Growth-dependent signals drive an increase in early G1 cyclin concentration to link cell cycle entry with cell growth

Version of Record

Accepted for publication after peer review and revision.

Download
Cite
Share
CommentOpen annotations (there are currently 0 annotations on this page).

Version of Record published: November 15, 2021 (This version)
Accepted Manuscript updated: November 1, 2021 (Go to version)
Accepted: October 29, 2021
Accepted Manuscript published: October 29, 2021 (Go to version)
Received: October 27, 2020
Preprint posted: October 1, 2020 (Go to version)

1. Of interest
Atypical peripheral actin band formation via overactivation of RhoA and nonmuscle myosin II in mitofusin 2-deficient cells

Yueyang Wang, Lee D Troughton ... Qing Deng

Research Article Oct 4, 2023
Further reading

Abstract
Introduction
Results
Discussion
Materials and methods
Data availability
References
Decision letter
Author response
Article and author information
Metrics

Abstract

Entry into the cell cycle occurs only when sufficient growth has occurred. In budding yeast, the cyclin Cln3 is thought to initiate cell cycle entry by inactivating a transcriptional repressor called Whi5. Growth-dependent changes in the concentrations of Cln3 or Whi5 have been proposed to link cell cycle entry to cell growth. However, there are conflicting reports regarding the behavior and roles of Cln3 and Whi5. Here, we found no evidence that changes in the concentration of Whi5 play a major role in controlling cell cycle entry. Rather, the data suggest that cell growth triggers cell cycle entry by driving an increase in the concentration of Cln3. We further found that accumulation of Cln3 is dependent upon homologs of mammalian SGK kinases that control cell growth and size. Together, the data are consistent with models in which Cln3 is a crucial link between cell growth and the cell cycle.

Introduction

Cell cycle progression is subservient to cell growth. Thus, key cell cycle transitions occur only when sufficient growth has occurred. A particularly important point at which cell growth controls cell cycle progression is at the end of G1 phase, when cells decide whether to commit to cell cycle entry. The key molecular event that marks cell cycle entry is transcription of late G1 phase cyclins, which bind and activate cyclin-dependent kinases (CDKs) to initiate cell cycle events. Transcription of the late G1 cyclins is initiated only when sufficient growth has occurred.

A key question concerns how growth triggers cell cycle entry. Most models suggest that growth-dependent changes in the number or concentration of key cell cycle regulatory proteins trigger cell cycle entry. If the number of molecules of a protein increases at the same rate as growth, the number of molecules per cell will increase but the concentration of the protein will stay constant. In this case, an increase in the number of protein molecules per cell could be sufficient to trigger cell cycle entry if the activity of the protein is titrated against an activity that does not scale with growth. Alternatively, cell cycle entry could require changes in the concentration of cell cycle regulatory proteins. In this case, if the number of molecules per cell increases faster than growth, the concentration and activity of the protein should also increase. Conversely, if the number of protein molecules stays constant as growth occurs the concentration of the molecule will decrease. It remains unclear whether changes in number or concentration of cell cycle regulators trigger cell cycle entry. In addition, it remains possible that changes in the catalytic rate of an enzyme, such as a kinase or phosphatase, play an important role.

In budding yeast, key regulators of the cell cycle that are thought to link cell cycle entry to cell growth were first identified by genetic analysis. An early G1 cyclin called Cln3 was found to be a dose-dependent regulator of cell cycle entry and cell size (Cross, 1988; Nash et al., 1988). Overexpression of CLN3 causes premature cell cycle entry at a reduced cell size, while loss of Cln3 causes delayed cell cycle entry at a large cell size. A poorly understood protein called Bck2 plays redundant roles with Cln3 (Epstein and Cross, 1994). Loss of Cln3 or Bck2 causes delayed cell cycle entry, whereas loss of both causes a failure in cell cycle entry.

The Cln3/Cdk1 complex triggers cell cycle entry by activating the SBF transcription factor, which drives transcription of late G1 phase cyclins, as well as hundreds of additional genes required for subsequent cell cycle events. The Cln3/Cdk1 complex is thought to activate SBF by directly phosphorylating and inactivating Whi5, a transcriptional repressor that keeps SBF inactive prior to cell cycle entry (Jorgensen et al., 2002; Costanzo et al., 2004; de Bruin et al., 2004). Loss of Whi5 causes premature cell cycle entry at a reduced cell size, which suggests that Whi5 plays an important role in the mechanisms that link cell cycle entry to cell growth (Jorgensen et al., 2002).

Early studies suggested a model in which accumulation of Cln3 protein drives cell cycle entry (reviewed in Jorgensen and Tyers, 2004b; Turner et al., 2012). Cln3 protein is rapidly turned over, which should make Cln3 levels highly sensitive to translation rate (Tyers et al., 1992). If translation rate increases with cell size, the number of Cln3 protein molecules in the cell should also increase with cell size, which could lead to cell cycle entry. Difficulties in detecting Cln3 protein initially made it difficult to test the prediction that Cln3 protein levels increase with cell growth during G1 phase. More recent studies found that levels of Cln3 increase gradually during G1 phase, reaching a peak around the time of bud emergence, consistent with the idea that gradually rising levels of Cln3 trigger cell cycle entry (Zapata et al., 2014). Another recent study used an in vivo reporter to measure the rate of Cln3 translation and found that the rate increases faster than the rate of growth in G1 phase, which suggests that an increase in the concentration of Cln3 could trigger cell cycle entry (Litsios et al., 2019). In contrast, another study that utilized a mutant stabilized overexpressed version of Cln3 concluded that the concentration of Cln3 does not change during growth (Schmoller et al., 2015). Overall, the relationship between Cln3 accumulation and cell growth remains poorly understood. It is unclear whether Cln3 provides a simple readout of overall translation rate or whether the rate of Cln3 accumulation is influenced by other growth-dependent signals. It is also unclear how regulation of Cln3 transcription, translation, and turnover influences accumulation of Cln3 during cell growth. Finally, it remains unknown whether or how a threshold level of Cln3 triggers cell cycle entry.

Another study suggested that dilution of Whi5 by cell growth is the critical event that links cell cycle entry to cell growth (Schmoller et al., 2015). In this study, it was found that the concentration of Cln3 remains constant during growth because the number of Cln3 molecules per cell increases in a manner that is commensurate with growth. In contrast, the concentration of Whi5 was found to decrease because no new Whi5 is produced during G1 phase. The model postulates that once Whi5 concentration drops below a threshold, the activity of Cln3 becomes sufficient to inactivate Whi5, thereby triggering cell cycle entry. A limitation of this model is that cells undergo little growth in G1 phase (Ferrezuelo et al., 2012; Leitao and Kellogg, 2017; Litsios et al., 2019), which means that Whi5 concentration is reduced by less than 50% and by as little as 10–20%. It is unclear how such small changes in Whi5 concentration could be translated into an effective cell size control decision. Another limitation is that the effects of Whi5 protein concentration on cell cycle entry were analyzed only in bck2∆ cells, and only under nutrient poor conditions. Finally, the behavior of Cln3 protein was analyzed using a mutant version of Cln3 that lacks amino acid sequences that target it for proteolytic destruction, which results in overexpression and loss of cell cycle-dependent changes in protein levels.

Another limitation of current models for size control in G1 phase is that they fail to explain proportional relationships between cell size and growth rate that hold across all orders of life. Yeast cells growing slowly in poor nutrients are nearly twofold smaller than cells growing rapidly in rich nutrients (Johnston et al., 1977). Moreover, a proportional relationship between cell size and growth rate holds even when comparing yeast cells growing at different rates under identical nutrient conditions (Ferrezuelo et al., 2012; Leitao and Kellogg, 2017). For example, cells in a population of wild-type yeast cells in early G1 phase show a threefold variance in growth rate. The slow-growing cells within the population undergo cell cycle entry at a reduced cell size compared to rapidly growing cells. Thus, it appears that signals linked to nutrients and growth rate modulate the threshold amount of growth required for cell cycle progression. Classic experiments in fission yeast suggest that the threshold amount of growth required for cell cycle progression is readjusted within minutes of a shift to new nutrient conditions (Fantes and Nurse, 1977). A model that could explain the link between cell size and growth rate is that global signals that set growth rate also set the threshold amount of growth required for cell cycle progression (Lucena et al., 2018). However, the mechanisms by which nutrient-dependent signals modulate growth thresholds remain largely unknown.

Distinguishing models will require a clear understanding of how changing concentrations of Cln3 and Whi5 influence cell cycle entry, and how the concentrations of Cln3 and Whi5 vary during growth and in response to changes in growth rate. However, there are conflicting reports on the behaviors of Cln3 and Whi5 during growth in G1 phase (Tyers et al., 1993; Zapata et al., 2014; Liu et al., 2015; Schmoller et al., 2015; Lucena et al., 2018; Dorsey et al., 2018; Blank et al., 2018; Litsios et al., 2019; Barber et al., 2020; Black et al., 2020). Most previous studies used quantitative fluorescence microscopy to analyze the behavior of tagged proteins in living cells. Differences in fluorescent tags, nutrient and imaging conditions, potential imaging artifacts, as well as the use of mutant proteins or mutant genetic backgrounds could explain differences in results. In addition, the low abundance of Cln3 has made it difficult to measure levels of wild-type Cln3 by fluorescence microscopy.

Here, we used quantitative western blotting in synchronized cells to carry out a comprehensive analysis of the function and regulation of Cln3 and Whi5 during growth in G1 phase. A simple model for nutrient modulation of cell size could be that poor nutrients cause a reduction in Whi5 protein levels, which would reduce the amount of growth required to dilute Whi5 below the threshold required for cell cycle entry. However, little is known about the relationship between Cln3 and Whi5 protein levels during growth in G1 phase in differing nutrient conditions. We therefore investigated how levels of the Whi5 and Cln3 proteins are modulated during growth in G1, and how nutrient availability modulates levels of both proteins. We also investigated the mechanisms that control accumulation of Cln3 during growth in G1 phase.

Results

Cells growing in poor carbon undergo cell cycle entry at a dramatically lower ratio of Cln3 to Whi5

Previous studies found that levels of Cln3 mRNA and protein are reduced in asynchronous cells growing under poor nutrient conditions (Gallego et al., 1997; Parviz and Heideman, 1998; Hall et al., 1998; Blank et al., 2018). However, more recent work discovered that Cln3 is also synthesized in mitosis (Landry et al., 2012; Zapata et al., 2014). Therefore, analysis of Cln3 in asynchronous cells could not define the behavior of Cln3 during growth in G1 phase. Here, we used synchronized cells to characterize the dynamics of Cln3 and Whi5 proteins during growth in G1 phase in rich versus poor carbon.

Cells were grown in a poor carbon source (YP media containing 2% glycerol and 2% ethanol) and centrifugal elutriation was used to isolate small newborn daughter cells in early G1 phase. The isolated cells were then resuspended in YP media containing rich carbon (2% dextrose) or poor carbon to initiate growth in G1 phase. An advantage of this approach is that the cells growing in rich versus poor carbon were previously grown under identical conditions and each culture contains identical numbers of cells, which allows comparison of levels of Cln3 and Whi5 between conditions. Another advantage is that the cells shifted to rich carbon undergo a prolonged interval of growth in G1 phase to reach the increased size that is characteristic of cells growing in rich carbon, which provides a longer interval for examining the behavior of Cln3 and Whi5.

Median cell size was determined with a Coulter Channelyzer and plotted as a function of time (Figure 1A, Figure 1—figure supplement 1A). Since bud emergence provides a marker of cell cycle entry, the percentage of cells with buds was plotted as a function of time (Figure 1B). Finally, a plot of the percentage of cells with buds versus cell size provided a measure of cell size at cell cycle entry (Figure 1C). As expected, the cells transferred to rich carbon grew faster and spent more time undergoing growth in G1 phase compared to cells in poor carbon. The cells in rich carbon also entered the cell cycle at a larger size than the cells in poor carbon. The cells that remained in poor carbon grew in size by only 24% before bud emergence (defined as the time at which 20% of cells were budded). In contrast, the cells shifted from poor to rich carbon increased in size by 75%.

Figure 1 with 2 supplements see all

Download asset Open asset

Dynamics of Cln3 and Whi5 proteins during G1 phase in cells growing in rich or poor carbon.

Wild-type cells were grown to mid-log phase in poor carbon (YPG/E), and small unbudded cells were isolated by centrifugal elutriation. Cells were released into either rich carbon (YPD) or poor carbon (YPG/E) at 25°C, and samples were taken at 10 min intervals. All data in the figure are from the same biological replicate. (A) Median cell size was measured using a Coulter counter at 10 min intervals and plotted as a function of time. (B) The percentage of budded cells as a function of time. (C) The percentage of budded cells as a function of median cell size at each time point. (D) The behavior of Cln3-6XHA and Whi5-3XHA was analyzed by western blot on the same blot for each carbon source. An anti-Nap1 antibody was used as a loading control. (E) Quantification of Cln3-6XHA protein levels from (D) as a function of time. At each time point, the relative Cln3 signal was calculated as a ratio to the signal of the 10 min time point. (F) The fold change in peak Cln3 protein signal was calculated as a ratio over the signal from the 10 min time point. The data represent the average of three biological replicates. The error bars represent SEM. (G) The difference in peak Cln3 between rich and poor carbon was calculated by first normalizing the peak Cln3 signal to the loading control in each carbon source, and then comparing the signal between carbon sources. The error bars represent SEM. (H) Quantification of Whi5-3XHA from western blots in (D). The change in relative protein abundance over time was measured by calculating the ratio over the zero time point for each time point. (I) The relative change in Cln3 concentration over time was calculated by taking the ratio of the Cln3 signal over cell size at each time point. The bud emergence data from panel (B) are included for reference. The experiment shown in this figure was repeated for three biological replicates, which gave similar results. In panels (E), (H), and (I), the Cln3 and Whi5 signals were not normalized to a loading control because the loading control signal increases with growth.

Figure 1—source data 1 Source data for Figure 1D.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig1-data1-v3.pdf
Download elife-64364-fig1-data1-v3.pdf
Figure 1—source data 2 Source data for Figure 1D.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig1-data2-v3.pdf
Download elife-64364-fig1-data2-v3.pdf

The behaviors of Whi5 and Cln3 were assayed by western blot in the same samples. A constant number of cells was collected for each time point so that western blot signals provide a measure of protein copy number per cell. Whi5 was detected with a 3XHA tag, whereas Cln3 was detected with a 6XHA tag. Cln3 tagged with 3XHA could not be detected, which necessitated use of the 6XHA tag. Tagging both Cln3 and Whi5 with 6XHA caused the proteins to co-migrate on SDS-PAGE, which prevented use of identical tags. The use of HA tags on both proteins allowed detection on the same western blot so that relative behaviors of Cln3 and Whi5 during growth could be directly compared. However, the use of tags with different detection sensitivities precluded comparison of absolute protein levels between Cln3-6XHA and Whi5-3XHA. Nevertheless, the fact that Whi5-3XHA is easily detectable, whereas Cln3-3XHA is undetectable, indicates that levels of Whi5 far exceed those of Cln3. Cln3-6XHA caused a slight increase in cell volume in rich carbon and had little effect on cell size in poor carbon (Figure 1—figure supplement 1B). A previous study found Cln3 tagged with protein A also causes a slight increase in cell volume but fully rescues loss of CLN3 function in genetic backgrounds that require CLN3 function for viability (Cross et al., 2002). Previous studies found that fluorescently tagged versions of Whi5 do not cause a decrease in cell size, which suggests that tagged versions of Whi5 are fully functional. Here, we found that Whi5-3XHA cells were slightly smaller than wild-type cells (Figure 1—figure supplement 1B). Whi5 tagged with a 13Myc tag did not cause a decrease in cell size, which suggests that the HA tag has a mild effect on Whi5 function. Whi5-3XHA and Whi5-13Myc showed the same behavior in elutriated cells undergoing growth in G1 phase. Use of Whi5-3XHA in these experiments allowed us to directly compare the behavior of Cln3-6XHA and Whi5-3XHA in the same samples and on the same western blots. Western blots comparing Cln3 and Whi5 in rich versus poor carbon were carried out simultaneously under identical conditions to allow direct comparison of protein levels between carbon sources. Here, we use ‘concentration’ to refer to the number of molecules per unit of cell volume and the more general term ‘protein levels’ to refer to the number of molecules in the cell.

In both conditions, Cln3 was not detectable at the 0 min time point (Figure 1D, see Figure 1—figure supplement 1C for a longer exposure). Cln3 was first detected at 10 min after elutriation and levels of Cln3 increased gradually during growth in both rich and poor carbon. Cln3 levels peaked slightly before the first buds could be detected in both conditions, consistent with a previous study that found that the rate of Cln3 translation peaks before Whi5 exits the nucleus (Litsios et al., 2019). Cln3 levels were dramatically lower in poor carbon. Quantification of the Cln3 signal relative to the 10 min time point in multiple biological replicates indicated that Cln3 protein levels increased approximately 10-fold before bud emergence in rich carbon and 2.5-fold in poor carbon (Figure 1E and F). Quantification of peak Cln3 levels in rich and poor carbon suggests that cells in poor carbon initiate cell cycle entry with approximately eightfold lower levels of Cln3 than cells in rich carbon (Figure 1G). Note that this population-level analysis could miss transient bursts in Cln3 translation rate that were previously observed using a translation reporter in single cells (Litsios et al., 2019). Thus, accumulation of Cln3 protein in single cells may not be as gradual as indicated in the population-level analysis.

