The spindle checkpoint is less persistent in meiosis compared to mitosis

To investigate differences in spindle checkpoint strength between mitosis and meiosis, we used live-cell fluorescence microscopy to monitor the duration of the spindle checkpoint delay. The cells expressed the spindle pole body (SPB) component Spc42 tagged with mCherry and a biosensor for the enzyme separase, which cleaves cohesin at the metaphase-to-anaphase transition [47]. The biosensor consists of a chromosomal locus tagged with a fluorescent focus that disappears upon separase activation; a LacO array was integrated into a chromosome and the cells expressed a GFP-tagged Lac repressor engineered with an Scc1 cleavage site (GFP-Scc1-LacI) for mitosis or a Rec8 cleavage site for meiosis (GFP-Rec8-LacI). With separase activation, the cohesin subunit Scc1 or Rec8 will be cleaved and a GFP focus will no longer be visible, although the nucleus will still have a GFP haze (Fig 1A and 1B). The advantage of using the separase biosensor is that we can then monitor anaphase onset in conditions or mutant backgrounds that may have depolymerized microtubules or spindle elongation prior to cohesin cleavage (Fig 1C and 1D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. The spindle checkpoint is less persistent in meiosis compared to mitosis.

(A-B) Representative time lapse images of a cell harboring Spc42-mCherry and the separase biosensor in mitosis (A) and meiosis I (B). Time from SPB separation to separase activation. Time 0 indicates SPB separation. Scale bar = 5 μm. (C-D) Cartoons depicting activation of the spindle checkpoint through loss of kinetochore (KT) tension and loss of kinetochore-microtubule (KT-MT) attachments in mitosis (C) and meiosis I (D) Wildtype = WT. E) Graph depicting mean time from start of movie to anaphase onset in mitosis following treatment with DMSO or nocodazole. Asterisk indicates statistically significant difference between DMSO-treated and nocodazole-treated cells (p <0.0001, Mann-Whitney test) and error bars show standard deviation (SD) n≥ 50 cells per genotype. F) Graph depicting mean time from start of movie to anaphase onset in meiosis following treatment with DMSO or nocodazole. (G-H) Graph showing the mean time from SPB separation to anaphase onset in mitosis (G) and meiosis I (H). Asterisks indicate statistically significant difference between wildtype and mutants (p < 0.0001, Mann-Whitney test). Error bars show SD. n≥100 cells per genotype.

https://doi.org/10.1371/journal.pgen.1010707.g001

We first measured the duration of a spindle checkpoint delay in a condition in which kinetochore-microtubule attachments are disrupted due to treatment with the microtubule-depolymerizing drug nocodazole (Fig 1C and 1D). In mitosis, we released haploid cells from an alpha-factor-induced G1 arrest and treated cells with 15 μg/mL of nocodazole after we observed separation of SPBs in most of the cells. We then started the movie and measured the duration from the start of the movie until loss of the GFP focus, which represents the time of separase cleavage of cohesin. We observed an average duration of 348 ± 93 minutes when cells were treated with nocodazole, which is on average a delay of 323 minutes compared to cells treated with the solvent control (DMSO) (average ± SD; Fig 1E). As previously shown for mitotic cells, we observed cell-to-cell variability in the duration of the delay [28]. This delay is dependent on the spindle checkpoint because mad3Δ cells treated with nocodazole progressed into anaphase with a similar duration as the DMSO control cells.

To assess the duration of a checkpoint delay in meiosis I, we treated cells with 30 μg/mL of nocodazole after release from a prophase I arrest using the GAL-NDT80 arrest and release system [48,49]. We treated cells with nocodazole 80-minutes after release from prophase I, which is when we observed SPB separation. We specifically waited to add the nocodazole until after prophase I because microtubule perturbation earlier causes a G2 arrest [50]. In the presence of nocodazole, cells had a metaphase I duration of 103 ± 56 minutes, which is 69 minutes longer than the control cells treated with DMSO alone (average ± SD; Fig 1F). The mad3Δ cells treated with nocodazole had a similar time of anaphase onset as the DMSO control, suggesting that the delay was spindle checkpoint dependent. We conclude that the spindle checkpoint is more persistent in mitosis than in meiosis when most kinetochore-microtubule attachments are disrupted.

Chromosome segregation errors that cause single chromosome aneuploidies often arise from the misattachment of kinetochores and microtubules, such that the two sister chromatids in mitosis or the two homologous chromosomes in meiosis I are attached to the same pole. The monopolar attachment creates a lack of intra-kinetochore stretch, or tension [51]. The kinetochore-microtubule attachments that are not under tension are preferentially released due to the activation of the Ipl1^{Aurora B} kinase, ultimately leading to spindle checkpoint activation [52–55]. We therefore wanted to compare the persistence of the spindle checkpoint delay between mitosis and meiosis I upon disruption of kinetochore tension. We engineered strains to create a condition in which all kinetochores can attach to spindle microtubules but cannot generate tension (Fig 1C and 1D).

To prevent kinetochore tension in mitosis, we depleted the essential protein Cdc6, which is required for initiation of DNA replication (Piatti, 1995). Because only one microtubule binds per kinetochore in budding yeast, without DNA replication, the single sister chromatid will be unable to biorient, creating a lack of kinetochore tension. To deplete Cdc6, CDC6 was placed under control of the GAL1 galactose-inducible promoter and fused to a Ubi degron (Cdc6 depletion or Cdc6-dp) [56]. When cells are grown in the presence of glucose, CDC6 expression is repressed, and any remaining Cdc6 is degraded. We assessed the duration of the spindle checkpoint delay in diploid Cdc6-dp cells by measuring the time from SPB separation to separase activation, as assessed by the loss of the GFP focus. Cdc6-dp cells were in metaphase for a duration of 352 ± 148 minutes, which is an average delay of 295 minutes when compared to wildtype cells (average ± SD; Fig 1G). There was also substantial cell-to-cell variability in the duration of the checkpoint delay in Cdc6-dp cells, similar to nocodazole-treated cells. The Cdc6-dp haploid cells had a similar delay as the diploid cells, with metaphase duration of 339 ± 157 minutes, suggesting that ploidy did not cause major differences in the delay time (S1A Fig). The observed delay in Cdc6-dp cells was due to spindle checkpoint activation because Cdc6-dp mad3Δ cells progressed through metaphase with similar timings as the wildtype control cells.

We note that the timings from the lack of tension experiments cannot be directly compared to the timings from the nocodazole experiments. In the nocodazole experiments, we did not measure the full duration of metaphase because time-lapse imaging was started after the addition of nocodazole, to prevent the drug from interfering with SPB separation (see Materials and Methods). In contrast, we measured the time from SPB separation to anaphase onset in the Cdc6-dp experiments. However, the results show that both nocodazole and Cdc6-dp cause an extended spindle checkpoint delay that cells can eventually escape.

