Decreased mitochondrial membrane potential delays G1-to-S phase progression

To study the effects of mtDNA loss, an mtDNA-devoid (ρ⁰) variant of the mtDNA-containing (ρ⁺) WT strain was made by standard treatment with ethidium bromide (12), and confirmed by quantitative real-time PCR. In accordance with previous reports, ρ⁰ cells grew slower than their ρ⁺ counterparts in a dextrose-containing rich medium, and exhibited an altered cell cycle profile (Figs 1A and S1A) (11, 13). Specifically, the percentage of G1 cells in an early logarithmic phase–unsynchronized culture of ρ⁰ cells was twice as high as in ρ⁺ cells grown under identical conditions; this difference remained constant for at least 2 h during exponential growth and was also observed in cells that had lost their mtDNA because of deficiency of known mtDNA maintenance factors such as Rim1, Mip1, and Mgm101 (Figs 1A and B and S1B and C). The increased frequency of G1 cells in the unsynchronized ρ⁰ culture was attributable to a slower G1-to-S transition rather than a faster progression through other cell cycle stages because α-factor–treated ρ⁰ cells released from G1 synchrony transitioned into the S phase later than their ρ⁺ counterparts (Fig 1C and D; compare timepoints when % of G1 cells starts to decrease or % of S cells peaks in panel D), corroborating previous findings (11). A G1-to-S transition delay of the ρ⁰ variant was observed even after release from G2 synchrony achieved by nocodazole treatment: although ρ⁺ and ρ⁰ cells showed comparable timing of G2 exit (Fig S1D and E), ρ⁰ cells accumulated in G1 and displayed delayed entry into the S phase when compared to ρ⁺ cells (Fig 1E). Using our mtDNA-deficient strain that exhibited the previously reported delay in G1/S transition (5, 11, 14), we proceeded to seek the identity of the proximal signal that triggers the cell cycle phenotype in mtDNA-deficient cells.

Yeast mtDNA encodes proteins that are essential for the ETC and OXPHOS. The impairment of these processes in ρ⁰ cells results in compromised mitochondrial ATP production, decreased ΔΨm, and increased generation of ROS (15, 16, 17), all of which are reasonable candidates for the initial signal that triggers the G1-to-S phase delay in ρ⁰ cells. In line with a previous report (11), treatment of ρ⁺ cells with the ATP synthase inhibitor oligomycin did not cause accumulation of cells in the G1 phase (Figs 2A and S2A); the potency of the oligomycin was confirmed by its effect on mitochondrial membrane potential measured using the fluorescent cationic dye tetramethylrhodamine methyl ester perchlorate (TMRE) (Fig S2B, lower panel). As expected based on the lack of a functional F_o subunit, ρ⁰ cells were not impacted by oligomycin treatment (Fig S2B, upper panel). Decreased mitochondrial ATP synthesis was therefore excluded as a cause of the G1-to-S transition delay observed in ρ⁰ cells.

We next tested the importance of the ΔΨm for normal cell cycle progression. To this end, we treated exponentially growing cells with BAM15, an uncoupler that efficiently dissipates the ΔΨm without depolarizing the plasma membrane (18). Interestingly, BAM15 treatment of ρ⁺ cells caused a transient accumulation of cells in the G1 phase (Figs 2B and S2C). The observed accumulation of cells in the G1 phase was accompanied by a corresponding decrease in the percentage of cells in the S phase. Treatment with another uncoupler, CCCP, caused a similar transient increase in G1 phase cells as BAM15 (Fig S2D and E).

To further corroborate the connection between ΔΨm and G1-to-S progression, we synchronized ρ⁺ cells in G1 with α-factor and released them into a medium containing BAM15. In analogy to the G1-to-S transition delay in ρ⁰ cells, BAM15 treatment delayed the progression of α-factor–synchronized ρ⁺ cells from the G1 into the S phase: although the percentage of G1 cells began to decline 20 min after release of ρ⁺ cells into a normal medium, indicating transition from the G1 into the S phase, ρ⁺ cells released into the BAM15-containing medium did not show signs of G1 exit until 70 min after release (Fig 2C and D). BAM15 treatment also delayed G1-to-S transition in ρ⁺ cells released from G2 synchrony: although the presence of BAM15 did not appreciably affect the transition from the G2 to the G1 phase, ρ⁺ cells released from G2 into the BAM15-containing medium showed an increased percentage of G1 cells and delayed S phase entry, consistent with a specific G1-to-S transition delay (Fig S2G–J).

