Candida albicans, a distinctive fungal model for cellular aging study

The unicellular eukaryotic organisms represent the popular model systems to understand aging in eukaryotes. Candida albicans, a polymorphic fungus, appears to be another distinctive unicellular aging model in addition to the budding yeast Saccharomyces cerevisiae and fission yeast Schizosaccharomyces pombe. The two types of Candida cells, yeast (blastospore) form and hyphal (filamentous) form, have similar replicative lifespan. Taking the advantage of morphologic changes, we are able to obtain cells of different ages. Old Candida cells tend to accumulate glycogen and oxidatively damaged proteins. Deletion of the SIR2 gene causes a decrease of lifespan, while insertion of an extra copy of SIR2 extends lifespan, indicating that like in S. cerevisiae, Sir2 regulates cellular aging in C. albicans. Interestingly, Sir2 deletion does not result in the accumulation of extra-chromosomal rDNA molecules, but influences the retention of oxidized proteins in mother cells, suggesting that the extra-chromosomal rDNA molecules may not be associated with cellular aging in C. albicans. This novel aging model, which allows efficient large-scale isolation of old cells, may facilitate biochemical characterizations and genomics/proteomics studies of cellular aging, and help to verify the aging pathways observed in other organisms including S. cerevisiae.


Introduction
Aging is usually defined as the progressive loss of function accompanied by decreasing fertility and increasing mortality with advancing age (Kirkwood & Austad, 2000). Lifespan regulation is evolutionarily conserved, and found in various species ranging from eukaryotic multicellular (e.g. humans) to unicellular (e.g. yeast), and to prokaryotic (e.g. Escherichia coli ) organisms (Stewart et al ., 2005). In unicellular species, cellular aging and organismal aging coincide. Previous studies in several model organisms have described many aging-related phenotypes, including morphological changes, accumulation of oxidatively damaged DNA and proteins, shortened telomeres, enriched extra-chromosomal rDNA circles (ERCs), etc. (reviewed in Sinclair et al ., 1998;Balaban et al ., 2005;Kirkwood, 2005;Lombard et al ., 2005). Nowadays, the elucidation of the molecular mechanism(s) in aging is of great interest to biologist.
The budding yeast Saccharomyces cerevisiae , which divides asymmetrically, is a unicellular eukaryotic organism with a short and easily studied lifespan, and enables researchers to chase an individual cell through many cell divisions. The yeast cells undergo a limited number of cell divisions before senescence, and small budding daughter cells come from the larger aging mother cells in most of the lifespan. The total number of daughter cells produced by a mother cell prior to senescence is defined as the replicative lifespan of the mother cell (Mortimer & Johnston, 1959). Aging in yeasts S. cerevisiae and Schizosaccharomyces pombe has also been measured as the length of time a population stay alive; it is referred to as chronological aging (Fabrizio & Longo, 2003;Roux et al ., 2006).
Recent studies reveal that some of the molecular mechanisms of aging existing in higher eukaryotes have also been unraveled in S. cerevisiae . For example, the human WRN helicase defect causes the premature aging Werner syndrome (Epstein et al ., 1966), while mutation of Sgs1 helicase, the WRN homolog in yeast, causes the shortening of replicative lifespan . Calorie restriction, which significantly promotes longevity in various organisms including worms, flies, fish and mammals, also lengthens yeast lifespan (Weindruch & Walford, 1988;Jiang et al ., 2000;Lin et al ., 2000;Lane et al ., 2001). Deletion of Sir2 histone deacetylase decreases the replicative longevity of yeast, whereas an extra copy of SIR2 integrated into the genome results in longer replicative lifespan (Kaeberlein to prevent DNA breaks, homologous recombination and ERC formation in a Fob1-mediated pathway . Therefore, it is suggested that ERCs are a cause of yeast aging , and Sir2 plays a specific role in promoting longevity. Since the budding yeast S. cerevisiae has been developed as an advantageous model of aging research, the development of techniques for large-scale isolation of old cells is extremely required in any biochemical or high-throughput analysis. As this problem is of major interest, several methods have been established in S. cerevisiae for obtaining large quantities of old cells (Park et al ., 2002), including yeast 'baby machine' (Helmstetter, 1991), centrifugation elutriation (Woldringh et al ., 1995), sucrose gradient centrifugation (Egilmez et al ., 1990), fluorescenceactivated cell sorting (Sloat & Pringle, 1978) and magnetic sorting (Smeal et al ., 1996). However, these approaches seem costly and laborious because (i) the mother and daughter cells are similar in shape; (ii) the proportion of old cells in any growing culture of logarithmic phase is miniscule (Park et al ., 2002); (iii) as S. cerevisiae yeast mother cells grow old, they enlarge and tend to produce large and short-lived daughter cells (Kennedy et al ., 1994), and at last several divisions, daughter and mother cells are usually indistinguishable in size and can hardly separate from each other (Kennedy et al ., 1995). Hence, we are encouraged to develop another model, Candida albicans , for aging study.
Candida albicans , which diverges from S. cerevisiae over 140 million years ago (Berman & Sudbery, 2002), has been evolved to be a polymorphic fungus and exists as a commensal of warmblooded animals including humans. It can grow in different ways: proliferating in yeast form and forming filamentous hyphal cells that can give birth to new mycelia or yeast form (ellipsoidal or blastospore) daughters (Molero et al ., 1998). The growth form can be controlled by changing culture conditions including temperature, pH and nutrient composition (Soll, 1986). C. albicans usually exists as a diploid and parasexual organism, which is different from S. cerevisiae . As these interesting features are concerned, we anticipated to establish C. albicans as a new aging model complementary to S. cerevisiae and S. pombe . The complete genome sequencing has revealed that many C. albicans open reading frames have obvious S. cerevisiae homologs and the well-established tools for molecular-genetic manipulations make it feasible to set up C. albicans as a model organism.
In the current study, we observed that C. albicans cells, like S. cerevisiae ones, have a finite lifespan. We have established a method to prepare a large number of old cells, facilitating the high-throughput analysis of cellular aging in C. albicans . We also characterized the old cells and found out that glycogen and oxidatively damaged proteins accumulate during the replicative aging of C. albicans . In addition, Sir2 regulates Candida lifespan in a dose-dependent manner, but the extra-chromosomal rDNA molecules show no obvious correlations with Sir2 dosage, suggesting that unlike in S. cerevisiae , extra-chromosomal rDNA molecules may not be the marker of aging in C. albicans .