We did not detect differences in Whi5 protein levels between the two carbon sources. In Figure 1D, a slightly longer exposure was used for the western blot from cells in poor carbon so that Cln3 could be detected, but quantification of Whi5 levels relative to a loading control showed no difference between rich and poor carbon. To test whether Whi5 levels change during G1 phase, we calculated a ratio of the Whi5 signal at each time point over the Whi5 signal at the 0 min time point (Figure 1H). We used the same approach to calculate the fold change in Whi5 levels between the zero time point and bud emergence in multiple biological replicates (Figure 1—figure supplement 1D). Both plots suggest that levels of the Whi5 protein do not change substantially before bud emergence in rich or poor carbon, consistent with several previous studies that used microscopy to analyze fluorescently tagged versions of Whi5 (Schmoller et al., 2015; Dorsey et al., 2018; Litsios et al., 2019; Black et al., 2020).

A fraction of Whi5 shifted to lower electrophoretic mobility forms at the time of bud emergence, which likely corresponds to the activity of a previously described positive feedback loop in which the late G1 cyclins Cln1 and Cln2 activate Cdk1 to phosphorylate and inactivate Whi5 (Cross and Tinkelenberg, 1991; Costanzo et al., 2004). Note that the Whi5 phosphorylation is first detected 20–30 min after Cln3 protein reaches peak levels, consistent with phosphorylation being an indirect consequence of accumulation of Cln3.

We also attempted to assay Cln3 and Whi5 protein levels during G1 phase in cells grown continuously in rich medium (YP+ dextrose). However, we found that it was not possible to isolate a uniform population of small unbudded cells because very little growth occurs in G1 phase in rich medium, which means that newborn cells are nearly the same size as mother cells (Leitao and Kellogg, 2017). This, combined with normal variation in cell size observed in wild-type yeast, means that elutriation yields a mix of unbudded and budded cells. Previous studies also encountered this issue (Futcher, 1999; Litsios et al., 2019). As an alternative, we grew cells in complete synthetic medium (CSM) containing 2% dextrose. Under these conditions, cells grow slowly and produce small newborn cells because CSM is limiting for nutrients other than dextrose. We isolated small unbudded cells and released them into the same medium. We found that Cln3 protein levels increased threefold, while Whi5 protein levels did not change substantially (Figure 1—figure supplement 2), similar to the results obtained in YP medium containing poor carbon.

Together, the data show that Cln3 protein levels are correlated with the extent of growth in volume in G1 phase in both rich and poor carbon, reaching a peak near the time of bud emergence. In addition, overall Cln3 protein levels in G1 phase are correlated with the growth rate set by carbon source. Thus, Cln3 levels are high in rich carbon and low in poor carbon. In contrast, Whi5 protein levels are not modulated by carbon source. As a result, cells growing in poor carbon enter the cell cycle at a much lower ratio of Cln3 to Whi5 protein compared to cells in rich carbon.

The large increase in Cln3 protein levels in cells shifted from rich to poor carbon likely reflects a rapid resetting of the threshold amount of Cln3 required for cell cycle entry. In this case, the cells shifted to rich carbon would require a higher threshold of Cln3 for cell cycle entry, which would require more growth to accumulate sufficient Cln3 protein to drive cell cycle entry. The fact that there is no difference in Whi5 protein levels between rich and poor carbon suggests that Whi5 protein levels have little to do with setting threshold amount of Cln3 required for cell cycle entry.

The concentration of Cln3 increases during growth in G1 phase

A previous study utilized quantitative fluorescence microscopy to analyze accumulation of a mutant stabilized overexpressed version of Cln3 and concluded that the concentration of Cln3 does not change during growth in complete synthetic media containing poor carbon (Schmoller et al., 2015). To estimate changes in wild-type Cln3 concentration before bud emergence, we divided the Cln3 signal from western blots by cell volume at each time point (Figure 1I). Using this approach, we quantified the increase in Cln3 concentration between the 10 min time point and the time of bud emergence for multiple biological replicates (Figure 1—figure supplement 1E). This showed that the concentration of Cln3 increased sevenfold prior to bud emergence in rich carbon and twofold in poor carbon. A rise in concentration of Cln3 protein during G1 phase was also detected in a previous study that used targeted proteomics to analyze the concentration of untagged wild-type Cln3 (Litsios et al., 2019).

We carried out a similar analysis for Whi5 and found that Whi5 concentration decreased by approximately 30% in rich carbon and 20% in poor carbon, consistent with the fact that Whi5 protein levels do not change substantially and there is only a small increase in cell size in G1 phase (Figure 1—figure supplement 1F). Our results are consistent with several previous studies that observed a relatively small increase in cell volume in G1 phase (Ferrezuelo et al., 2012; Leitao and Kellogg, 2017) and little or no change in Whi5 concentration (Dorsey et al., 2018; Litsios et al., 2019; Black et al., 2020). A previous study found that a small fraction of cells growing in complete synthetic media containing poor carbon show a 30–40% decrease in Whi5 concentration, but most cells showed a decrease of 20–30% or less (Schmoller et al., 2015). Population-level analysis of Cln3 and Whi5 protein levels by western blotting could miss changes in Whi5 or Cln3 concentration driven by changes in localization to specific subcellular compartments.

Cln3 protein levels respond rapidly to changes in nutrient availability

In both budding yeast and fission yeast, cells rapidly readjust the threshold amount of growth required for cell cycle progression when shifted to nutrient conditions that support different growth rates (Fantes and Nurse, 1977; Johnston et al., 1979; Kief and Warner, 1981; Lucena et al., 2018). In fission yeast, the threshold amount of growth appears to be readjusted within minutes. Therefore, cell cycle regulators that link cell cycle progression to cell growth should show a similar rapid response to changes in nutrients.

A previous study found that Cln3 protein levels decrease within 30 min of a shift from rich nitrogen medium to medium that completely lacks nitrogen, but did not test shorter time points or the effects of carbon source (Gallego et al., 1997). We therefore analyzed the behavior of Cln3 and Whi5 following a shift from rich to poor carbon. Cln3 underwent rapid changes in abundance when asynchronous cells were shifted from rich to poor carbon (Figure 2). The amount of Cln3 protein transiently increased within 5 min and then rapidly decreased to become nearly undetectable. In contrast, Whi5 levels remained constant (Figure 2). Whi5 phosphorylation was lost during the time course, consistent with previous studies showing that cells in poor nutrients spend more time in early G1 phase, which is when Whi5 is found in a dephosphorylated state (Hartwell and Unger, 1977). These observations show that levels of Cln3, but not Whi5, respond rapidly to nutrient-dependent signals that modulate cell growth and size.

Figure 2

Download asset Open asset

Cln3 protein levels respond rapidly to changes in nutrient availability.

Rapidly growing cells in rich carbon (YPD) were shifted to poor carbon (YPG/E) at 25°C, and the behavior of Cln3-6XHA and Whi5-3XHA was analyzed by western blot on the same blot. The levels of Cln3 and Whi5 protein were quantified relative to the 0 min time point. An anti-Nap1 antibody was used as a loading control. To account for any differences in total protein after shifting to poor carbon, Cln3 and Whi5 signals were calculated as a ratio to that of the loading control, which was quantified relative to the 0 min time point.

Figure 2—source data 1 Source data for Figure 2C.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig2-data1-v3.pdf
Download elife-64364-fig2-data1-v3.pdf

Increased dosage of WHI5 does not cause large effects on cell size

The discovery that cells in rich or poor carbon undergo cell cycle entry with a large difference in the ratio of Cln3 to Whi5 suggests that Whi5 protein concentration may not have a major influence on cell cycle entry. To investigate further, we carried out additional tests of how the concentration of Whi5 influences cell cycle entry and cell size. A previous study found that an additional copy of WHI5 causes an increase in cell size (Schmoller et al., 2015). However, the effects of an additional copy of WHI5 were only tested under nutrient-poor synthetic media conditions in cells that lack Bck2, a poorly understood inducer of cell cycle entry that becomes essential in cells that lack Cln3. In addition, the extra copy of WHI5 was tagged, which could influence function, and it included both the promoter and part of the coding sequence of a gene neighboring WHI5.

To avoid potential effects of tags, we constructed an integrating vector that includes only the wild-type untagged WHI5 gene with the normal upstream and downstream control regions. The WHI5 construct fully rescued the size defects of whi5∆ (Figure 3—figure supplement 1A). Integration of the plasmid into wild-type cells to introduce an extra copy of WHI5 had no effect on the size of cells growing in YP medium containing rich or poor carbon (Figure 3A). When cells were grown in nutrient-poor synthetic medium, the extra copy of WHI5 caused a 7% increase in median cell size in rich carbon and a 10% increase in poor carbon (Figure 3B). We also tested the effects of a twofold decrease in dosage of WHI5 by deleting one copy of WHI5 in diploid cells and comparing the size of WHI5/whi5∆ cells to wild-type WHI5/WHI5 cells in both rich and poor carbon media (Figure 3C). Loss of one copy of WHI5 caused little change in cell size.

Figure 3 with 1 supplement see all

Download asset Open asset

Increased dosage of Whi5 does not cause strong effects on cell size.

(A) Wild-type and *2xWHI5* cells were grown to log phase in rich (YPD) or poor carbon (YPG/E), and cell size was measured using a Coulter counter. (B) Wild-type and *2xWHI5* cells were grown to log phase in complete synthetic medium containing dextrose (CSM-rich carbon) or glycerol/ethanol (CSM-poor carbon). Cell size was measured using a Coulter counter. (C) Wild-type diploid cells and heterozygous *WHI5/whi5∆* cells were grown to log phase in rich (YPD) or poor carbon (YPG/E), and cell size was measured using a Coulter counter. (D) Wild-type and *TEF1-WHI5* cells were grown to log phase in rich (YPD) or poor carbon (YPG/E), and cell size was measured using a Coulter counter.

To test the effects of a larger increase in Whi5 protein levels, we put WHI5 under the control of the TEF1 promoter. Quantitative western blot analysis of Whi5-3XHA or Whi5-13MYC showed that the TEF1 promoter drives a 12-fold increase in Whi5 protein levels (Figure 3—figure supplement 1B, and data not shown), as well as a gradual increase in Whi5 levels throughout G1 phase, which would prevent dilution (Figure 3—figure supplement 1C and D). Expression of untagged WHI5 from the TEF1 promoter increased median cell size by 35% in rich carbon medium and 58% in poor carbon medium (Figure 3D). A previous study found that overexpression of WHI5 from the GAL1 promoter in otherwise wild-type cells has only mild effects on cell size and cell cycle progression (Costanzo et al., 2004), while two other studies observed stronger effects of GAL1-WHI5 (de Bruin et al., 2004; Barber et al., 2020).

It remains controversial whether Whi5 undergoes substantial dilution during G1 phase (Dorsey et al., 2018; Litsios et al., 2019). However, the maximal dilution that has been reported for most cells is 20-30% or less (Schmoller et al., 2015). Therefore, if dilution of Whi5 plays a role in cell size control one would expect a twofold increase or decrease in WHI5 dosage to have a substantial effect on cell size. The fact that twofold changes in WHI5 dosage have little effect suggests that dilution of Whi5 is unlikely to play a major role in control of cell size. The finding that a twofold increase in WHI5 has mild effects on cell size in nutrient-poor synthetic media is likely due to the large reduction in Cln3 protein levels in poor nutrients. The finding that a greater than 10-fold increase in Whi5 protein levels does not cause a lethal cell cycle arrest or a massive increase in cell size indicates that there must be dilution-independent mechanisms for overcoming Whi5 inhibition. Finally, the Whi5 dilution model cannot explain how cells in poor carbon initiate cell cycle entry despite a nearly 10-fold decrease in the ratio of Cln3 to Whi5. A recent study that utilized a heterologous promoter to drive Whi5 synthesis during G1 phase also reached the conclusion that dilution of Whi5 does not play a significant role in cell size control (Barber et al., 2020).

Blocking membrane trafficking events required for cell growth prevents accumulation of Cln3

Together, the preceding experiments suggest that changes in the concentration of Whi5 do not play a substantial role in mechanisms that link cell cycle entry to cell growth. Rather, the data are more consistent with a model in which accumulation of Cln3 could be the critical readout of growth that triggers cell cycle entry. We therefore investigated the relationship between accumulation of Cln3 and cell growth. As a first step, we sought to test whether accumulation of Cln3 protein in G1 phase is dependent upon growth. To do this, we first searched for ways to enforce a rapid block to growth during G1 phase. In budding yeast, vesicular traffic that drives bud growth after cell cycle entry occurs along actin cables, and depolymerization of actin causes rapid cessation of bud growth. However, we discovered that addition of the actin depolymerizing drug latrunculin A had no effect on growth during G1 phase, even though bud emergence was completely blocked. We considered the possibility that growth in G1 phase requires vesicular transport on microtubules, but addition of both latrunculin A and the microtubule depolymerizing drug nocodazole had no effect on cell growth in G1 phase (Figure 4A and B).

Figure 4 with 1 supplement see all

Download asset Open asset

Blocking membrane trafficking events required for cell growth prevents accumulation of Cln3.

(**A, B**) Wild-type cells were grown to mid-log phase in poor carbon (YPG/E), and small unbudded cells were isolated by centrifugal elutriation. The cells were split into two cultures and were then released into rich carbon (YPD) at 25°C. After 20 min, 100 µM of latrunculin A and 20 µM nocodazole were added to one culture (arrow). (A) Median cell size was measured at 20 min intervals using a Coulter counter and plotted as a function of time. (B) The percentage of budded cells as a function of time. (**C–E**) *SEC7-AID* cells were grown to mid-log phase in poor carbon, and small unbudded cells were isolated by centrifugal elutriation. The cells were split and then released into rich carbon at 25°C. After 30 min, 1 mM auxin was added to one culture (arrow). (C) Median cell size was measured at 15 min intervals using a Coulter counter and plotted as a function of time. (D) The percentage of budded cells as a function of time. (E) The behavior of Cln3-6XHA in the *SEC7-AID* cells was analyzed by western blot. An anti-Nap1 antibody was used as a loading control. (F) Autoradiogram of ³⁵S-methionine labeling to detect de novo protein synthesis. *SEC7-AID* cells were grown to mid-log phase in -MET synthetic media containing dextrose. After auxin or vehicle addition, samples were labeled with ³⁵S-methionine for 15 min intervals starting every 20 min to measure the rate of protein synthesis within the 15 min intervals.

Figure 4—source data 1 Source data for Figure 4E.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig4-data1-v3.pdf
Download elife-64364-fig4-data1-v3.pdf
Figure 4—source data 2 Source data for Figure 4E.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig4-data2-v3.pdf
Download elife-64364-fig4-data2-v3.pdf
Figure 4—source data 3 Source data for Figure 4F.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig4-data3-v3.pdf
Download elife-64364-fig4-data3-v3.pdf

We next tested whether proteins that drive early steps in membrane trafficking pathways are required for growth in G1 phase. Sec7 is required for ER-to-Golgi transport (Lupashin et al., 1996; Deitz et al., 2000). Inactivation of Sec7 using an auxin-inducible degron (sec7-AID) caused a rapid and complete block of growth in G1 phase (Figure 4C), as well as a complete block of bud emergence (Figure 4D). Inactivation of Sec7 completely blocked the gradual accumulation of Cln3 protein that normally occurs during growth in G1 phase (Figure 4E). We obtained similar results using a temperature-sensitive allele of SEC7 (sec7-1) (not shown).

One explanation for the effects of inactivating Sec7 could be that blocking membrane traffic causes a general decrease in the rate of protein synthesis, which would lead to a failure to accumulate Cln3. However, previous work found that protein synthesis continues after membrane traffic is blocked (Novick and Schekman, 1979). Moreover, we found that levels of a loading control protein increased normally, which argues against a general decrease in the rate of protein synthesis (Figure 4E, Figure 4—figure supplement 1A). Finally, we directly measured the rate of protein synthesis in SEC7-AID cells via incorporation of ³⁵S-methionine. To do this, auxin was added to asynchronously growing SEC7-AID cells and the rate of incorporation of ³⁵S-methionine was measured during 15 min intervals after addition of auxin (Figure 4F). We detected a few changes in the pattern of proteins synthesized in the SEC7-AID cells, which is likely caused by a failure in post-translational processing of proteins in the secretory pathway. However, we did not detect a decrease in the rate of protein synthesis. We obtained similar results when we induced destruction of SEC7-AID in synchronized cells undergoing growth in G1 phase (Figure 4—figure supplement 1B–D). These data suggest that failure to accumulate Cln3 after inactivation of Sec7 is not due to a general decrease in the rate of protein synthesis.

Together, these observations show that the gradual rise in Cln3 levels during G1 phase is dependent upon membrane trafficking events that drive plasma membrane growth, consistent with the idea that Cln3 levels could provide a readout of cell growth.

The budding yeast homologs of mammalian SGK kinases are required for accumulation of Cln3 during G1 phase

To further explore the relationship between cell growth and gradual accumulation of Cln3, we searched for signals that influence cell growth and size in G1 phase, with the goal of testing whether these signals also influence Cln3 accumulation. Polar bud growth that occurs after G1 phase requires CDK activity (McCusker et al., 2007). We therefore first tested whether CDK activity is required for growth in G1 phase. There are two CDKs that play overlapping roles in G1 phase: Cdk1 and Pho85. We utilized a strain that is dependent upon analog-sensitive alleles of both kinases (cdk1-as pho85-as) to test whether CDKs are required for growth in G1 phase. Inhibition of both kinases had no effect on growth in G1 phase (Figure 5—figure supplement 1A). This observation, along with the discovery that actin is not required for growth in G1 phase (Figure 4A), suggests that there may be substantial differences between the mechanisms that drive growth in G1 phase and those that drive growth of the daughter bud.