To prevent intra-kinetochore tension during meiosis I, we disrupted proteins needed for crossover formation. Normally, crossovers, held in place by arm cohesin, link homologs together and resist the spindle forces when homologous chromosomes biorient, which creates intra-kinetochore tension [33]. We prevented crossover formation by using a catalytically inactive allele of SPO11, spo11-Y135F, which is a mutant form of the enzyme required to make double-stranded breaks for crossover formation [57]. Without crossovers, homologs are not physically linked together, causing a lack of tension across the kinetochores. In these cells, chromosomes move to the spindle pole prematurely in an anaphase-like prometaphase I because they cannot resist spindle forces [58]. The spo11-Y135F cells exhibited a metaphase I duration of 194 ± 81 minutes, which is an average delay of 140 minutes compared to the wildtype control cells, although there is substantial cell-to-cell variability (average ± SD; Fig 1H). The spo11-Y135F mad3Δ cells showed similar timings to the wildtype control cells, suggesting that the delay was spindle checkpoint dependent.

Like the mitotic studies, we cannot directly compare the nocodazole-treated cells with the spo11-Y135F cells. In the nocodazole experiment, we started the time-lapse imaging after addition of the nocodazole to cells in early metaphase I, just after SPB separation. In contrast, the spo11-Y135F cells were imaged throughout meiosis, providing the full duration of the metaphase I delay.

To determine if the duration of the spindle checkpoint delay is similar in other mutants with reduced crossovers, we deleted MEK1, which biases repair of double-strand breaks (DSBs) to the homolog, allowing for crossover formation. In cells harboring MEK1 mutants, crossover formation is reduced by ~85%, so some homologs will retain crossovers [59–61]. The mek1Δ cells displayed an average metaphase I delay of approximately 130 minutes with cell-to-cell variability, which was similar to spo11-Y135F cells (Fig 1H). Overall, our results demonstrate that the spindle checkpoint is not as persistent in meiosis as it is in mitosis, signaled through either addition of nocodazole or through genetic backgrounds that cause a lack of kinetochore tension.

We wondered whether the shorter duration of the spindle checkpoint in meiosis was due to the size differences between budding yeast meiotic and mitotic cells. Mouse oocytes have a weaker spindle checkpoint than mitotic cells, which is thought to be due to their larger size [62–69]. Similarly, checkpoint strength is thought to be at least partially dependent on cell size in the oocyte and developing C. elegans embryo [70–72]. To determine if a size difference could be causing a weaker spindle checkpoint in meiotic budding yeast cells, we measured the volume of budding yeast cells at the time point just before anaphase onset and found that mek1Δ cells in meiosis had a volume of 254 ± 46 fL, while Cdc6-dp cells had an average volume of 343 ± 89 fL (S1B Fig). From this we conclude that cells arrested in mitosis have a larger average volume compared to cells arrested in meiosis, so it is not likely that the inherent weakness of the meiotic checkpoint is due to a larger cell volume.

The meiotic spindle checkpoint prevents some chromosome mis-segregation when crossover formation is drastically reduced

Previous work revealed that the spindle checkpoint proteins Mad1, Mad2, and Mad3 have different functions in meiosis. Mad3 provides a checkpoint delay, but this delay is not needed for faithful chromosome segregation in normal growth conditions [58,73,74]. The checkpoint components Mad1 and Mad2 are also needed for a delay but have additional roles in cell cycle regulation and chromosome biorientation [73–76]. Interestingly, the spindle checkpoint delay strongly reduces the mis-segregation of a pair of non-exchange homeologous chromosomes [74,77]. We asked whether the spindle checkpoint could decrease chromosome mis-segregation in mek1Δ cells, which have a much higher fraction of non-exchange chromosomes than the strains previously tested. To address this question, we compared chromosome segregation fidelity between mek1Δ cells and mek1Δ mad3Δ cells, which lacked spindle checkpoint activity. We monitored the segregation of a single pair of homologs with LacO repeats near the centromere of chromosome IV in cells that expressed LacI-GFP under the control of a copper-inducible CUP1 promoter [78]. Upon the addition of copper sulfate, LacI-GFP was expressed and bound the LacO repeats, which allowed visualization of each homolog as a single focus (Fig 2A). In wildtype and mad3Δ cells, the percent of chromosome IV mis-segregation was less than 2%. In contrast, 24% of mek1Δ and 38% of mek1Δ mad3Δ cells mis-segregated chromosome IV in anaphase I (Fig 2B). Thus, the meiotic spindle checkpoint prevented some chromosome mis-segregation in anaphase I, despite not having as long of a delay as the mitotic checkpoint. We note that this is experiment is only monitoring one of the 16 pairs of homologous chromosomes, suggesting that each mek1Δ cell likely contains many chromosomes that are not bioriented when they escape the spindle checkpoint arrest.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Three nonexchange chromosomes signal a similar average duration of the spindle checkpoint delay as 32 non-exchange chromosomes.

A) Representative images of a cell in which both homologs of chromosome IV are tagged with a GFP focus at anaphase I. Scale bar = 5 μm. B) Graph of percent mis-segregation of chromosome IV during meiosis I. Asterisk indicates a statistically significant difference from wildtype (Fisher’s Exact test; p = 0.0327). n≥ 100 cells per genotype. C) Graph of mean time from SPB separation to separase biosensor activation. Numbers above bars indicate the number of non-exchange chromosomes in each genotype. Asterisk represents statistically significant difference between wildtype and the indicated genotype (p <0.0001, Mann-Whitney test), and error bars show SD. n≥100 cells per genotype. KT = kinetochore. SLC = short linear chromosome.

https://doi.org/10.1371/journal.pgen.1010707.g002

Three non-crossover chromosomes are needed to signal the same duration of a meiotic spindle checkpoint delay as 32 non-crossover chromosomes

Previous work in animal cells and fission yeast mitosis have shown that the spindle checkpoint strength depends on the number of unattached kinetochores and the amount of spindle checkpoint components recruited to kinetochores [31,32,79]. Given the similarity in checkpoint duration between spo11-Y135F and mek1Δ cells in meiosis I, we questioned whether a threshold number of non-exchange chromosomes was needed for the full duration of the spindle checkpoint delay in meiosis or if a single chromosome without a partner could cause a full delay. To this end, we engineered budding yeast strains with increasing numbers of non-exchange chromosomes and measured the time from SPB separation to separase activation. First, we introduced one short linear chromosome (SLC) into an otherwise wildtype yeast strain (WT + SLC). This SLC has telomere sequences at each end and is approximately 27kb, containing a marker as well as 256 repeats of the LacO array. SLCs are thought not to form crossovers due to their short size and inability to hold the arm cohesins in place to stabilize the crossover [80–82]. But, the SLC has a centromere on which a kinetochore can build. The WT + SLC strain displayed a metaphase I duration of 75 ± 20 minutes, which is 24 minutes longer than wildtype cells (average ± SD; Fig 2C).