Both yeast and hyphal form of C. albicans cells have finite lifespan
Candida albicans is a prevalent opportunistic fungal pathogen in humans, and can grow in budding form or filamentous form (Fig. 1A). Like S. cerevisiae , C. albicans cells of yeast form propagate by budding at 30 ° C (Fig. 1A, left panel). In response to serum at 37-40 ° C, the yeast form cells can be induced to filamentous hyphae that usually contain a few cells in tandem (Fig. 1A, middle panel). The filamentous hyphal cells can also give rise to yeast form daughters when they are grown at 30 ° C in the absence of serum (Fig. 1A, right panel) (Molero et al., 1998).
To find out whether C. albicans cells have finite lifespan, we performed replicative lifespan assay of both yeast and hyphal form cells. Since yeast cells produce daughters by budding, we analyzed their replicative lifespan as described in S. cerevisiae (Kaeberlein et al., 2004). Wild-type cells of SC5314 strain have a mean lifespan of 19-23 generations ( Fig. 1B and Fig. S4). To measure the lifespan of hyphal cells, we induced the growth of hyphal mothers containing four to six single cells in tandem that can produce yeast daughters ( Fig. 2A). The total daughters produced by a hyphal mother were counted. The ratio of the total daughter number to the cell number of a hyphal mother represents the replicative lifespan of a single hypha cell. Figure 1B shows that hyphal form cells share similar replicative lifespan with yeast form cells, indicating the morphological change does not affect the lifespan. These results also demonstrate that both yeast and hyphal form cells have finite replica-tive lifespan. Our data suggest that C. albicans could be used as a model to study cellular aging.