The results of Fig 2 demonstrate that a ΔΨm collapse induced by uncoupler treatment causes cells to accumulate in the G1 phase by triggering a G1-to-S phase transition delay similar to the one observed in ρ⁰ cells. These findings implicate ΔΨm in modulating cell cycle progression in ρ⁺ cells.

The cell cycle defect in ρ⁰ cells is exacerbated by a further decrease in membrane potential

In contrast to uncoupler-treated cells where the ΔΨm is largely dissipated, ρ⁰ cells have the ability to maintain a membrane potential that, although lower than in ρ⁺ cells, is still sufficient to support various ΔΨm-dependent processes such as the import of nuclear-encoded mitochondrial proteins (9, 19). Next, we examined whether the decreased ΔΨm could underlie the cell cycle phenotype of ρ⁰ cells and act as the initial signal to trigger the G1-to-S transition delay. We measured the ΔΨm of WT ρ⁺ and ρ⁰ cells using TMRE, and corrected the signal for potential changes in strain mitochondrial mass as detected by nonyl acridine orange (NAO), a green-fluorescent dye that localizes to the mitochondria in a ΔΨm-independent manner (20). Microscopy analysis confirmed the uptake of NAO, and thus its suitability as a readout of mitochondrial mass, into both ρ⁺ and ρ⁰ cells (Fig S3A). Uncoupler-treated samples analysed in parallel provided a measure of the background fluorescence. As expected, the ΔΨm of WT ρ⁰ cells was significantly lower than that of WT ρ⁺ cells, but clearly higher than the baseline value of uncoupler-treated cells (Figs 3A and S3B and C).

In additional support of a role for ΔΨm in regulating cell cycle progression, the accumulation of ρ⁰ cells in the G1 phase was further exacerbated when ΔΨm was fully dissipated by BAM15 or CCCP treatment (Figs 3B and S2C, D, and F). The effect of BAM15 was more sustained in ρ⁰ than in ρ⁺ cells and was evident even 2 h after the addition of the compound (compare Figs 2B and 3B). Expectedly, time-course experiments after release from α-factor– or nocodazole-induced synchrony revealed a more severe G1-to-S phase progression delay in ρ⁰ cells in the presence of BAM15 (G1 exit starting 40 min and 80 min after release from α-factor synchrony into the medium lacking and containing BAM15, respectively) (Figs 3C and D and S2G–J). These observations establish a quantitative correlation between ΔΨm and cell cycle progression, where the extent of the G1-to-S delay is governed by the severity of ΔΨm loss.

Increasing ΔΨm rescues the cell cycle phenotype of ρ⁰ cells

The metabolism of ρ⁰ cells differs significantly from that of ρ⁺ cells given that respiratory-deficient ρ⁰ cells must compensate for the loss of a subset of the citric acid cycle reactions, the products of which are central to many anabolic pathways (21). If the G1-to-S phase progression delay is indeed driven by decreased ΔΨm and not by other functional or metabolic differences between ρ⁺ and ρ⁰ cells, we hypothesized that it should be rescued by increasing ΔΨm in ρ⁰ cells. In the absence of ETC activity, mtDNA-deficient cells maintain ΔΨm by an alternative mechanism that involves the hydrolysis of glycolytically produced ATP by the “reverse” action of the F₁ subunit of the mitochondrial ATP synthase and the electrogenic exchange of ATP⁴⁻ for ADP³⁻ over the inner mitochondrial membrane (7, 9). The ATP1-111 mutation in the Atp1 subunit of the mitochondrial F₁-ATPase results in a hyperactive enzyme that generates a higher ΔΨm in ρ⁰ cells than the one maintained in WT ρ⁰ cells, and improves their growth (5, 19, 22). Accordingly, the ΔΨm of ATP1-111 ρ⁰ cells was comparable to that of WT ρ⁺ cells, confirming the ATP1-111 ρ⁰ strain as a suitable model to test our hypothesis in Figs 4A and S3B and C.

