Dnt1 is involved in facilitating proper mitotic chromosome segregation

In the course of our characterizing Dnt1 as a negative regulator of Dma1 in early mitosis [44], we noticed that dnt1Δ cells were extremely sensitive to microtubule-destabilizing drug thiabendazole (TBZ), and deletion of dma1⁺ in these cells only slightly reversed this phenotype (Fig 1A). Since the disruption of genes participating in chromosome segregation, such as the kinetochore- or spindle checkpoint-related genes, usually results in TBZ sensitivity [48,49], we assumed that dnt1⁺ gene might have dma1⁺-irrelevant function that is essential for chromosome segregation.

Download:

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

Fig 1. Characterization on the involvement of Dnt1 in facilitating proper mitotic chromosome segregation.

(A) Tenfold serial dilution analyses of the indicated yeast strains grown on the indicated media to measure the TBZ sensitivity. (B) Minichromosome loss rate per division was measured in strains bearing the Ch16 (ade6-M216) minichromosome and ade6-M210 allele (n > 8,000). (C) (Left) Chromosome segregation was observed in anaphase in the indicated cells carrying mCherry-tagged α-tubulin (atb2⁺) (red) after being fixed and stained with DAPI (blue). (Right) Quantitative analysis of chromosome missegregation and lagging chromosome phenotypes. n, numbers of anaphase cells analyzed. Scale bar, 5 μm. (D-F) Anaphase B is slightly delayed in dnt1Δ cells. (D) Time-lapse microscopy of a wild-type and a dnt1Δ cell carrying ndc80-GFP and sad1-mCherry during mitosis. Images were acquired at 2-min intervals. Arrows indicate a lagging chromosome. The enlarged image corresponding to the time frame indicated by red dot is shown in inset. Scale bar, 5 μm. (E) Distance between SPBs (SPB to SPB, marked by Sad1-mCherry) was measured at 2-min intervals. Each line represents data collected from an individual cell. Dashed lines indicate 3 μm, which roughly marks the timing for start of anaphase B in most cases. (F) Box-and-whiskers representation of anaphase B (during which pole-to-pole distance increases with separated sister chromatids) duration, in which boxes indicates median and upper and lower quartile and whiskers indicates range of data. The data were extracted and quantified from spindle dynamics measurements in (D). ****, p<0.0001.

https://doi.org/10.1371/journal.pgen.1010397.g001

Consistent with this assumption, we found that dnt1Δ cells lost minichromosomes (Ch16, ade6-M216) at an elevated rate that is almost 100 times higher than that of the wild-type (Fig 1B), and displayed increased frequency of lagging chromosomes and chromosome mis-segregation at mitotic anaphase (Fig 1C). By following the kinetochore separation and spindle dynamics during mitosis under time-lapse microscopy, we found that dnt1Δ cells spent almost the same length of time at late prometaphase/metaphase and anaphase A as wild-type cells except for the occasionally observed lagging chromosomes (2 cases in 14 dnt1Δ cells but 0 in 17 wild-type cells) at early anaphase, but dnt1Δ cells stayed for extended length of time at anaphase B (Fig 1D–1F). These data strongly suggested that Dnt1 is involved in facilitating proper chromosome segregation.

Chromosome segregation is a precisely regulated process involving many proteins, including kinetochore proteins, monopolins, cohesins, chromosome passenger proteins, centromeric heterochromatin proteins, microtubule-binding proteins, and regulators of kinetochore-microtubule attachment [50]. In order to dissect the mechanism behind how Dnt1 is involved in maintaining the fidelity of chromosome segregation, we systematically tested the genetic interactions between dnt1Δ and some representative mutants that have been previously reported to cause chromosome missegregation. Strikingly, dnt1Δ showed mild to strong negative genetic interactions with almost all these mutants (S1 Fig), indicating that Dnt1 became essential when any of these proteins was absent or mutated. Thus, Dnt1 should be involved in faithful mitotic chromosome segregation in a previously unrecognized manner.

Dnt1 is dispensable for activating the SAC in the absence of kinetochore-microtubule attachment or tension

The spindle checkpoint is another safeguard mechanism that ensures proper chromosome segregation, which is rapidly activated when the kinetochore-spindle microtubule attachment or the tension generated by this attachment is absent or compromised [16]. We then examined whether the SAC is properly activated in dnt1Δ cells when the kinetochore-spindle microtubule attachment or the tension was compromised by either the cold-sensitive β-tubulin mutation nda3-KM311 [51] or the temperature-sensitive cohesin subunit mutation psc3-1T [52,53], respectively. In these two experimental set-ups, the accumulation of two major APC/C substrates cyclin B (Cdc13-GFP) and securin (Cut2-GFP) at SPBs or within nuclei, respectively, served as an indicator of SAC activation and mitotic arrest (S2 Fig). Our quantified data showed that the Cdc13-GFP or Cut2-GFP were accumulated at SPBs or within nuclei in nda3-KM311 or pst3-1T cells, respectively, with indistinguishable rate in the presence and absence of dnt1⁺ (S2B and S2D Fig), indicating that Dnt1 is largely dispensable for activating SAC in the absence of either attachment or tension.