We next tested Tor kinase signaling pathways, which play conserved roles in control of cell growth. Tor kinases are assembled into two distinct multi-protein signaling complexes called TORC1 and TORC2 (Loewith et al., 2002). A key downstream target of TORC1 is the Sch9 kinase, a member of the AGC kinase family that is thought to be the functional ortholog of vertebrate S6 kinase. Sch9 mediates TORC1-dependent control of ribosome biogenesis (Urban et al., 2007). Cells that lack Sch9 are viable but proliferate slowly and show a large decrease in cell size (Jorgensen et al., 2002; Jorgensen and Tyers, 2004a). To investigate the effects of a loss of function of Sch9, we utilized an analog-sensitive allele of SCH9 (sch9-as) (Jorgensen and Tyers, 2004a). Previous work found that sch9-as cells are smaller than wild-type cells in the absence of inhibitor, consistent with the fact that most analog-sensitive kinases have partially compromised function. Addition of the inhibitor makes the cells very small, consistent with complete or nearly complete inhibition of sch9-as kinase activity (Jorgensen and Tyers, 2004a Jorgensen and Tyers, 2004a). Inhibition of sch9-as in G1 phase had no effect on growth rate (Figure 5A) but caused a slight delay in bud emergence (Figure 5B). Inhibition of sch9-as also caused a reduction in Cln3 levels but did not block gradual accumulation of Cln3 during growth in G1 phase (Figure 5C). Addition of inhibitor to wild-type cells did not have substantial effects on wild-type control cells (Figure 5—figure supplement 1B and C). The finding that acute inhibition of sch9-as does not have strong immediate effects on growth rate and cell size suggests that the effects of sch9∆ could be a long-term consequence of a decreased rate of ribosome biogenesis. A caveat is that the sch9-as protein could retain a low level of activity in the presence of inhibitor.

Figure 5 with 1 supplement see all

Download asset Open asset

Inhibition of Sch9 does not cause strong effects on accumulation of Cln3 during G1 phase.

*sch9-as* cells were grown to mid-log phase in poor carbon (YPG/E), and small unbudded cells were isolated via centrifugal elutriation. The cells were split into two cultures and were released into rich carbon (YPD, not supplemented with additional adenine; see Materials and methods). After 45 min, 250 nM 1-NM-PP1 was added to one culture (arrow). (A) Median cell size was measured at 15 min intervals using a Coulter counter and plotted as a function of time. (B) The percentage of budded cells was plotted as a function of time. (C) The behavior of Cln3-6XHA was analyzed by western blot. An anti-Nap1 antibody was used for a loading control.

Figure 5—source data 1 Source data for Figure 5C.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig5-data1-v3.pdf
Download elife-64364-fig5-data1-v3.pdf
Figure 5—source data 2 Source data for Figure 5C.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig5-data2-v3.pdf
Download elife-64364-fig5-data2-v3.pdf

We next tested whether components of a TORC2 signaling network are required for growth in G1 phase. A key function of TORC2 is to activate a pair of redundant kinase paralogs called Ypk1 and Ypk2, which are the budding yeast homologs of vertebrate SGK kinases (Casamayor et al., 1999; Kamada et al., 2005). Previous work found that Ypk1/2 are required for normal control of cell growth and size (Lucena et al., 2018). For example, loss of Ypk1 causes a large decrease in cell size, as well as a reduced rate of proliferation. Loss of both Ypk1 and Ypk2 is lethal.

To test whether Ypk1/2 influence cell growth and accumulation of Cln3 in G1 phase, we utilized an analog-sensitive allele of YPK1 in a ypk2∆ background (ypk1-as ypk2∆) (Sun et al., 2012). Inhibition of ypk1-as in early G1 phase caused a decrease in growth rate and a complete failure in bud emergence (Figure 6A and B). It also caused a rapid and complete loss of Cln3 protein (Figure 6C). In wild-type control cells, the inhibitor caused a slight decrease in growth rate and a slight decrease in Cln3 levels (Figure 6—figure supplement 1A–C). The loss of Cln3 after inhibition of ypk1-as did not appear to be caused by a general shutdown of protein synthesis because growth continued and levels of a loading control protein increased (Figure 6A and B, Figure 6—figure supplement 1D). Direct measurement of the rate of protein synthesis confirmed that inhibition of ypk1-as did not cause a general shutdown of protein synthesis (Figure 6—figure supplement 1E). Previous work found that ypk1∆ alone causes a large decrease in cell size (Lucena et al., 2018). Here, we found that ypk1∆ caused a large reduction in Cln3 levels in asynchronous cells in both rich and poor carbon (Figure 6D).

Figure 6 with 1 supplement see all

Download asset Open asset

The budding yeast homologs of mammalian SGK kinases are required for accumulation of Cln3 during G1 phase.

(**A–C**) *ypk1-as ypk2∆* cells were grown overnight in poor carbon (YPG/E) to mid-log phase, and small unbudded cells were isolated by centrifugal elutriation. The cells were divided into two cultures and were then released into rich carbon (YPD, not supplemented with additional adenine; see Materials and methods). After 30 min, 25 µM 3-MOB-PP1 was added to one culture (arrow). (A) Median cell size was measured at 15 min intervals using a Coulter counter and plotted as a function of time. (B) The percentage of budded cells as a function of time. (C) The behavior of Cln3-6XHA in the *ypk1-as ypk2∆* cells was analyzed by western blot. An anti-Nap1 antibody was used for a loading control. (D) Wild-type and *ypk1∆* cells were grown to mid-log phase in rich or poor carbon. The levels of Cln3-6XHA protein were analyzed by western blot. An anti-Nap1 was used for a loading control. Levels of Cln3-6XHA protein were quantified relative to levels of Cln3-6XHA in wild-type cells in rich carbon. Cln3 protein levels were first normalized to the loading control. (E) Wild-type cells were grown to mid-log phase in rich carbon (YPD) and then shifted to poor carbon (YPG/E) at 25°C. The behavior of Cln3-6XHA, TORC2-dependent phosphorylation of Ypk1/2, and total Ypk1 protein was assayed by western blot. A phospho-specific antibody was used to detect a TORC2-dependent phosphorylation site present on both Ypk1 and Ypk2 (referred to as anti-Ypk-pT662). Total Ypk1 protein was detected with an anti-Ypk1 antibody. Asterisk indicates a background band that also serves as a loading control.

Figure 6—source data 1 Source data for Figure 6C.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig6-data1-v3.pdf
Download elife-64364-fig6-data1-v3.pdf
Figure 6—source data 2 Source data for Figure 6C.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig6-data2-v3.pdf
Download elife-64364-fig6-data2-v3.pdf
Figure 6—source data 3 Source data for Figure 6D.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig6-data3-v3.pdf
Download elife-64364-fig6-data3-v3.pdf
Figure 6—source data 4 Source data for Figure 6D.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig6-data4-v3.pdf
Download elife-64364-fig6-data4-v3.pdf
Figure 6—source data 5 Source data for Figure 6D.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig6-data5-v3.pdf
Download elife-64364-fig6-data5-v3.pdf
Figure 6—source data 6 Source data for Figure 6D.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig6-data6-v3.pdf
Download elife-64364-fig6-data6-v3.pdf
Figure 6—source data 7 Source data for Figure 6E.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig6-data7-v3.pdf
Download elife-64364-fig6-data7-v3.pdf
Figure 6—source data 8 Source data for Figure 6E.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig6-data8-v3.pdf
Download elife-64364-fig6-data8-v3.pdf
Figure 6—source data 9 Source data for Figure 6E.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig6-data9-v3.pdf
Download elife-64364-fig6-data9-v3.pdf

Ypk1/2 undergo complex regulation and are phosphorylated by multiple kinases. We found that changes in Cln3 protein levels that occur shortly after a shift from rich to poor carbon closely paralleled changes in the phosphorylation state of Ypk1 that can be detected via shifts in electrophoretic mobility (Figure 6E). Thus, the transient increase in Cln3 levels immediately after a shift to poor carbon was correlated with a loss of Ypk1 phosphorylation, and the decrease in Cln3 levels starting at 10 min was correlated with an increase in Ypk1 phosphorylation. A reappearance of Cln3 at the end of the time course was accompanied by another decrease in Ypk1 phosphorylation. The phosphorylation of Ypk1 that can be detected by electrophoretic mobility shifts appears to be due primarily to related kinase paralogs called Fpk1 and Fpk2; however, the functions of these phosphorylation events are poorly understood (Roelants et al., 2010). Nevertheless, the correlation between changes in Ypk1 phosphorylation and changes in Cln3 protein levels provides further correlative evidence for a connection between Ypk1 signaling and Cln3 protein levels.

We also assayed TORC2-dependent phosphorylation of Ypk1 and Ypk2 in the same samples, which can be detected with a phospho-specific antibody that recognizes a site found on both kinases (referred to as T662 in Ypk1) (Niles et al., 2012). It is unknown whether TORC2-dependent phosphorylation influences the electrophoretic mobility of Ypk1/2, and TORC2-dependent phosphorylation of Ypk1/2 is measured as the intensity of the signal generated by the anti-pT662 antibody. Previous studies found that a shift from rich to poor carbon causes rapid loss of TORC2-dependent phosphorylation of Ypk1/2 (Lucena et al., 2018). Here, we found that loss of TORC2-dependent phosphorylation of Ypk1/2 upon a shift to poor carbon was immediate and preceded the decrease in Cln3 levels by approximately 10 min (Figure 6E, bottom panel). The increase in Cln3 phosphorylation that occurred later in the time course as cells adapted to the new carbon source was correlated with an increase in TORC2-dependent phosphorylation of Ypk1/2. Thus, changes in Cln3 levels were not strongly correlated with TORC2-dependent phosphorylation of Ypk1/2.

Together, the data show that Ypk1/2, which were previously shown to play roles in control of cell growth and size, also play a role in regulating levels of Cln3. The data further suggest that Cln3 protein levels are unlikely to be influenced only by translation rate. Rather, the data indicate that poorly understood signals associated with plasma membrane growth and the TORC2 network strongly influence Cln3 levels.

Sphingolipid-dependent signals influence Cln3 levels

A key function of Ypk1/2 is to control the activity of a biosynthetic pathway that builds sphingolipids and ceramide lipids (Aronova et al., 2008; Roelants et al., 2011; Sun et al., 2012; Muir et al., 2014). A simplified overview of the role of Ypk1/2 is shown in Figure 7—figure supplement 1A. Ypk1/2 stimulate the enzyme that initiates production of sphingolipids and also stimulates ceramide synthase, which builds ceramide from sphingolipid precursors. Sphingolipids and ceramides play poorly understood roles in signaling. In previous work, we found that Ypk1/2-dependent control of the ceramide synthesis pathway is required for normal control of cell growth and size (Lucena et al., 2018). For example, myriocin, a small molecule inhibitor of sphingolipid synthesis, causes a dose-dependent decrease in growth rate in G1 phase and a corresponding decrease in cell size at cell cycle entry. Similarly, loss of ceramide synthase causes a large reduction in cell size, as well as a complete failure in nutrient modulation of cell size, growth rate, and TORC2 activity.

To test whether Ypk1/2 control Cln3 levels via sphingolipids or ceramides, we first tested the effects of myriocin, which inhibits the first step in production of sphingolipids. In previous work, we analyzed the effects of sublethal doses of myriocin on cell growth and size in G1 phase (Lucena et al., 2018). Here, we used a higher concentration of myriocin that blocks proliferation. Addition of myriocin caused a large reduction in growth rate in G1 phase and delayed bud emergence but caused only a modest decrease in Cln3 levels (Figure 7A–D). Myriocin also caused cells to initiate bud emergence at a smaller size (Figure 7C). Thus, the complete loss of Cln3 caused by inhibition of Ypk1/2 is not due to a loss of sphingolipid-dependent signals.

Figure 7 with 1 supplement see all

Download asset Open asset

Sphingolipid-dependent signals influence Cln3 levels.

(**A–D**) Wild-type cells were grown overnight in poor carbon (YPG/E), and small unbudded cells were isolated via centrifugal elutriation. The cells were split into two cultures and released into rich carbon (YPD) at 25°C. After taking the initial time point, 5 µg/mL myriocin was added to one culture (arrow). (A) Median cell size was measured at 20 min intervals using a Coulter counter and plotted as a function of time. (B) The percentage of budded cells as a function of time. (C) The percentage of budded cells plotted as a function of cell size at each time point. (D) The levels of Cln3-6XHA protein were analyzed by western blot. An anti-Nap1 antibody was used for a loading control. (E) Wild-type and *lac1∆ lag1∆* cells were grown to mid-log phase in rich (YPD) and poor carbon (YPG/E). The levels of Cln3-6xHA protein were analyzed by western blot. An anti-Nap1 antibody was used as a loading control. Levels of Cln3-6XHA protein were quantified relative to levels of Cln3-6XHA in wild-type cells in rich carbon. Cln3 protein levels were first normalized to the loading control. Error bars represent SEM of three biological replicates. (F) Wild-type and *lac1∆ lag1∆* cells were grown to mid-log phase in rich carbon (YPD). 20 µM phytosphingosine (PHS) was added to each culture, and the cultures were incubated at 25°C. The levels of Cln3-6XHA protein were analyzed by western blot. An anti-Nap1 antibody was used as a loading control. (G) A model depicting how Cln3 levels could be modulated by Ypk1/2.

Figure 7—source data 1 Source data for Figure 7D.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig7-data1-v3.pdf
Download elife-64364-fig7-data1-v3.pdf
Figure 7—source data 2 Source data for Figure 7D.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig7-data2-v3.pdf
Download elife-64364-fig7-data2-v3.pdf
Figure 7—source data 3 Source data for Figure 7E.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig7-data3-v3.pdf
Download elife-64364-fig7-data3-v3.pdf
Figure 7—source data 4 Source data for Figure 7E.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig7-data4-v3.pdf
Download elife-64364-fig7-data4-v3.pdf
Figure 7—source data 5 Source data for Figure 7F.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig7-data5-v3.pdf
Download elife-64364-fig7-data5-v3.pdf
Figure 7—source data 6 Source data for Figure 7F.: https://cdn.elifesciences.org/articles/64364/elife-64364-fig7-data6-v3.pdf
Download elife-64364-fig7-data6-v3.pdf

In previous work, we found that ceramide-dependent signals modulate TORC2 activity via a feedback loop and are also required for normal control of cell growth and size (Lucena et al., 2018; Alcaide-Gavilán et al., 2018). We therefore also tested the effects of inactivating ceramide synthase, which builds ceramide from sphingolipid precursors. The catalytic subunit of ceramide synthase is encoded by a pair of redundant paralogs called LAC1 and LAG1 (Figure 7—figure supplement 1A). In contrast to myriocin, we found that lac1∆ lag1∆ caused a substantial decrease in levels of Cln3 in asynchronous cells (Figure 7E). This was surprising because both myriocin and inactivation of ceramide synthase block production of ceramide, yet they appear to have different effects on Cln3. An explanation could be that inactivation of ceramide synthase leads to a build-up of sphingolipids that inhibit production of Cln3. To test this, we added exogenous sphingolipids to cells. In previous work, we found that the sphingolipid phytosphingosine (PHS) causes a rapid and dramatic decrease in TORC2 signaling to Ypk1/2 (Lucena et al., 2018). The decrease in TORC2 signaling was dependent upon ceramide synthase, which provided evidence for ceramide-dependent feedback signaling to TORC2. Here, we found that exogenously added PHS caused a rapid loss of Cln3 in both wild-type and lac1∆ lag1∆ cells (Figure 7F). Levels of Cln3 recovered in wild-type cells, most likely due to conversion of the added PHS to ceramide. In contrast, Cln3 levels did not recover in lac1∆ lag1∆ cells.

These data show that sphingolipid-dependent signals strongly influence Cln3 levels and provide further evidence that complex signaling mechanisms control accumulation of Cln3 protein.

Discussion

The concentration of Cln3 protein increases during growth in G1 phase

One class of models for cell size control suggests that accumulation of Cln3 is a key reporter of cell growth during G1 phase (Jorgensen and Tyers, 2004b; Turner et al., 2012). However, there have been conflicting reports about the behavior of Cln3 in G1 phase (Zapata et al., 2014; Schmoller et al., 2015; Litsios et al., 2019). Here, we used quantitative western blotting to analyze the behavior of Cln3 during growth in G1 phase in rich versus poor carbon. Under the growth conditions used for our experiments, we found that the number of Cln3 molecules per cell increases approximately 10-fold before cell cycle entry in rich carbon and 2.5-fold in poor carbon. When combined with growth data, these data suggest that Cln3 concentration increases sevenfold before bud emergence in rich carbon and twofold in poor carbon. The large increase in Cln3 concentration observed for cells growing in rich carbon may be due partially to the fact that the cells were grown in poor carbon prior to synchronization and therefore underwent an unusually extensive growth interval during G1 phase to reach the larger size that is characteristic of cells in rich carbon.

In both rich and poor carbon, synthesis of Cln3 protein outpaces the overall rate of protein synthesis that would be needed for the 20–30% increase in cell size that occurs during growth in the same interval. Thus, increasing Cln3 concentration could generate a signal with substantial dynamic range that could be used to measure growth and trigger cell cycle entry. Similar conclusions were reached in a previous study that used an in vivo reporter to measure Cln3 translation rates and targeted mass spectrometry to measure Cln3 protein levels (Litsios et al., 2019).

The threshold amount of Cln3 required for cell cycle entry is dramatically reduced in poor carbon

Simple models based on previous studies would suggest that cell cycle entry at a reduced cell size in poor carbon could be achieved by reducing levels of Whi5 protein or by increasing levels of Cln3. However, we found that Cln3 levels are reduced nearly 10-fold in poor carbon, while levels of Whi5 do not change. Thus, cells in poor carbon undergo cell cycle entry despite a nearly 10-fold decrease in the ratio of Cln3 to Whi5. We also found that Cln3 protein levels are exquisitely sensitive to changes in carbon source that influence growth rate. Thus, a shift from rich to poor carbon, which causes a rapid decrease in growth rate, causes a transient increase in Cln3 protein levels within 5 min, followed by a dramatic reduction in Cln3 levels. These data show that Cln3 is highly responsive to nutrient-dependent signals that control cell growth and size, as expected for a critical regulator of cell size in G1 phase.