Dramatic morphological difference facilitates large-scale preparation of replicatively old cells
Biochemical investigation of the age-associated changes often requires a large-scale preparation of sufficiently pure and terminally senescent cells. However, old cells only account for a negligible fraction of a logarithmic growing culture. The isolation of replicatively aged cells has proved to be one of the most difficult tasks in yeast aging research. Since the hyphal cells of C. albicans can produce yeast form daughters, we thought that the dramatic difference in size between hyphal and yeast cells provided an opportunity to feasibly separate the old cells from their progenies. With this in mind, we established a method that allows separation of hyphal and yeast form cells based on sucrose gradient centrifugation. The separation efficiency was determined by the percentage of hyphal cells in the population. Most of the hyphae in our culturing system have four to six hyphal cells and one original yeast cell ( Fig. 2A). Hyphal cells account for ~80% of the population after separation, so we think the separation is efficient enough. The separated hyphal cells are shown in Fig. S1.
To separate the old hyphal mothers from their yeast daughters and study the aging associated phenotypes in C. albicans, we designed a procedure to obtain certain amount of old mother cells, schematically shown in Fig. 2B. Briefly, yeast form cells were incubated at 30 °C overnight, and hyphal cells were induced at 37 °C in the presence of serum for about 5 h. The hyphal cells in this culture should be at various replicative ages. The purified hyphae were then shifted back to YPD medium (1% yeast extract, 2% peptone, 2% glucose) to produce yeast cells at 30 °C for about 2.5 h. The relatively short time of culture ensured that the newborn yeast cells experienced no more than two cell divisions and could be regarded as young cells. These 'synchronized' young yeast cells were isolated from the culture by sucrose gradient centrifugation and then used to generate young hyphal mothers. When the hyphal mothers were cultured at 30 °C, they produced yeast form daughters and became older and older with prolonged culture time. Therefore, the hyphal cells are old cells and separated from yeast cells by sucrose gradient centrifugation (Fig. 2B). In this experiment, there are rare opportunities of young yeast daughter contamination in old hyphal mothers due to the dramatic size difference between hyphal and yeast form cells.
In order to verify the separation strategy, we measured the replicative lifespan of the hyphal mothers separated at various time points (0 h, 38 h, 75 h and 114 h). As shown in Fig. 2D, the lifespan of the hyphal cells decreases along with growing old and the division capacity of the oldest cells is diminished largely, indicating that the hyphal mothers age gradually after they continuously give birth to yeast daughters, and the oldest cells we obtained are presumably not far from replicative senescence. In the following sections (Figs 3 and 5), we used the cell samples prepared at the same four time points. The old cells can also be distinguished from the young cells morphologically: hyphae are enlarged and broken into single cells (Fig. 2C). The procedures described here offer an opportunity to prepare old cells of different replicative ages in large-scale and examine the progressive changes associated with cellular aging.

Glycogen and oxidatively damaged proteins accumulate during replicative aging of C. albicans
Glycogen accumulation has been described as a marker of aging, and reported in S. cerevisiae (Lin et al., 2001). Genomics study has also suggested that genes involved in gluconeogenesis and glycogen production are induced in old cells (Lesur & Campbell, 2004). We analyzed the glycogen level in C. albicans hyphal cells of different replicative ages, and the results showed that glycogen level increased during the aging period (Fig. 3A). The accumulation of glycogen in old cells may reflect a turnover of carbohydrates catabolism and energy metabolism.
It is increasingly evident that the oxidative modification of cellular components, including oxidation of proteins, lipids and nucleic acids, is another apparent phenotype in old cells over many species (Stadtman, 2006). To characterize the old C. albicans cells, we examined the overall oxidation of proteins according to the methods reported by Levine et al. (Levine et al., 1994;Cabiscol et al., 2000). The carbonyl groups (aldehydes and ketones) are introduced into oxidatively modified proteins irreversibly (Stadtman, 1992). Based on the reaction of carbonyl groups with 2,4-dinitrophenylhydrazine to form a 2,4-dinitrophenylhydrazone, we could immunodetect protein carbonylation using antibodies to the 2,4-dinitrophenyl moiety (Levine et al., 1994). Figure 3B shows that the level of carbonylated proteins was increasing remarkably as cells age, and the level of tubulin served as an internal control. These data demonstrate that oxidatively damaged proteins accumulate during replicative aging of C. albicans. Our results are consistent with the work in S. cerevisiae reported previously (Aguilaniu et al., 2003;Reverter-Branchat et al., 2004;Grzelak et al., 2006). It has been postulated that the age-related accumulation of oxidized proteins is a complex process that may integrate increases of reactive oxygen species generation rates, decreases in antioxidant activities or losses in the capacity to degrade oxidized proteins (Stadtman, 2006).
In addition to glycogen and oxidatively damaged proteins that accumulate in old C. albicans cells, we also examined telomere length and chromosome integrity by Southern blot and pulsed field gel electrophoresis (PFGE), respectively. No obvious differences between young and old cells have been observed (Figs S2 and S3).