Dnt1 facilitates efficient anaphase initiation upon SAC inactivation

The foregoing results rendered us to hypothesize that Dnt1 might have a role in turning off the spindle checkpoint in mitosis. To explore this, we adopted two well-established SAC silencing assays [39,54]. In these assays, the SAC was first robustly activated by the nda3-KM311 mutant simultaneously carrying ark1-as3 and then inactivated by addition of ATP analogue 1-NM-PP1 or simply by shifting mitotically arrested cells back to permissive temperature (30°C) as previously described (Figs 2A and S3A). Because Cdc13 (cyclin B) and Cut2 (securin) localize to the spindle pole bodies (SPBs) or nucleus, respectively, in early mitosis and should be degraded by APC/C to promote metaphase-anaphase transition, the disappearance rate of Cdc13-GFP spot or nuclear Cut2-GFP after 1-NM-PP1 addition or shifting back to 30°C reflects the SAC inactivation efficacy (Figs 2B and S4A). We found that dnt1Δ cells retained high amounts of SPB-localized Cdc13-GFP and nuclear Cut2-GFP for much prolonged period compared to wild-type cells, almost to the same degree as previously identified SAC-inactivation defective mutant dis2Δ [39] (Figs 2C, S3B, and S4B). Also, the total cellular protein of Cdc13 or Cut2 was indeed degraded much slower in dnt1Δ cells than in wild-type cells recovered from nda3-mediated checkpoint arrest (Figs 2D, 2E and S4B). All above data suggested that Dnt1 is required for the timely inactivation of SAC to efficiently initiate anaphase.

Download:

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

Fig 2. Dnt1 facilitates timely degradation of Cyclin B and dissociation of Mad2 and Mad3 from APC/C complex upon SAC inactivation.

(A) Schematic depiction of the experiment design for (B-G and I). (B-E) Dnt1 is required for timely degradation of Cyclin B upon SAC inactivation. Example pictures of cells with Cdc13-GFP signals enriched or disappeared at spindle pole bodies (SPBs) are shown in (B). The percentage of cells with Cdc13-GFP on SPBs was assessed at each time point after shift to 30°C (C), which served as an indicator for mitotic exit. Total protein levels of Cdc13-GFP were detected by Western blotting (D) and normalized to those of total Cdc2 at each time point, with the relative ratio between Cdc13-GFP and Cdc2 at 0 min set as 1.0 (E). Scale bar, 5 μm. (F-G) Dissociation of the Mad2 and Mad3 checkpoint proteins from the APC/C during mitotic arrest-and-release is delayed in dnt1Δ cells. The association of Mad2 and Mad3 to Apc4/Lid1 was assessed by immunoprecipitation of Apc4/Lid1-TAP and Western blot (F). The amount of co-immunoprecipitated Mad2 and Mad3 was normalized to those of total immunoprecipitated Apc4/Lid1 at each time point, with the relative ratio between Mad2-GFP or Mad3-GFP and Apc4/Lid1-TAP at 0 min set as 1.0 (G). (H) Schematic summary of results shown in C-G. Prolonged MCC-APC/C association in dnt1Δ cells anti-correlates with decelerated anaphase entry and mitotic exit upon spindle checkpoint inactivation. (I) Depletion of SAC signaling by deletions of mad2⁺ or bub1⁺ abrogates the delayed SAC inactivation after release from checkpoint arrest in dnt1Δ cells. Cells of indicated strains bearing Cdc13-GFP were treated and assessed as in (B and C). The experiments were repeated two (D-G) or three (C and I) times. Error bars correspond to standard deviation (SD).

https://doi.org/10.1371/journal.pgen.1010397.g002

Dnt1 is required for timely dissociation of MCC from APC/C during SAC inactivation

In fission yeast, the key SAC components Mad2 and Mad3 and one molecule of Slp1^Cdc20 form mitotic checkpoint complex (MCC) which binds to APC/C through another molecule of Slp1^Cdc20 upon checkpoint arrest, and the recovery from mitotic arrest accompanies the loss of MCC-APC/C binding [22,39,55,56]. To examine whether delayed anaphase initiation upon recovery from mitotic arrest in dnt1Δ cells was due to persistent MCC-APC/C binding, we analyzed the rate of dissociation of the MCC from the APC/C by immunoprecipitations of the APC/C subunit Apc4/Lid1 from mitosis-exiting cells. Indeed, the levels of MCC bound to the APC/C in dnt1Δ cells stayed high for longer period than in wild-type cells (Fig 2F and 2G), suggesting that Dnt1 functions for timely and efficient MCC-APC/C dissociation and spindle checkpoint inactivation (Fig 2H), which is required for activating APC/C to degrade Cdc13 and Cut2. Deletions of mad2⁺ or bub1⁺ significantly lowered the percentage of dnt1Δ cells with Cdc13 at SPBs upon nda3-mediated checkpoint activation (Fig 2I), indicating that sustained SAC activation and consequent prometaphase- or metaphase-arrest in the absence of Dnt1 relies on prior SAC activation.

