Hypertrophy hypothesis as an alternative explanation of the phenomenon of replicative aging of yeast


Correspondence: Tomasz Biliński, Department of Biochemistry and Cell Biology, University of Rzeszów, Zelwerowicza 4, 35-601 Rzeszów, Poland. Tel.: +48 17 7855406; fax: +48 17 8721265; e-mail: tomasz.bilinski91@gmail.com


This paper summarizes numerous arguments demonstrating that the hypothesis of accumulation of the senescence factor, which was the basis for introducing yeast to the group of model organisms of gerontology, finds no experimental support. Among several candidates for the role of the causative agents of replicative aging, only one – hypertrophy – always accompanies symptoms of aging, not only in Saccharomyces cerevisiae, but also in Schizosaccharomyces pombe.

The phenomenon known as replicative aging was discovered more than 50 years ago (Mortimer & Johnston, 1959). It was shown that a single yeast cell is able to form only a limited number of daughter cells during its life. It was postulated that this limit is a consequence of accumulation of bud scars on the cell wall of the yeast mother cell. The original explanation of the phenomenon called yeast replicative aging did not raise interest the of gerontologists, because of its group-specific character. Yeast became a model organism of gerontology when the group of S.M. Jazwinski postulated that accumulation of the hypothetical ‘senescence factor’ is a causative factor of replicative aging of yeast (Egilmez & Jazwinski, 1989).

Since then, it has been assumed by a considerable part of the research community that studies of replicative aging of the yeast should unravel basic mechanisms governing aging of higher organisms, including humans. The tacit assumption was that the cessation of budding in the yeast is a general model of aging and the hypothetical ‘senescence factor(s)’ to be revealed in the yeast can account for aging of higher organisms (Steinkraus et al., 2008; Kaeberlein, 2010; Unal et al., 2011). However, having in mind that the essence of science is critical testing and not believing, let us ask whether this assumption is valid?

Observations made during the first decade of the studies revealed the existence of the following new age-related phenotypes:

  1. Daughters of young mothers have identical replicative life spans (rls) as the their mothers; (their ‘cell division counter’ is reset to zero) (Jazwinski et al., 1989; Kennedy et al., 1994);
  2. Daughters formed at the end of life of mothers have shortened rls (incomplete reset of the ‘cell division counter’) (Jazwinski et al., 1989; Kennedy et al., 1994);
  3. At the end of their life, some mothers may form daughters of size equal to the size of the mother cell (‘symmetric division’). Rls of both mothers and daughters is the same and equal to the residual rls of the mother cell (Jazwinski et al., 1989; Kennedy et al., 1994);
  4. Daughters born with shortened rls can produce their daughters and granddaughters, which have normal life span. Therefore, this symptom is not a consequence of irreversible damage (Jazwinski et al., 1989);
  5. Cells approaching their reproduction limit reach a very big size (become hypertrophic) (Mortimer & Johnston, 1959; Egilmez & Jazwinski, 1989).

To explain the origin of these phenotypes, it has been postulated by Jazwinski et al., that the senescence factor is soluble, diffusible, and degradable, at least in daughters of young mothers.

Papers concerning replicative aging of yeast postulated (Sinclair et al., 1998ab; Bitterman et al., 2003; Henderson & Gottschling, 2008) that the senescence factor is a molecule or cellular structure, which exerts negative effects on the cell and its accumulation results eventually in cell death. It also has a universal character, accounting also for aging of humans.

Various groups of investigators postulated that they had discovered the nature of the factor. The most commonly postulated senescence factor could be rDNA circles (Sinclair & Guarente, 1997; Sinclair et al.,1998ab). However, the accumulation of rDNA circles is a group-specific phenomenon, similar to accumulation of bud scars, because these episomes do not accumulate in animal or human cells. Additionally, the inability to form rDNA circles, instead of extending life span of rad52 mutants, shortens it (Park et al., 1999). Similar conclusions drawn from other observations (Heeren et al., 2009) and theoretical considerations (Gillespie et al., 2004; Steinkraus et al., 2008) make the crucial role of these episomes doubtful. Other candidates for the role of the senescence factor are as follows: protein aggregates, oxidatively damaged proteins and aged mitochondria (Lai et al., 2002; Aguilaniu et al., 2003). However, hypoxia or anoxia does not influence the replicative life span of yeast cells (Wawryn et al., 2002). Twenty years after formulation of the ‘senescence factor’ hypothesis, its protagonists suggest (Henderson & Gottschling, 2008) that its detection and identification will be possible in the future, thus admitting that the causative role of any of the postulated ‘senescence factors’ has not been finally documented.