Lower glucose concentration delays cellular aging in C. albicans
Calorie restriction extends lifespan in a variety of species and has been proved a general mechanism to enhance longevity (Weindruch & Walford, 1988;Masoro, 2002). In the yeast form cells of C. albicans, we examined the effect of low glucose concentration (Lin et al., 2000;Kaeberlein et al., 2002) on replicative lifespan. Cells grown in the presence of 0.1% glucose exhibited a longer lifespan in both SC5314 and BWP17 strains (Fig. 4A). Although the increase of lifespan in strain SC5314 seems weaker than that in strain BWP17 or S. cerevisiae, the calorie-restriction effect in strain SC5314 is significant because the p-value is less than 0.05. These data indicate that like in other species, calorie restriction extends lifespan in C. albicans. Notably, the two wild-type strains, SC5314 and BWP17, exhibited different lifespan: SC5314 is short-lived, while BWP17 is long-lived (Fig. S4). The divergence may be resulted from their different genetic background.

Replicative lifespan of C. albicans correlates with the dosage of Sir2
It is known that the human Werner syndrome is caused by a defect in WRN helicase. Mutation of Sgs1 helicase, the WRN homolog, which is involved in DNA metabolism, causes the shortening of yeast lifespan . In order to study a Candida counterpart of the SGS1 gene, we deleted one copy of SGS1 in C. albicans, and measured the replicative lifespan of the mutant cells. The shortened lifespan of the sgs1Δ/SGS1 heterozygote reveals a haplo-insufficiency, indicating that as in S. cerevisiae, Sgs1 is required for the regular lifespan of C. albicans (Fig. 4B, left panel).
Previous studies in S. cerevisiae demonstrate that Fob1/Sir2 pathway is involved in regulating cellular aging via affecting ERCs Kaeberlein et al., 1999). sir2Δ mutant exhibits shorter lifespan, whereas an extra copy of SIR2 integrated into the genome results in longer lifespan . The deletion of FOB1, a gene required for blocking rDNA replication, suppresses rDNA recombination and ERC formation, extends lifespan in wild-type cells, and restores lifespan in sir2 mutant cells (Defossez et al., 1999). Despite the fact that the Fob1 homolog in C. albicans has not been identified, the Sir2 homologue in C. albicans has been reported to be essential for the maintenance of chromosome stability (Perez-Martin et al., 1999). We suspected that Sir2 also affects Candida longevity. Because C. albicans always lives as a diploid, we examined the lifespan of sir2Δ/SIR2 and sir2Δ/sir2Δ cells. Like S. cerevisiae, the reduced dosage of Sir2 resulted in a decrease of replicative lifespan (Fig. 4B, middle panel) . When an extra copy of SIR2 was integrated into the genome at the ADE2 locus, the lifespan increased compared with the insertion of the empty vector (Fig. 4B, right panel), suggesting that a conserved Sir2 pathway that regulates cellular aging exists in C. albicans. It is not clear why the introduction of a control empty vector at ADE2 locus in BWP17 shortened its lifespan (Fig. S4A). One possible explanation is that the disruption of ADE2 has a negative role on replicative lifespan.