Dnt1 positively regulates the protein levels of Slp1^Cdc20

Given that MCC functions as a potent inhibitor of APC/C upon spindle checkpoint activation and Dnt1 is required to promote timely dissociation of MCC from APC/C, we were suspicious that one direct consequence of the prolonged MCC-APC/C association in dnt1Δ cells might be the inhibition of APC/C activity. Consistent with this assumption, we found that dnt1Δ enhanced the growth defects of some temperature-sensitive mutants of essential APC/C subunits, such as nuc2-663, cut9-234, cut20-100 and cut23-547, though dnt1Δ did not have any negative genetic interactions with two other APC/C mutants apc15Δ and slp1-mr63, both are defective in spindle checkpoint arrest [55, 57] (Fig 3A). Most strikingly, we observed that dnt1Δ rescued slp1-362 surprisingly well (Fig 3A). This specific genetic interaction suggested a possible role of Dnt1 in MCC-APC/C dissociation through regulating Slp1^Cdc20. To test this hypothesis, we examined the protein levels of the full-length Slp1^Cdc20 in nda3-KM311 dnt1⁺ and nda3-KM311 dnt1Δ cells after being released from metaphase-arrest. Surprisingly, Slp1^Cdc20 was slightly, but appreciably and reproducibly, less abundant (ranging from roughly 20% to 50% at different time points) in dnt1Δ cells than in wild-type cells (Fig 3B), suggesting Dnt1 may indeed positively regulate the levels of intact Slp1^Cdc20. In addition, this regulation of Slp1^Cdc20 stability by Dnt1 is Dma1-independent, as the dnt1Δ dma1Δ double mutant has a similar level and degradation profile of Slp1^Cdc20 as dnt1Δ single mutant (S5 Fig).

Download:

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

Fig 3. Dnt1 is required for maintaining the protein levels of Slp1^Cdc20 upon SAC activation.

(A) dnt1Δ rescues the temperature-sensitivity of slp1-362 but not other loss-of-function APC/C mutants. Serial dilutions (10-fold) of the indicated strains were spotted on YE plates and incubated at the indicated temperatures. Note that temperature-sensitivity of most loss-of-function APC/C mutants is exacerbated by dnt1Δ, but dnt1Δ rescues slp1-362. In addition, dnt1Δ does not have any negative genetic interactions with two other APC/C mutants, apc15Δ and slp1-mr63, both are defective in spindle checkpoint arrest. (B) Slp1^Cdc20 levels are slightly reduced in dnt1Δ cells compared to wild-type cells during metaphase arrest and SAC inactivation. Strains with indicated genotypes were grown, treated and synchronized as in Fig 2A. Mid-log phase samples at 30°C were also collected as asynchronous cultures (Asyn.). (Left) Samples were subjected to Western blot analyses using anti-Slp1 and anti-Cdc2 antibodies. (Right) Slp1^Cdc20 levels were normalized to those of total Cdc2 at each time point, with the relative ratio between Slp1^Cdc20 and Cdc2 at time 0 min in nda3-KM311 set as 1.0. (C) Schematic depiction of the genomic positions of endogenous slp1⁺ locus and ectopic slp1⁺ at lys1⁺ on chromosome I and at leu1⁺ on chromosome II. Red bars, two centromeres 1 and 2. (D) Western blot analyses of Slp1^Cdc20 protein levels in metaphase-arrested cells. Slp1^Cdc20 levels were normalized to those of total Cdc2 for each sample, with the relative ratio between Slp1^Cdc20 and Cdc2 at 0 min in nda3-KM311 without extra copies of slp1⁺ set as 1.0. Note that in the strain carrying both endogenous slp1⁺ and lys1Δ::slp1⁺, Slp1^Cdc20 level is close to 2.5 instead of 2 times of endogenous level, most likely due to its flanking adh1 terminator sequence. Error bars correspond to standard deviation. p values were calculated against the strain of nda3-KM311 without extra copies of slp1⁺. ***, p<0.001; **, p<0.01; n.s., not significant. (E) Elevated Slp1^Cdc20 levels rescue the TBZ sensitivity of dnt1Δ cells. (F) Elevated Slp1^Cdc20 levels reverse the anaphase initiation defect upon SAC inactivation in dnt1Δ cells. (G) Schematic summary of data from (B-F). The balance of relative abundance between Slp1^Cdc20 and checkpoint proteins governs the robustness of spindle checkpoint and thus mitotic exit rate. Lighter or darker shading depicts lower or higher levels of Slp1^Cdc20 present in MCC or APC/C in dnt1Δ or slp1⁺-overexpression (slp1^oe) cells.