We first analysed the cell cycle profile of unsynchronized WT and ATP1-111 ρ⁺ and ρ⁰ cells. As expected based on the relatively high ΔΨm in ATP1-111 ρ⁺ cells, the strain showed a cell cycle profile that was indistinguishable from WT ρ⁺ cells (Fig 4B and C). Moreover, analysis of ⍺-factor–synchronized cells revealed that the ATP1-111 mutation was largely able to rescue the delay in G1-to-S transition observed in ρ⁰ cells: although the percentage of G1 cells in cultures of WT and cox4Δ ρ⁰ cells was still above 90% 30 min after release from synchrony, only 67% of ATP1-111 ρ⁰ cells remained in the G1 phase (Figs 4D and S3D). Finally, the growth of ATP1-111 ρ⁰, as judged by colony size or cell density measurements in liquid media, was partially restored relative to WT ρ⁰ cells (Figs 4E and S3E), as previously reported by others (5, 22).

The earlier experiments carried out on WT cells revealed a more sustained effect of BAM15 on the cell cycle profile of ρ⁰ than ρ⁺ cells (Figs 2B and 3B). We attributed this to the slower recovery of ΔΨm in ρ⁰ cells that lack a functional ETC. To explore this aspect further, we monitored cell cycle progression in ρ⁺ cells lacking Cox4, a nuclear-encoded subunit of complex IV. Unperturbed cox4Δ ρ⁺ cells were able to maintain a ΔΨm that did not significantly differ from that of WT ρ⁺ cells (Figs 4A and S3B and C). Accordingly, deletion of the COX4 gene did not have a pronounced effect on the cell cycle profile of unsynchronized ρ⁺ cells (Fig 4B and C). However, cox4Δ ρ⁺ cells exhibited a somewhat delayed transition from G1 to S in synchrony experiments (Figs 4D and S3D) and a stronger, more sustained cell cycle response to BAM15 than WT or ATP1-111 ρ⁺ cells, indicating that a functional ETC aids in timely recovery from uncoupling (Figs 4F and S3F). In line with this notion, all three ρ⁰ strains, including the ATP1-111 ρ⁰ with the restored ΔΨm, failed to recover from the BAM15-induced accumulation of G1 cells within the 120-min follow-up period, although the percentage of G1 cells was lower in ATP1-111 ρ⁰ than in WT or cox4Δ ρ⁰ cells (Figs 4G and S3F).

Based on the findings in Fig 4, we conclude that the delayed G1-to-S phase progression in ρ⁰ cells can be attributed to a decreased ΔΨm. Moreover, the cell cycle defect and the consequent petite phenotype and slow growth of ρ⁰ cells can be rescued by increasing ΔΨm through the ATP1-111 mutation, implicating membrane potential—rather than other metabolic adaptations in ρ⁰ cells—as the main determinant of the cell cycle delay. Accordingly, recovery from transient uncoupler-induced cell cycle delay is expedited by a functional ETC.

Altered ROS is not the signal for the G1-to-S phase progression delay in ρ⁰ cells

Mitochondria are a major source of ROS, and mitochondrial dysfunction is often associated with increased levels of intracellular ROS (3). Furthermore, ROS are documented to influence cell proliferation (1). Therefore, we wanted to determine whether ROS levels play a role in the ρ⁰ G1-to-S phase progression delay. Rather than increasing the percentage of cells in G1, however, H₂O₂ treatment primarily caused the accumulation of WT ρ⁺ cells in the early S phase, as previously reported by others (Figs 5A and B and S4A) (23). In contrast, decreasing ROS by treatment of WT ρ⁺ cells with the antioxidants N-acetylcysteine (NAC) or reduced glutathione (GSH) caused a transient accumulation of cells in the G1 phase along with a decrease in the percentage of cells in the S phase (Fig 5C, left panel; Fig S4B). A similar but more sustained effect of treatment with either NAC or GSH was observed in WT ρ⁰ cells (Fig 5C, right panel; Fig S4B). On their own, these data would suggest that neutralizing ROS by antioxidant treatment inhibits G1-to-S transition. However, measurement of ΔΨm in antioxidant-treated ρ⁺ cells revealed considerable effects of the antioxidants on ΔΨm, with a decrease comparable to that after uncoupler treatment (Figs 5D and S4C and D). Similar effects of NAC on ΔΨm have been reported in mammalian cells (24, 25). The impact of antioxidant treatment on S. cerevisiae cell cycle progression observed in Fig 5C may therefore be indirect and attributable to decreased ΔΨm rather than lower ROS levels per se.