One estimate suggests that there could be as few as 100 molecules of Cln3 per cell in rich carbon (Cross et al., 2002), which would suggest that there could be as few as 10–20 molecules of Cln3 per cell in poor carbon. In contrast, Whi5 could be present at approximately 2500 molecules per cell, although there is considerable variance in estimates of the abundance of Whi5 protein (Ho et al., 2018; Dorsey et al., 2018). The mechanism that sets the threshold amount of Cln3 required to overcome Whi5 inhibition is key to understanding how cell cycle progression is linked to cell growth yet remains one of the central enigmas of cell size control. Recent work suggests that the transcription factors that promote cell cycle entry are present at higher concentrations in poor carbon, which could help promote cell cycle entry at lower levels of Cln3 (Dorsey et al., 2018).

The concentration of Whi5 protein has little influence on cell size

The effects of Whi5 overexpression on cell size have been the primary mechanistic test of the Whi5 dilution model; however, increased dosage of Whi5 could also influence a positive feedback loop in which the late G1 cyclins Cln1 and Cln2 promote cell cycle entry via inactivation of Whi5 (Cross and Tinkelenberg, 1991). Thus, current data do not distinguish whether any effects of Whi5 overexpression are due to inhibition of the initial initiation of late G1 cyclin transcription versus inhibition Whi5 inactivation via positive feedback.

To further test the Whi5 dilution model, we investigated the effects of increased or decreased expression of WHI5 on cell size. An extra copy of untagged Whi5 expressed from its own promoter had little effect on cell size in wild-type cells growing in YP media containing rich or poor carbon. Similarly, loss of one copy of WHI5 in diploid cells had little effect on cell size. We further found that greater than 10-fold overexpression of WHI5 from the TEF1 promoter had a relatively modest effect on cell size in rich carbon, which indicates that there are dilution-independent mechanisms for inactivation of Whi5. Furthermore, a key test of the model is to determine the effects of synthesizing Whi5 during G1 phase, which would prevent dilution. A previous study tested this by using a heterologous promoter to drive Whi5 synthesis during G1 phase and did not observe the effects predicted by the Whi5 dilution model (Barber et al., 2020). We conclude that growth-dependent dilution of Whi5 is unlikely to play a major role in measuring cell size in G1 phase. Multiple recent studies reached similar conclusions (Dorsey et al., 2018; Litsios et al., 2019; Barber et al., 2020).

Dilution models for cell size control present a number of unresolved mechanistic issues. For example, dilution models require that key factors show a concentration-dependent activity. The only known activity of Whi5 is to bind and inhibit the SBF transcription factor. For Whi5 to repress transcription at SBF promoters, it must be present in excess over SBF binding sites and bind with sufficient affinity to achieve high occupancy at promoters. Moreover, to achieve high occupancy at SBF promoters, Whi5 must be present at a concentration higher than the K_d for its interaction with SBF. One study estimated the concentration of Whi5 in the nucleus to be 120 nM (Dorsey et al., 2018). Independent studies that measured the total number of Whi5 molecules in the cell and nuclear volume suggest that the concentration could be much higher (Jorgensen et al., 2007; Ho et al., 2018). The K_d for the interaction between Whi5 and SBF is unknown but could easily be in the low nM range. For comparison, the Rb transcription inhibitor binds to the E2F transcription factor with a K_d of 40 nM (Burke et al., 2010). Thus, the concentration of Whi5 in the nucleus may be substantially higher than the K_d for binding to SBF. In this case, a 20–30% change in cell volume would have no effect on occupancy of Whi5 at SBF promoters, which is the only known activity of Whi5. It has been proposed that growth-dependent dilution causes a change in the ratio of Cln3 to Whi5. However, under the poor carbon conditions used to test the dilution model, the concentration of Whi5 is clearly much higher than the concentration of Cln3, so a less than twofold change in Whi5 concentration will have little effect on the ratio. Phosphorylation of Whi5 by Cdk1 could reduce the affinity of Whi5 for SBF, but in that case there may be no reason to hypothesize that anything other than rising Cln3 levels leads to inactivation of Whi5. It is certainly possible that there are biochemical parameters for the mechanisms that inactivate Whi5 that would be compatible with dilution models. However, in the absence of any data on those parameters it remains unclear whether dilution models are mechanistically viable.

Accumulation of Cln3 during G1 phase is dependent upon membrane trafficking events that drive cell growth

Previous studies have proposed that accumulation of Cln3 protein could be a key readout of cell growth. To provide a measure of growth, accumulation of Cln3 must be dependent upon growth. To test this, we used an auxin-inducible degron allele of SEC7 to block membrane trafficking events that drive cell growth. Inactivation of Sec7 in early G1 phase blocked cell growth, as well as accumulation of Cln3. This observation, combined with the results of our analysis of Cln3 accumulation in wild-type cells, suggests that gradual accumulation of Cln3 in G1 phase is dependent upon growth and correlated with the extent of growth, as expected for a protein that provides a readout of growth.

In previous work, we found additional evidence that membrane trafficking events that drive cell growth generate signals that are correlated with growth, which could be used to measure growth. For example, delivery of vesicles to the plasma membrane in growing daughter buds generates signals that lead to phosphorylation of the kinase Pkc1 as well as two related kinases called Gin4 and Hsl1 (Anastasia et al., 2012; Jasani et al., 2020). In each case, multi-site phosphorylation of these kinases appears to gradually increase during bud growth, and phosphorylation is dependent upon growth and correlated with the extent of growth. The signals relayed by these kinases are required for normal regulation of the duration and extent of bud growth, but there is no evidence that they influence growth in G1 phase. Several previous studies also found connections between membrane traffic, ribosome biogenesis, and cell size (Mizuta and Warner, 1994; Nanduri and Tartakoff, 2001; Lempiäinen et al., 2009; Singh and Tyers, 2009).

Cln3 protein levels are controlled by the budding yeast homologs of mammalian SGK kinases

We also searched for signals that drive growth-dependent accumulation of Cln3. Growth of the daughter bud after cell cycle entry is dependent upon Cdk1 activity (McCusker et al., 2007). However, growth in G1 phase and accumulation of Cln3 showed no dependence on Cdk1 activity. Surprisingly, growth in G1 phase was also not dependent upon actin filaments, which are required for bud growth. The fact that bud growth is strongly dependent upon Cdk1 and actin filaments, whereas growth in G1 phase is not, suggests that there may be substantial differences in the mechanisms that drive cell growth at different stages of the cell cycle.

Inhibition of Ypk1/2, the budding yeast homologs of human SGK kinases, caused a rapid and complete loss of Cln3 protein. Previous studies have shown that Ypk1/2 relay nutrient-dependent signals and are required for normal control of cell growth and size (Lucena et al., 2018). Thus, the discovery that Ypk1/2 control Cln3 protein levels establishes a link between Ypk1/2 and Cln3 that could help explain how Ypk1/2 and nutrient-dependent signals modulate cell size. The mechanism by which Ypk1/2 influence Cln3 protein levels is unknown. Since Cln3 is a rapid turnover protein, the mechanism could work at the level of transcription, translation, or protein stability. Further investigation of the mechanisms that link Cln3 levels to Ypk1/2 activity should provide important clues to how cell growth and size are controlled.

Ypk1/2 are direct targets of TORC2 and the TORC2 signaling network is strongly modulated by carbon source, which suggests that TORC2 could play a role in modulating Cln3 levels (Kamada et al., 2005; Niles et al., 2012; Lucena et al., 2018; Alcaide-Gavilán et al., 2018). However, there are additional kinases that phosphorylate Ypk1/2 and regulation of Ypk1/2 remains poorly understood (Casamayor et al., 1999; Roelants et al., 2010). Thus, it is also possible that there are signals that act in parallel with TORC2 to modulate Cln3 levels via Ypk1/2. A further complication is that Ypk1/2 relay poorly understood feedback signals that regulate their upstream kinases, which can make it difficult to clearly delineate direct versus indirect effects of manipulating the signaling network (Roelants et al., 2010; Roelants et al., 2011; Berchtold et al., 2012; Lucena et al., 2018; Alcaide-Gavilán et al., 2018).

There was an important difference between the effects of inhibiting Ypk1/2 and the effects of inhibiting Sec7. Inhibition of Ypk1/2 caused a reduced rate of growth and a rapid and complete loss of Cln3. In contrast, inhibition of Sec7 completely blocked cell growth as well as further accumulation of Cln3 but did not cause a loss of Cln3. These observations suggest that the effects of inhibiting Ypk1/2 are not a simple consequence of a slowdown in overall growth, and that Ypk1/2 could play a more direct role in controlling Cln3 levels. A key function of Ypk1/2 is to stimulate production of sphingolipids and ceramides, which relay signals that influence cell growth and size. Myriocin, an inhibitor of the first step of sphingolipid synthesis, did not block accumulation of Cln3. Similarly, loss of ceramide synthase reduced but did not eliminate Cln3. Thus, the disappearance of Cln3 upon inhibition of Ypk1/2 is not due to a failure in synthesis of sphingolipids and ceramides. Rather, another target of Ypk1/2 likely exists that influences Cln3 levels.

The different effects of Sec7, Ypk1/2 and myriocin on Cln3 levels are puzzling. Previous studies found that myriocin, inhibition of Ypk1/2, and inhibition of membrane traffic all cause increased TORC2-dependent phosphorylation of Ypk1/2, which is thought to lead to increased activity of Ypk1/2 (Niles et al., 2014; Clarke et al., 2017; Lucena et al., 2018). In each case, the increased TORC2-dependent phosphorylation of Ypk1/2 is thought to be due to poorly understood negative feedback loops in the TORC2 network. We confirmed that each of these experimental manipulations also cause increased TORC2-dependent phosphorylation of Ypk1/2 during growth in G1 phase (not shown). Although each experimental manipulation causes a decrease in growth rate, they each have different effects on accumulation of Cln3. We also found that addition of exogenous sphingolipids to cells causes a rapid decline in Cln3 levels that is not dependent upon conversion of sphingolipids to ceramides.

What kind of model could explain these observations? One possibility is that there is a pool of Ypk1/2 at the endoplasmic reticulum that promotes accumulation of Cln3, and sphingolipids produced at the endoplasmic reticulum inhibit this pool of Ypk1/2 (Figure 7G). In this model, loss of Ypk1/2 activity would cause rapid loss of Cln3 because Ypk1/2 directly promote production of Cln3. Inactivation of Sec7 would lead to accumulation of sphingolipids in the endoplasmic reticulum because they fail to be transported to the Golgi apparatus, which would inhibit the Ypk1/2-dependent increase in Cln3 during cell growth. Myriocin would lead to loss of inhibitory sphingolipids at the endoplasmic reticulum so that Ypk1/2 can drive an increase in Cln3, even though growth has stopped. Finally, addition of exogenous sphingolipids would lead to an increase in sphingolipid-dependent signals that inhibit Ypk1/2, leading to loss of Cln3. Several previous studies have suggested that Cln3 is localized to the endoplasmic reticulum and that regulation of Cln3 is closely associated with endoplasmic reticulum function (Vergés et al., 2007; Yahya et al., 2014).

Data from previous studies are consistent with a model in which there are multiple differentially regulated pools of TORC2 and Ypk1/2. For example, Ypk1/2 must carry out functions at the endoplasmic reticulum because the Orm1/2 proteins and ceramide synthase, two well-established direct targets of Ypk1/2, are localized there. Yet, TORC2 and Ypk1/2 are also found at the plasma membrane (Berchtold and Walther, 2009; Niles et al., 2012) and transport of sphingolipids or ceramide to the plasma membrane appears to be essential for relaying negative feedback signals that inhibit TORC2-dependent phosphorylation of Ypk1/2, consistent with the existence of signaling events at the plasma membrane (Clarke et al., 2017). Thus, there may be distinct, differentially regulated pools of Ypk1/2 at the endoplasmic reticulum and the plasma membrane.

At this point, numerous alternative models are possible, and our ability to construct detailed models is constrained by our limited understanding of the function and regulation of TORC2 and Ypk1/2. Nevertheless, the fact that Cln3, a critical dose-dependent regulator of cell size, is strongly influenced by signaling from key components of the TORC2 network suggests that a full understanding of cell size control will require a much better understanding of growth control and the TORC2 signaling network.

Materials and methods

Yeast strains, media, plasmids, and inhibitors

Request a detailed protocol

All strains are in the W303 background (leu2-3, 112 ura3-1 can1-100 ade2-1 his3-11,15 trp1-1 GAL+ ssd1-d2). All strains are derived from the parent strain DK186 with the exception of DK2961, which is derived from Y3507 (mat α) (gift from James R. Broach, Penn State). Table 1 shows additional genetic features. One-step, PCR-based gene replacement was used for making deletions and adding epitope tags at the endogenous locus (Longtine et al., 1998; Janke et al., 2004). A plasmid carrying a copy of WHI5 was made by amplifying the WHI5 gene with 329 bp upstream of the start codon and 159 bp downstream of the stop codon from genomic DNA with oligos GCGGGATCCCCGTCTTCCTTGTGCTGTTTATG and GCGGAATTCGCGGATGTTGATCGGCGGAT. The PCR fragment was cloned into the BamH1 and EcoR1 sites of YIplac211 to create plasmid pRAS1. To integrate an extra copy of WHI5 into the genome, the plasmid was cut with StuI to target integration at the URA3 locus. The empty vector YIplac211 was integrated to create isogenic control strains.

Table 1

Strains with multiple strain numbers indicate multiple independent isolates of the same strain.

Strain	Genotype	Reference or source	Figures
DK186	MATa bar1	Altman and Kellogg, 1997	Figure 3
DK3092	MATa bar1 WHI5-3xHA::HPHNT1 CLN3-6xHA::HIS3	This study	Figure 1 and Figure 1—figure supplement 1, Figure 1—figure supplement 2, Figure 2, Figure 3—figure supplement 1B
DK4395	MATa bar1 WHI5-13MYC::HIS5	This study	In support of Figure 1, data not shown
DK3669DK3670	MATa bar1 URA3::YIplac211 (URA3)	This study	Figure 3A and B and Figure 3—figure supplement 1A
DK3722DK3875DK3876	MATa bar1 URA3::pRAS1 (WHI5::URA3)	This study	Figure 3A and B and Figure 3—figure supplement 1A
DK3726DK3885	MATa bar1 whi5∆::KANMX URA3::pRAS1 (WHI5::URA3)	This study	Figure 3—figure supplement 1A
DK3728	MATa bar1 whi5∆::KANMX URA3::YIplac211 (URA3)	This study	Figure 3—figure supplement 1A
DK3716DK3717DK3718	MATa/α WHI5/whi5::KANMX	This study	Figure 3C
DK3712	KANMX::TEF1-WHI5	This study	Figure 3D
DK3323	KANMX::TEF1-WHI5-3xHA::HPHNT1 CLN3-6xHA::HIS3	This study	Figure 3—figure supplement 1B-D
DK2017	MATa bar1 CLN3-6xHA::HIS3	Zapata et al., 2014	Figure 4A and B, Figure 5—figure supplement 1C and D, Figure 6D and E and Figure 6—figure supplement 1A-C, Figure 7A–F
DK2907	MATa bar1 SEC7-V5-AID::KANMX CLN3-6xHA::NATNT2 pTIR2::HIS3 pTIR4::LEU2	This study	Figure 4C–F and Figure 4—figure supplement 1
DK2936	MATa bar1 cdc28-as1(F88G) pho85-as1(F82G) CLN3-6xHA::HIS3	This study	Figure 5—figure supplement 1A and B
DK2961	MATα sch9-as(T492G) CLN3-6xHA::NATNT2	This study	Figure 5
DK3000	MATa bar1 ypk1-as (L424A) ypk2∆::HIS3 CLN3-6xHA::NATNT2	This study	Figure 6A–C and Figure 6—figure supplement 1D and E
DK3525	MATa bar1 ypk1∆::HIS3 CLN3-6xHA::NATNT2	This study	Figure 6D
DK3578	MATa bar1 lac1∆::HIS3 lag1∆::KANMX CLN3-6xHA::NATNT2	This study	Figure 7E and F

Cells were grown in YP medium (1% yeast extract, 2% peptone) that contained 40 mg/L adenine and a carbon source, except where noted in the figure legends. Rich carbon medium (YPD) contained 2% dextrose, while poor carbon medium (YPG/E) contained 2% glycerol and 2% ethanol. In experiments using the ATP analog inhibitors 1-NM-PP1, 1-NA-PP1, or 3-MOB-PP1, no additional adenine was added to the media. CSM (MP Biosciences) contained either 2% dextrose (rich carbon) or 2% glycerol and 2% ethanol (poor carbon) with an additional 40 mg/L adenine. CSM without methionine was used for ³⁵S-methionine labeling experiments.

Myriocin (Sigma) was solubilized in 100% methanol to make a 500 µg/mL stock solution. We have observed significant batch-to-batch differences in the effective concentration of myriocin from the same supplier. All ATP analog inhibitors were solubilized in 100% DMSO. 3-MOB-PP1 was a gift from Jack Stevenson and Kevan Shokat lab (UCSF). Auxin (indole-3-acetic acid) (Aldrich) stock was prepared at 50 mM in 100% EtOH and used at 1 mM.