https://doi.org/10.1371/journal.pgen.1010397.g003

Artificially increased Slp1 abundance mitigates TBZ sensitivity of dnt1Δ cells

Based on the above results, we wondered whether the decreased levels of Slp1^Cdc20 may be the major cause of the TBZ sensitivity and the defective anaphase initiation upon SAC inactivation in dnt1Δ cells. To test this possibility, we artificially increased Slp1^Cdc20 abundance by expressing one or two extra copies of slp1⁺ under its endogenous regulatory sequences (Fig 3C). Interestingly, two and three copies of slp1⁺ restored abundance of Slp1^Cdc20 in metaphase-arrested dnt1Δ cells to close to or slightly above the endogenous level in wild-type cells (Fig 3D). Consequently, sensitivity of dnt1Δ cells to TBZ was largely but not completely suppressed by excessive Slp1^Cdc20 expression achieved by three copies of slp1⁺ (Fig 3E). It should be noted that although the ectopic expression of leu1-32::slp1⁺ restored Slp1^Cdc20 in dnt1Δ cells almost to the same levels as in wild-type cells, it still could not rescue sensitivity of dnt1Δ cells to TBZ (Fig 3E). Nevertheless, increased Slp1^Cdc20 abundance compromised the checkpoint activation in both wild-type and dnt1Δ cells, as revealed by lower percentages of cells with Cdc13-GFP at SPBs and metaphase-arrest upon cold-shock treatment, and the extent of failed checkpoint response was largely proportional to Slp1^Cdc20 levels (Fig 3F and 3G). This result is consistent with previous reports in both fission and budding yeast that have underlined the importance of accurate relative abundance between checkpoint proteins and Cdc20/Slp1^Cdc20, which sets an important determinant of checkpoint robustness [58, 59]. However, we noticed that the disappearance rate of Cdc13-GFP from SPBs was not accelerated in either wild-type or dnt1Δ background strains when Slp1^Cdc20 abundance was artificially increased, regardless of the copy number of Slp1^Cdc20 present (Fig 3F), indicating that the SAC silencing tempo was not altered by Slp1^Cdc20 overexpression.

Together, these results suggested that the excessive Slp1 by three copies of slp1⁺ antagonizes the negative effects of dnt1 deletion on TBZ sensitivity and lowered Slp1^Cdc20 protein levels may only partly contribute to the TBZ sensitivity of dnt1Δ cells. The failure of complete rescue of the TBZ sensitivity of dnt1Δ cells by artificially increased Slp1^Cdc20 abundance is possibly due to the retained and prolonged MCC-APC/C association, which still poses inhibition on APC/C activity.

Enhanced MCC-APC/C association and lowered Slp1^Cdc20 abundance in dnt1Δ cells can be reversed by depletion of Apc15

It has been shown in both human and fission yeast cells that Apc15 mediates MCC binding to APC/C and is required for Cdc20/Slp1 autoubiquitylation and its turnover by APC/C [22,55,60,61]. Since our above data suggested that the TBZ sensitivity of dnt1Δ cells is likely caused by at least two aspects of defects in this mutant: one is the prolonged MCC-APC/C association, and the other is the lowered Slp1^Cdc20 protein level, we reasoned that the absence of Apc15 may reverse the positive effect of deletion of dnt1⁺ on MCC-APC/C association and its negative effect on Slp1^Cdc20 levels.

We first investigated how the absence of apc15⁺ affected the MCC-APC/C interaction in dnt1Δ cells. By immunoprecipitation of Apc4/Lid1-TAP, we found that the absence of Dnt1 enhanced MCC-APC/C interaction as more Mad2 and Mad3 were co-immunoprecipitated in dnt1Δ cells when compared to those in wild-type cells, whereas the deletion of apc15⁺ abolished the MCC-APC/C interaction both in wild-type and dnt1Δ cells (Fig 4A). Quite unexpectedly, although the total cellular level of Slp1^Cdc20 was significantly attenuated in dnt1Δ cells compared to that in wild-type cells (see Input in Fig 4A and 4B), we observed much more Slp1^Cdc20 was co-immunoprecipitated by Apc4/Lid1-TAP but significantly reduced in Mad3-GFP immunocomplexes in dnt1Δ cells (Fig 4A and 4B). Interestingly, apc15Δ restored Slp1^Cdc20 levels bound to MCC in dnt1Δ cells (Fig 4B). These observations seem to support the idea that the efficiency of Slp1^Cdc20 being co-immunoprecipitated by APC/C subunit or MCC component differs significantly. Also, these data suggested that the increased APC/C-associated Slp1^Cdc20 is likely “trapped” in the interface between MCC and APC/C and is unable to fulfill its function as the activator for APC/C.

Download:

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

Fig 4. apc15Δ reverses enhanced MCC-APC/C association and lowered Slp1^Cdc20 abundance in dnt1Δ cells.