We next constructed a strain in which we introduced a single pair of homeologous chromosomes, which are chromosomes from two different yeast species that cannot form crossovers due to their sequence divergence [74,83,84]. The spindle checkpoint helps homeologous chromosomes segregate accurately, but the mechanism that ensures their disjunction is unknown [74,77]. The strain harboring the homeologs displayed a checkpoint delay of 62 minutes compared to wildtype cells (Fig 2C). Next, we created a strain with three non-exchange chromosomes, by adding a SLC to the strain harboring the homeologs (homeologs +1SLC). These cells underwent metaphase I after 176 ± 41 minutes, which is a checkpoint delay of 125 minutes, similar to that of spo11-Y135F cells (Fig 2C). We conclude that there is a graded checkpoint response, and 3 non-exchange chromosomes are sufficient to signal a similar average duration of the spindle checkpoint delay as cells with 32 non-exchange chromosomes.

Mps1 and Ipl1 maintain the spindle checkpoint delay

In meiosis, Ipl1^{Aurora B} releases improper attachments and Mps1 helps form force-generating attachments [85]. Previous work has shown that inhibition of Mps1 and Ipl1^{Aurora B} shortened metaphase I in spo11Δ cells, suggesting that both kinases are needed for spindle checkpoint activity [86]. We asked if the kinases are also needed to maintain the meiotic spindle checkpoint delay after it has been initiated in mek1Δ cells that have a reduced number of crossovers. We used the anchor away technique to deplete Ipl1^{Aurora B} or Mps1 from the nucleus [87] (S2A Fig). We tagged endogenous IPL1 or MPS1 at its C-terminus with FRB in a strain with the ribosome protein RPL13A tagged with FKBP12. Because Rpl13A shuttles from the nucleus into the cytosol during ribosome biogenesis, the addition of rapamycin, which allows the stable interaction between FRB and FKBP12, depleted Ipl1-FRB or Mps1-FRB from the nucleus. We added rapamycin at 8 hours after cells were switched to sporulation media, a timepoint when many of the cells were in metaphase I, and then began live-cell imaging. We measured the time of anaphase I onset only in the cells that were in metaphase I at the start of imaging. We found that anchoring away of either Ipl1 or Mps1 in mek1Δ cells shortened metaphase by 42 minutes and 53 minutes, respectively (S2B Fig). Therefore, both Ipl1^{Aurora B} and Mps1 are needed to maintain the spindle checkpoint delay in mek1Δ cells.

Chromosomes undergo error correction attempts throughout the delay

We hypothesized that perhaps anaphase onset occurred prior to having all the chromosomes bioriented in mek1Δ cells because either Ipl1^{Aurora B} localization or error correction activity decreased over time in mek1Δ cells. We first measured the level of Ipl1-3GFP that co-localized with the kinetochore protein Mtw1-mRuby2 at anaphase I onset. We found no significant difference in Ipl1-3GFP kinetochore levels in wildtype compared to mek1Δ cells (S3A and S3B Fig).

Because Ipl1^{Aurora B} kinetochore-localization did not decline, we tested our next hypothesis that Ipl1^{Aurora B} error correction activity declined during the delay. We monitored the number of error correction attempts of one pair of homologs throughout a mek1Δ delay. To visualize the chromosomes, we added a LacO array approximately 12kb from the centromere of both copies of chromosome IV. In the absence of Ipl1^{Aurora B} in meiosis, improper kinetochore-microtubule attachments are not corrected and both homologs stay attached to the same SPB [85,88]. Therefore, to ask whether Ipl1 activity was attenuated during the meiotic spindle checkpoint delay, we counted the number of error correction attempts during a mek1Δ spindle checkpoint delay by following the detachment and reattachment of a single pair of homologs between SPBs, taking images every 5 minutes. We scored chromosomes as bioriented and non-bioriented and noted when a chromosome traversed between two SPBs. We considered any traversing from one pole to another or switching from non-bioriented to bioriented as an error correction attempt (Fig 3A). In wildtype cells, biorientation occurs very quickly when the spindle is short in prometaphase I, making it difficult to monitor re-orientation [85]. But, biorientation is prolonged in mek1Δ cells with non-exchange chromosomes.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Chromosomes undergo error correction attempts throughout the spindle checkpoint delay in mek1Δ cells.

A) Cartoon showing an example of a mek1Δ cell undergoing four error correction attempts. B) Representative time-lapse images of a mek1Δ cell undergoing two error correction events in the first 50 minutes of the delay. Time 0 is the time at which SPBs separate. Arrowheads show movement of chromosome between SPBs. Scale bar = 5 μm. C) Time-lapse images of a mek1Δ cell undergoing error correction after 50 minutes of the metaphase I delay. Time 0 indicates SPB separation. Scale bar is 5 μm. Arrowheads show error correction event. D) Percent of cells that underwent error correction in the first 50 minutes of metaphase I and after 50 minutes of metaphase I. n≥ 100 cells. E) Graph of individual error correction attempts per cell within the first 50 minutes and after the first 50 minutes of metaphase I. n≥ 100 cells. Error bars show SD. F) Graph showing spindle length, measured 10 minutes prior to anaphase onset. n≥ 50 cells per genotype. Error bars show SD. Asterisk indicates a statistically significant difference between wildtype and mek1Δ (p < 0.0001, Mann-Whitney test). G) Time-lapse images of a cell that does not undergo error correction after 50 minutes of metaphase I. Time 0 is the time at which SPBs separate. Arrowheads show chromosome movement between SPBs. Asterisk indicates chromosomes straying from SPB. Scale bar = 5 μm.

https://doi.org/10.1371/journal.pgen.1010707.g003

Because the time of anaphase onset was approximately 50 minutes after SPB separation in wildtype cells, we reasoned that the additional spindle checkpoint delay in mek1Δ cells was beyond 50 minutes. We hypothesized that we may see a dampening of error correction events during the delay beyond 50 minutes. Therefore, we split our analysis into two categories, the number of error correction events in the first 50 minutes of metaphase I and the number of events after cells were delayed for 50 minutes (Fig 3B and 3C). We found that 78% of mek1Δ cells underwent error correction events in the first 50 minutes of the delay, with an average of 2.3 ± 1.7 error correction attempts per cell (average ± standard deviation) (Fig 3D and 3E). After 50 minutes, 47% of mek1Δ cells still underwent error correction attempts, with a total average of 2.1 ± 3 error correction attempts. Unfortunately, the 47% may be an underestimation because there may have been more cells that attempted error correction, but the event did not hold up to our criteria. For example, the mek1Δ cells had a longer spindle length than wildtype cells prior to anaphase I onset, likely due to fewer crossovers that resist the forces to keep the appropriate spindle size (3.3um ± 0.9 μm in wildtype compared to 4.5um ± 1.2 μm in mek1Δ cells; average ± SD; Fig 3F). The increased spindle length may have made it harder for kinetochores to attach to microtubules emanating from the opposite SPB, and instead, the released kinetochore reattached to the same SPB. We noticed that the LacI-GFP focus often drifted from the SPB, suggesting that the chromosome was released from the SPB, but may not have been able to traverse or biorient (Fig 3G). Therefore, we conclude that because we observed error correction throughout the spindle checkpoint delay in mek1Δ cells, premature attenuation of Ipl1 activity is likely not responsible for the short meiotic spindle checkpoint delay.