To further explore the relationship between ROS, ΔΨm, and cell cycle progression, we gauged the levels of oxidative stress in our strains of interest by measuring lipid peroxidation, a reliable indicator of cellular oxidative damage (26, 27). This analysis revealed comparable levels of lipid peroxidation in all six strains, while accurately detecting an increase in lipid peroxidation in control samples treated with H₂O₂ (Figs 5E and S4E). Also, levels of mitochondrial superoxide as indicated by MitoSOX Red fluorescence were comparable across our strains of interest (Fig 5F; see Fig S4F for controls). Based on these results, we conclude that the cell cycle phenotype observed in ρ⁰ cells does not correlate with the cellular levels of oxidative stress and that the rescued cell cycle phenotype in ATP1-111 ρ⁰ cells cannot be attributed to normalized ROS levels. Taken together, the findings of this study confirm that diminished ΔΨm, rather than ROS or defective mitochondrial ATP synthesis, underlies the G1-to-S transition delay observed in ρ⁰ cells and that loss of ΔΨm inhibits cell cycle progression even in ρ⁺ cells.

ΔΨm modulates G1/S transition through cell size

The G1/S transition is controlled by a late-G1 checkpoint called Start, passage through which requires attainment of a critical cell size and commits cells to cell division (28, 29). In brief, progression through G1 is driven by the G1 cyclin Cln3 in complex with the cyclin-dependent kinase Cdk1 that phosphorylates and partly inactivates the transcriptional inhibitor Whi5, leading to the induction of a transcriptional program encompassing over 200 genes that promotes G1/S transition (30). This transcriptional wave is controlled by two transcription factors, SBF and MBF. SBF regulates the expression of genes involved in cell morphogenesis and the timing of cell cycle commitment such as those encoding the G1 cyclins Cln1 and Cln2. MBF targets, including the S phase cyclins Clb5 and Clb6, drive DNA replication and repair (31). Once expressed, the association of Cln1/2 with Cdk1 promotes features such as bud emergence and pheromone resistance, and Clb5/6-Cdk1 complexes induce DNA replication. A prominent model for how Start transition is coupled to attainment of a critical cell size involves the growth-dependent dilution of the Whi5 repressor (32, 33), although other mechanisms have also been proposed (e.g., Refs. (34, 35)). Moreover, cell size, G1/S transcription, and S phase entry are influenced by additional factors and pathways including the cAMP/protein kinase A, TOR, and Snf1/AMPK pathways, the full extent and interaction of which are currently not understood (36). In fact, passage through Start is regulated by the integration of multiple internal and external signals including, but not limited to, cell size, nutrient availability, and stress (37).

To gain insight into the mechanisms connecting mitochondrial metabolism and ΔΨm to Start passage, we first explored the cell size of WT versus ATP1-111 ρ⁺ and ρ⁰ cells. Given that Start passage is coupled to cell size, we asked whether the rescued G1/S transition in ATP1-111 ρ⁰ cells is merely a consequence of larger cell size and thus a shorter “pre-Start” G1 period compared with WT ρ⁰ cells, or whether it is explained by changes to some other aspect regulating Start passage, as is observed upon switching to non-fermentable carbon sources when the critical cell size required for G1/S transition decreases (38). Microscopy-based measurement of the area of G1 phase (=unbudded) cells revealed that although WT and cox4Δ ρ⁰ cells were smaller than their ρ⁺ counterparts, the size of ATP1-111 ρ⁰ cells was comparable to that of ATP1-111 ρ⁺ and WT ρ⁺ cells (Figs 6B and S5A). Thus, the rescued G1/S transition in ATP1-111 ρ⁰ cells with a partially recovered ΔΨm may be explained by their larger size and the subsequently shorter pre-Start G1 period.