(A) apc15Δ abolishes the elevated binding of MCC to APC/C in dnt1Δ cells. The association of Mad2, Mad3 and Slp1^Cdc20 to Apc4/Lid1 was assessed by immunoprecipitation of Apc4/Lid1-TAP in checkpoint-arrested cells as in Fig 2F. Note that more Mad2, Mad3 and Slp1^Cdc20 was co-immunoprecipitated in dnt1Δ cells compared to wild-type cells, although the amount of Slp1^Cdc20 was less abundant in dnt1Δ cells than that in wild-type cells. Results shown are the representative of three independent experiments. (B) apc15Δ restores the amount of Slp1^Cdc20 bound to MCC in dnt1Δ cells. The assembly of MCC was assessed by immunoprecipitation of Mad3-GFP in checkpoint-arrested cells as in (A). Note that less Slp1^Cdc20 was co-immunoprecipitated in dnt1Δ cells compared to wild-type cells. Results shown are the representative of three independent experiments. (C) apc15Δ restores the abundance of Slp1^Cdc20 in dnt1Δ cells. Strains with indicated genotypes were grown and treated as in (A) to enrich checkpoint-arrested cells. Slp1^Cdc20 levels were quantified with the relative ratio between Slp1^Cdc20 and Cdc2 in wild-type strain set as 1.0. The experiments were repeated 3 times and the mean value for each sample was calculated as in Fig 3D. Error bars correspond to standard deviation. ***, p<0.001; **, p<0.01; n.s., not significant. (D) apc15Δ suppresses the delayed APC/C activation defect in dnt1Δ cells. Cells were synchronized by HU and then arrested at 18°C for 6 hours before being released at 30°C. The percentage of cells with Cdc13-GFP on SPBs was assessed at each time point as in Fig 2C. (E) Schematic summary of data from (A-D). White wavy lines depict the absence of Apc15. Lighter or darker shading depicts lower or higher levels of Slp1^Cdc20 present in MCC or APC/C in mutant cells compared to wild-type cells.

https://doi.org/10.1371/journal.pgen.1010397.g004

Next, we examined whether depletion of Apc15 could rescue decreased Slp1^Cdc20 abundance in dnt1Δ cells. As expected, we indeed observed that the absence of apc15⁺ restored Slp1^Cdc20 to a level that was even higher than that in wild-type cells (Fig 4C). This quantitative data suggested that the attenuated Slp1^Cdc20 levels in dnt1Δ cells were most likely due to Apc15-faciltated degradation once Slp1^Cdc20 is incorporated in MCC.

Furthermore, the absence of Apc15 also compromised the spindle assembly checkpoint response to disruption of spindles in dnt1Δ cells due to the loss of MCC-APC/C association, as the percentages of cells with Cdc13-GFP at SPBs were sharply reduced from over 90% to only about 70% in apc15Δ cells after nda3-mediated SAC activation (Fig 4D). These results are consistent with previous report showing that the relatively higher amount of Slp1^Cdc20 can override the inhibitory effect of checkpoint proteins on APC/C activation [58]. Our results also suggested that Dnt1 is required for antagonizing the APC/C-mediated Slp1^Cdc20 degradation and maintaining Slp1^Cdc20 above its threshold level necessary for activating APC/C once the inhibitory signal from SAC is removed (Fig 4E).

Human CUEDC2 can partially rescue the TBZ sensitivity and spindle checkpoint inactivation defects of dnt1Δ cells

In humans, CUEDC2 mediates the release of APC/C^Cdc20 activity from Mad2 inhibition, and depletion of CUEDC2 causes a checkpoint-dependent delay of the metaphase-anaphase transition [27]. Since deletion of dnt1⁺ in fission yeast also causes a checkpoint-dependent delay in anaphase entry, we wondered whether human CUEDC2 and fission yeast Dnt1 might share similar functions. To test this possibility, we ectopically expressed human CUEDC2 in fission yeast cells, and examined whether it could rescue the TBZ sensitivity and anaphase initiation defects upon SAC inactivation in dnt1Δ cells. Interestingly, the expression of nuclear localized CUEDC2 [i.e. CUEDC2 tagged with two copies of SV40 nuclear localization sequence (2×NLS)] slightly restored their growth at low concentrations of TBZ and efficiency of anaphase entry after SAC inactivation (S7 Fig). Cytoplasmic CUEDC2 did not have the same effect (S7B Fig), which was likely due to the fact that fission yeast undergoes “closed” mitosis, during which APC/C and MCC function only inside the nucleoplasm. On the contrary, similarly expressed human p31^comet, another factor with verified function in timely spindle checkpoint silencing by promoting the release of Mad2 from MCC [32–35, 37], failed to exert a rescuing effect on TBZ sensitivity of dnt1Δ cells (S7B Fig). Interestingly, we found limited homology between N-terminal portion of Dnt1 and CUEDC2 (S8 Fig). These data raised a possibility that fission yeast Dnt1 might be a functional homologue of human CUEDC2 and they may share a similar function in releasing APC/C^Cdc20 from checkpoint inhibition during mitotic exit.