The spindle checkpoint is primarily silenced in meiosis I despite non-bioriented chromosomes

We wondered why the spindle checkpoint delay was shorter in meiosis compared to mitosis. To address this question, we asked how cells escaped the spindle checkpoint arrest. Progression past the spindle checkpoint occurs through various mechanisms, which can be classified into two major categories: checkpoint silencing and mitotic slippage. Checkpoint silencing is how the spindle checkpoint is turned off under normal conditions with the establishment of correct kinetochore-microtubule attachments. Checkpoint proteins are removed from the kinetochore, which ultimately leads to APC/C activation [89]. For example, the normal mechanism of turning off the checkpoint after chromosomes have bioriented is through PP1’s dephosphorylation of the kinetochore protein Spc105^KNL1 during mitosis in yeast, worms, and animal cells [3,21,22,90–95]. Dephosphorylation of Spc105^KNL1 reduces the binding affinity of Bub3, such that the kinetochore can no longer serve as a scaffold to build the MCC. During checkpoint slippage, cells can escape the spindle checkpoint arrest through various mechanisms despite the persistence of checkpoint proteins at the kinetochore [25,28,30].

We asked if differences in how cells escaped the spindle checkpoint could provide an explanation for why spindle checkpoint delays are shorter in meiosis than in mitosis. To distinguish between accelerated kinetochore silencing or slippage, we monitored spindle checkpoint proteins localized at the kinetochore at anaphase onset. We reasoned that with silencing, cells would disperse checkpoint proteins from the kinetochore prior to anaphase, while with slippage, cells would retain checkpoint proteins at the kinetochore during anaphase. A previous study using a similar assay found that mitotic budding yeast cells retained spindle checkpoint proteins at the kinetochore upon escaping the spindle checkpoint after the addition of microtubule depolymerizing drugs [28]. However, whether the same mechanism is used in cells that signal the checkpoint through another mechanism or during meiosis I is not known. We tagged the spindle checkpoint protein Bub3 with three copies of mCherry (Bub3-3mCherry) in cells containing the separase biosensor (Fig 4A). Using live-cell imaging, we monitored Bub3-3mCherry at the kinetochore and categorized its localization at anaphase onset.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. The spindle checkpoint is inappropriately silenced in meiosis I but not in mitosis.

A) Representative time-lapse images show cells entering anaphase, as indicated by the dispersal of separase biosensor focus. Time 0 indicates anaphase onset. Scale bars = 5 μm. Arrowhead shows Bub3-3mcherry focus after anaphase onset. B) Quantification of Bub3-3mCherry localization at anaphase onset in mitosis. n≥50 cells per genotype. C) Representative time-lapse images of a nocodazole-treated cell in meiosis I. Arrowhead shows the absence of Bub3 focus at the kinetochore. Scale bars = 5 μm. D) Quantification of Bub3-3mCherry localization at anaphase onset in meiosis I. n≥50 cells per genotype. E) Representative time-lapse images of a mek1Δ cell entering anaphase I. Time 0 indicates anaphase onset. Arrowhead shows the absence of Bub3 focus at the kinetochore. F-G) Quantification of fluorescence intensity at the kinetochore. At each time point indicated, the fluorescence intensity of the separase biosensor and Bub3-3mcherry were measured and compared to the measurements taken at -15 minutes. Time 0 indicates anaphase onset. Graphs display the percent of cells that, at each time point, retained at least 50% of the fluorescence intensity measured 15 minutes prior to anaphase onset. n≥30 cells per genotype.

https://doi.org/10.1371/journal.pgen.1010707.g004

We monitored the localization of Bub3-3mCherry during mitosis, using two different methods of activating the spindle checkpoint, through nocodazole addition, which disrupts kinetochore-microtubule attachments, and through depletion of Cdc6, which prevents replication of sister chromatids. We observed three phenotypes. First, Bub3-3mCherry dispersed from the kinetochore prior to anaphase onset. Second, Bub3-3mCherry dispersed in the same five-minute time interval in which anaphase onset occurred. Third, Bub3-3mCherry stayed localized during anaphase onset, not dispersing until after anaphase onset. When haploid cells were treated with nocodazole in mitosis, 64% retained Bub3-3mCherry at anaphase onset, as scored by separase biosensor cleavage (Fig 4B). These results supported previous findings of mitotic slippage in the presence of nocodazole [28]. Similarly, we observed that 68% of diploid cells depleted for Cdc6 retained Bub3-3mCherry at the kinetochore at anaphase onset, suggesting that most mitotic cells underwent mitotic slippage after prolonged checkpoint activation (Fig 4A and 4B).

We performed similar experiments in meiosis, using either nocodazole to disrupt kinetochore-microtubule attachments or deletion of MEK1 to decrease crossovers between homologous chromosomes. Strikingly, with nocodazole treatment, 58% of cells dispersed Bub3-3mCherry from the kinetochore prior to anaphase I onset (Fig 4C and 4D). In mek1Δ cells, Bub3-3mCherry localized to kinetochore foci, but also to the spindle, especially at anaphase onset (Fig 4D and 4E). Similar to the nocodazole-treated cells, the focus of Bub3 at the kinetochore dispersed from the kinetochore prior to anaphase I onset in 54% of mek1Δ cells. Although, we note that there is still a population on the microtubules.

To further analyze our results, we measured the fluorescence intensity of both kinetochore-localized Bub3-3mcherry and the separase biosensor foci every five minutes, from 15 minutes prior to anaphase until 10 minutes after anaphase. We plotted the percentage of cells that retained at least half of the fluorescence measured at 15 minutes prior to anaphase. In Cdc6-dp mitotic cells, 87% of cells retained at least half of their localized Bub3 fluorescence at anaphase onset (Fig 4F). In contrast, only 38% of mek1Δ cells retained at least half of their Bub3 fluorescence intensity at anaphase I onset (Fig 4G). These results are consistent with the conclusion that cells in mitosis primarily undergo mitotic slippage, while cells in meiosis I primarily silence the checkpoint at the kinetochore after a prolonged delay.

