Rb analog Whi5 regulates G1 to S transition and cell size but not replicative lifespan in budding yeast

Abstract

An increase in cell size with age is a characteristic feature of replicative aging in budding yeast. Deletion of the gene encoding Whi5 results in shortened duration of G1 and reduced cell size, and has been previously suggested to increase replicative lifespan. Upon careful analysis of multiple independently derived haploid and homozygous diploid whi5Δ mutants, we find no effect on lifespan, but we do confirm the reduction in cell size. We suggest that instead of antagonizing lifespan, the elongated G1 phase of the cell cycle during aging may actually play an important role in allowing aged cells time to repair accumulating DNA damage.

Introduction

Appropriate regulation of the cell cycle is essential for cellular and organismal fitness. The G1 to S phase transition represents a critical point in the cell cycle of budding yeast Saccharomyces cerevisiae that is tightly regulated [1]. This regulation ensures that daughter cells achieve optimal cell volume and biomass prior to replication and transition to becoming a mother cell. This transition is regulated by the transcriptional repressor Whi5 (analogous to mammalian Rb), which when localized to the nucleus acts to inhibit the heterodimeric transcription factor SBF (Swi4/Swi6) [2]. Upon reaching the critical cell size, the Cln3/Cdc28 complex phosphorylates Whi5, which drives Whi5 out of the nucleus, alleviating SBF inhibition and promoting the G1 to S transition. Deletion of Whi5 reduces the length of G1 in the first cell division of a virgin daughter cell, and causes cells within the population to be smaller in size due to prematurely transitioning from G1 into S phase during the first cell cycle [3].

Budding yeast have served as a useful model for cellular aging, with several genetic factors and molecular mechanisms having been found to be shared between aging in yeast and higher eukaryotes [4, 5]. Replicative lifespan in yeast is defined as the number of daughter cells produced by a mother cell prior to irreversible cell cycle arrest [6, 7]. Replicative lifespan varies depending on the strain background from around 15 generations in the short-lived X2180–1A haploid background [8] to around 29 generations in the long-lived S288c MATα haploid background [9]. The yeast ORF deletion collection is derived from the S288c background with the BY4741 and BY4742 strains defining the MATa and MATα haploids, respectively [10]. Both of these strains have average lifespans of about 26–27 generations [11]. Interestingly, the corresponding diploid BY4743 strain has an average lifespan of about 33–39 generations, significantly longer than either isogenic haploid derivative [11–13]. The mechanisms underlying this effect of ploidy on lifespan remain unknown.

One of the most striking morphological characteristics of replicative aging in yeast is a dramatic increase in cell size accompanied by prolonged G1 [14]. It has been speculated that this increase in cell size may limit replicative lifespan [15, 16] and some long-lived mutants have reduced cell size [17, 18]. This relationship remains correlative, however, and direct causal evidence linking cell size to longevity are lacking. Two recent reports showed that cell size at birth was only weakly correlated with replicative lifespan in wildtype cells [12, 19], and both studies concluded that the data do not support a model where a maximal cell size limits lifespan in wildtype yeast. One recent report created a strain with reduced replicative lifespan by using a synthetic circuit to cause an extended G1 arrest and a dramatic increase in cell size [20]. The authors interpreted this to support a model where cell size limits replicative lifespan during normative aging, but in order to achieve a significant reduction in replicative lifespan, cell volume was increased several-fold above what cells reach during normal aging.

To further explore the relationship between cell size and longevity, we examined the effect of deleting Whi5 on replicative lifespan in the BY4741, BY4742, and BY4743 strains. It was previously reported that WHI5 deletion increases lifespan in the deletion set background [16]. We were unable to reproduce this finding. Instead we observed that deletion of WHI5 has no significant effect on lifespan in either haploid or diploid cells by testing multiple genotyped and phenotypically confirmed isolates. As expected, deletion of WHI5 did result in a reduction both in cell size and G1 duration. Thus, our data are inconsistent with the model that cell size or regulation of G1 to S transition by Whi5 limits yeast replicative lifespan.

Results

To understand whether cell size plays a limiting role during replicative aging, we examined the effect of deleting WHI5 on lifespan, as whi5Δ cells are significantly smaller than wild type cells [3]. One prior report suggested that deletion of WHI5 in the BY4743 diploid background increased replicative lifespan [16]; however, we noted that the control used in that study was substantially shorter-lived than expected, based on prior reports [11–13] and our own experience (Figure 1). In that report, the putative diploid BY4743 control strain had a mean lifespan of 26.5 generations, within the range expected for a haploid strain in this background (Figure 1). The whi5Δ/whi5Δ diploid strain used in the prior study had a mean lifespan of 34.5 generations [16], which is comparable to what other labs have reported for the diploid BY4743 control strain as well as our own historical data on over 1,500 diploid mother cells (Figure 1).