Dnt1 associates with APC/C upon SAC activation and during anaphase initiation

It has been previously shown that Cdk1-phosphorylated CUEDC2 directly binds to human Cdc20, and mediates the release of APC/C^Cdc20 activity from Mad2 inhibition [27]. To further dissect the possible mechanism that how Dnt1 promotes SAC inactivation, facilitates APC/C activation and maintains Slp1^Cdc20 stability, we examined whether Dnt1 also similarly associated with Slp1^Cdc20 or APC/C. By immunoprecipitation of sfGFP-Slp1 from SAC-activated cells arrested by nda3-KM311, we found limited amount of co-purified Dnt1 in addition to APC/C subunit Apc4/Lid1 and MCC component Mad2 (Fig 5A). Weak association between Dnt1 and APC/C was also detected in mitotically arrested but not asynchronized cultures by immunoprecipitation of Apc4/Lid1-TAP (Fig 5B and 5C). On the contrary, despite our intensive efforts, we were unable to detect the interaction between Mad2 or Mad3 and Dnt1 by co-immunoprecipitation (Fig 5D). These data suggested that very likely Dnt1 physically interacts more directly with APC/C but not MCC. Supporting this scenario, the interaction between sfGFP-Slp1 and Dnt1 was enhanced by depletion of Mad3 (Fig 5E), possibly because more APC/C assemblies devoid of MCC are available, thus Dnt1 gains better access to APC/C in mad3Δ cells. Unfortunately, we failed to detect the direct interaction between Dnt1 and APC/C subunits including Slp1^Cdc20 by yeast two-hybrid analyses (S6 Fig). These data suggested that possibly Dnt1 physically interacts with APC/C through a “mediating” factor, which remains to be identified in the future studies.

Download:

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

Fig 5. Dnt1 associates with APC/C upon SAC activation and during anaphase initiation.

(A) Slp1^Cdc20 co-immunoprecipitates Dnt1 in metaphase-arrested and mitosis-exiting cells. nda3-KM311 mad2-13myc strains simultaneously carrying either sfGFP-slp1 or dnt1-13myc or both were arrested in mitosis at 18°C and then released to 30°C, samples were collected at 0 and 40 min. Whole-cell extracts (input) were incubated with GFP-Trap beads and immunoprecipitated (IP) fractions were analyzed by immunoblotting. (B, C) Dnt1 can be co-immunoprecipitated by Apc4/Lid1 in metaphase-arrested but not asynchronous interphase cells. nda3-KM311 strains with indicated genotypes were either arrested at metaphase at 18°C or grown as asynchronous cultures at 30°C. Apc4/Lid1-GFP or Apc4/Lid1-HA was immunoprecipitated with GFP-Trap beads or anti-HA antibodies respectively and purified fractions were analyzed by immunoblotting. (D) Dnt1 cannot be co-immunoprecipitated by Mad2 and Mad3 in metaphase-arrested cells. nda3-KM311 strains with indicated genotypes were arrested at metaphase as in (B). Both Mad2-GFP and Mad3-GFP were immunoprecipitated with GFP-Trap beads and purified fractions were analyzed by immunoblotting. The around 120 kDa band corresponding to Dnt1-13myc was not detectable in co-immunoprecipitated sample (lane #3). Note that one strain carrying apc4/lid1-TAP but without dnt1-13myc served as a positive control, and anti-myc antibodies cross-reacted with Apc4/Lid1-TAP. (E) Mad3 imposes inhibitory effect on association between Slp1 and Dnt1. cdc25-22 strains expressing the indicated tagged proteins in the presence or absence of mad3⁺ were first arrested at 36°C for 3.5 hours and then released to 25°C, samples were collected at 60 min after release. Immunoprecipitations were performed as in (A). In Fig 5A–5E, all asterisks indicate the unspecific band recognized by anti-myc antibodies, and all results shown are the representative of three independent experiments.

https://doi.org/10.1371/journal.pgen.1010397.g005

Presence of Dnt1 is beneficial to maintaining APC/C activity and cell survival upon spindle checkpoint activation

Our above data suggested that the absence of Dnt1 causes prolonged and enhanced SAC activation. We wondered whether this defective SAC silencing is beneficial to cells. Very intriguingly, we noticed that individually isolated nda3-KM311 dnt1Δ cells formed far fewer colonies on solid rich medium at the permissive temperature 30°C than nda3-KM311 dnt1⁺ cells (47% vs. 68% viability) when they were treated at 18°C for 10 hours to activate SAC (Fig 6A). In contrast, the viability of dnt1Δ cells was not affected by this transient cold-shock (Fig 6A). This data indicated that Dnt1 is required to maintain cell viability specifically upon spindle stress, although it is largely dispensable for spindle checkpoint activation (S2 Fig).

Download:

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

Fig 6. Dnt1 involves in antagonizing SAC and maintaining cell survival upon spindle checkpoint activation.

(A) nda3-KM311 dnt1Δ cells lose viability after transient cold-shock. nda3-KM311 dnt1⁺, nda3-KM311 dnt1Δ or dnt1Δ cells were first grown at 30°C in liquid cultures and then being shifted to 18°C for 10 hours. Individual cells (n ≥80 for each strain) were isolated using tetrad dissection manipulator and placed on solid rich media at the permissive temperature of 30°C (Left). The number of colonies formed before and after cold-treatment was quantified (Right). (B) Overexpressed Dnt1 antagonizes negative effect of Mad2 overexpression on loss-of-function APC/C mutants. Serial dilutions (10-fold) of the indicated strains were spotted on YE plates and incubated at the indicated temperatures. Overexpression of Mad2 and Dnt1 was achieved under nmt1 promoter (P_nmt1) or adh1 promoter (P_adh1), respectively. Note that cut20-100 P_nmt1-mad2⁺ mutant was lethal and thus not included in the spot assay.