The larger size of ATP1-111 ρ⁰ cells is expected to lead to a faster activation of G1 transcription because of earlier dilution of the transcriptional inhibitor Whi5 (Fig 6A). We analysed the timing of the G1/S transcriptional wave in WT versus ATP1-111 ρ⁺ and ρ⁰ cells released from ⍺-factor synchrony, an experimental approach previously used by others (39, 40). As expected (41), no fluctuation of the levels of the non-cycling G1 cyclin CLN3 was observed after release, and transcript levels of CLN3 or WHI5 did not significantly differ across the analysed strains (Figs 6C and S5B and C). In contrast, the induction of the late-G1 cyclins CLN1 and CLN2 was delayed in WT ρ⁰ cells as compared to WT ρ⁺ cells. A similar but somewhat weaker tendency was observed for the S phase cyclins CLB5 and CLB6 (Fig S5C). Unexpectedly, and in discordance with the rescued G1/S transition phenotype, the transcriptional pattern of these cyclins was not restored in ATP1-111 ρ⁰ cells (Fig 6C; compare ATP1-111 ρ⁰ with WT ρ⁰). The rescued cell cycle phenotype of ATP1-111 ρ⁰ cells does therefore not appear to be explained by an earlier accumulation of CLN1/2 or CLB5/6 gene products.

Taken together, the findings in Fig 6 indicate that ATP1-111 ρ⁰ cells are larger in size than WT ρ⁰ cells, likely explaining their rescued cell cycle phenotype. However, the rescued timing of their Start passage appears to involve improvements downstream of G1 and S cyclin transcription. Signalling through numerous pathways, including the PKA, TOR, or Snf1/AMPK pathways, is known to impinge on Start timing, and their involvement, and that of any Cdk inhibitors acting on the Cln1/2-Cdk1 complex, should be addressed in future work to gain further insight into cell cycle regulation in cells with mitochondrial dysfunction.

Yeast strains and growth conditions

Unless otherwise indicated, all S. cerevisiae strains used in this study are congenic to W4069-4c, which is in a W303 background (51), and are listed in Table 1. To ensure fitness before each experiment, all strains were streaked out from glycerol stocks onto a YPDA solid medium (1% yeast extract, 2% peptone, 50 mg/l adenine, 2% glucose, and 2% agar). Unless otherwise indicated, yeast cells were grown in YPDA at 30°C with shaking at 180 rpm.

Ρ⁰ strains lacking mtDNA were generated by growth in the YPDA medium supplemented with 25 μM ethidium bromide for 4–15 d depending on the strain (12). Loss of mtDNA was confirmed by lack of growth on the YPDA medium (1% yeast extract, 2% peptone, 50 mg/l adenine, and 3% glycerol) in combination with the absence of amplification of a region of the mitochondrially encoded COX1 gene by quantitative real-time PCR in samples where a region of the nuclear-encoded ACT1 gene did amplify. Real-time PCRs contained 0.2 μM forward and reverse primers targeting COX1 or ACT1 (Table 2) and 10 μl 2x SyGreen Mix (PCR Biosystems), and were run on a LightCycler 96 instrument (Roche) using the following program: 95°C for 180 s, 45 cycles of (95°C for 10 s, 56°C for 10 s, and 72°C for 1 s with signal acquisition), and melting curve (95°C for 5 s, 65°C for 60 s, and heating to 97°C at 1 s with continuous signal acquisition).

The strain bearing the ATP1-111 mutation was generated by introduction of the ATP1 variant carrying a T>G substitution at position 331 from a pRS316 integration vector as previously described (19, 22). Briefly, the ATP1-111 insert in pRS306 was excised with BamHI and XbaI and re-cloned into the pRS316 integration vector. Yeast cells were transformed using a standard lithium acetate procedure (53), and the presence of the mutation in cells that grew in a medium supplemented with 5-FOA was confirmed by sequencing.

Deletion strains were constructed by replacing the entire ORF with the kanMX4 or natMX cassette (54). Gene deletion was confirmed by growth in geneticin or nourseothricin, respectively, and by PCR using both a primer pair that flanks the ORF and a pair internal to it.

Colony size and growth curves

Yeast strains grown in a YPDA liquid medium for at least 24 h were counted using Neubauer Improved Hemocytometer and adjusted to 100 cells per 100 μl dH₂O. 100 μl of the cell suspension was spread on a single-partitioned YPDA plate and incubated at 30°C for 48 h before imaging on a ChemiDoc imaging system (Bio-Rad). The area of all colonies, except for ones with another colony attached to them, was analysed using Fiji (55), and individual values were normalized to the average of all WT ρ⁺ cell colonies.