Figure 1:

Deletion of WHI5 reduces cell size and G1 duration, but does not increase lifespan in diploid budding yeast.

A) Overlay of historical data from our lab with lifespan traces from previously published work on whi5Δ/whi5Δ diploid cells [16]. The diploid control lifespans from prior work overlap with our historical haploid lifespans.

B) Newly generated diploid whi5Δ/whi5Δ cells exhibit expected reduction of cells in G1 and an increase in cells in G2/M. Log-phase cells were fixed and stained with Sytox Green to visualize DNA content by flow cytometry (n=10,000 both strains).

C) Newly generated diploid whi5Δ/whi5Δ cells exhibit expected reduction in cell size. Cell sized determined by forward scatter (FSC-A) using flow cytometry of log-phase cells. The whi5Δ/whi5Δ diploid cells are significantly different from BY4741 cells as determined by Student’s t-test with p<0.0001 (n=10,000 both strains).

D) Compared with experimentally matched BY4743 cells, the whi5Δ/whi5Δ mutants do not have an increase in replicative lifespan (p=0.20 Wilcoxon rank sum test).

Given the apparent inconsistencies in lifespan data from the prior study, we independently tested the effect of deleting WHI5 in the BY4743 diploid background. We first generated whi5Δ knockouts in both the BY4741 MATa and BY4742 MATα haploid mating types. We confirmed deletion of WHI5 by PCR amplification and observed the predicted whi5Δ phenotypes of reduced cell size and shortened G1 in both mating types (see Materials and Methods). We then generated diploid whi5Δ/whi5Δ cells by mating and confirmed that diploid whi5Δ/whi5Δ cells also showed the predicted changes in cell cycle (Figure 1B) and cell size (Figure 1C). We next carried out lifespan analysis and found no significant difference in lifespan between BY4743 wild type and experiment-matched whi5Δ/whi5Δ homozygous deletion cells (Figure 1D). We speculate that the control strain used in the prior study may have been a haploid strain.

We next set out to determine whether deletion of WHI5 extends lifespan in haploid cells. We derived several independent whi5Δ strains in both the BY4742 and BY4741 haploid mating types. After confirming their genotype by PCR and phenotypically confirming the reduction in both cell size (Figure 2A) and G1 duration (Figure 2B), we measured replicative lifespan for five independently-derived lines. In every case, there was no significant difference between the corresponding wild type strain and the whi5Δ deletion strain (Figure 2C–G). When all haploid whi5Δ mutants are pooled with the experiment-matched wildtype controls, there is no significant change in replicative lifespan (Figure 2H).

Figure 2:

Deletion of WHI5 reduces cell size and G1 duration in haploid yeast, but does not increase replicative lifespan.

A) When compared with wild-type haploid yeast (BY4741), all of the whi5Δ haploid strains have a significant reduction in cell size. Cell sized determined by forward scatter (FSC-A) using flow cytometry of log-phase cells. (n=10,000 all strains, p<0.001 all strains, Student’s t-test).

B) When compared with wild-type haploid yeast (BY4741), all of the whi5Δ haploid strains have a reduction in the fraction of cells in G1. Cell cycle stage determined using log-phase asynchronous cultures fixed, and stained with Sytox Green to determine DNA content (n=10,000 all strains).

C-G) Lifespan curves for five different whi5Δ haploid strains. (panel C n=300 WT, n=299 whi5Δ; panels D-G n=40 WT, N=40 whi5Δ). p-value determined by Wilcoxon rank sum test.

H) When all haploid whi5Δ experiments are pooled with the experimentally matched haploid wildtype cells, there is no difference in replicative lifespan (p>0.05 Wilcoxon rank sum test).

Discussion

Taken together, our data indicate that Whi5 does not play a major role in determining replicative lifespan in either haploid or diploid cells of the ORF deletion collection background. This contradicts a prior report for diploid cells [16], but the control strain in that report was abnormally short-lived (Figure 1A), which may have contributed to the apparent lifespan extension. A pre-print posted on BioRxiv reported that the whi5Δ strain from the MATa diploid deletion collection was long-lived relative to the parental strain [21]. This experiment was performed using a microfluidic device that traps cells based on cell size. It is possible that differences between microfluidic lifespan analysis and microdissection, which was employed here, may contribute to different effects of WHI5 deletion on lifespan, and we are performing additional studies to address this possibility.