Extra-chromosomal rDNA molecules do not increase in sir2Δ Δ Δ Δ mutants of C. albicans
Sir2 in S. cerevisiae has been indicated to affect the recombination of rDNA to regulate cell longevity (Guarente, 2000). Homologous recombination between adjacent rDNA repeats which locate on chromosome XII results in the formation of ERCs. Sir2 is essential for the silenced chromatin formation in rDNA loci (Bryk et al., 1997). In C. albicans, the rDNA units are clustered as tandem repeats on both chromosome R, and each unit is about 11.5 to 12.5 kb (Wickes et al., 1991;Iwaguchi et al., 1992;Rustchenko et al., 1993). The insertion of an exogenous URA3 gene in rDNA loci did not affect the growth of cells in the 5'-FOA containing medium, indicating that the URA3 gene was silenced as observed in S. cerevisiae (Fig. 4C, row 2). In contrast, single copy deletion of SIR2 liberated the expression of the URA3 gene, demonstrating a reduction of rDNA silencing (Fig. 4C, row 3). The silenced chromatin inhibits homologous recombination between rDNA repeats, and reduces the formation of ERCs. Several lines of evidence suggest that the ERC accumulation may be a cause of cell aging in S. cerevisiae, and introduction of artificial ERCs shortens lifespan of yeast cells . In C. albicans, the extra-chromosomal rDNA containing molecules have also been reported, but their size is ranging from 50 to 100 kb (Perez-Martin et al., 1999), which are much larger than the ones detected in S. cerevisiae. To investigate whether the extra-chromosomal rDNA molecules increases when the SIR2 gene is deleted in C. albicans, we employed PFGE to separate the whole genome, and did Southern blot using an rDNA probe. As shown in Fig. 4D, the extra-chromosomal rDNA molecules were not increased in sir2Δ/SIR2 or sir2Δ/sir2Δ mutant as compared with the wild-type strain BWP17. These data indicate that the decrease of lifespan in 2 mutant cells may not be associated with the amount of extra-chromosomal rDNA molecules.
Previous work by Perez-Martin et al. (1999) reported that extra-chromosomal rDNA molecules were not detected in wildtype strain 3153 A, but enriched in sir2Δ/sir2Δ cells of C. albicans. In our experiments, the extra-chromosomal rDNA molecules have been detected in all examined strains including the wild-type SC5314 cells (Fig. 4D). The discrepancy may be resulted from a strain-specific genetic background.

Extra-chromosomal rDNA molecules do not accumulate in aging Candida cells
ERCs have been shown to accumulate during cellular aging in S. cerevisiae, so we wondered whether the extra-chromosomal rDNA molecules also accumulated along with the aging process in C. albicans. Thus, we performed PFGE and examined rDNA with Southern blot. In both wild-type strain and the two sir2 deletion mutants, the extra-chromosomal rDNA molecules did not display any obvious changes as the replicative age increased (Fig. 5A). The extra-chromosomal rDNA molecules in hyphal cells seemed not to be as homogenous as that in yeast cells (compare the lane marked 'SC5314 (yeast)' with other lanes). We also examined the rDNA copy number in wild-type, sir2Δ/ SIR2 and sir2Δ/sir2Δ cells with different ages, and no increase was detected (Fig. 5B). All these findings further suggest that the extra-chromosomal rDNA molecules may not be associated with cellular aging of C. albicans.

Candida Sir2 functions in the distribution of oxidatively damaged proteins during cell division
Since the correlations between Sir2 and extra-chromosomal rDNA molecules were not observed in C. albicans, and previous studies in S. cerevisiae elucidated that Sir2 regulates the asymmetric inheritance of oxidatively damaged proteins during cell division (Aguilaniu et al., 2003), we wanted to know whether Candida Sir2 also affects the distribution of oxidatively damaged proteins in mother and daughter cells. In wild-type strain, more oxidatively damaged proteins were retained in mother cells as compared to its daughter budding cell. In contrast, in sir2Δ/sir2Δ strain, the oxidatively damaged proteins were distributed evenly between mother and daughter cells (Fig. 6A). The result illuminates that cells fail to segregate oxidized proteins when Sir2 is absent. To ask whether the accumulation of oxidatively damaged proteins is increased in the sir2Δ mutants, we performed the carbonyl assay to detect the overall level of oxidatively damaged proteins. Interestingly, the overall level of oxidatively damaged proteins in sir2Δ/SIR2 or sir2Δ/sir2Δ cells seemed not to be higher than that in wild-type cells (Fig. 6B). These data indicate that Sir2 controls the ability of mother cells to retain oxidatively damaged proteins during cell division. Our results are consistent with the previous findings in S. cerevisiae (Aguilaniu et al., 2003).