For analysis of growth in liquid media, exponentially growing cultures were re-inoculated into fresh YPDA liquid media with a starting OD₆₀₀ = 0.1. OD₆₀₀ readings were recorded every hour by spectrophotometry.

Cell cycle analysis by flow cytometry

Cells grown overnight in YPDA were diluted to a fresh medium to a starting OD₆₀₀ of 0.1. Cells were harvested at OD₆₀₀ 0.35–0.5, fixed with cold 70% ethanol, and incubated at 4°C overnight. The fixed cells were washed with distilled water once, resuspended in 50 mM Tris–HCl, pH 7.5, with 15 mM NaCl, and treated with 2 mg/ml RNase A at 37°C overnight. The next day, samples were treated with 1.8 mg/ml proteinase K at 50°C for 1 h, spun down, and resuspended in 50 mM Tris, pH 7.5. An aliquot was added to 50 mM Tris, pH 7.5, containing a 1:10,000 dilution of SYBR Green nucleic acid stain (Invitrogen). The cell suspension was sonicated for 10 s with an amplitude of 20% before analysis (Q500; QSonica). DNA content was detected in the FL1 channel using the Cytomics FC500 (Beckman Coulter). Data analysis and quantification of cells in G1, S, and G2 phases was performed using FCS Express 7 Flow (De Novo Software). The y-axis scale of the cell cycle histograms was adjusted to fit the highest cell count.

For time-course analysis of unsynchronized cells, 30 ml culture (with a starting OD₆₀₀ of 0.1) was grown to an OD₆₀₀ of 0.4–0.6 and split into two portions: the specified drug added to one portion, whereas the other served as a control. Cell aliquots were harvested and immediately fixed every 15 min for 2 h. Time 0 indicates the addition of the drug.

For analysis of synchronized cells, cultures were synchronized in G1 by the addition of 5 μg/ml α-factor pheromone every hour for 2 h. For G2 synchronization, cells were treated with 10–25 μg/ml nocodazole for 2 h. Cells were washed twice with one volume of ice-cold distilled water and released from synchrony by resuspension in the fresh medium. Cells were harvested and immediately fixed every 10 min for 90–120 min after release.

Mitochondrial membrane potential

Cells grown overnight in YPDA were diluted in the fresh medium to a starting OD₆₀₀ of 0.1 and grown for 24 h; 1 ml of the culture was harvested and washed twice with one volume of PBS buffer (0.14 M NaCl, 0.0027 M KCl, and 0.010 M phosphate buffer, pH 7.4). Cells were reconstituted in 5 ml PBS to an OD₆₀₀ of 0.05. A 1 ml aliquot was treated with 2 μM tetramethylrhodamine methyl ester perchlorate (TMRE) (Molecular Probes) and incubated at 37°C for 30 min. As a control, another aliquot of the cell suspension was treated with the same concentration of TMRE for 20 min followed by the addition of 20 μM carbonyl cyanide 3-chlorophenyl hydrazone (CCCP) for 10 min to collapse the membrane potential. When different strains were compared, mitochondrial mass of each strain was estimated by adding 250 nM NAO (Invitrogen) to a third aliquot of the cell suspension and incubated at 37°C for 30 min. Both mitochondrial membrane potential and mitochondrial mass were determined by flow cytometry on a Beckman Coulter Cytomics FC500. ΔΨm was calculated by subtracting the TMRE fluorescence of the untreated and uncoupled sample, and, when comparing across strains, normalized for mitochondrial mass (19).

Lipid peroxidation analysis and detection of mitochondrial ROS

Lipid peroxidation was determined by measurement of the amount of thiobarbituric acid reactive substance using QuantiChrom Thiobarbituric Acid Reactive Substance Assay Kit (BioAssay Systems) following the manufacturer’s instructions. The concentration of malondialdehyde (MDA) was measured on a Synergy H4 hybrid microplate reader (BioTek) with a fluorescence intensity of (λ_em/ex = 530 nm/550 nm). Lipid peroxidation was expressed as moles of MDA equivalent equilibrated to the total amount of protein (nM of MDA equivalent/ng protein). Protein concentration was measured using the Quick Start Bradford protein assay (Bio-Rad).