We also monitored Bub3-eGFP kinetochore localization in cells expressing mCherry-Tub1, scoring anaphase I onset as the time at which spindle elongation occurred in wildtype and mek1Δ cells (S4A Fig). Similar to wildtype, most mek1Δ cells dispersed Bub3-eGFP from the kinetochore prior to anaphase I onset (S4B Fig). Although both Cdc6-depletion and loss of MEK1 prevent chromosomes from achieving biorientation in either mitosis or meiosis, respectively, our results suggest that unlike mitotic cells, most meiosis I cells underwent inappropriate spindle checkpoint silencing after a delay.

We next wondered if the checkpoint in meiosis II more closely resembled that of meiosis I or mitosis. To disrupt kinetochore tension in meiosis II, we deleted SPO12. Previous experiments showed that spo12Δ cells undergo meiosis I normally but fail to duplicate SPBs in meiosis II, producing two half-spindles or one weak spindle in between the two old SPBs [96–98]. We then monitored the cleavage of the centromeric Rec8-GFP at anaphase II. Similar to previous findings, the spo12Δ cells were delayed in cohesin cleavage, with an average duration from cohesin cleavage in anaphase I to cohesin cleavage in anaphase II of 156 ± 54 minutes (average ± SD), which is 93 minutes longer than that observed in wildtype cells (S5A–S5C Fig) [98]. The delay was dependent on the spindle checkpoint because disruption of MAD3 resulted in a decreased time to anaphase II onset. We conclude that the spindle checkpoint delay is also shorter in meiosis II compared to mitosis.

One hypothesis for a faster anaphase II onset in meiosis II compared to mitosis is that precocious APC/C activation occurs through the meiosis-specific co-activator the APC/C, Ama1, which is active at the end of meiosis II [98]. Ama1 is impervious to the spindle checkpoint, which inhibits the APC/C coactivator Cdc20. To test this hypothesis, we deleted AMA1 in spo12Δ cells and measured the duration of anaphase II onset. The spo12Δ ama1Δ cells had a delay with a duration similar to that of spo12Δ cells (S5C Fig). Therefore, the shorter duration of the spindle checkpoint delay in meiosis II compared to mitosis was not dependent on Ama1.

To determine if the meiosis II cells underwent mitotic slippage or checkpoint silencing, we monitored Bub3-3mcherry kinetochore localization at anaphase II onset. Surprisingly, we found that 87% of cells retained Bub3-3mcherry at the kinetochore at anaphase II onset, suggesting that they underwent mitotic slippage (S5D and S5E Fig). Similarly, 83% of spo12Δ ama1Δ cells retained Bub3-3mcherry at the kinetochore at anaphase II onset. Therefore, in meiosis II, slippage was not dependent on the activation of APC/C^Ama1. Overall, we conclude that meiosis I has a mechanism for spindle checkpoint silencing after a prolonged delay, whereas cells in mitosis or meiosis II undergo mitotic slippage.

Kinetochore-localized PP1 silences the spindle checkpoint in meiosis I in the presence of non-bioriented chromosomes

We hypothesized that with prolonged spindle checkpoint activity in meiosis, cells prematurely triggered the normal kinetochore checkpoint silencing pathway through PP1. The current model for spindle checkpoint silencing at the kinetochore in budding yeast is that once proper kinetochore-microtubule interactions are established, PP1 binds Spc105^KNL1 and dephosphorylates Mps1 and Ipl^{Aurora B} substrates, releasing Bub3-Bub1 from the kinetochore (Fig 5A) [1]. When chromosomes are not properly attached, Ipl1^{Aurora B} counteracts PP1’s kinetochore binding by phosphorylating the RVSF motif on Spc105^KNL1. Previous reports have shown that in vegetative cells, the spc105^RASA mutation, which disrupts PP1’s kinetochore binding, is lethal, hypothesized to be due to persistent spindle checkpoint activity (Fig 5B) [22].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Mitotic slippage occurs in spc105^RASA mitosis.

A) Cartoon representation of Spc105^WT in metaphase (left) and at the metaphase-to-anaphase transition (right). To signal the spindle checkpoint, Mps1 phosphorylates Spc105 for binding of spindle checkpoint proteins at the kinetochore. At the metaphase-to-anaphase transition, PP1 binds Spc105 at the RVSF motif to dephosphorylate Spc105 and silence the spindle checkpoint. B) Cartoon representation of spc105^RASA mutation, which prevents PP1 from binding at the RVSF motif on Spc105 and prevents PP1 from silencing the spindle checkpoint. C) Representative time-lapse images of a spc105^RASA cell undergoing delayed spindle elongation. Time 0 is the time at which SPBs separate. Scale bar = 5 μm. D) Graph depicts mean duration of mitosis, measured from SPB separation to spindle elongation. n≥100 cells per genotype. Error bars show SD, * indicates statistical significance, p<0.0001, Mann-Whitney test. E) Representative time-lapse images of a spc105^RASA cell expressing Cdc14-GFP. Time 0 indicates SPB separation. Scale bar = 5 μm. F) Quantification of the percentage of cells that release or retain Cdc14-GFP upon spindle elongation. n = 100 cells. G) Representative time-lapse images of a spc105^RASA cell undergoing mitotic slippage. Time 0 is the time at which SPBs separated. Scale bar = 5 μm. H) Percentage of cells that retained Bub3-eGFP at the kinetochore during anaphase I (red), dispersed Bub3-eGFP prior to anaphase I onset (blue), and dispersed Bub3-eGFP from the kinetochore at the same time frame as anaphase I onset (gray). n≥50 cells per genotype.

https://doi.org/10.1371/journal.pgen.1010707.g005

We first asked if the lethality of the spc105^RASA cells in mitosis was caused by a persistent checkpoint arrest, or if the cells were able to undergo mitotic slippage. We used the anchor away system to deplete Spc105 from the nucleus in cells expressing spc105^RASA under the SPC105 promoter. We tagged endogenous SPC105 at its C-terminus with a FRB tag in a strain with RPL13A tagged with FKBP12. Addition of rapamycin allows the stable interaction between FRB and FKBP12, depleting Spc105-FRB from the nucleus but leaving spc105^RASA in the nucleus. As a control, we integrated a wildtype copy of SPC105 into the genome of strains in which we anchored away endogenous Spc105. We found that spc105^RASA cells underwent anaphase onset, after a prolonged metaphase delay, with anaphase onset approximately 356 minutes later than Spc105-FRB cells expressing wildtype SPC105 or mad3Δ cells expressing spc105^RASA (Fig 5C and 5D). As another marker of anaphase, we found that the Cdc14 phosphatase was released from the nucleolus in 95% of the cells that elongate their spindles at anaphase (Fig 5E and 5F). The spc105^RASA cells died after the first or second cell cycle following rapamycin addition. Furthermore, we observed that Bub3 remained localized at the kinetochore at anaphase onset in 98% of cells, suggesting persistent checkpoint signaling (Fig 5G and 5H). Therefore, we conclude that the cells that cannot localize PP1 underwent mitotic slippage after a prolonged checkpoint arrest.