Nobody questions that these structures accumulate. However, neither negative effects of various noxious factors on the cell nor their accumulation within this cell is proof that they are responsible for replicative aging of individual yeast cells.

Discovery that replicative aging is also observed in another yeast species Schizosaccharomyces pombe casts doubt on whether its senescence factor is of the same nature. The conclusion of these studies (Barker & Walmsley, 1999) is that the inability of these cells to divide is associated with the weakening of cell wall structures, which are neither soluble nor diffusible and hardly degradable. Therefore, both universal and molecular character of the yeast senescence factor postulated in most studies seems doubtful.

After several years of studies, some basic principles of yeast gerontology studies have been challenged. First, it has been shown that the yeast ‘survival curves’ are fecundity curves (Gershon & Gershon, 2000). Fecundity, however, correlates rather negatively with longevity (Kirkwood & Holliday, 1979; Shanley et al., 2007). Secondly, the term per cent survival used on the ordinates is also incorrect; it is really per cent of cells able to reproduce. However, the cells that lost their ability to reproduce are still alive (Minois et al., 2005; Zadrag et al., 2008). As a consequence, mutants that are able to produce more daughter cells than the corresponding parent strain should not be called ‘longevity’ mutants.

The senescence factor hypothesis cannot explain the origin of most phenotypes listed, without creating auxiliary hypotheses, and it ignores phenotype 5. However, it is the only phenotype that always accompanies the symptoms of replicative aging. According to the rules of methodology of sciences, the natural phenomenon, which always accompanies the studied one, is most probably its causative factor if other accompanying phenomena can be eliminated without elimination of the phenomenon studied (Ducheyne, 2008). We have postulated (Bilinski & Bartosz, 2006; Zadrag et al., 2006) that simply reaching excessive volume by yeast cells at the end of their reproductive period is responsible for their inability to undergo further reproduction. The only a priori assumption, on which all hypotheses are based, is that the level of any senescence factor cannot increase infinitely. Therefore, also cells with too high a volume may lose their homeostasis and stop budding. Our conclusions were confirmed by the results of recent studies in which relationships between cell size, the rate of cell growth in successive reproductive cycles and rls have been subject to extensive analysis (Yang et al., 2011).

Our hypothesis of hypertrophy (illustrated in Fig. 1) is founded on the following experimental data.

  1. Virgin yeast cells have to increase their volume to a certain level, before entering their first cell cycle (Hartwell & Unger, 1977). We propose to call this minimal volume the Hartwell's threshold volume.
  2. The cells of strains of the same genetic background stop budding after reaching identical volume, irrespective of the number of completed cell divisions (Zadrag et al., 2006; Zadrag-Tecza et al., 2009; Yang et al., 2011). We propose to call this volume the maximal volume.
  3. The yeast cells increase their volume during each cell cycle (Hartwell & Unger, 1977; Woldringh et al., 1993; Zadrag et al., 2005; Zadrag-Tecza et al., 2009).
  4. The volume of the daughter cells increases during the consecutive buddings, but at a much lower rate than the volume of mother cells (Zadrag-Tecza et al., 2009), except for the so-called symmetric cell division (Kennedy et al., 1994).
Figure 1.

Volume growth of yeast mother cell and volume of daughters produced in successive cell cycles of the mother cell. As long as the volume of the daughter cell is lower than the Hartwell's threshold volume, the cell must grow to enter the first cell cycle and retains the full replicative lifespan. The size of daughters increases with the number of cell cycles accomplished by the mother cell. At some point (cycle 23 on the example shown), the volume of the daughter cell is equal to the Hartwell's threshold volume. Daughter cells produced in subsequent cycles have initial volume above this value and reach the maximal volume after smaller number of cell cycles, therefore their replicative lifespan is shortened. Scheme drawn on the basis of cell volumes measured experimentally (Zadrag-Tecza et al., 2009).