Cell size analysis and bud emergence

Request a detailed protocol

Cells were grown in YPD or YPG/E medium overnight at 22°C to mid-log phase (OD₆₀₀ less than 0.7). Cells were fixed with 3.7% formaldehyde for 30 min and were then washed with PBS + 0.02% Tween-20 + 0.1% sodium azide (PBTA) before measuring cell size using a Z2 Coulter Channelyzer and Z2 AccuComp v3.01a software as previously described (Lucena et al., 2018). For measurement of log phase cultures, each cell size plot is an average of at least three independent biological replicates in which each biological replicate is the average of three technical replicates. Biological replicates are measurements obtained from independent cultures grown on different days, whereas technical replicates are independent measurements of cultures grown on the same day. Between 30,000 and 40,000 cells are counted for each measurement. For 2xWHI5 measurements, three independent strain isolates generated by integration of the WHI5 plasmid (pRAS1) were measured in more than three independent biological replicates and averaged. Similarly, the isogenic control strain integrated with empty vector is an average of two independent isolates measured in more than three independent biological replicates and averaged. The percentage of budded cells was calculated by counting >200 cells at each time point using a Zeiss Axioskop 2 (Carl Zeiss).

Centrifugal elutriation

Request a detailed protocol

Strains were grown in YPG/E medium overnight to OD₆₀₀ 0.4–0.8 at 30°C except for strains harboring temperature-sensitive alleles and the ypk1-as ypk2∆ strain, which were grown overnight at 22°C. Cells were harvested by spinning at 4000 rpm in a JLA 8.1 rotor at 4°C for 6 min. Cell pellets were resuspended in ~100 mL cold YPG/E, then sonicated for 1 min at duty cycle 0.5 using a Braun-Sonic U sonicator with a Braun 2000U probe at 4°C. Cells were loaded onto a Beckman JE-5.0 elutriator in a Beckman Coulter J6-MI centrifuge spinning at 2900 rpm at 4°C. After all cells were loaded, fluid flow was continued for 10 min to allow equilibration. Pump speed was then increased gradually to collect small unbudded cells, which were collected and pelleted by spinning in a JA-14 rotor at 5000 rpm for 5 min. Cells were resuspended into fresh medium at OD₆₀₀ 0.4–0.6 and grown at 25°C in a shaking water bath. For western blotting, 1.6 mL samples were collected at each time point as cells progressed through G1 phase. Since the cells do not divide during the time course, a constant number of cells is collected at each time point so that signals are intrinsically normalized on a per cell basis. Thus, western blot signals represent protein copy number per cell, rather than protein concentration. For Coulter counter analysis, samples were collected and fixed with 3.7% formaldehyde for 15–30 min for cell size measurements and to calculate the percentage of budded cells. Median cell size was calculated by the Coulter AccuComp software for each time point to plot cell size during growth in G1. We found that cell size measurements of ypk1-as ypk2∆ cells isolated from centrifugal elutriation were more consistent when cells were not fixed with formaldehyde but were measured immediately after washing once with PBTA to remove YP media.

Western blotting

1.6 mL samples of cells taken from cultures were pelleted in a microfuge at 13,200 rpm for 15 s before aspirating the supernatant and adding 250 µL of glass beads and freezing on liquid nitrogen. Cells were lysed in 140 µL of 1× SDS-PAGE sample buffer (65 mM Tris-HCl, pH 6.8, 3% SDS, 10% glycerol, 100 mM β-glycerophosphate, 50 mM NaF, 5% β-mercaptoethanol, 2 mM PMSF, and bromophenol blue) by bead beating in a Biospec Mini-Beadbeater-16 at 4°C for 2 min. The lysate was centrifuged for 15 s to bring the sample to the bottom of the tube and was then incubated in a 100°C water bath for 5 min followed by centrifugation for 5 min at 13,200 rpm. Depending on the protein blot, 7.5–15 µL of the lysate were loaded into 10% acrylamide SDS-PAGE gels, then run at a constant current setting of 20 mA per gel at a maximum of 165 V. Gels were transferred to nitrocellulose membrane in a Bio-Rad Trans-Blot Turbo Transfer system. Blots were probed overnight at 4°C in 4% milk in western wash buffer (1× PBS + 250 mM NaCl +0.1% Tween-20) with mouse monoclonal anti-HA antibody (12CA5, gift of David Toczyski, University of California, San Francisco), polyclonal anti-Nap1 antibody (unpublished), polyclonal anti-Ypk1 antibody (Alcaide-Gavilán et al., 2018), or polyclonal rabbit anti-T662P antibody (gift from Ted Powers, University of California, Davis). Western blots using anti-T662P antibody were first blocked using TBST (10 mM Tris-Cl, pH 7.5, 100 mM NaCl, and 0.1% Tween-20) + 4% milk, followed by one wash with TBST, then incubated overnight with anti-T662P antibody in TBST +4% BSA. Western blots were incubated in secondary donkey anti-mouse (GE Healthcare NA934V) or donkey anti-rabbit (GE Healthcare NXA931 or Jackson Immunoresearch 711-035-152) antibody conjugated to HRP at room temperature for 60–90 min before imaging with Advansta ECL chemiluminescence reagents in a Bio-Rad ChemiDoc imaging system.

Log phase nutrient shift

Request a detailed protocol

Cells were grown to mid-log phase (OD₆₀₀ 0.4–0.7) in rich carbon (YPD) at 22°C. Cells were shifted to poor carbon (YPG/E) using a vacuum filter apparatus with a 0.45 µM HAWP mixed cellulose ester filter. Cells were washed three times with room temperature YPG/E and then resuspended in room temperature YPG/E by pipetting and grown in a shaking water bath at 25°C.

Western blot quantification

Request a detailed protocol

Western blots were quantified using Bio-Rad Imagelab software v6.0.1. For the elutriation experiments, relative signal was calculated as a ratio of the signal of each time point over the signal at either the zero time point (Whi5) or the 10 min time point (Cln3) (see figure legends for details). The signal was not normalized to the loading control because the loading control signal increased with growth during G1 phase. No Cln3 signal was detected at the 0 min time point in elutriated cells, so the 10 min time point was used as reference for that set of samples. Log phase samples were quantified by setting all signals relative to either the 0 min time point (Figure 2) or wild-type cells in rich carbon (Figures 3B, 6D and 7E). The signal was normalized to signal for the loading control band to correct for differences in total protein between samples.

³⁵S-Methionine pulse labeling

Request a detailed protocol

Cells were grown overnight at room temperature to mid-log phase in CSM-methionine media containing 2% dextrose. The culture was concentrated to an OD₆₀₀ of 0.7 and incubated in a shaking water bath at 25°C. The vehicle or drug was added to the appropriate culture flask. A 1.2 mL sample of culture was transferred to a 1.6 mL screw top tube to label proteins made during the next 15 min. The labeling reaction was initiated by adding 1 µL of EasyTag L-[³⁵S]-Methionine from PerkinElmer at a stock concentration of 1 µCi/µL. The sample was mixed by vortexing and then placed into the 25°C shaking water bath. Labeling reactions were allowed to progress for 15 min, during which time samples were mixed by inversion every ~5 min. Labeling reactions were performed during 15 min intervals starting every 20 min. Samples were centrifuged at 13,200 rpm for 30 s, the supernatant was removed, and 250 µL of acid-washed beads were added before freezing on liquid nitrogen. Samples were prepared for SDS-PAGE similarly to samples described for western blotting in 140 µL of 1× sample buffer. After running the gel to the dye front, polyacrylamide gels were stained in R-250 Coomassie stain and then dried on Whatman paper using a Bio-Rad Gel Dryer. The dried gel was exposed film.

Experimental replicates

Request a detailed protocol

All experiments were repeated for a minimum of three biological replicates, with the exception of Figure 4—figure supplement 1D, which is based on a single biological replicate. Biological replicates are defined as experiments carried out on different days with different starting cultures.

Data availability

Figures present data from biological replicates that are representative of multiple biological replicates. Coulter counter data show the average of multiple biological replicates.

References

(2018) Modulation of TORC2 Signaling by a Conserved Lkb1 Signaling Axis in Budding Yeast
Genetics 210:155–170.
https://doi.org/10.1534/genetics.118.301296
- PubMed
- Google Scholar
1. Altman R
2. Kellogg D
(1997) Control of mitotic events by Nap1 and the Gin4 kinase
The Journal of Cell Biology 138:119–130.
https://doi.org/10.1083/jcb.138.1.119
- PubMed
- Google Scholar
1. Anastasia SD
2. Nguyen DL
3. Thai V
4. Meloy M
5. MacDonough T
6. Kellogg DR
(2012) A link between mitotic entry and membrane growth suggests a novel model for cell size control
The Journal of Cell Biology 197:89–104.
https://doi.org/10.1083/jcb.201108108
- PubMed
- Google Scholar
1. Aronova S
2. Wedaman K
3. Aronov PA
4. Fontes K
5. Ramos K
6. Hammock BD
7. Powers T
(2008) Regulation of ceramide biosynthesis by TOR complex 2
Cell Metabolism 7:148–158.
https://doi.org/10.1016/j.cmet.2007.11.015
- PubMed
- Google Scholar
(2020) Cell-size regulation in budding yeast does not depend on linear accumulation of Whi5
PNAS 117:14243–14250.
https://doi.org/10.1073/pnas.2001255117
- PubMed
- Google Scholar
1. Berchtold D
2. Walther TC
(2009) TORC2 plasma membrane localization is essential for cell viability and restricted to a distinct domain
Molecular Biology of the Cell 20:1565–1575.
https://doi.org/10.1091/mbc.e08-10-1001
- PubMed
- Google Scholar
1. Berchtold D
2. Piccolis M
3. Chiaruttini N
4. Riezman I
5. Riezman H
6. Roux A
7. Walther TC
8. Loewith R
(2012) Plasma membrane stress induces relocalization of Slm proteins and activation of TORC2 to promote sphingolipid synthesis
Nature Cell Biology 14:542–547.
https://doi.org/10.1038/ncb2480
- PubMed
- Google Scholar
1. Black L
2. Tollis S
3. Fu G
4. Fiche J-B
5. Dorsey S
6. Cheng J
7. Ghazal G
8. Notley S
9. Crevier B
10. Bigness J
11. Nollmann M
12. Tyers M
13. Royer CA
(2020) G1/S transcription factors assemble in increasing numbers of discrete clusters through G1 phase
The Journal of Cell Biology 219:3619–3622.
https://doi.org/10.1083/jcb.202003041
- PubMed
- Google Scholar
(2018) Scaling of G1 Duration with Population Doubling Time by a Cyclin in Saccharomyces cerevisiae
Genetics 210:895–906.
https://doi.org/10.1534/genetics.118.301507
- PubMed
- Google Scholar
(2010) Phosphorylation-induced conformational changes in the retinoblastoma protein inhibit E2F transactivation domain binding
The Journal of Biological Chemistry 285:16286–16293.
https://doi.org/10.1074/jbc.M110.108167
- PubMed
- Google Scholar
(1999) Functional counterparts of mammalian protein kinases PDK1 and SGK in budding yeast
Current Biology 9:186–197.
https://doi.org/10.1016/s0960-9822(99)80088-8
- PubMed
- Google Scholar
1. Clarke J
2. Dephoure N
3. Horecka I
4. Gygi S
5. Kellogg D
(2017) A conserved signaling network monitors delivery of sphingolipids to the plasma membrane in budding yeast
Molecular Biology of the Cell 28:2589–2599.
https://doi.org/10.1091/mbc.e17-01-0081
- PubMed
- Google Scholar
1. Costanzo M
2. Nishikawa JL
3. Tang X
4. Millman JS
5. Schub O
6. Breitkreuz K
7. Dewar D
8. Rupes I
9. Andrews B
10. Tyers M
(2004) CDK activity antagonizes Whi5, an inhibitor of G1/S transcription in yeast
Cell 117:899–913.
https://doi.org/10.1016/j.cell.2004.05.024
- PubMed
- Google Scholar
1. Cross FR
(1988) DAF1, a mutant gene affecting size control, pheromone arrest, and cell cycle kinetics of Saccharomyces cerevisiae
Molecular and Cellular Biology 8:4675–4684.
https://doi.org/10.1128/mcb.8.11.4675-4684.1988
- PubMed
- Google Scholar
1. Cross FR
2. Tinkelenberg AH
(1991) A potential positive feedback loop controlling CLN1 and CLN2 gene expression at the start of the yeast cell cycle
Cell 65:875–883.
https://doi.org/10.1016/0092-8674(91)90394-E
- PubMed
- Google Scholar
(2002) Testing a Mathematical Model of the Yeast Cell Cycle
Molecular Biology of the Cell 13:52–70.
https://doi.org/10.1091/mbc.01-05-0265
- PubMed
- Google Scholar
(2004) Cln3 activates G1-specific transcription via phosphorylation of the SBF bound repressor Whi5
Cell 117:887–898.
https://doi.org/10.1016/j.cell.2004.05.025
- PubMed
- Google Scholar
(2000) Sec7p directs the transitions required for yeast Golgi biogenesis
Traffic 1:172–183.
https://doi.org/10.1034/j.1600-0854.2000.010209.x
- PubMed
- Google Scholar
1. Dorsey S
2. Tollis S
3. Cheng J
4. Black L
5. Notley S
6. Tyers M
7. Royer CA
(2018) G1/S Transcription Factor Copy Number Is a Growth-Dependent Determinant of Cell Cycle Commitment in Yeast
Cell Systems 6:539–554.
https://doi.org/10.1016/j.cels.2018.04.012
- Google Scholar
1. Epstein CB
2. Cross FR
(1994) Genes that can bypass the CLN requirement for Saccharomyces cerevisiae cell cycle START
Molecular and Cellular Biology 14:2041–2047.
https://doi.org/10.1128/mcb.14.3.2041-2047.1994
- PubMed
- Google Scholar
1. Fantes P
2. Nurse P
(1977) Control of cell size at division in fission yeast by a growth-modulated size control over nuclear division
Experimental Cell Research 107:377–386.
https://doi.org/10.1016/0014-4827(77)90359-7
- PubMed
- Google Scholar
(2012) The critical size is set at a single-cell level by growth rate to attain homeostasis and adaptation
Nature Communications 3:1012.
https://doi.org/10.1038/ncomms2015
- PubMed
- Google Scholar
1. Futcher B
(1999) Cell cycle synchronization
Methods in Cell Science 21:79–86.
https://doi.org/10.1023/a:1009872403440
- PubMed
- Google Scholar
1. Gallego C
2. Garí E
3. Colomina N
4. Herrero E
5. Aldea M
(1997) The Cln3 cyclin is down-regulated by translational repression and degradation during the G1 arrest caused by nitrogen deprivation in budding yeast
The EMBO Journal 16:7196–7206.
https://doi.org/10.1093/emboj/16.23.7196
- PubMed
- Google Scholar
(1998) Regulation of the Cln3-Cdc28 kinase by cAMP in Saccharomyces cerevisiae
The EMBO Journal 17:4370–4378.
https://doi.org/10.1093/emboj/17.15.4370
- PubMed
- Google Scholar
1. Hartwell LH
2. Unger MW
(1977) Unequal division in Saccharomyces cerevisiae and its implications for the control of cell division
The Journal of Cell Biology 75:422–435.
https://doi.org/10.1083/jcb.75.2.422
- PubMed
- Google Scholar
(2018) Unification of Protein Abundance Datasets Yields a Quantitative Saccharomyces cerevisiae Proteome
Cell Systems 6:192–205.
https://doi.org/10.1016/j.cels.2017.12.004
- Google Scholar
1. Janke C
2. Magiera MM
3. Rathfelder N
4. Taxis C
5. Reber S
6. Maekawa H
7. Moreno-Borchart A
8. Doenges G
9. Schwob E
10. Schiebel E
11. Knop M
(2004) A versatile toolbox for PCR-based tagging of yeast genes: new fluorescent proteins, more markers and promoter substitution cassettes
Yeast 21:947–962.
https://doi.org/10.1002/yea.1142
- PubMed
- Google Scholar
(2020) Growth-Dependent Activation of Protein Kinases Suggests a Mechanism for Measuring Cell Growth
Genetics 215:729–746.
https://doi.org/10.1534/genetics.120.303200
- PubMed
- Google Scholar
(1977) Coordination of growth with cell division in the yeast Saccharomyces cerevisiae
Experimental Cell Research 105:79–98.
https://doi.org/10.1016/0014-4827(77)90154-9
- PubMed
- Google Scholar
(1979) Regulation of cell size in the yeast Saccharomyces cerevisiae
Journal of Bacteriology 137:1–5.
https://doi.org/10.1128/jb.137.1.1-5.1979
- PubMed
- Google Scholar
(2002) Systematic identification of pathways that couple cell growth and division in yeast
Science 297:395–400.
https://doi.org/10.1126/science.1070850
- PubMed
- Google Scholar
1. Jorgensen P
2. Tyers M
(2004a) A dynamic transcriptional network communicates growth potential to ribosome synthesis and critical cell size
Genes & Development 18:2491–2505.
https://doi.org/10.1101/gad.1228804
- Google Scholar
1. Jorgensen P
2. Tyers M
(2004b) How Cells Coordinate Growth and Division
Current Biology 14:R1014–R1027.
https://doi.org/10.1016/j.cub.2004.11.027
- Google Scholar
(2007) The size of the nucleus increases as yeast cells grow
Molecular Biology of the Cell 18:3523–3532.
https://doi.org/10.1091/mbc.e06-10-0973
- PubMed
- Google Scholar
1. Kamada Y
2. Fujioka Y
3. Suzuki NN
4. Inagaki F
5. Wullschleger S
6. Loewith R
7. Hall MN
8. Ohsumi Y
(2005) Tor2 directly phosphorylates the AGC kinase Ypk2 to regulate actin polarization
Molecular and Cellular Biology 25:7239–7248.
https://doi.org/10.1128/MCB.25.16.7239-7248.2005
- PubMed
- Google Scholar
1. Kief DR
2. Warner JR
(1981) Coordinate control of syntheses of ribosomal ribonucleic acid and ribosomal proteins during nutritional shift-up in Saccharomyces cerevisiae
Molecular and Cellular Biology 1:1007–1015.
https://doi.org/10.1128/mcb.1.11.1007-1015.1981
- PubMed
- Google Scholar
(2012) F-box protein specificity for g1 cyclins is dictated by subcellular localization
PLOS Genetics 8:e1002851.
https://doi.org/10.1371/journal.pgen.1002851
- PubMed
- Google Scholar
1. Leitao RM
2. Kellogg DR
(2017) The duration of mitosis and daughter cell size are modulated by nutrients in budding yeast
The Journal of Cell Biology 216:3463–3470.
https://doi.org/10.1083/jcb.201609114
- PubMed
- Google Scholar
1. Lempiäinen H
2. Uotila A
3. Urban J
4. Dohnal I
5. Ammerer G
6. Loewith R
7. Shore D
(2009) Sfp1 Interaction with TORC1 and Mrs6 Reveals Feedback Regulation on TOR Signaling
Molecular Cell 33:704–716.
https://doi.org/10.1016/j.molcel.2009.01.034
- PubMed
- Google Scholar
(2019) Differential scaling between G1 protein production and cell size dynamics promotes commitment to the cell division cycle in budding yeast
Nature Cell Biology 21:1382–1392.
https://doi.org/10.1038/s41556-019-0413-3
- PubMed
- Google Scholar
1. Liu X
2. Wang X
3. Yang X
4. Liu S
5. Jiang L
6. Qu Y
7. Hu L
8. Ouyang Q
9. Tang C
(2015) Reliable cell cycle commitment in budding yeast is ensured by signal integration
eLife 4:e03977.
https://doi.org/10.7554/eLife.03977
- PubMed
- Google Scholar
1. Loewith R
2. Jacinto E
3. Wullschleger S
4. Lorberg A
5. Crespo JL
6. Bonenfant D
7. Oppliger W
8. Jenoe P
9. Hall MN
(2002) Two TOR complexes, only one of which is rapamycin sensitive, have distinct roles in cell growth control
Molecular Cell 10:457–468.
https://doi.org/10.1016/s1097-2765(02)00636-6
- PubMed
- Google Scholar
1. Longtine MS
2. McKenzie A
3. Demarini DJ
4. Shah NG
5. Wach A
6. Brachat A
7. Philippsen P
8. Pringle JR
(1998) Additional modules for versatile and economical PCR-based gene deletion and modification in Saccharomyces cerevisiae
Yeast 14:953–961.
https://doi.org/10.1002/(SICI)1097-0061(199807)14:10<953::AID-YEA293>3.0.CO;2-U
- PubMed
- Google Scholar
1. Lucena R
2. Alcaide-Gavilán M
3. Schubert K
4. He M
5. Domnauer MG
6. Marquer C
7. Klose C
8. Surma MA
9. Kellogg DR
(2018) Cell Size and Growth Rate Are Modulated by TORC2-Dependent Signals
Current Biology 28:196–210.
https://doi.org/10.1016/j.cub.2017.11.069
- PubMed
- Google Scholar
(1996) Biochemical requirements for the targeting and fusion of ER-derived transport vesicles with purified yeast Golgi membranes
The Journal of Cell Biology 132:277–289.
https://doi.org/10.1083/jcb.132.3.277
- PubMed
- Google Scholar
(2007) Cdk1 coordinates cell-surface growth with the cell cycle
Nature Cell Biology 9:506–515.
https://doi.org/10.1038/ncb1568
- PubMed
- Google Scholar
1. Mizuta K
2. Warner JR
(1994) Continued functioning of the secretory pathway is essential for ribosome synthesis
Molecular and Cellular Biology 14:2493–2502.
https://doi.org/10.1128/mcb.14.4.2493-2502.1994
- PubMed
- Google Scholar
(2014) TORC2-dependent protein kinase Ypk1 phosphorylates ceramide synthase to stimulate synthesis of complex sphingolipids
eLife 3:e03779.
https://doi.org/10.7554/eLife.03779
- Google Scholar
1. Nanduri J
2. Tartakoff AM
(2001) The arrest of secretion response in yeast: signaling from the secretory path to the nucleus via Wsc proteins and Pkc1p
Molecular Cell 8:281–289.
https://doi.org/10.1016/s1097-2765(01)00312-4
- PubMed
- Google Scholar
1. Nash R
2. Tokiwa G
3. Anand S
4. Erickson K
5. Futcher AB
(1988) The WHI1+ gene of Saccharomyces cerevisiae tethers cell division to cell size and is a cyclin homolog
The EMBO Journal 7:4335–4346.
https://doi.org/10.1002/j.1460-2075.1988.tb03332.x
- PubMed
- Google Scholar
1. Niles BJ
2. Mogri H
3. Hill A
4. Vlahakis A
5. Powers T
(2012) Plasma membrane recruitment and activation of the AGC kinase Ypk1 is mediated by target of rapamycin complex 2 (TORC2) and its effector proteins Slm1 and Slm2
PNAS 109:1536–1541.
https://doi.org/10.1073/pnas.1117563109
- PubMed
- Google Scholar
(2014) TOR complex 2-Ypk1 signaling maintains sphingolipid homeostasis by sensing and regulating ROS accumulation
Cell Reports 6:541–552.
https://doi.org/10.1016/j.celrep.2013.12.040
- Google Scholar
1. Novick P
2. Schekman R
(1979) Secretion and cell-surface growth are blocked in a temperature-sensitive mutant of Saccharomyces cerevisiae
PNAS 76:1858–1862.
https://doi.org/10.1073/pnas.76.4.1858
- PubMed
- Google Scholar
1. Parviz F
2. Heideman W
(1998) Growth-independent regulation of CLN3 mRNA levels by nutrients in Saccharomyces cerevisiae
Journal of Bacteriology 180:225–230.
https://doi.org/10.1128/JB.180.2.225-230.1998
- PubMed
- Google Scholar
1. Roelants FM
2. Baltz AG
3. Trott AE
4. Fereres S
5. Thorner J
(2010) A protein kinase network regulates the function of aminophospholipid flippases
PNAS 107:34–39.
https://doi.org/10.1073/pnas.0912497106
- PubMed
- Google Scholar
(2011) Protein kinase Ypk1 phosphorylates regulatory proteins Orm1 and Orm2 to control sphingolipid homeostasis in Saccharomyces cerevisiae
PNAS 108:19222–19227.
https://doi.org/10.1073/pnas.1116948108
- PubMed
- Google Scholar
(2015) Dilution of the cell cycle inhibitor Whi5 controls budding-yeast cell size
Nature 526:268–272.
https://doi.org/10.1038/nature14908
- PubMed
- Google Scholar
1. Singh J
2. Tyers M
(2009) A Rab escort protein integrates the secretion system with TOR signaling and ribosome biogenesis
Genes & Development 23:1944–1958.
https://doi.org/10.1101/gad.1804409
- Google Scholar
1. Sun Y
2. Miao Y
3. Yamane Y
4. Zhang C
5. Shokat KM
6. Takematsu H
7. Kozutsumi Y
8. Drubin DG
(2012) Orm protein phosphoregulation mediates transient sphingolipid biosynthesis response to heat stress via the Pkh-Ypk and Cdc55-PP2A pathways
Molecular Biology of the Cell 23:2388–2398.
https://doi.org/10.1091/mbc.E12-03-0209
- PubMed
- Google Scholar
(2012) Cell size control in yeast
Current Biology 22:R350–R359.
https://doi.org/10.1016/j.cub.2012.02.041
- PubMed
- Google Scholar
1. Tyers M
2. Tokiwa G
3. Nash R
4. Futcher B
(1992) The Cln3-Cdc28 kinase complex of S. cerevisiae is regulated by proteolysis and phosphorylation
The EMBO Journal 11:1773–1784.
https://doi.org/10.1002/j.1460-2075.1992.tb05229.x
- PubMed
- Google Scholar
(1993)
Comparison of the Saccharomyces cerevisiae G1 cyclins: Cln3 may be an upstream activator of Cln1, Cln2 and other cyclins