It has long been known that a majority of wild type mother cells senesce in G1 [13, 22]. This observation has led some to propose that increased G1 durations during aging are maladaptive and that G1 arrest is the result of the Whi5 circuit going awry [21, 23]. These interpretations should be reconsidered in light of our data indicating that deletion of WHI5 does not significantly extend lifespan. Another report found that while overexpression of CLN2 reduced both the cell size and the G1 duration it also failed to extend lifespan [23]. This is consistent with our model that cell size is not a primary factor limiting lifespan in yeast.

An alternative model, which we favor, is that the cell cycle elongation during aging is an adaptive response to repair age-associated damage. In support of this, we note that individual wild type cells which terminally arrest in G1 are longer lived than those that arrest outside of G1 [13], while in short-lived mutants with constitutively high levels of genome instability a majority of the cells fail to arrest in G1 [22]. Intriguingly, the reduced G1 duration resulting from CLN2 overexpression leads to an improved ability in older cells to respond to DNA damage and perform single strand annealing [24]. That this improved DNA damage response fails to increase replicative lifespan suggests that the G1 lengthening during aging may serve a purpose that has yet to be identified. For example, a short G1 phase imposes constitutive replication stress in cycling stem cells, and delaying the G1/S transition results in reduced fork slowing and reversals during S phase [25]. Thus, a longer G1 may reduce the chances of DNA damage during S-phase even while it reduces DNA repair capacity. It is also noteworthy that, while many long-lived mutants have reduced cell size, they are smaller because of reduced mRNA translation [26], which unlike deletion of WHI5 results in a longer G1 period [27, 28]. Thus, the reason for a reduction in cell size in the long-lived mutants, as well as the length of G1 in those mutants, is fundamentally different from whi5Δ cells, which are not long-lived. Of course, the fact that deletion of WHI5 does not substantially shorten lifespan, suggests that Whi5 is also unlikely to play a major role as a protective factor during aging of wild type cells. It will be of interest to determine whether WHI5 deletion has effects on lifespan in other genetic backgrounds with altered levels of age-related genome instability.

Materials and Methods

Yeast strains, culture conditions, and strain validation

The whi5Δ was generated from a BY4742 strain by knocking out the WHI5 gene and replacing it with a URA3 selection marker. This was confirmed by colony PCR and the reduction in cell size by flow cytometry. MC435–438 were generated by crossing JD156 with BY4741, sporulating the resulting diploid, and dissecting tetrads to obtain individual spore clones. Deletion of WHI5 in these strains was confirmed by selectable marker, PCR genotyping, and phenotyping (cell size and cell cycle progression). Complete list of strains available in Table 1. All flow cytometry was performed using a FACS CantoII. Determination of cell size using flow cytometry was done using forward scattering which is a widely used measurement that scales with cell volume [29]. Cells were inoculated into YPD for overnight growth, and then diluted 1:200 the following morning and allowed to grow for 5h. Determination of DNA content and G1 duration was determined by fixing and staining for DNA content using Sytox Green as detailed in [30]. PCR genotyping was done using the following primers:

• WHI5 3’UTR CHK R TCGATCAACAAAACCTTGTC

• WHI5 5’UTR CHK F TTACGGGAGAGAGTCTTGTG

• WHI5 WT CHK F AGCAGGACTAGCGATGAAC

• WHI5 WT CHK R TCTTCTTCATTTTCGTCGTC

Table 1.

Strains used in this study.

Strain Name	Genotype	Source
BY4741	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0	ATCC
BY4742	MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0	ATCC
BY4743	MATa/α his3Δ/his3Δ leu2Δ/leu3Δ lys2Δ/LYS2 ura3Δ/ura3Δ met15Δ/MET15	ATCC
JD156	MATα whi5Δ::URA3 his3 leu2 ura3 MET15 lys2	this study
JD157	MATα whi5Δ::URA3 his3 leu2 ura3 MET15 lys2	this study
MC435	MATα whi5Δ::URA3 lys2 met15 his3 leu2 ura3	this study
MC436	MATa whi5Δ::URA3 MET15 lys2 his3 leu2 ura3	this study
MC437	MATα whi5Δ::URA3 lys2 met15 his3 leu2 ura3	this study
MC438	MATa whi5Δ::URA3 LYS2 MET15 his3 leu2 ura3	this study
MC476	MATa/α whi5Δ::URA3/ whi5Δ::URA3 his3/his3 leu2/leu2 met15/MET15 lys2/LYS2	this study
MC477	MATa/α whi5Δ::URA3/ whi5Δ::URA3 his3/his3 leu2/leu2 met15/MET15 lys2/LYS2	this study

Replicative lifespan analysis

Microdissection experiments to determine replicative lifespans were done as previously described [31, 32]. Briefly, cells were patched onto YPD plates and allowed to grow overnight. Then, cells were arrayed, and virgin daughters were selected for use in the lifespan. New daughters were manually removed from mothers until mother cells died.