Discussion
The unicellular eukaryotic organism, S. cerevisiae has been considered as one of the representative model organisms in aging research. Previous studies have indicated that S. cerevisiae and C. albicans, the two well-studied yeast, bear many resemblances genetically and biochemically (Berman & Sudbery, 2002). In our current work, we characterized C. albicans, the polymorphic fungus that has a finite lifespan (Fig. 1B). Like S. cerevisiae, the old C. albicans exhibits some aging-associated phenotypes, including morphologically larger in size (Fig. 2C), accumulation of glycogen and oxidatively damaged proteins (Fig. 3), and down-regulation of nascent protein synthesis (data not shown). In addition, the old Candida cells show no changes in telomere length and chromosome integrity (Figs S2 and S3). Moreover, C. albicans appears to share some conserved aging pathways with S. cerevisiae. For example, calorie restriction prolongs lifespan in C. albicans (Fig. 4A). Decreasing the dosage of Sir2 shortens lifespan, and overexpression of Sir2 leads to lifespan extension in C. albicans (Fig. 4B). These observations indicate that C. albicans could be considered as a new model system for aging study.
In C. albicans, deletion of one copy of SIR2 affects the rDNA silencing (Fig. 4C), indicating that like S. cerevisiae, Candida SIR2 is involved in the maintenance of silent chromatin. In S. cerevisiae, it has been shown that Sir2 may delay aging by repressing recombination in the rDNA loci so as to inhibit the formation of ERCs (Gottlieb & Esposito, 1989;. Interestingly, in C. albicans, the reduction of rDNA silencing in sir2Δ/SIR2 mutant does not result in an increase of extra-chromosomal rDNA molecules (Fig. 4D). We speculate that the mechanism of regulating cellular aging by Sir2 in C. albicans may not be exactly the same as in S. cerevisiae. In fact, although the longevity-promoting role of Sirtuin family has been proved in a variety of species, including S. cerevisiae, C. elegans, and Drosophila melanogaster Tissenbaum & Guarente, 2001;Rogina & Helfand, 2004), the extra-chromosomal rDNA molecule (circle) associated aging has not been observed in other organisms except S. cerevisiae. Therefore, it remains elusive whether the molecular mechanism by which the extra-chromosomal rDNA molecules (circles) cause aging is evolutionarily conserved. This idea is also supported by the observation that extrachromosomal rDNA molecules do not accumulate in the aging C. albicans cells in either wild-type or sir2 mutant strains (Fig. 5A). In addition, plasmids are indeed not stable in Candida cells (reviewed in De Backer et al., 2000). Thus, we postulate that the decrease of rDNA silencing in sir2Δ/SIR2 cells accelerates aging through a mechanism other than promoting rDNA recombination to form extra-chromosomal rDNA molecules (circles) (Kennedy et al., 1997;. Although it is not clear how Sir2 affect C. albicans longevity, oxidative damage accumulation seems to be a function of replicative age in C. albicans. It has been proved in S. cerevisiae that mother cells retain oxidatively damaged proteins during cell division, and the uneven distribution of these proteins between mother and daughter cells is not due to the different rate in production of reactive oxygen species or degradation of carbonylated proteins, but a result of asymmetric segregation of the oxidized proteins during cytokinesis in a Sir2-dependent manner (Aguilaniu et al., 2003). Given that in C. albicans, oxidatively damaged proteins accumulate with replicative aging (Fig. 3B), and evenly distribute in mother and daughter cells when Sir2 is absent (Fig. 6A), Candida Sir2 acts more likely as Saccharomyces Sir2 to protect the daughters from inheriting oxidative damage accumulated in the aged mothers. Therefore, Sir2 may promote longevity in C. albicans through regulating the distribution of oxidatively damaged proteins.
Besides the similar aging-associated phenotypes between C. albicans and S. cerevisiae, C. albicans is able to provide some advantages, potentiating it a complementary model. To prepare large amount of old cells seems to be a bottleneck for agingrelated biochemical or genomics/proteomics study. Because C. albicans cells can grow as yeast form and hyphal form, which are dramatically different in shape and size, someone would be able to obtain a large amount of old cells. The method we have established for acquirement of old cells is relatively simple and effective. In addition, the separation procedure allows harvesting both young and old cells at discretionary time points, and provides an opportunity to perform the phenotypic time-course analysis of aging progression.
In conclusion, we have established a new cellular aging model of C. albicans, which allows large-scale isolation of replicative old cells. Our work also suggests that the extra-chromosomal rDNA molecules may not be associated with aging in C. albicans. This novel aging model will facilitate to find out novel aging markers and investigate the mechanisms of cellular aging.