To determine if PP1 is needed for the premature checkpoint silencing seen in meiosis I, we integrated mutant spc105^RASA (expressed under the meiosis-specific REC8 promoter) into a strain expressing SPC105-FRB. As a control, we integrated a wildtype copy of REC8 promoted SPC105 into the genome in strains that expressed SPC105-FRB (SPC105^WT). The control cells with SPC105^WT completed both meiotic divisions with normal timings (54 ± 20 minutes for meiosis I and 42 ± 12 minutes for meiosis II in the SPC105^WT strain, compared to 45 ± 15 minutes and 51 ± 17 minutes in a wildtype strain; Fig 6A and 6B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Kinetochore-localized PP1 prematurely silences the spindle checkpoint in meiosis.

A) Representative time-lapse images of a SPC105^WT cell completing both meiotic divisions. Time 0 indicates SPB separation. Scale bar = 5 μm. B) Graph showing the duration of metaphase I and metaphase II in SPC105^WT cells with and without the anchor away background. Error bars show SD. C) Quantification of cell cycle completion in Spc105^WT and spc105^RASA cells. Phenotypes of cells that entered prometaphase I were scored. n≥100 cells per genotype. D) Representative time-lapse images of a spc105^RASA cell arresting after anaphase I. E) Graph showing the mean time from SPB separation to anaphase I spindle elongation. Error bars show SD. * indicates statistical significance as determined by Mann-Whitney test (p <0.0001). n≥100 cells per genotype. F) Time-lapse images of a spc105^RASA cell releasing Cdc14-GFP from nucleolus at the time of spindle elongation. Time 0 is the time at which SPBs separate. Scale bar = 5 μm. G) Representative time-lapse images of a SPC105^WT cell. Scale bar = 5 μm. H) Percentage of cells that retained Bub3-eGFP at the kinetochore during anaphase I (red), dispersed Bub3-eGFP prior to anaphase I onset (blue), and dispersed Bub3-eGFP from the kinetochore at the same time frame as anaphase I onset (gray). n≥50 cells per genotype. I) Representative time-lapse images of a spc105^RASA cell. Scale bar = 5 μm.

https://doi.org/10.1371/journal.pgen.1010707.g006

The spc105^RASA cells that entered metaphase I exhibited three phenotypes: 7% arrested in metaphase I for the duration of the imaging; 77% completed anaphase I after a delay but did not progress into metaphase II; and 16% completed both meiotic divisions (Fig 6C and 6D). Those that progressed into anaphase I had a metaphase I duration of 192 ± 32 minutes (mean ± SD; Fig 6E). The failure to complete meiosis was due to spindle checkpoint activity; deletion of the spindle checkpoint protein MAD3 in the spc105^RASA cells resulted in 98% of cells that completed both meiotic divisions (Fig 6C). We find that 100% of the cells that underwent anaphase I also released the phosphatase Cdc14 from the nucleolus, serving as another marker for anaphase I onset (Fig 6F). These results suggest that the cells are undergoing a transition from metaphase I to anaphase I, not just inappropriate spindle elongation.

Similar to spc105^RASA cells, the mek1Δ spc105^RASA cells that progressed into metaphase I also displayed the three phenotypes with 11% arrested in metaphase I for the duration of the movie, 88% arrested after anaphase I, and 9% completed both divisions (Fig 6C). Interestingly, the mek1Δ spc105^RASA cells that underwent anaphase I did so after a delay, in that metaphase I was 212 ± 52 minutes, which is 157 minutes longer than wildtype cells expressing wildtype SPC105 and 36 minutes longer than mek1Δ cells expressing wildtype SPC105 (mean ± SD; Figs 6E and 1H). These results demonstrate that spindle checkpoint activity causes a delay but does not completely block anaphase I onset in cells that cannot localize PP1 to the kinetochore.

We were surprised to find that the spc105^RASA and mek1Δ spc105^RASA cells were able to undergo anaphase I after the delay. We questioned whether these cells were silencing or undergoing slippage in response to spindle checkpoint activity. We monitored the localization of Bub3-eGFP at the onset of spindle elongation. We found that 80% of cells with wildtype SPC105 dispersed Bub3-eGFP from the kinetochore prior to anaphase I spindle elongation, as expected (Fig 6G and 6H). In contrast, in greater than 85% of spc105^RASA and mek1Δ spc105^RASA cells, Bub3 persisted at the kinetochore after spindle elongation (Fig 6H and 6I). These results are consistent with mitotic slippage, in that in the absence of PP1 kinetochore localization, the spindle checkpoint proteins were not released from the kinetochore, but the cells were able to undergo anaphase through another mechanism. These results suggest that in meiosis I there are two mechanisms to ensure that cells can escape a permanent checkpoint arrest: i) after a delay, cells will silence the spindle checkpoint through PP1-dependent removal of the checkpoint proteins at the kinetochore, and ii) if silencing is abrogated, cells will undergo slippage in response to persistent spindle checkpoint activity. However, most of the spc105^RASA and mek1Δ spc105^RASA cells that escape spindle checkpoint activity were unable to enter meiosis II, suggesting that the cell cycle is mis-regulated due to persistent checkpoint activation. We conclude that with prolonged spindle checkpoint activity in meiosis I, cells use a PP1-dependent mechanism to silence the checkpoint with non-bioriented chromosomes, whereas in mitosis, cells undergo mitotic slippage.

PP1 kinetochore binding limits the duration of the spindle checkpoint delay

When chromosomes are not properly attached to microtubules, Ipl1^{Aurora B} counteracts PP1’s kinetochore binding by phosphorylating the RVSF motif in Spc105^KNL1. The phosphorylation can be abrogated by changing the RVSF motif to RVAF [22,94] (Fig 7A). The spc105^RVAF cells have an accelerated metaphase I and metaphase II, but not an accelerated mitosis [22,88]. We hypothesized that if PP1 binding at the kinetochore sets the timing of checkpoint silencing, then mek1Δ cells expressing spc105^RVAF should silence the checkpoint more quickly than mek1Δ SPC105^WT cells. Indeed, we found that mek1Δ spc105^RVAF cells progressed from prometaphase I to anaphase I at an average time of 106 ± 24 minutes, compared to 190 minutes in mek1Δ SPC105^WT cells (Fig 7B). Furthermore, Bub3-eGFP dispersed from kinetochores before anaphase onset in 86% of mek1Δ spc105^RVAF cells analyzed compared to the 54% observed in mek1Δ cells (S4A and S4B Fig). Thus, premature PP1 binding accelerates the time of anaphase onset through checkpoint silencing in mek1Δ cells. We conclude that PP1 limits the duration of the spindle checkpoint delay in meiosis I by silencing kinetochore-based signaling.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. PP1 kinetochore binding sets the duration of the spindle checkpoint delay.