https://doi.org/10.1371/journal.pgen.1010397.g006

It has been previously shown that overexpression of spindle checkpoint protein Mad2 can activate SAC and block anaphase entry even in the absence of spindle defects [55,62], this is due to the fact that SAC proteins function as potent inhibitors of APC/C. Consistently, it is not surprising that fission yeast cells with temperature-sensitive mutations in the APC/C subunits were sensitive to elevated levels of Mad2 (P_nmt1-mad2⁺) [62] (Fig 6B). Given that Dnt1 is required to promote timely disassembly of MCC and cell survival, we examined whether elevated levels of Dnt1 can reverse the deleterious effects of overexpressed Mad2 on temperature-sensitive APC/C mutants. As expected, ectopic expression of dnt1⁺ under the control of adh1 promoter (P_adh1-dnt1⁺) rescued the growth defects of tested APC/C mutants overexpressing Mad2 (Fig 6B). Most strikingly, higher levels of Dnt1 allowed the survival of cut20-100 P_nmt1-mad2⁺ mutant, which is lethal by its own (Figs 6B and S9).

Together, these results revealed that Dnt1 is required for maintaining cell viability, and also for antagonizing SAC thus resulting timely and efficient anaphase initiation and mitotic exit, especially when the SAC needs to be inactivated.

Fission yeast strains and genetic analyses

Standard media (either YE (yeast extract) rich medium or EMM (Edinburgh minimal medium) and culturing methods were used [68,69]. G418 disulfate (Sigma-Aldrich), hygromycin B (Sangon Biotech) or nourseothricin (clonNAT; Werner BioAgents) was used at a final concentration of 100 μg/ml and thiabendazole (TBZ) (Sigma-Aldrich) at 5–15 μg/ml in YE media. For serial dilution spot assays, 10-fold dilutions of a mid-log-phase culture were plated on the indicated media and grown for 3 to 5 days at indicated temperatures. To examine the possible synthetic lethality of double mutants, at least 20 complete tetrads were dissected after each genetic cross. Yeast strains containing Cdc13-mCherry, Mad2-13myc and Mad3-13myc were generated by a PCR-based module method [70], with the DNA sequence information obtained from PomBase (https://www.pombase.org/). To create strains with an extra copy of slp1⁺ or dnt1⁺ at lys1 locus, the slp1⁺ genomic region from 1,504 bp 5’ to stop codon of the open reading frame or 1852 bp of the open reading frame of dnt1⁺ was first cloned into the vector pUC119-P_adh1-MCS-hphMX6-lys1* with adh21 promoter (P_adh21) maintained or removed using the ‘T-type’ enzyme-free cloning method [71]. Similar procedures were employed to construct pUC119-P_adh21-MCS-hphMX6-lys1*-based plasmids carrying sequences corresponding to human p31^comet or CUEDC2 (with plasmid or cDNA kindly provided by Hongtao Yu or Jiahuai Han respectively). Then, GFP or two tandem SV40 NLS or both sequences was introduced in front of dnt1⁺, p31^comet or CUEDC2 coding sequences by Quikgene method [72]. Finally, the resultant plasmids were linearized by ApaI and integrated into the lys1 locus, generating the strains lys1Δ::P_slp1-slp1⁺-T_slp1::hphMX6, lys1Δ::P_slp1-slp1⁺-T_adh1::hphMX6, lys1Δ::P_adh21-dnt1⁺-T_adh1::hphMX6, lys1Δ::P_adh21-GFP-dnt1⁺-T_adh1::hphMX6, lys1Δ::P_adh21-GFP-p31^comet-T_adh1::hphMX6, lys1Δ::P_adh21-2×NLS-p31^comet-T_adh1::hphMX6, lys1Δ::P_adh21-2×NLS-GFP-p31^comet-T_adh1::hphMX6, lys1Δ::P_adh21-GFP-CUEDC2-T_adh1::hphMX6, lys1Δ::P_adh21-2×NLS-CUEDC2-T_adh1::hphMX6, and lys1Δ::P_adh21-2×NLS-GFP-CUEDC2-T_adh1::hphMX6. The other strain with an extra copy of slp1⁺ at leu1 locus has been described previously [58]. Strains used in this study are listed in S1 Table.

Minichromosome loss assay

The minichromosome loss assay was performed in cells bearing the Ch16 (ade6-M216) minichromosome and ade6-M210 allele as previously described [54,73]. The rate of minichromosome loss per division was calculated by dividing the number of at least half red-sectored colonies grown on YE containing 12 mg/L adenine at 30°C by the number of total colonies (red colonies were excluded from the count).

Cell synchronization methods

For cdc25-22 strains, cells were grown at 25°C until mid-log phase and arrested at late G2 phase by shifting to 36°C for 3.5 hr and released at 25°C. For nda3-KM311 strains, cells were grown at 30°C to mid-log phase, synchronized at S phase by adding HU (Sangon Biotech) to a final concentration of 12 mM for 2 hr followed by a second dose of HU (6 mM final concentration) for 3.5 hr. HU was then washed out and cells were released at specific temperatures as required by subsequent experiments.