Strains and growth conditions
The C. albicans strains used in this study are SC5314 (wild-type) and its derivative BWP17 (ura3Δ::λimm434/ura3Δ::λimm434 his1::hisG/his1::hisG arg4::hisG/arg4::hisG). To make sgs1Δ/ SGS1 or sir2Δ/SIR2 strain, one copy of SGS1 or SIR2 was deleted, respectively, in BWP17 using a PCR-based gene disruption method (Wilson et al., 1999). The sir2Δ/sir2Δ strain was generated similarly on the basis of sir2Δ/SIR2 strain. BES116 plasmid was used to construct the Sir2 overexpression strain (Feng et al., 1999). The plasmid contains two ADE2 homologous arms (0.6 kb and 0.9 kb, respectively) for the recombination at ADE2 locus and a URA3 gene for nutrition selection. Integration of SIR2 at ADE2 locus was accomplished by transforming BWP17 with the linear fragment of BES116-SIR2 digested with AscI. In addition to the entire coding region of SIR2, ~1000 bp of upstream sequence and ~500 bp of downstream sequence are included. A vector control was obtained by transforming the AscI digested product of BES116 into BWP17. Yeast cells were propagated in YPD medium at 30 °C for routine culture. Hyphal cells were induced with an initial density of 10 6 yeast cells mL -1 and cultured in YPD plus 15% bovine serum for 5 h at 37 °C.

Replicative lifespan analysis
Replicative lifespan assay of yeast cells was performed as described previously (Kaeberlein et al., 2004). Unless otherwise specified, the replicative lifespan experiments were performed on yeast form cells. Lifespan of hyphal cells was determined as follows. Hyphal cells were separated from liquid culture (refer to the 'Large-scale preparation of old and young cells' section) and subjected to lifespan analysis immediately. A cohort of randomly selected hyphal cells was patched onto the lifespan plate and arrayed using a Singer MSM micromanipulator (Singer Instruments, Roadwater, Watchet, Somerset, UK). The budded yeast daughter cells from hyphal mothers were removed and counted every one to two generations. If a hyphal mother contains more than one cell, the average number of daughter cells was calculated. The plates were incubated at 30 °C during the day and stored at 4 °C at night. All lifespan experiments were carried out on standard YPD plates (2% glucose) except for CR analysis. Each experiment has been repeated at least twice, and data shown represent single experiment results. Statistical significance was determined by Wilcoxon rank-sum test, and significant differences were stated for p < 0.05.

Calcofluor staining
The

Separation of hyphal and yeast cells
Discontinuous sucrose density gradient was made by layering 30 mL of 30% sucrose at the bottom of a 50 mL centrifuge tube (Corning Inc., Corning, NY, USA) under 15 mL of distilled H 2 O. Five milliliter of a mixture of hyphal and yeast cells was loaded onto a sucrose gradient, and centrifuged at 150 g for 3 -5 min. Hyphal cells stayed at the bottom of the tubes, and yeast cells stayed in the upper layer of the gradient. The upper layer supernatant was collected to pellet yeast cells by centrifugation (3000 g, 3 min). The bottom pellet was visualized under a Nikon Eclipse E200 (Tokyo, Japan) microscope to check the purity of separated hyphal cells. More sucrose gradient centrifugation(s) would be performed until no dissociative yeast cells were mixed in the hyphal population. Hyphal cells were finally pelleted by centrifugation (3000 g, 3 min).

Large-scale preparation of old and young cells
A fresh colony was inoculated into YPD medium and incubated at 30 °C overnight. Hyphal cells were induced in the presence of 15% serum at 37 °C for 5 h, and then cultured in YPD for 2.5 h at 30 °C to produce yeast form cells. The newborn yeast cells in this short time were harvested by sucrose gradient centrifugation and used to induce the growth of young hyphae. The young hyphae were cultured in YPD at 30 °C to produce yeast form daughters continuously. During the cultivation sucrose gradient centrifugation was employed to get rid of yeast daughter cells, and hyphal mother cells were transferred to fresh YPD every 10 -12 h. At the indicated time points, hyphal cells were separated thoroughly and subjected to replicative lifespan assay immediately.