The EMBO Journal 12:1955–1968.
- PubMed
- Google Scholar
1. Urban J
2. Soulard A
3. Huber A
4. Lippman S
5. Mukhopadhyay D
6. Deloche O
7. Wanke V
8. Anrather D
9. Ammerer G
10. Riezman H
11. Broach JR
12. De Virgilio C
13. Hall MN
14. Loewith R
(2007) Sch9 is a major target of TORC1 in Saccharomyces cerevisiae
Molecular Cell 26:663–674.
https://doi.org/10.1016/j.molcel.2007.04.020
- PubMed
- Google Scholar
1. Vergés E
2. Colomina N
3. Garí E
4. Gallego C
5. Aldea M
(2007) Cyclin Cln3 Is Retained at the ER and Released by the J Chaperone Ydj1 in Late G1 to Trigger Cell Cycle Entry
Molecular Cell 26:649–662.
https://doi.org/10.1016/j.molcel.2007.04.023
- PubMed
- Google Scholar
1. Yahya G
2. Parisi E
3. Flores A
4. Gallego C
5. Aldea M
(2014) A Whi7-Anchored Loop Controls the G1 Cdk-Cyclin Complex at Start
Molecular Cell 53:115–126.
https://doi.org/10.1016/j.molcel.2013.11.015
- Google Scholar
1. Zapata J
2. Dephoure N
3. Macdonough T
4. Yu Y
5. Parnell EJ
6. Mooring M
7. Gygi SP
8. Stillman DJ
9. Kellogg DR
(2014) PP2ARts1 is a master regulator of pathways that control cell size
The Journal of Cell Biology 204:359–376.
https://doi.org/10.1083/jcb.201309119
- PubMed
- Google Scholar

Decision letter

Naama Barkai

Senior and Reviewing Editor; Weizmann Institute of Science, Israel

Our editorial process produces two outputs: (i) public reviews designed to be posted alongside the preprint for the benefit of readers; (ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "Growth-dependent signals drive an increase in early G1 cyclin concentration to link cell cycle entry with cell growth" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Naama Barkai as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

As you will see below, all reviewers are highly supportive of the work, but asked for some clarification and some new data. In your revision, please address all points described below.

Reviewer #1 (Recommendations for the authors):

The experiments are performed to a high technical standard and in numerous biological replicates so the main conclusions are sound. My main concerns pertain to a few particular aspects of the results that are not explicitly discussed by the authors. While there are obviously many further experiments to be done to sort out the mechanism whereby Yck1/2 and sphingolipid metabolism control Cln3, these lie outside the reasonable scope of this study in my opinion. Nevertheless, some further discussion of possible mechanisms seems warranted. Finally, a few additional published observations should be mentioned and cited.

1. It is somewhat surprising that cells grown in glucose don't grow faster and that budding upon inoculation in rich is slower than in poor nutrients as a function of time. I would have expected the much faster growth rate in glucose to result in temporally faster budding than in glycerol/ethanol medium. What are the doubling times observed in rich versus poor medium?

2. It is also a bit puzzling that there is a transient burst of Cln3 immediately after inoculation into poor medium. This could be due to residual glucose in the yeast extract used to formulate YP medium, i.e., cells metabolize the small amount of glucose in the yeast extract before shifting to non-fermentative metabolism. This can be avoided by depleting YP of glucose by pre-culture with yeast in the absence of added carbon source (e.g., as a 5X YP stock, such that the yeast quickly consume all glucose then cease growth, the stock is then clarified and sterilized before use in subsequent experiments). Alternatively, to maintain complete nutrient homeostasis, cells can be elutriated in the clarified medium in which they were grown prior to elutriation and then diluted into the same medium. The transient burst doesn't affect the authors conclusions but a similar burst has been observed in other studies where residual glucose was not properly accounted for. The authors may wish to mention this.

3. It appears that the Cln3 signal reaches a maximum well before Start (e.g., 30-50 min before 20% budding is reached in most experiments) but the authors do not comment on this finding. This result aligns with recent single cell analysis from the Heinemann lab, which also suggests that Cln3 peaks well before Start (Litsios et al., 2019). It would be worth addressing this point in the discussion, including the caveat that population level studies can obscure single cell effects.

4. Why is bulk protein synthesis is not affected in the SEC7-AID strain despite the complete block to growth and what does this mean? Is the protein concentration going up in the arrested G1 phase cells? Or is there a higher rate of protein degradation to compensate for on-going synthesis? It does appear that the loading control increases throughout the experiment consistent with the former possibility. It would be useful for the authors to show the 35S incorporation data for the elutriated cultures so that the effect in G1 phase can be assessed (this is currently cited as data not shown). If the protein concentration in the arrested cells is higher, do the arrested cells rapidly grow and pass Start quickly upon release from the AID block?

5. A more precise description of key technical details would help the reader understand the exact experimental set-up and results.

(a) How were cell volumes estimated? half peak height on the daughter size of the size plot or mode size?

(b) For all elutriation experiments, the volume of each fraction and the percent budding should be indicated above each lane so that readers can estimate the time of Start. For at least one elutriation experiment, representative size plots and FACS profiles of the starting size fraction and progressive timepoints should be shown to indicate uniformity of the starting distribution and synchronous progression through the cell cycle.

(c) If I interpret the methods correctly, each elutriation-release timepoint represents a fixed volume of cells and samples are not normalized for total protein content prior to loading on gels. This should be stated explicitly when the elutriation experiments are first described so that readers understand that the signals are intrinsically normalized on a per cell basis (for cells still in G1 phase), i.e., the signal on the blots represents copy number, not concentration. Also, what fraction of each 1.6 mL timepoint was loaded on the gels?

6. The main unanswered question is the mechanism whereby Cln3 accumulation is prevented by sphingolipid accumulation. If, for example, Cln3 translation is dependent on Ypk1/2 activity, might this be mediated by uORFs in the CLN3 mRNA? Testing effects on a version of CLN3 that lacks the uORFs would answer this question. Or is Cln3 itself a substrate for Ypk1/2, perhaps stabilizing Cln3 to allow its accumulation? Or is there an effect on CLN3 mRNA levels? Also, does the feedback inhibition from sphingosine affect TORC2 activity, which in turn may affect Cln3? Does a reduction in TORC2 activity affect Cln3 accumulation? While these experiments are probably beyond the scope of the study, some discussion is warranted.

7. The results of this study underscore the major question of how cells in poor nutrients with low levels of Cln3 (and also Cln1/2, shown by Dorsey et al., 2018) activate Start at a smaller cell size, i.e., despite low Cln-Cdc28 activity. One possible contribution is the higher levels of G1/S transcription factors (Swi4, Swi6, Mbp1) in poor nutrients described by Dorsey et al., (2018), particularly for Mbp1, which is not inhibited by Whi5. This result should be cited along with the observation from Dorsey et al., that G1/S TF abundance is limiting with respect to promoter/TF binding site copy number in early G1 phase cells. Also, on a note for citation completeness, Dorsey et al., should be cited in the sentence "…consistent with several previous studies that used fluorescence microscopy to analyze Whi5 levels (Schmoller et al., 2015; Litsios et al., 2019)." on p5.

8. Other literature considerations:

(a) Importantly, the authors point out that some of the key experiments in the Schmoller paper were obtained with a bck2 strain that is defective for Start (which they replicate) and then show these effects do not hold in wild type cells. Costanzo et al., (2004) reported that overexpression of GAL1-WHI5 in a cln3 strain causes a G1 arrest, and that overexpression in a wild type strain causes a G1 delay. Similarly, overexpression of GAL1-WHI5 causes a severe growth defect in cln1 cln2 strain background. In addition, Costanzo et al., found that GAL1-WHI5 had only moderate effects on size and cell cycle distribution in wild type cells. These results are consistent with the authors' WHI5 dosage results and should be cited appropriately.

(b) The proposed links between the secretory system and size are intriguing. In this regard, it is worth noting that two previous studies uncovered a direct connection between the secretory system, ribosome biogenesis and cell size (Lempiäinen et al., 2009; Singh and Tyers 2009). The authors cite previous studies purporting ER localization of Cln3 as a possible mechanism for sphingolipid effects but this data should be interpreted cautiously given the difficulty in detecting low abundance Cln3 signals at the single cell level.

(c) A key previous result often overlooked in nutrient shift experiments is that rates of rRNA transcription reach maximal levels within a few minutes of shift from ethanol to glucose medium whereas protein synthetic rate does not achieve a maximum until nearly an hour after the shift (Kief and Warner 1981). Because cells adjust their critical cell-size threshold very rapidly upon nutrient shifts (Lorincz and Carter 1979), this result suggests that the rate of ribosome biogenesis and not protein synthesis per se sets the size threshold. The work of Keif and Warner could be cited in the sentences "In both budding yeast and fission yeast, cells rapidly readjust the threshold amount of growth required for cell cycle progression when shifted to nutrient conditions that support different growth rates (Fantes and Nurse 1977; Johnston et al., 1979; Lucena et al., 2018). In fission yeast, the threshold appears to be readjusted within minutes." on p8.

(d) In addition to previous single cell studies on Whi5 abundance as a function of size using live cell imaging in either time course studies or fixed single images, a recent publication using PALM imaging also showed that Whi5 is not diluted as a function of size (Black et al., 2019).

Reviewer #2 (Recommendations for the authors):

My major recommendations are included in the public review. I would be particularly interested to see the authors test the bck2∆ strain from the Schmoller paper with by western blot experiments as in Figure 1.

Reviewer #3 (Recommendations for the authors):

1. Cln3 was made detectable on western blots by tagging with 6xHA. It is crucial to demonstrate that tinkering with the 3'UTR of CLN3 does not lead to changes in Cln3 levels or activity. Indeed, it is published that a tagged version of Cln3 (CLN3-PrA) exhibits about a 10% increase in cell volume (Cross et al., 2002). This can be done easily by comparing cell size distribution profiles of WT and CLN3-6xHA cells, in rich and poor medium, as done for other constructs in Figure 3A, 3C and Figure 3_SF1. The same should be done with Whi5-3xHA.