A) Cartoon depicting spc105^RVAF allele. Ipl1 cannot phosphorylate the RVSF motif, which allows PP1 to bind Spc105 prematurely. B) Graph showing the mean time from SPB separation to separase activation in meiosis in the indicated genotypes. Error bars show SD. Asterisk indicates statistical significance relative to wildtype (p< 0.0001, Mann-Whitney test). n≥100 cells per genotype. C) Graph showing the mean time from SPB separation to separase activation in mitosis in the indicated genotypes. Error bars show SD. Asterisk indicates statistical significance relative to wildtype (p< 0.0001, Mann-Whitney test). n≥100 cells per genotype. D-E) Model of PP1 prematurely silencing the spindle checkpoint in meiosis I, but not in mitosis. D) Activation of the spindle checkpoint in mitosis causes a relatively long delay, which cells eventually escape through mitotic slippage. E) Activation of the spindle checkpoint in meiosis causes a relatively short delay, which cells escape through a PP1-dependent checkpoint silencing. Failure of PP1 to localize at Spc105’s RVSF motif causes meiotic cells to undergo slippage. Premature PP1 kinetochore recruitment accelerates meiotic checkpoint silencing.

https://doi.org/10.1371/journal.pgen.1010707.g007

Although we observed that Cdc6-dp cells underwent mitotic slippage in response to a prolonged checkpoint delay, we wondered if forcing PP1’s premature kinetochore localization could accelerate the time at which Cdc6-dp cells escaped the checkpoint. We expressed spc105^RVAF in Cdc6-dp cells and measured the duration of SPB separation to separase activation. We found that Cdc6-dp spc105^RVAF cells were in metaphase for 195 ± 104 minutes, which is 155 minutes faster than Cdc6-dp cells (average ± SD; Fig 7C). Thus, allowing PP1 to prematurely bind the kinetochore in mitosis or meiosis reduces the duration of the spindle checkpoint delay.

Strains and manipulations

S. cerevisiae strains used in this study are derivatives of W303 (S1 and S2 Tables). Gene deletions and gene tagging were performed using standard PCR-based lithium acetate transformation [104,105]. Genotypes of transformed strains were verified by PCR. Plasmids containing the separase biosensor (P_CUP1-LacI-Rec8-GFP:His3 and P_CUP1-LacI-Scc1-GFP:His3) were gifts from David Morgan [47]. Strains containing P_GAL1-CDC6 and cdc6::ura3 were gifts from Andrew Murray [56]. Strains harboring homeologous chromosome V were S288C-derived strains and were back-crossed into W303 at least six times to obtain W303-derived strains harboring the homeologs. SLCs were introduced into yeast strains by restriction digest with BamHI, which linearizes the plasmid. All strains harboring SLCs were grown in media lacking leucine to maintain selection for SLCs. Spc105 point mutations were made by site-directed mutagenesis of a yeast integrating plasmid containing Spc105 with its endogenous promoter and terminator (~700 bp upstream and downstream of the SPC105 ORF) for mitosis experiments. For meiosis experiments, the same site-directed mutagenesis was performed on yeast integrating plasmids containing SPC105 with its endogenous terminator but containing the REC8 promoter (900 bp upstream of the REC8 start codon). Mutant and wildtype control alleles of SPC105 were integrated at the LEU2 or TRP1 locus in anchor away strains in which endogenous Spc105 was tagged with FRB at its C-terminus such that endogenous Spc105 was depleted from the nucleus upon addition of rapamycin. The anchor away strains also have tor1-1 and fpr1Δ to prevent rapamycin toxicity.

Growth conditions

All cultures were kept on a roller drum at the indicated temperature. All liquid media was supplemented with 1% tryptophan and 0.5% adenine. For all meiosis experiments, unless otherwise noted, cells were grown in 2 mL liquid media containing 1% yeast extract, 2% peptone, and 2% dextrose (YPD) for 10–24 hours at 30°C, then 40 μL of this culture was transferred to 2 mL of 1% yeast extract, 2% peptone, and 2% potassium acetate (YPA) for 12–15 hours at 30°C (S3 Table). Cells were washed twice with water and transferred to 1% potassium acetate (1% KAc) for incubation at 25°C. Time-lapse imaging of meiosis, except NDT80-in strains, began at 6–8 hours post-KAc transfer. For Spc105 anchor away movies, rapamycin was added to a final concentration of 1μM 6–7 hours after KAc transfer such that Spc105 remained unperturbed for prophase I. P_GAL1,10-NDT80 GAL4-ER strains were kept in 1% KAc at 25°C for 10–12 hours, before release from prophase I by addition of beta-estradiol to a final concentration of 1mM. Cells harboring a SLC were grown in media lacking leucine to maintain selection for the SLC. These cells were prepared for imaging by inoculating 2 mL of 2XSC+glucose–leucine (0.67% yeast nitrogen base without amino acids, 0.2% dropout mix containing all amino acids except leucine, 2% glucose) and letting the culture grow for 10–24 hours at 30°C. 40 μL of this culture was transferred to 2 mL 2XSCA—leucine (0.67% yeast nitrogen base without amino acids, 0.2% dropout mix containing all amino acids except leucine, 2% potassium acetate) for 12–16 hours at 30°C, then washed twice with water before resuspending in 1% KAc for 6–8 hours of incubation at 25°C. For all mitosis experiments, unless otherwise noted, cells were grown overnight in synthetic complete media (2XSC + glucose). For nocodazole experiments, cells were grown in 2XSC + 1% peptone, because the presence of peptone increases the effect of nocodazole in synthetic media [28,106]. For all Cdc6-dp experiments, cells were grown in 2XSC+ 20% galactose until time of experiment.

