Nup132 modulates meiotic spindle attachment in fission yeast by regulating kinetochore assembly

Meiosis is an important process for sexually reproducing eukaryotes and generates inheritable haploid gametes from a parental diploid cell. Through meiosis, genome ploidy is halved via two consecutive rounds of chromosome segregation after a single round of DNA replication. The first meiotic division is an uncommon chromosome segregation step in which homologous chromosomes are segregated but sister chromatids remain attached. To maintain the attachment between sister chromatids during meiosis I, the sister centromeres are rearranged into a meiosis-specific mono-oriented structure so that the sister centromeres are attached to microtubules that originate from the same poles (monopolar spindle attachment; Hauf and Watanabe, 2004; Parra et al., 2004). In contrast, the second meiotic division, like mitosis, segregates sister chromatids. In meiosis II, the sister centromeres bind to microtubules that are derived from opposite spindle poles (bipolar spindle attachment). Understanding the mechanisms underlying faithful segregation of chromosomes is clinically important because chromosome missegregation during meiosis is a major cause of human miscarriage and trisomy disorders.

Studies in the fission yeast Schizosaccharomyces pombe have revealed that establishment of the mono-oriented centromere in meiosis I requires the meiosis-specific cohesin Rec8 and that monopolar spindle attachment is supported by chiasmata formed between the homologous chromosomes (Watanabe and Nurse, 1999; Molnar et al., 2001a,,b; Yamamoto and Hiraoka, 2003; Hirose et al., 2011). During meiotic prophase, Rec8 holds sister chromatids together along the chromosome arms and centromeric regions (Watanabe and Nurse, 1999; Ding et al., 2006). Homologous chromosome recombination that is initiated by programmed meiotic double-strand breaks (DSBs) results in chiasma formation and physically connects the two homologues. In cooperation with the meiosis-specific centromere protein Moa1, Rec8 mono-orients the sister centromeres (Watanabe and Nurse, 1999; Yokobayashi and Watanabe, 2005). The pulling force that the spindles exert on the mono-oriented sister centromeres is stabilized by the tension generated via chiasmata formation between the homologues (Hauf and Watanabe, 2004; Parra et al., 2004; Hirose et al., 2011). Once the tension is sensed, cohesion of the chromosome arms is lost because of separase cleavage of Rec8, which allows homologous chromosome separation (Watanabe and Nurse, 1999; Molnar et al., 2001a,,b; Kitajima et al., 2003; Yamamoto and Hiraoka, 2003; Hirose et al., 2011). S. pombe shugoshin (Sgo1) protects centromeric Rec8 from cleavage, which keeps the sister chromatids together at anaphase I (Watanabe and Nurse, 1999; Kitajima et al., 2004; Ding et al., 2006). In the rec8Δ and recombination mutants, the frequency of untimely sister chromatid segregation in meiosis I is increased (Watanabe and Nurse, 1999; Molnar et al., 2001a; Yamamoto and Hiraoka, 2003; Yokobayashi and Watanabe, 2005). When the tension at the centromeres is lost, such as in the achiasmate rec12 mutant, the spindle assembly checkpoint (SAC) is activated, which halts cells in metaphase I (Yamamoto et al., 2008). Components of the SAC are localized to the centromeres during metaphase and function as a surveillance system to ensure that the centromeres are attached to the microtubules under the appropriate tension (Uchida et al., 2009). Activation of the SAC inhibits the activity of the anaphase-promoting complex (APC) to prevent anaphase from proceeding until the tension at the centromeres is generated (Kitajima et al., 2003).