2. The most dramatic increase in Cln3 levels occurs when G1 cells previously grown in poor carbon medium are transferred to rich carbon source (dextrose). It is well established that a sudden rewiring of carbon metabolism ('sugar shock') delays CLN1,2 transcription and requires higher Cln3 accumulation (e.g. Baroni et al., 1994; Tokiwa et al., 1994; Flick et al., 1998). It is important to determine how much Cln3 has to accumulate before Start occurs when elutriated small G1 cells pre-grown in rich medium are transferred to the same rich medium.

3. Figure 5 seems to rule out that cell size at Start is controlled by the Sch9 kinase regulated by TORC1, which is at odds with previous reports showing reduced cell growth and smaller size of sch9∆ cells. What is the rationale of adding 1-NM-PP1 45 minutes after inoculation of small G1 cells in the medium and not earlier. Moreover, the experiment in Figure 5 does not control for Sch9-as inactivation by 1-NM-PP1. This could be done by looking at phosphorylation of the downstream target Rsp6 (González et al., 2015). As the TOR1/Sch9 pathway promotes protein synthesis (Eltschinger and Loewith, 2016), it is not surprising that inhibition, even partial, of Sch9 leads to a decrease in Cln3 levels. It is important that the authors show the status of the loading control that they use throughout the manuscript and quantify the level of Cln3 as it is performed on figure 1 for instance. It is also important to control that 1-NM-PP1 has no off-target effects on growth or budding by monitoring its effect on WT cells (as done for 3-MOB in Figure 6 SF1). As it is, the experiment is unconclusive and should be put in SupFigure at best.

4. Figure 6E is used to indicate that Ypk1 phosphorylation by TORC2 might favour Cln3 accumulation. This conclusion seems premature as slow-migrating Ypk1 bands appear at the time (10 min) when Cln3 also up-shifts and disappears, suggesting in fact that Ypk1 phosphorylation might correlate with Cln3 proteolysis. The same figure 6E is described in two paragraphs (on page 11 and 12) that should be merged. It is puzzling to read that the conclusions written in each paragraph differ from one another. In the first one the authors claim that 'A reappearance of Cln3 at the end of the time course was accompanied by another decrease in Ypk1 phosphorylation'. In the second one, they claim that 'The increase in Cln3 that occurred later in the time course as cells adapted to the new carbon source was correlated with an increase in TORC2-dependent phosphorylation of Ypk1/2.' The latter seems to be supported by the WB displayed. Nevertheless, a quantification of Cln3 level, Ypk1 phosphorylation level along with the level of the loading control would be helpful to support the claim. To this reviewer knowledge, Cln3 ubiquitin-dependent proteolysis is triggered by Cdk1 mediated phosphorylation. The positive arrow connecting Ypk1,2 and Cln3 in the model presented in Figure 7 is not sustained conclusively by the data.

Comments and suggestions

1. There is lots of emphasis on Cln3 and Whi5 levels, but what counts for Start is the phosphorylation status or nuclear exclusion of Whi5. Whi5 appears as a doublet on some western blots (e.g. Figure 2, Figure 3B,D), a putative proxy for Whi5 phosphorylation. This could be used to determine which other kinases or phosphatases control Whi5's phosphorylation status when Cln3 levels are very low in cells grown on poor carbon medium.

2. It is said at several places in the manuscript that Cln3 accumulates gradually during G1. This may be true when looking at large cell populations, but Litsios et al., have shown that Cln3 is synthesized in bursts during G1, following metabolic oscillations. These bursts do not take place at exactly the same time/size in all cells, but may translate as a gradual/linear increase when looking at the population of cells. This could be worth mentioning in the manuscript.

3. In all but one experiment (Figure 3 —figure supplement 1B), cells were grown in a rich-nutrient medium (YP, 1% yeast, 2% peptone) that contained 2 types of carbon sources, either dextrose or 2% glycerol/2% ethanol referred to as rich and poor carbon medium, respectively. In the experiment displayed on Figure 3 —figure supplement 1B, cells were grown in a poor-nutrient synthetic complete medium. However, the authors mix rich or poor carbon sources with rich or poor nutrients. For the sake of clarity, they should use rich or poor carbon on all figures panels and keep rich or poor nutrients only for Figure 3 —figure supplement 1B.

4. Based on their data the authors argue that it is the increase of Cln3 concentration rather than the dilution of Whi5 that would control cell cycle entry. The key quantification of this is shown in Figure 1 SF1C,D and should be moved to the main panel.

5. p17-18: To support the claim that accumulation of sphingolipid would inhibit the Ypk1/2-dependent increase in Cln3, the authors could monitor the phosphorylation status of Ypk1/2 in sec7 mutants and/or in cells incubated with phytosphingosine. Alternatively, they could analyze whether addition of phytosphingosine on ypk1∆ or ypk1-as ypk2∆ mutant leads to a further decrease of Cln3 level or not.

Data presentation

1. Figure 1E: It should be mentioned that it represents the quantification of the WB displayed in D, and if the signal was normalized to a loading control. Please explain why this was not done if it was not.

2. Figure 3 and figure 3 —figure supplement 1: Two different genotypes are used 2xWHI5 and Wild type + WHI5. It is not clear whether there is a difference between them. If there is none, the use of only 1 genotype would help for the comprehension. It there is a difference, it should be explained.

3. Figure 3D: Contrary to the authors sentence (p9 3rd paragraph) Cln3 level is not displayed on the panel. If it is relevant, it should be shown. If not, the sentence should be changed. The authors should add a loading control for this panel, which is not shown in figure 3 —figure supplement 1E as stated in the legend of the figure.

4. Figure 5: the panel of the loading control is missing in the figure, while it is mentioned in the legend. It should be added.

5. Figure 6 —figure supplement 1B is not cited in the text. A loading control should be added to the western blot.

6. p12, 4th paragraph; figure 7C and 7D are not cited accurately in the text.

7. Table 1: Some strains are not referenced correctly to the corresponding figures. For DK3572, it should be S3C, for DK3743, it should be S3D. The strain 2xWHI5 used in figure 3 —figure supplement 1 is not referenced in the table 1.

8. Overall, the manuscript is quite long with some repeats, and would benefit from some shortening.

– Figure 4 A,B and the figure 5 —figure supplement 1 could be merged in a unique figure and the corresponding text simplified, as the conclusions drawn from these results are similar.

– p12, 3rd paragraph; the first and third sentence are redundant. It should be rephrased.

– The Discussion could be shortened as results or interpretations already presented in Results are merely rephrased, for instance p15 1st paragraph, p16 4th and 5th paragraphs…

https://doi.org/10.7554/eLife.64364.sa1

Author response

Essential revisions:

As you will see below, all reviewers are highly supportive of the work, but asked for some clarification and some new data. In your revision, please address all points described below.

Reviewer #1 (Recommendations for the authors):

The experiments are performed to a high technical standard and in numerous biological replicates so the main conclusions are sound. My main concerns pertain to a few particular aspects of the results that are not explicitly discussed by the authors. While there are obviously many further experiments to be done to sort out the mechanism whereby Yck1/2 and sphingolipid metabolism control Cln3, these lie outside the reasonable scope of this study in my opinion. Nevertheless, some further discussion of possible mechanisms seems warranted. Finally, a few additional published observations should be mentioned and cited.

1. It is somewhat surprising that cells grown in glucose don't grow faster and that budding upon inoculation in rich is slower than in poor nutrients as a function of time. I would have expected the much faster growth rate in glucose to result in temporally faster budding than in glycerol/ethanol medium. What are the doubling times observed in rich versus poor medium?

For the experiment shown in Figure 1, cells were first grown overnight in YP media containing a poor carbon source (2% glycerol/2% ethanol). Under these conditions, cells are born at a very small size and spend a longer time growing in G1 phase, which facilitates isolation of a uniform population of small unbudded cells in early G1 phase. After isolation of small unbudded cells by elutriation, the cells were divided and half were grown in the same media, while the other half were shifted to a rich carbon source (2% glucose). The cells shifted to rich carbon grow faster but they also immediately increase the growth requirement for cell cycle entry and therefore must undergo an unusually long G1 phase to reach the larger cell size required for cell cycle entry in rich carbon. In contrast, the cells that remain in poor carbon can undergo cell cycle entry at a much smaller size and therefore spend less time in G1 phase. We have edited the text to highlight these issues and how the influence interpretation of the results. See also our response to Reviewer 2.

In a previous publication, we provided detailed information regarding growth rates during each phase of the cell cycle in cells grown continuously in rich or poor carbon (PMID: 28939614). This showed that growth rate varies during the cell cycle, and that cells in rich carbon grow 2-3 fold faster than cells in poor carbon.

2. It is also a bit puzzling that there is a transient burst of Cln3 immediately after inoculation into poor medium. This could be due to residual glucose in the yeast extract used to formulate YP medium, i.e., cells metabolize the small amount of glucose in the yeast extract before shifting to non-fermentative metabolism. This can be avoided by depleting YP of glucose by pre-culture with yeast in the absence of added carbon source (e.g., as a 5X YP stock, such that the yeast quickly consume all glucose then cease growth, the stock is then clarified and sterilized before use in subsequent experiments). Alternatively, to maintain complete nutrient homeostasis, cells can be elutriated in the clarified medium in which they were grown prior to elutriation and then diluted into the same medium. The transient burst doesn't affect the authors conclusions but a similar burst has been observed in other studies where residual glucose was not properly accounted for. The authors may wish to mention this.

In Figure 2, the cells are shifted from a very rich carbon medium (YP + 2% glucose) into a very poor carbon medium (YP + 2% glycerol/ethanol). We are careful to ensure that the cells pre-grown in rich glucose media are kept at low density so that they have not significantly depleted the glucose. Therefore, even if there is residual glucose in the yeast extract used to formulate YP medium, the concentration of that residual glucose must be far lower than the 2% glucose in the rich carbon medium so that the cells experience a dramatic reduction in the availability of glucose. We therefore think that it is unlikely that the transient burst in Cln3 is due to residual glucose, but we would be happy to mention this possibility if requested by the reviewer.

3. It appears that the Cln3 signal reaches a maximum well before Start (e.g., 30-50 min before 20% budding is reached in most experiments) but the authors do not comment on this finding. This result aligns with recent single cell analysis from the Heinemann lab, which also suggests that Cln3 peaks well before Start (Litsios et al., 2019). It would be worth addressing this point in the discussion, including the caveat that population level studies can obscure single cell effects.

We agree and have added a reference to the Litsios et al., results in our discussion of the data shown in Figure 1. We have also included new text to highlight the fact that population-level analysis can miss effects seen in single cells.

4. Why is bulk protein synthesis is not affected in the SEC7-AID strain despite the complete block to growth and what does this mean? Is the protein concentration going up in the arrested G1 phase cells? Or is there a higher rate of protein degradation to compensate for on-going synthesis? It does appear that the loading control increases throughout the experiment consistent with the former possibility. It would be useful for the authors to show the 35S incorporation data for the elutriated cultures so that the effect in G1 phase can be assessed (this is currently cited as data not shown). If the protein concentration in the arrested cells is higher, do the arrested cells rapidly grow and pass Start quickly upon release from the AID block?

Classic work carried out by the Schekman lab found that protein production continues after an arrest of membrane traffic, and the resulting increase in cell density was a key tool in the genetic screens that led to isolation of the sec mutants. The western blot data in Figure 4E and the S³⁵ incorporation data for log phase cells shown in Figure 4F are consistent with the Schekman results. We have included S³⁵ incorporation data for the synchronized sec7-AID cells in Figure 4 —figure supplement 1D, which are also consistent with the Schekman results.

The idea that accumulation of extra protein in growth-arrested sec7-AID cells could lead to faster growth and cell cycle entry upon release from the arrest is intriguing. However, we have not yet tested this due to concerns that we would lack a suitable reference control that would allow us to determine if the cells pass Start more quickly. There is also a concern that the sec7-AID cells could require extra time to recover from the disruption of membrane traffic.

5. A more precise description of key technical details would help the reader understand the exact experimental set-up and results.

(a) How were cell volumes estimated? half peak height on the daughter size of the size plot or mode size?

The reported cell volumes are median cell size. This information has been included in the Methods section.

(b) For all elutriation experiments, the volume of each fraction and the percent budding should be indicated above each lane so that readers can estimate the time of Start. For at least one elutriation experiment, representative size plots and FACS profiles of the starting size fraction and progressive timepoints should be shown to indicate uniformity of the starting distribution and synchronous progression through the cell cycle.

We have added a new supplemental figure that shows Coulter Counter plots for successive time points for the elutriation experiment in Figure 1, which shows a tight starting distribution that increases in size with time (Figure 1 —figure supplement 1A).

Space is very tight in Figure 1 so we found that we were not able to add volumes and percent budding above each lane without disrupting the layout of the figure. Adding the numbers also made the panel look more crowded, and it seemed to repeat data in the other plots showing percent bud emergence and change in volume. We would therefore prefer to keep the data about percent budding and volume in separate plots, but would be happy to try to rearrange the data if requested by the reviewer.

(c) If I interpret the methods correctly, each elutriation-release timepoint represents a fixed volume of cells and samples are not normalized for total protein content prior to loading on gels. This should be stated explicitly when the elutriation experiments are first described so that readers understand that the signals are intrinsically normalized on a per cell basis (for cells still in G1 phase), i.e., the signal on the blots represents copy number, not concentration. Also, what fraction of each 1.6 mL timepoint was loaded on the gels?

This is a great suggestion that would help clarify the manuscript. We have added information to the main text to explain that a constant number of cells is collected for each time point. For each time point, cells are lysed into 140 ul of sample buffer and 15 ul are loaded on the gel.

6. The main unanswered question is the mechanism whereby Cln3 accumulation is prevented by sphingolipid accumulation. If, for example, Cln3 translation is dependent on Ypk1/2 activity, might this be mediated by uORFs in the CLN3 mRNA? Testing effects on a version of CLN3 that lacks the uORFs would answer this question. Or is Cln3 itself a substrate for Ypk1/2, perhaps stabilizing Cln3 to allow its accumulation? Or is there an effect on CLN3 mRNA levels? Also, does the feedback inhibition from sphingosine affect TORC2 activity, which in turn may affect Cln3? Does a reduction in TORC2 activity affect Cln3 accumulation? While these experiments are probably beyond the scope of the study, some discussion is warranted.

We agree that these are amongst the most important next questions and that they are beyond the scope of the current study. We have added new text to highlight these questions.

7. The results of this study underscore the major question of how cells in poor nutrients with low levels of Cln3 (and also Cln1/2, shown by Dorsey et al., 2018) activate Start at a smaller cell size, i.e., despite low Cln-Cdc28 activity. One possible contribution is the higher levels of G1/S transcription factors (Swi4, Swi6, Mbp1) in poor nutrients described by Dorsey et al., (2018), particularly for Mbp1, which is not inhibited by Whi5. This result should be cited along with the observation from Dorsey et al., that G1/S TF abundance is limiting with respect to promoter/TF binding site copy number in early G1 phase cells. Also, on a minor note for citation completeness, Dorsey et al., should be cited in the sentence "…consistent with several previous studies that used fluorescence microscopy to analyze Whi5 levels (Schmoller et al., 2015; Litsios et al., 2019)." on p5.

We have added new text to cite the results and model of Dorsey et al., We have also added the missing Dorsey et al., citation.

8. Other literature considerations:

(a) Importantly, the authors point out that some of the key experiments in the Schmoller paper were obtained with a bck2 strain that is defective for Start (which they replicate) and then show these effects do not hold in wild type cells. Costanzo et al., (2004) reported that overexpression of GAL1-WHI5 in a cln3 strain causes a G1 arrest, and that overexpression in a wild type strain causes a G1 delay. Similarly, overexpression of GAL1-WHI5 causes a severe growth defect in cln1 cln2 strain background. In addition, Costanzo et al., found that GAL1-WHI5 had only moderate effects on size and cell cycle distribution in wild type cells. These results are consistent with the authors' WHI5 dosage results and should be cited appropriately.

We agree and have included the citation for the Costanzo data showing that GAL1-WHI5 has modest effects. We were uncertain how the GAL1-WHI5 expression data in cln3∆ and cln1/2∆ cells is relevant in the context of the dilution model and therefore have not mentioned those results, but please let us know if we are missing something.

(b) The proposed links between the secretory system and size are intriguing. In this regard, it is worth noting that two previous studies uncovered a direct connection between the secretory system, ribosome biogenesis and cell size (Lempiäinen et al., 2009; Singh and Tyers 2009). The authors cite previous studies purporting ER localization of Cln3 as a possible mechanism for sphingolipid effects but this data should be interpreted cautiously given the difficulty in detecting low abundance Cln3 signals at the single cell level.

We agree and have added the additional citations. We were unaware of these papers and found them to be quite interesting.

(c) A key previous result often overlooked in nutrient shift experiments is that rates of rRNA transcription reach maximal levels within a few minutes of shift from ethanol to glucose medium whereas protein synthetic rate does not achieve a maximum until nearly an hour after the shift (Kief and Warner 1981). Because cells adjust their critical cell-size threshold very rapidly upon nutrient shifts (Lorincz and Carter 1979), this result suggests that the rate of ribosome biogenesis and not protein synthesis per se sets the size threshold. The work of Keif and Warner could be cited in the sentences "In both budding yeast and fission yeast, cells rapidly readjust the threshold amount of growth required for cell cycle progression when shifted to nutrient conditions that support different growth rates (Fantes and Nurse 1977; Johnston et al., 1979; Lucena et al., 2018). In fission yeast, the threshold appears to be readjusted within minutes." on p8.

The Johnston et al., 1979 article that we cited in the original version includes Lorincz and Carter, as well as the data showing that critical cell size is rapidly reset after a nutrient shift. We were not able to find another Lorincz and Carter article on PubMed, so we believe that the Johnston article is the relevant article, but please let us know if we missed an article. We have also included a citation to the Kief and Warner article.

