion mutants in the SP1 background was very similar to their counterparts in the DBY746 background (Fig. 3).
Stress resistance, superoxide dismutases and longevity
Several studies support a major role for multiple stress resistance in the regulation of longevity in yeast: (a) to identify long-lived mutants we preselected for heat/oxidative stress-resistant yeast and isolated mutants that are highly resistant to both heat and oxidative damage and live up to three times longer than wild type; (b) the deletion of RAS2 increases longevity as well as thermotolerance and resistance to oxidative stress; (c) the life span extension of the ras2 and cyr1 mutants was shown to be mediated by stress-induced transcription factors Msn2/Msn4 (Fabrizio et al., 2001); (d) the deletion of the SOD2 gene, encoding for the mitochondrial superoxide dismutase dramatically reduced the life span extension of the ras2, cyr1 and sch9 mutants (Fabrizio et al., 2003); and (e) the up-regulation of the heat shock response produced by decreasing the activity of the Hsp90 chaperone extends the life span (Harris et al., 2001).
Superoxide resistance seems to be particularly important in extending the chronological life span. In fact, mutants lacking either SOD1 (cytoplasmic superoxide dismutase) or SOD2 or both are short-lived compared to wild type (Fig. 2). Furthermore, our studies on the activity of the mitochondrial enzyme aconitase in post-diauxic yeast strongly suggest a role for superoxide toxicity in aging (Longo et al., 1999). Aconitase is a citric acid cycle enzyme, which contains a 4Fe−4S cluster that is particularly sensitive to superoxide. An age-dependent reversible inactivation of this enzyme caused by superoxide has been shown during survival (Longo et al., 1999). This inactivation contributes to the mitochondrial damage that precedes cellular death. Interestingly, both cyr1 and sch9 mutants showed a delay in the aconitase inactivation, confirming the role played by superoxide in aging (Fabrizio et al., 2001) (Fig. 4).
The age-dependent inactivation of mitochondrial aconitase and the role of superoxide dismutases in extending the chronological life span in yeast indicate that superoxide and other toxic oxygen species play a major role in aging. In 1956 Harman proposed that oxygen species with one unpaired electron (free radicals) may cause aging. The free radical theory of aging became one of the most widely accepted theories after the overexpression of anti-oxidant enzymes was shown to extend longevity and after most long-lived model organisms were shown to be resistant to oxidative stress (Longo, 1999; Finkel & Holbrook, 2000). Although superoxide toxicity contributes to aging and death in model organisms, mutations in the yeast Sch9/Cyr1 and in the worm and fly insulin/IGF-I-like pathways appear to extend longevity by regulating the expression of many genes. In fact, the overexpression of Sod1, Sod2 and catalase in yeast and flies can extend longevity by up to 30%, whereas mutation in signal transduction genes can extend longevity up to three-fold (Clancy et al., 2001; Fabrizio et al., 2001; Tatar et al., 2001). Therefore, increasing anti-oxidant protection appears to be important but not sufficient for the major longevity extension caused by mutations in glucose or insulin/IGF-I signalling pathways.
From yeast to mammals
Thanks to studies by several laboratories, a significant proportion of the genes that function in longevity pathways in yeast and worms have been characterized. In yeast, the down-regulation of glucose signalling by ras2, cyr1 and sch9 mutations increase longevity and stress resistance (Fig. 5). In the cyr1 mutants chronological life span extension is mediated by stress resistance transcription factors Msn2 and Msn4, which induce the expression of genes encoding for several heat shock proteins, catalase (CTT1), the DNA damage inducible gene DDR2, and SOD2 (Fig. 5) (Thevelein & de Winde, 1999; Fabrizio et al., 2001). Analogously, in worms, the inactivation of the insulin/IGF-I-like/daf-2 pathway extends survival (Johnson, 1990; Kenyon et al., 1993) and increases thermotolerance and anti-oxidant defences (Larsen, 1993; Lithgow et al., 1995; Kimura et al., 1997), by activating stress resistance transcription factor DAF-16 (Fig. 5) (Lin et al., 1997). In yeast, chronological life span extension is associated with decreased superoxide generation and aconitase inactivation in the mitochondria (Fabrizio et al., 2001). Furthermore, life span extension in ras2, cyr1 and sch9 mutants requires SOD2 and the overexpression of superoxide dismutases extends longevity (Fabrizio et al., 2003). In worms, among the genes regulated by the daf-2 pathway are mitochondrial MnSOD and several heat shock proteins (Boy-Marcotte et al., 1998; Cherkasova et al., 2000). The yeast Ras/Cyr1/PKA pathway down-regulates glycogen storage and genes involved in the switch to the hypometabolic stationary phase and to the dormant spore state (Werner-Washburne et al., 1996; Boy-Marcotte et al., 1998). The worm daf-2 pathway also down-regulates the storage of reserve nutrients (fat and glycogen) and the switch to the hypometabolic dauer larvae state (Fig. 5) (Kenyon et al., 1993; Morris et al., 1996; Kimura et al., 1997). Thus, in addition to the high sequence similarities between the yeast SCH9 and the worm AKT-1/AKT-2 genes, these distantly related organisms appear to regulate stress resistance and longevity by modulating the activity of similar proteins and pathways (Fig. 5). Remarkably, life span regulation of more complex organisms such as Drosophila and mice appears to depend on the activity of insulin/IGF-I-stimulated pathways similar to those identified in yeast and worms. These pathways include serine/threonine kinases (Akt/PKB) and regulate the activity of several stress-resistant proteins including SODs, catalase and Hsps (Longo & Fabrizio, 2002).
The analogous role of glucose or hormone/growth factor signalling in stress resistance and aging in the major genetics model systems suggests that longevity can be extended by similar mechanisms in many organisms. Life span extension appears to be caused by a shift from a reproductive phase to a non-reproductive maintenance phase, accompanied by changes in the expression of many genes normally induced by starvation. The yeast chronological and replicative life span paradigms have provided evidence for the conservation of longevity regulation in eukaryotic organisms. These systems also play an important role in the dissection of the fundamental molecular mechanisms leading to both chronological and replicative aging in eukaryotes.