Yeast cells are able to reproduce, when their volume is confined between the Hartwell's threshold and maximal volumes. Thus, the numeric value of rls, being the measure of yeast cell longevity, is a quotient of the difference between maximal and the Hartwell's threshold values and the increase in the cell volume per one cell cycle.

How does the hypertrophy hypothesis explain known age-related phenotypes?

Phenotype 1. Resetting the ‘cell division counter’. Daughter cells formed by young mothers have volumes below or equal to the Hartwell's threshold volume. Therefore, irrespective of their initial volume they all enter their first S phase of the cell cycle after reaching this threshold value. Taking into account that maximal volume is constant for cells of the same genetic background (Zadrag et al., 2005, 2006; Zadrag-Tecza et al., 2009; Yang et al., 2011), the maximal number of daughter cells formed (maximal rls) depends mainly on the rate of cell volume increase during one cycle.

Phenotype 2. At the end of their reproduction period, mothers produce daughters, whose size exceeds the Hartwell's threshold volume. They are born as large as their mother cells and therefore the number of cell cycles they can perform before reaching maximal volume is lower (‘incomplete reset of cell division counter’).

Phenotype 3. Daughter cells formed during ‘symmetric cell divisions’ have life spans, identical to that of their mothers, simply because their volumes are identical.

Phenotype 4. Cells with shortened life span at the moment of their separation from old mothers can generate daughters and granddaughters with normal life span, because their cell volume is reduced as compared to their mothers during each cell cycle (Jazwinski et al., 1989; Kennedy et al., 1994; Zadrag-Tecza et al., 2009) (asymmetric cytokinesis) Consequently, after two cycles, the volume of the descendent cells drops to, or below, the Hartwell's threshold value.

Phenotype 5. An unavoidable increase in volume of the mother cell during each cell cycle (Hartwell & Unger, 1977) is a direct consequence of budding as the mechanism of cytokinesis in Saccharomyces cerevisiae, whereas in Schizosaccharomyces pombe, hyperthrophy is a consequence of the occurrence of asymmetry of cell fission.


The idea that the increase in size of yeast cells is the cause of replicative aging of this organism is not new. This observation accompanies all studies, and the fact of reaching exceptionally large volume by cells approaching their reproductive limit is accepted by all experimenters. The same phenomenon was also observed during replicative aging of S. pombe (Barker & Walmsley, 1999).

The question arises as to why the role of hypertrophy, the only phenomenon always accompanying replicative aging, is never discussed in scientific papers? The authors of one experimental paper (Kennedy et al., 1994) rejected this explanation, because size increase in haploid MATa cells by treatment with alpha pheromone for 4 h did not change their rls significantly. They concluded that ‘an increase in cell size does not necessarily lead to a shortening in life span’. However, the results of our studies brought quite different results, fully supporting the hypertrophy hypothesis (Zadrag et al., 2005, 2006; Zadrag-Tecza et al., 2009). Moreover, experimental elevation of cell volume by nocodazole treatment or by growing the cdc28 mutant at restrictive temperature (Yang et al., 2011) also shortened rls.

The hypertrophy hypothesis seems to explain the phenomenon of replicative aging of yeast much better than the hypothesis of the senescence factor accumulation. The same conclusion could be drawn from recent studies (Yang et al., 2011). The hypertrophy hypothesis proposed by us explains more phenomena than the previous one, without the necessity of creating any auxiliary hypotheses. It means that the hypertrophy hypothesis meets better the standard of simplicity of William of Ockham. Its predictive power is also the highest. These two arguments suggest that the hypertrophy hypothesis should be at least taken into account as worth serious reconsideration.

Further studies are necessary to explain the molecular nature of the processes going on in oversized cells, including those of mammalian cells which also become oversized when grown in vitro.


This work was supported by Grant No. N N303 376436 of the Polish Ministry of Science and Higher Education.