(d) In addition to previous single cell studies on Whi5 abundance as a function of size using live cell imaging in either time course studies or fixed single images, a recent publication using PALM imaging also showed that Whi5 is not diluted as a function of size (Black et al., 2019).

We agree and have added the Black et al., citation in several places in the manuscript.

Reviewer #2 (Recommendations for the authors):

My major recommendations are included in the public review. I would be particularly interested to see the authors test the bck2∆ strain from the Schmoller paper with by western blot experiments as in Figure 1.

We are curious about this too, and there are a number of other experiments that we have considered carrying out to reconcile our results with the Schmoller et al., results. However, we’ve had our hands full with the other experiments so we found it necessary to put these experiments on hold. In addition, we have wanted to focus as much as possible on how cell size control works in wild type cells, rather than in mutant backgrounds.

Reviewer #3 (Recommendations for the authors):

1. Cln3 was made detectable on western blots by tagging with 6xHA. It is crucial to demonstrate that tinkering with the 3'UTR of CLN3 does not lead to changes in Cln3 levels or activity. Indeed, it is published that a tagged version of Cln3 (CLN3-PrA) exhibits about a 10% increase in cell volume (Cross et al., 2002). This can be done easily by comparing cell size distribution profiles of WT and CLN3-6xHA cells, in rich and poor medium, as done for other constructs in Figure 3A, 3C and Figure 3_SF1. The same should be done with Whi5-3xHA.

We agree that this is a good test and the data are shown in a new figure panel (Figure 1 —figure supplement 1B). We found that Cln3-6XHA causes a slight increase in cell size in YP medium containing rich carbon, as observed previously for Cln3 tagged with protein A (Cross, 2002). Cln3-6XHA caused little effect on cell size in YP medium containing poor carbon. A previous study found that Cln3 tagged with protein A rescues a complete loss of function of CLN3 in genetic backgrounds that require CLN3 function, which suggests that tagged versions of Cln3 retain most of the normal functions of Cln3 (Cross, 2002). It would be ideal to have an antibody that recognizes Cln3. We have successfully raised antibodies that recognize over 25 proteins in our lab, but Cln3 has been uniquely difficult. We attempted to express multiple regions of the protein in E. coli, but in every case the proteins have been completely insoluble and refractory to purification. We are still working on making a Cln3 antibody and our current approach is to purify large amounts of stabilized Cln3 from yeast using methods similar to those used by the Diffley lab to purify DNA replication proteins. However, even in a best case scenario an antibody is at least 6 months away and the approach may not work.

Previous studies found that fluorescently tagged versions of Whi5 do not cause a decrease in cell size, which suggested that tagged versions of Whi5 are fully functional. We found that Whi5-3XHA cells are slightly smaller than wild type cells (Figure 1 —figure supplement 1B). We also found that Whi5-13Myc tag does not cause a decrease in cell size, which suggests that the HA tag sequence has a mild effect on Whi5 function. We further found that Whi5-3XHA and Whi5-13Myc showed the same behavior in elutriated cells undergoing growth in G1 phase. We therefore used the Whi5-3XHA so that the behavior of Cln3-6XHA and Whi5-3XHA could be directly compared in the same western blot samples. We have included new text in the Results section describing the properties of the Cln3 and Whi5 fusion proteins. In addition, we repeated the TEF1-WHI5 expression experiments in Figure 3 using untagged versions of WHI5.

This important issue was a blind spot for us in the original version, so we thank the reviewer for pointing it out.

2. The most dramatic increase in Cln3 levels occurs when G1 cells previously grown in poor carbon medium are transferred to rich carbon source (dextrose). It is well established that a sudden rewiring of carbon metabolism ('sugar shock') delays CLN1,2 transcription and requires higher Cln3 accumulation (e.g. Baroni et al., 1994; Tokiwa et al., 1994; Flick et al., 1998). It is important to determine how much Cln3 has to accumulate before Start occurs when elutriated small G1 cells pre-grown in rich medium are transferred to the same rich medium.

We attempted to grow cells in YPD, isolate small unbudded cells, and then release the cells back into YPD. However, we found that it was not possible to isolate a uniform population of small unbudded cells under these conditions. The problem is that very little growth occurs in G1 phase in YPD so that newly born cells are nearly the same size as mother cells (PMID: 28939614). This, combined with the normal variation in cell size observed in wild type yeast, means that elutriation yields a mix of unbudded and budded cells. Others have faced the same problem (PMID: 31685990, 10728640). The fact that so little growth occurs in G1 phase in YPD is an additional argument against the idea that dilution of Whi5 plays a substantial and general role in cell size control.

As an alternative, we grew cells in complete synthetic medium (CSM) containing 2% glucose. Under these conditions, cells grow more slowly and are smaller because CSM is limiting for nutrients other than glucose. We isolated small unbudded cells and released them into the same medium so that there would not be shift in carbon source. We found that Cln3 levels increase 3-fold, while Whi5 levels do no change substantially, similar to the results for cells growing in YP medium containing poor carbon. These data are shown in a new figure (Figure 1 —figure supplement 2).

3. Figure 5 seems to rule out that cell size at Start is controlled by the Sch9 kinase regulated by TORC1, which is at odds with previous reports showing reduced cell growth and smaller size of sch9∆ cells. What is the rationale of adding 1-NM-PP1 45 minutes after inoculation of small G1 cells in the medium and not earlier. Moreover, the experiment in Figure 5 does not control for Sch9-as inactivation by 1-NM-PP1. This could be done by looking at phosphorylation of the downstream target Rsp6 (González et al., 2015). As the TOR1/Sch9 pathway promotes protein synthesis (Eltschinger and Loewith, 2016), it is not surprising that inhibition, even partial, of Sch9 leads to a decrease in Cln3 levels. It is important that the authors show the status of the loading control that they use throughout the manuscript and quantify the level of Cln3 as it is performed on figure 1 for instance. It is also important to control that 1-NM-PP1 has no off-target effects on growth or budding by monitoring its effect on WT cells (as done for 3-MOB in Figure 6 SF1). As it is, the experiment is unconclusive and should be put in SupFigure at best.

The logic for adding inhibitor at 45 minutes was that we wanted to see whether inhibition of sch9 caused rapid loss of Cln3 levels, as we observed for inhibition of ypk1-as. We were also concerned that adding inhibitor earlier would stop the cell cycle before Cln3 accumulation had been initiated.

The sch9-as allele was first reported by Jorgensen et al., 2004, who showed that sch9-as cells are smaller than wild type cells in the absence of inhibitor, consistent with the fact that most analog-sensitive kinases have partially compromised function. Addition of the inhibitor made the cells very small, consistent with complete or nearly complete inhibition of sch9-as kinase activity. We obtained the same results for our sch9-as strain, which verified the version of the strain in our collection (see below). The fact that addition of inhibitor causes a transient loss of Cln3 in sch9-as cells provides further evidence that the inhibitor is working.

We have included the loading control for the sch9-as experiment and quantified Cln3 levels (Figures 5C and Figure 5 —figure supplement 1D). Our data are not necessarily at odds with the previous sch9∆ data. Both sch9∆ and prolonged inhibition of sch9-as cause large reductions in growth rate and cell size. Our data suggest that in both cases the reductions could be a consequence of long term effects on the rate of ribosome biogenesis.

4. Figure 6E is used to indicate that Ypk1 phosphorylation by TORC2 might favour Cln3 accumulation. This conclusion seems premature as slow-migrating Ypk1 bands appear at the time (10 min) when Cln3 also up-shifts and disappears, suggesting in fact that Ypk1 phosphorylation might correlate with Cln3 proteolysis. The same figure 6E is described in two paragraphs (on page 11 and 12) that should be merged. It is puzzling to read that the conclusions written in each paragraph differ from one another. In the first one the authors claim that 'A reappearance of Cln3 at the end of the time course was accompanied by another decrease in Ypk1 phosphorylation'. In the second one, they claim that 'The increase in Cln3 that occurred later in the time course as cells adapted to the new carbon source was correlated with an increase in TORC2-dependent phosphorylation of Ypk1/2.' The latter seems to be supported by the WB displayed. Nevertheless, a quantification of Cln3 level, Ypk1 phosphorylation level along with the level of the loading control would be helpful to support the claim. To this reviewer knowledge, Cln3 ubiquitin-dependent proteolysis is triggered by Cdk1 mediated phosphorylation. The positive arrow connecting Ypk1,2 and Cln3 in the model presented in Figure 7 is not sustained conclusively by the data.

We agree that our interpretation of the data in Figure 6E were confusing in the original version. Part of the confusion may arise from a lack of clarity in our writing and in the literature about the different phosphorylation inputs into Ypk1/2. The literature suggests that changes in the electrophoretic mobility of Ypk1 could be due largely to the Fpk1/2 kinases. TORC2 also influences Ypk1/2 phosphorylation, as detected by a phosphospecific antibody, but it remains unclear whether TORC2 also influences the electrophoretic mobility of Ypk1/2. The data suggest phosphorylation of Ypk1/2 that can be detected via electrophoretic mobility shifts is correlated with Cln3 levels, while TORC2-dependent phosphorylation with a phosphospecific antibody is not well correlated with Cln3 levels. We have edited the manuscript to make this more clear and to clarify what can and cannot be concluded from the data.

After editing this section for greater clarity, it seems to make sense to keep the two paragraphs separate to emphasize that electrophoretic mobility shifts of Ypk1/2 and the phosphospecific antibody are likely measuring different inputs into Ypk1/2.

Comments and suggestions

1. There is lots of emphasis on Cln3 and Whi5 levels, but what counts for Start is the phosphorylation status or nuclear exclusion of Whi5. Whi5 appears as a doublet on some western blots (e.g. Figure 2, Figure 3B,D), a putative proxy for Whi5 phosphorylation. This could be used to determine which other kinases or phosphatases control Whi5's phosphorylation status when Cln3 levels are very low in cells grown on poor carbon medium.

We agree, but feel that further experiments along these lines are beyond the scope of this study.

2. It is said at several places in the manuscript that Cln3 accumulates gradually during G1. This may be true when looking at large cell populations, but Litsios et al., have shown that Cln3 is synthesized in bursts during G1, following metabolic oscillations. These bursts do not take place at exactly the same time/size in all cells, but may translate as a gradual/linear increase when looking at the population of cells. This could be worth mentioning in the manuscript.

We agree and have edited the manuscript to mention that our population-level studies could miss the transient bursts in translation detected by Litsios.

3. In all but one experiment (Figure 3 —figure supplement 1B), cells were grown in a rich-nutrient medium (YP, 1% yeast, 2% peptone) that contained 2 types of carbon sources, either dextrose or 2% glycerol/2% ethanol referred to as rich and poor carbon medium, respectively. In the experiment displayed on Figure 3 —figure supplement 1B, cells were grown in a poor-nutrient synthetic complete medium. However, the authors mix rich or poor carbon sources with rich or poor nutrients. For the sake of clarity, they should use rich or poor carbon on all figures panels and keep rich or poor nutrients only for Figure 3 —figure supplement 1B.

We agree and have edited to clarify which media are used in each experiment.

4. Based on their data the authors argue that it is the increase of Cln3 concentration rather than the dilution of Whi5 that would control cell cycle entry. The key quantification of this is shown in Figure 1 SF1C,D and should be moved to the main panel.

We agree that these data are important, but we’re out of space in Figure 1. We therefore think it is best to leave the data in the supplement. Another option would be to split Figure 1 into two figures – we would be happy to do this if requested by the reviewer.

5. p17-18: To support the claim that accumulation of sphingolipid would inhibit the Ypk1/2-dependent increase in Cln3, the authors could monitor the phosphorylation status of Ypk1/2 in sec7 mutants and/or in cells incubated with phytosphingosine. Alternatively, they could analyze whether addition of phytosphingosine on ypk1∆ or ypk1-as ypk2∆ mutant leads to a further decrease of Cln3 level or not.

In a previous study, we showed that addition of PHS to cells causes a rapid loss of TORC2-dependent phosphorylation of Ypk1/2, consistent with a model in which accumulation of PHS inhibits Ypk1/2-dependent signals that promote production of Cln3 (PMID: 29290562). However, we also found that inactivation of Sec7 during growth in G1 phase leads to an increase in TORC2-dependent phosphorylation of Ypk1/2, consistent with a previous study that found similar results for Sec6 in cells synchronized with mating pheromone (PMID: 28794263). We include a discussion of both of these results in the Discussion. To explain the difference in the results, we hypothesize that there are distinct pools of Ypk1/2 at different locations in the cell that undergo differential regulation. One pool of Ypk1/2 could be at the ER, where it controls sphingolipid synthesis and Cln3 levels, and is subject to feedback control by excess sphingolipids. A distinct pool at the plasma membrane could become hyperactive in response to inactivation of Sec7 but not have the capability to influence Cln3 levels. We include a discussion of evidence for two distinct pools of Ypk1/2, as well as evidence for connections between Cln3 regulation and the ER. The complex connections between TORC2, Ypk1/2, sphingolipids, the ER, and Cln3 suggest that we have long ways to go to understand how cell growth and cell cycle progression are mechanistically linked.

Data presentation

1. figure 1E: It should be mentioned that it represents the quantification of the WB displayed in D, and if the signal was normalized to a loading control. Please explain why this was not done if it was not.

We have now mentioned that the data in Figure 1E are a quantification of the western blot data in Figure 1D. We wanted the data to provide a measure of the increase in the number of Cln3 molecules per cells as a function of growth. Therefore, we did not normalize to the loading control because levels of the loading control protein increase with growth, so normalizing to the loading control would partially cancel out the growth-dependent increase in Cln3 protein molecules per cell.

2. Figure 3 and figure 3 —figure supplement 1: Two different genotypes are used 2xWHI5 and Wild type + WHI5. It is not clear whether there is a difference between them. If there is none, the use of only 1 genotype would help for the comprehension. It there is a difference, it should be explained.

We agree that it was unclear. The 2xWHI5 genotypes are the same in Figure 3 and Figure 3 —figure supplement 1. The figure inset has been changed to “2xWHI5” to avoid confusion. The strain table also denotes which genotypes are used in each figure panel to help avoid confusion.

3. Figure 3D: Contrary to the authors sentence (p9 3rd paragraph) Cln3 level is not displayed on the panel. If it is relevant, it should be shown. If not, the sentence should be changed. The authors should add a loading control for this panel, which is not shown in figure 3 —figure supplement 1E as stated in the legend of the figure.

The text has been changed to reflect that Cln3 is not in this figure, and a loading control is now included. This figure is now Figure 3 —figure supplement 1C.

4. Figure 5: the panel of the loading control is missing in the figure, while it is mentioned in the legend. It should be added.

We have included a loading control in Figure 5.

5. Figure 6 —figure supplement 1B is not cited in the text. A loading control should be added to the western blot.

Figure 6 —figure supplement 1B is now cited in the text, and we have included the loading control in Figure 6 —figure supplement 1B

6. p12, 4th paragraph; figure 7C and 7D are not cited accurately in the text.

This has been corrected.

7. Table 1: Some strains are not referenced correctly to the corresponding figures. For DK3572, it should be S3C, for DK3743, it should be S3D. The strain 2xWHI5 used in figure 3 —figure supplement 1 is not referenced in the table 1.

The strains listed above were corrected in Table 1 and the other strain figure references were double checked and corrected as necessary.

8. Overall, the manuscript is quite long with some repeats, and would benefit from some shortening.

– Figure 4 A,B and the figure 5 —figure supplement 1 could be merged in a unique figure and the corresponding text simplified, as the conclusions drawn from these results are similar.

We tried to do this but found that merging the two figures disrupted the logic in the text describing the figures.

– p12, 3rd paragraph; the first and third sentence are redundant. It should be rephrased.

We agree and have edited accordingly.

– The Discussion could be shortened as results or interpretations already presented in Results are merely rephrased, for instance p15 1st paragraph, p16 4th and 5th paragraphs…

We agree and have edited the section that starts on p. 15. We also deleted the section on Sch9 in the Discussion because it repeated information from the Results section.

https://doi.org/10.7554/eLife.64364.sa2

Article and author information

Author details

Robert A Sommer

Department of Molecular, Cell, and Developmental Biology, University of California, Santa Cruz, Santa Cruz, United States

Contribution
Formal analysis, Investigation, Methodology, Writing - original draft, Writing - review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0003-4529-0542
Jerry T DeWitt

Department of Molecular, Cell, and Developmental Biology, University of California, Santa Cruz, Santa Cruz, United States

Contribution
Investigation, Methodology

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0002-9077-643X
Raymond Tan

Department of Molecular, Cell, and Developmental Biology, University of California, Santa Cruz, Santa Cruz, United States

Contribution
Investigation

Competing interests
No competing interests declared
Douglas R Kellogg

Department of Molecular, Cell, and Developmental Biology, University of California, Santa Cruz, Santa Cruz, United States

Contribution
Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Writing - original draft, Writing - review and editing

For correspondence
dkellogg@ucsc.edu

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0002-5050-2194

Funding

National Institutes of Health (GM053959)

Douglas R Kellogg

National Institutes of Health (GM131826)

Douglas R Kellogg

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Andrew Murray, Jan Skotheim, Kurt Schmoller, Matthew Swaffer, Matthias Heinemann, Andreas Milias-Argeitis, and members of the lab for helpful discussions. We also thank Jack Stevenson and Kevan Shokat for providing 3-MOB-PP1. This work was funded by NIH grants GM053959 and GM131826.

Senior and Reviewing Editor

Naama Barkai, Weizmann Institute of Science, Israel

Version history

Preprint posted: October 1, 2020 (view preprint)
Received: October 27, 2020
Accepted: October 29, 2021
Accepted Manuscript published: October 29, 2021 (version 1)
Accepted Manuscript updated: November 1, 2021 (version 2)
Version of Record published: November 15, 2021 (version 3)

Copyright

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.