Iodine staining of glycogen
Hyphae of different ages were sonicated to break into single cells and then counted using hemacytometer. About 5 × 10 5 cells for each sample were dotted on GF/C filters (Whatman International Ltd, Maidstone, UK) triply. The filters were stained for 2 min by exposing to iodine vapor at 37 °C, and photographed with a digital camera. Images were quantitated with Bio-Rad Quantity-One program (Bio-Rad Laboratories, Hercules, CA, USA).

Determination of oxidatively damaged proteins
Total proteins were prepared using the glass beads protocol, and the protein concentrations were determined by Bradford (Bio-Rad Laboratories Inc., Hercules, CA, USA). Oxidative proteins were reacted with 2,4-dinitrophenylhydrazine in sodium dodecyl sulfate (Levine et al., 1994) and detected by Western blot with rabbit anti-2,4-dinitrophenyl antibody (Zymed Laboratories, South San Francisco, CA, USA).

Separation of C. albicans chromosomes
The whole chromosomes were separated by PFGE as described previously (Chen & Fonzi, 1992).

Detecting extra-chromosomal rDNA molecules and determination of rDNA repeat number
The extra-chromosomal rDNA molecules were detected by PFGE using a CHEF-DRIII system (Bio-Rad Laboratories). About 10 7 cells were harvested and digested with 40 μL of 2 mg mL -1 Zymolyase 20 T (MP Biomedicals, LLC, Solon, OH, USA) at 37 °C, then mixed with 40 μL of 2% low-melting agarose (cool to 50 °C). Plugs were prepared at 4 °C by adding 80 μL of the samples into a mold chamber, then treated with LET buffer (0.5 M EDTA, 0.01 M Tris (pH 7.5), > 0.5% β-mercaptoethanol), NDS buffer (0.5 M EDTA, 0.01 M Tris (pH 7.5), 1% N-laurylsarcosine, 1 mg mL -1 Protease K) and 0.05 M EDTA for 16-24 h, respectively. The plugs were then filled in a 0.8% (w/v) agarose gel for electrophoresis. The running conditions for detecting extra-chromosomal rDNA molecules are as following: 0.5 × TBE buffer at 14 °C; constant voltage of 3.6 V cm -1 ; a 120S pulse interval for 20 h. DNA was visualized by ethidium bromide staining. After capillary transfer to Hybond-N+ membrane, the rDNA signals were detected by hybridization with a 32 P labeled 1.7 kb PCR product of C. albicans 18S subunit of rDNA (forward primer: 5′-GGT TGATCCTGCCAGTAGTC-3′; reverse primer: 5′-GTTCACCTACGGAAACCTTG-3′). To determine the copy number of rDNA, genomic DNA was mini-prepared and digested with EcoRV and SalI, and then separated on a 0.9% agarose gel. DNA was blotted to Hybond-N+ membrane and probed with rDNA. Subsequently the membrane was stripped and hybridized to a 32 P labeled 1.1 kb PCR product of C. albicans ACT1 (forward primer: 5′-ATGGACGGTGAAGTT-GCTGC-3′; reverse primer: 5′-GTGGTGAACAATGGATGGACC-3′) for normalization. The copy number of rDNA was determined with Image Quant software. The young cells are assumed as 55 rDNA copies per cell (Jones et al., 2004).

Immunofluorescent detection of oxidatively damaged proteins
The oxidatively damaged proteins were detected as described in (Aguilaniu et al., 2003). Cells were digested with Zymolyase in the digestion buffer (50 mM Tris-HCl, pH 7.5; 10 mM MgCl 2 ; 1 M Sorbitol). Primary and secondary antibodies were Rabbit anti-DNP (Zymed Laboratories) and Cy TM 3-conjugated Affini-Pure Donkey Anti-Rabbit IgG (H+L) (Jackson ImmunoResearch Laboratories, West Grove, PA, USA), respectively. The fluorescence density was quantified using Image-Pro Plus software.

Supporting Information
Additional supporting information may be found in the online version of this article.

Fig. S1
Hyphal cells are separated from yeast cells by sucrose gradient centrifugation. The photo was taken under a Zeiss Axioplan 2 microscope.   Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the Aging Cell Central Office.