Time-lapse microscopy

Time-lapse imaging was performed as follows, unless otherwise noted. 200 μL of cells were concentrated and spread on a 35 mm x 40 mm coverslip coated with 1 mg/mL Concanavilin A (ConA) and fitted to a chamber [107]. A 5% agar plug was made by cutting off the tip of an Eppendorf tube and pipetting ~100 μL of melted 5% agar into the tube sitting on a glass slide. The dried agar plug was placed on top of the cells sitting on the ConA for 12 mins, such that the cells had time to adhere to the ConA. The rest of the KAc (or 2XSC for mitotic cells) culture (~ 2 mL) was spun down, and the supernatant was added drop-wise to the chamber in order to gently float and remove the agar pad. After cells were adhered to the coverslip and the chamber filled with medium, the chamber was immediately placed on the microscope for imaging. Two microscopes were used for this study. Images were acquired on a Delta Vision Elite (eDV) microscope, equipped with a PCO Edge5.5 sCMOS camera and Olympus 60X oil-immersion objective lens (Plan Apo N, 1.42). Cells were imaged at room temperature using FITC and mCherry filters. FITC exposure was at 2% for 15–25 msec and mCherry exposure was at 2–5% for 100 msec. Brightfield images were acquired at 5% intensity for 50 msec. Images were acquired using the SoftWorx Version 7.0 software (G.E. Healthcare), and data analysis was performed using Fiji and Image J. Nocodazole and Spc105 mitosis experiments were performed using a Nikon Ti2 microscope equipped with a Photometrics camera and 60X oil immersion objective lens. GFP and Ruby filters were used. Images were acquired with exposure times of 20–40 msec for GFP and Ruby filters with neutral density (ND) filters transmitting 2–5% of light intensity. Brightfield images were acquired at 70 msec at 5% ND. Images were acquired and data were analyzed using NIS Elements Viewer Version 4.20 software.

All images were collected with 5, 1–1.2-μM z-stacks, in 10-minute time increments, for 10–12 hours, unless otherwise noted. For experiments on the Nikon Ti2, NIS elements was used for analysis. Fiji software (NIH) was used to create final images with adjustment of brightness and contrast.

Image analysis and Fluorescence intensity measurement

Movie analysis was conducted in Fiji software (data analysis in S1 Data). Unless otherwise noted, z-stacks were combined into a single maximum-intensity projection for analysis and time-lapse images.

Fluorescence intensity measurement of Bub3-3mcherry and the separase biosensor focus was done with Fiji. Images were projected in SUM slices, and a small circle was drawn around each focus at the time points indicated. A background circle was drawn just beside the measured point for each channel, in each cell, at each time point, and subtracted from the fluorescence intensity measurement to generate each value. Fluorescence intensity measurement of Ipl1-3GFP and Mtw1-mRuby2 was performed using SUM slice projection of images. A small circle was drawn around Mtw1-mRuby2 using the circle tool in Fiji, and the fluorescence intensity of the mCherry channel and the GFP channel within the drawn circle were recorded.

Spindle length measurement

Spindle length was calculated by identifying the location of one SPB in a single, 1 μm z-stack out of a 5-stack image. The opposite SPB was detected in the same way, and the number of 1 μm stacks between the two SPBs was noted. A line was draw between the two SPBs and the Pythagorean theorem was used to calculate the distance between the two SPBs.

Cell volume measurement

Cell volume was measured using BudJ: an ImageJ plugin to analyze fluorescence microscopy images of budding yeast cells (http://www.ibmb.csic.es/home/maldea; [108]). Time-lapse imaging files were cropped such that the frames from 20 mins prior to anaphase until 30 minutes after anaphase were included in the analysis (the data reported are from 10 mins prior to anaphase onset). For cells in mitosis, the volume reported is the volume of the mother plus the volume of the bud.

Cdc6 depletion

For Cdc6-dp experiments, cells were grown overnight in 2X SC + galactose medium and diluted 1:20 the following day. After growth for 3 h, cells were spun down, rinsed 3 times with 2X SC + glucose, and resuspended in 2X SC + glucose containing 20 μM copper sulfate to induce expression of the separase biosensor. Cells were immediately spread on a coverslip containing Concanavilin A (ConA) and resuspended in the 2X SC + glucose containing 20 μM copper sulfate. For analysis, only the cells which were in S phase through early M phase at the start of imaging were analyzed, as determined by spindle length. These cells were monitored through the rest of the cell cycle, and only the following cell cycle was analyzed to allow for depletion of Cdc6 and to avoid analyzing cells which had undergone an entire cell cycle without Cdc6.

Nocodazole treatment

For all nocodazole (Sigma M1404) experiments, 30 mg/mL stocks in DMSO were kept at -20C in 20 μL aliquots to avoid multiple freeze-thaw cycles of the drug. Because nocodazole interferes with SPB separation, the drug was not added until two SPBs were observed. Additionally, image analysis of all nocodazole-treated cells and their respective controls was performed such that the time at which the imagining began was regarded as time 0.

For mitosis experiments, cells were grown in 2X SC + 1% peptone. Addition of peptone has been shown to increase the effect of nocodazole in synthetic media [28,106]. An overnight culture of cells was grown in 2XSC + 1% peptone+ 20 μM copper sulfate and diluted to 1:20 the following day. After 3 h of growth cells were arrested with 25 μM α-factor (Zymo Research) for 3 h. Cells were washed 5 times with 2X SC + 1% peptone + 20 μM copper sulfate to release from α -factor arrest. Following release, cells were grown for 30 mins at 30°C, and imaged under the microscope to confirm separation of SPBs. Cells were then immediately re-suspended in 2X SC + 1% peptone + 20 μM copper sulfate to which nocodazole was added drop-wise to a final concentration of 15 ug/mL. Cells were immediately spread on a coverslip containing ConA for imaging, and 2X SC + 1% peptone containing 15 μg/mL nocodazole and 20 μM copper sulfate was added to the chamber surrounding the coverslip.

For all meiosis nocodazole experiments, pre-conditioned media was prepared by taking a diploid strain through the sporulation process (12 h YPD, 12 h YPD, 24 h 1% KAc). The sporulation culture was then spun down and the supernatant filter sterilized. 1–2 h before the start of imaging, beta-estradiol was added to a final concentration of 1 mM, copper sulfate was added to a final concentration of 20μM, and nocodazole was added dropwise to a final concentration of 30 μg/mL. All meiosis nocodazole experiments were performed in cells harboring the NDT80 gene under control of the GAL1,10 promoter with the transcriptional activator GAL4-ER integrated into the genome, such that addition of beta-estradiol releases Gal4 into the nucleus to activate Ndt80 expression. Cells were grown in YPD for 8–12 h, after which 40 μL of culture was transferred to 3 mL YPA for 12–15 h, and then transferred to 3 mL 1% KAc for 12 h to induce a prophase I arrest. Cells were released from prophase I arrest by addition of beta-estradiol to a final concentration of 1 mM. Just after SPB separation was detected (~80 mins after release from prophase I), cells were resuspended in pre-conditioned media described above (1% KAc with 1 mM beta-estradiol, 20 μM copper sulfate, and 30 μg/mL nocodazole, and immediately spread on a coverslip containing ConA for imaging. 2 mL of the pre-conditioned 1% KAc + 30 μg/mL nocodazole + 20 μM copper sulfate was added to the chamber surrounding the coverslip.