The kinetochore at the centromere plays an important role in microtubule attachment. In addition to Rec8-mediated mono-orientation of centromeres, reorganization of the outer kinetochore components, known as the KNL1–Spc7-Mis12-Nuf2 (KMN) proteins, is also thought to assist in monopolar spindle attachment to the centromeres at meiosis I (Asakawa et al., 2005; Hayashi et al., 2006). The KMN complex functions as a molecular mechanical sensor that monitors microtubule-kinetochore attachment (Rago and Cheeseman, 2013). The KNL1/Spc7 family proteins are required for recruiting the SAC components as well as the factors that activate or deactivate SAC (Desai et al., 2003; Espeut et al., 2012; Shepperd et al., 2012). The Nuf2 complex binds directly to microtubules (Cheeseman et al., 2006), and the Mis12 complex links the Nuf2 complex to the inner kinetochores (Przewloka et al., 2011; Screpanti et al., 2011). The stable association between the Nuf2 complex and microtubules silences SAC activity. Although the KMN proteins are constitutive components of the outer kinetochores throughout the mitotic cell cycle, they transiently dissociate from the meiotic kinetochores during meiotic prophase in S. pombe (Hayashi et al., 2006). During the mitotic cell cycle, the centromeres cluster at the spindle pole body (SPB; a centrosome-equivalent structure in yeast), whereas the telomeres are located away from the SPB (Hou et al., 2012). In contrast, during meiotic prophase, the telomeres are brought to the SPB (Chikashige et al., 1994, 2006). At the same time, the centromeres detach from the SPB because of disassembly of the KMN proteins from the centromeres (Chikashige et al., 1997; Asakawa et al., 2005; Hayashi et al., 2006). Interestingly, when mating pheromone is absent, the KMN proteins persist and localize at the centromeres during meiotic prophase, and precocious sister chromatid segregation at meiosis I is often observed (Yamamoto and Hiraoka, 2003; Chikashige et al., 2004; Asakawa et al., 2005; Hayashi et al., 2006). This observation suggested that mating pheromone signaling, probably through transient dissociation of the KMN proteins, promotes kinetochore reorganization in favor of monopolar spindle attachment of sister chromatids. However, loading of the Rec8 protector Sgo1 onto the centromeres is also mating pheromone-signaling dependent (Hayashi et al., 2006); thus, it is unclear whether the KMN proteins regulate monopolar sister kinetochores through Sgo1.

S. pombe is an excellent system for studying meiosis because, in this organism, meiosis is easily induced and takes only 6–8 h to complete, and the entire process can be followed under a microscope. Normal yeast meiosis yields four spores (yeast gametes) in an ascus. Errors during meiotic progression result in an ascus with an abnormal number of spores, as is seen in rec8 and moa1 deletion mutants and recombination mutants (Watanabe and Nurse, 1999; Molnar et al., 2001b; Davis and Smith, 2003; Yokobayashi and Watanabe, 2005). In our attempt to identify spore formation mutants by deleting genes encoding nonessential nucleoporins, we identified nup132⁺: loss of Nup132 results in an aberrant number of spores after meiosis (Asakawa et al., 2014). Nup132, the S. pombe homologue of mammalian Nup133, is a component of the Nup107-Nup160 subcomplex that constitutes the core structure of the nuclear pore complex (Baï et al., 2004; Asakawa et al., 2014). In this study, we used a live-cell imaging system to monitor the meiotic progression of a nup132Δ mutant and found that Nup132 modulates monopolar spindle attachment of sister kinetochores at metaphase I by regulating outer kinetochore assembly in the preceding meiotic prophase.