Detection of mtROS was as previously described with some modifications (56). Briefly, overnight cell cultures were diluted to an OD₆₀₀ of 0.1 in a fresh YPDA medium and grown until an OD₆₀₀ of 0.5–0.6. Cells were harvested and washed twice with PBS. Cells were resuspended in PBS to an OD₆₀₀ of 0.5 and incubated with 5 μM MitoSOX Red (Invitrogen) at 30°C for 15 min before analysis by flow cytometry.

Gene expression analysis (RT–qPCR)

Cells released from α-factor–induced G1 synchrony were harvested every 10 min and frozen in liquid nitrogen. Cells were lysed using 0.5 mm Glass Beads (BioSpec) in a FastPrep-24 bead beater (MP Biomedical), and RNA was extracted using RNeasy Mini Kit (QIAGEN) according to the manufacturer’s instructions. 1 μg of the harvested RNA was treated with 1U DNase I (Thermo Fisher Scientific), and cDNA was synthesized using SuperScript III Reverse Transcriptase (Invitrogen). Gene expression was analysed in technical duplicates by quantitative real-time PCR using 12.5 ng of cDNA in a 20 μl reaction containing 0.4 μM forward and reverse primers (Table 2) and 10 μl 2x SyGreen Mix (#PB20.14-05; qPCRBIO) in a LightCycler instrument (Roche). The 180 s of preincubation at 95°C was followed by 40 cycles of amplification consisting of 95°C for 10 s, 56°C for 10 s, and 72°C for 1 s with single signal acquisition, and melting curve of 95°C for 5 s, 65°C for 60 s, and 97°C at 0.2°C/sec with continuous signal acquisition. Cq values determined by LightCycler 96 software (Roche) were used to calculate the expression of the cyclin genes relative to actin using the Pfaffl method (57). Statistical comparisons between selected timepoints of individual strains were performed by a standard two-tailed t test in GraphPad Prism (GraphPad Software, Inc.).

Cell imaging and size analysis

For mitochondrial staining, yeast cells grown in YPDA until an early stationary phase were harvested and washed twice with PBS. Cells were then resuspended in 1 ml PBS at a total cell density of OD₆₀₀ = 1.0. The cell suspension was co-stained with 250 nM NAO (manufacturer) and 250 nM MitoTracker Deep Red (Thermo Fisher Scientific) and incubated at 37°C for 10 min. 10 μl of the stained cells was mixed with a 10 μl UltraPure LMP Agarose (37°C, 1% [wt/vol] in PBS) (Invitrogen) and applied to a standard microscopy glass slide and coverslip (VWR). For yeast cell wall staining, early stationary phase cells were harvested and washed twice with PBS. Cells were then resuspended in 1 ml PBS with a total cell concentration of OD₆₀₀ = 1.0, stained with 20 μg/ml trypan blue (Invitrogen), and applied to a standard microscopy glass slide and coverslip. A Leica Widefield Thunder microscope equipped with an HC PL FLUOTAR 100x/1.32 OIL PH3 objective, with filters and laser settings for DAPI, FITC, TRITC, and Alexa Fluor 674, was used for all microscopic studies. Images were analysed using Fiji software (55). The area of unbudded (i.e., G1) cells was determined by drawing a region of interest around the stained cell wall and measuring the mean grey value using Fiji.

Author Contributions

• CM Gorospe: data curation, formal analysis, validation, investigation, visualization, methodology, project administration, and writing—original draft, review, and editing.

• G Carvalho: formal analysis, investigation, methodology, and writing—original draft, review, and editing.

• A Herrera Curbelo: formal analysis, investigation, and writing—review and editing.

• L Marchhart: formal analysis, investigation, and writing—review and editing.

• IC Mendes: investigation and writing—review and editing.

• K Niedźwiecka: investigation and writing—review and editing.

• PH Wanrooij: conceptualization, data curation, formal analysis, supervision, funding acquisition, validation, investigation, methodology, project administration, and writing—original draft, review, and editing.

Conflict of Interest Statement

The authors declare that they have no conflict of interest.