Spindle checkpoint activation assay

For metaphase arrest due to spindle checkpoint activation by disrupting mitotic spindles and abolishing kinetochore-microtubule attachment using the cold-sensitive nda3-KM311 mutation, cells synchronized by HU at 30°C were released to 18°C for up to 9 hr. For spindle checkpoint activation by the absence of tension generated by a temperature-sensitive mutation of a cohesin subunit, psc3-1T, cells were first arrested at S phase by HU at 25°C and then released to 37°C. For both methods, cells were withdrawn at certain time intervals and fixed with cold methanol and stained with DAPI (4’, 6-diamidino-2-phenylindole, Sigma-Aldrich). 200–300 cells were analyzed for each time point.

Spindle checkpoint silencing assay

For checkpoint silencing assay in the absence of microtubules, mid-log phase ark1-as3 cdc13-GFP nda3-KM311 cultures were first synchronized with HU at 30°C and then arrested in early mitosis by shifting to 18°C for 6 hr. 5μM ATP analog 1-NM-PP1 (Toronto Research Chemicals) was added to inactivate ark1-as3 and therefore spindle checkpoint. For checkpoint inactivation by shifting microtubule-depolymerized nda3-KM311 cells back to permissive temperature, cells were first synchronized with HU at 30°C and then arrested in early mitosis by shifting to 18°C for 6 hr, followed by incubation at 30°C to allow spindle reformation and therefore spindle checkpoint silencing. For both methods, cells were withdrawn at certain time intervals and fixed with cold methanol and stained with DAPI. 200–300 cells were analyzed for each time point. Each experiment was repeated at least three times.

Yeast two-hybrid assay

For yeast two-hybrid analysis, the Matchmaker system (Clontech) was used. Bait plasmids were constructed into pGBKT7 vector. Prey plasmids were constructed into pGADT7 vector. Bait and prey plasmids were co-transformed into the AH109 strain, and transformants were selected on double dropout medium (SD-Leu-Trp). The bait-prey interaction, which would activate the HIS3 and ADE2 reporter genes, was assessed by the growth on the triple (SD-Leu-Trp-His) or quadruple (SD-Leu-Trp-Ade-His) dropout media.

Immunoblotting, immunoprecipitation and antibodies

Western blot and immunoprecipitation experiments were performed essentially as previously described [44]. Proteins were immunoprecipitated by IgG Sepharose beads (GE Healthcare) (for Apc4/Lid1-TAP), GFP-Trap beads (ChromoTek) (for sfGFP-Slp1 and Apc4/Lid1-GFP) or anti-HA antibody-coupled protein A/G beads (for Apc4/Lid1-HA). When necessary, 300 mM instead of 150 mM NaCl was used in lysate buffer to remove unspecific binding of proteins to beads. The primary antibodies used for immunoblot analysis of cell lysates and immunoprecipitates were peroxidase-anti-peroxidase (PAP) soluble complex (Sigma-Aldrich), rabbit polyclonal anti-Myc (GeneScript), mouse monoclonal anti-GFP (Beijing Ray Antibody Biotech), rat monoclonal anti-HA (Roche), mouse monoclonal anti-Cdc13 antibodies (Novus Biologicals) or rabbit polyclonal anti-Slp1 (generated at Xiamen University antibody facility using recombinant N-terminal 290 amino acids region of Slp1 (6His-Slp1(1-290aa) as antigens, same as previously described [57]). Cdc2 was detected using rabbit polyclonal anti-PSTAIRE (sc-53, Santa Cruz Biotechnology) as loading controls. Secondary antibodies were anti-mouse or anti-rabbit HRP conjugates (Thermo Fisher Scientific) and were read out using chemiluminescence.

Fluorescence microscopy and live-cell imaging

GFP- and mCherry-fusion proteins (such as Cdc13-GFP, Cut2-GFP, Dnt1-GFP, CUEDC2-GFP, Cdc13-mCherry, and mCherry-Atb2) were observed in cells after fixation with cold methanol. Cells were washed in PBS and resuspended in PBS plus 1 μg/ml DAPI. Photomicrographs were obtained using a Nikon 80i fluorescence microscope coupled to a cooled CCD camera (Hamamatsu, ORCA-ER). Time-lapse imaging of live cells was performed at 30°C using a Perkin Elmer spinning-disk confocal microscope (UltraVIEW VoX) with a 100x NA 1.49 TIRF oil immersion objective (Nikon) coupled to a cooled CCD camera (9100–50 EMCCD; Hamamatsu Photonics) and spinning disk head (CSU-X1, Yokogawa). Image processing, analysis and spindle length measurement were carried out using Element software (Nikon), ImageJ software (National Institutes of Health) and Adobe Photoshop.

Statistical analysis

For quantitative analyses of each experiment, at least 200 cells were counted for each time point or sample, and each experiment was conducted at least three times. Error bars correspond to standard deviation (SD) throughout.