To determine what causes the aberrant numbers of spores observed in the nup132Δ mutant, we first monitored meiotic progression in nup132Δ cells coexpressing fluorescently tagged tubulin and histones. Tubulin forms spindles during nuclear division, and therefore the duration from spindle formation to depolymerization was determined. As shown in Fig. 1 A, in the wild-type cell, the durations of meiosis I and meiosis II were 36 min each (0–36 min and 54–78 min, respectively). In contrast, nup132Δ cells frequently spent more time at each meiotic division (Fig. 1 B; 0–54 min at meiosis I and 72–126 min at meiosis II). This prolonged nuclear division is meiosis specific because the duration of mitotic division in the nup132Δ mutant (29 ± 3 min) was similar to that of the wild type (31 ± 3 min). By plotting spindle length over time, we identified the three phases of nuclear divisions as previously stated (Nabeshima et al., 1998; Zimmerman et al., 2004): First, spindles attach to the kinetochores and grow stably during prometaphase. Then, they are maintained at a constant length during metaphase, and, finally, they elongate at anaphase (Fig. 1 C). It appeared that in most nup132Δ cells, prometaphase was not affected but the duration of the constant spindle length phase (i.e., metaphase) was longer than that in the wild-type cells (Fig. 1, B and C). In the nup132Δ mutant, chromosomes, as observed by histone H3-mRFP, could often segregate during the prolonged meiosis I; however, chromosome segregation at meiosis II sometimes failed (~25% of the cells; Fig. 1 B, arrows). This explains a possible cause for the previously reported nontetrad formation, which also occurs at ∼25% in this mutant; however, such nontetrad spores are viable (Asakawa et al., 2014; see Fig. 9).

The extended meiotic metaphase observed in the nup132Δ mutant infers activation of the SAC. Cells with a single deletion of one of the genes encoding the SAC components Mad2 or Bub1 underwent meiosis I at a similar time interval as wild-type cells (Fig. 1 D). In contrast, deletion of mad2⁺ or bub1⁺ in the nup132Δ background significantly shortened the prolonged meiosis I observed in the nup132Δ single mutant. However, prolonged meiosis II was not dependent on the SAC (Fig. 1 D). We reasoned that in the nup132Δ mutant, activated SAC inhibits the activity of the APC and delays the transition from metaphase I to anaphase I. To verify this, activity of the APC was monitored by using GFP-labeled Cut2/securin because it is known to localize to spindles and is abruptly degraded by the APC at the metaphase-anaphase transition in S. pombe (Funabiki et al., 1996). Disappearance of Cut2/securin from the metaphase I spindles was significantly delayed when nup132⁺ was depleted (Fig. S1). This result suggests that the activated SAC causes the metaphase I extension observed in the nup132Δ mutant.

S. pombe undergoes closed nuclear divisions with the intact nuclear envelope. As nup132⁺ encodes a scaffold protein of the nuclear pore complex, it is possible that depletion of nup132⁺ disrupts the nucleocytoplasmic barrier and broadly affects the factors that regulate meiotic progression. To test this, 3GFP-NLS was used as a nuclear marker to monitor the nucleocytoplasmic barrier. In a wild-type cell, 3GFP-NLS remains in the nucleus throughout meiosis (Fig. S2 A), except for transient dispersion to the cytoplasm in anaphase II (referred to as virtual nuclear envelope breakdown in Asakawa et al., 2010). In the nup132Δ mutant, the nucleocytoplasmic barrier was intact during the prolonged period of metaphase I although leakage of the 3GFP-NLS signals was observed at the onset of anaphase I and during interkinesis (Fig. S2, B and C). Therefore, based on the results of the nuclear reporter assay, there is no evidence that unregulated nucleocytoplasmic transport leads to prolongation of metaphase I in the nup132Δ mutant.

SAC is activated when the tension at sister kinetochores is absent. To analyze tension at the sister kinetochores, mCherry-Atb2 and cen2-lacO/lacI-GFP (called cen2-GFP hereafter) were introduced into a heterothallic strain of a nup132Δ mutant. The resultant strain was then mated with GFP-negative heterothallic partners lacking cen2-GFP. A pair of chromosome II sister chromatids was labeled with cen2-GFP and distinguished from the unlabeled homologous chromosomes. The sister cen2-GFP quickly moved back and forth on the spindles at prometaphase I in both the wild-type and nup132Δ mutant cells (Fig. 2, A and B, arrowheads). However, the sister cen2-GFP of most wild-type cells quickly settled at one end of the spindles as meiosis I proceeded (Fig. 2, A and C), suggesting that after tension resulting from chiasmata and monopolar spindle attachment was satisfied at the sister kinetochores, the sister centromeres soon moved together to the same spindle pole. On the contrary, the sister cen2-GFP of the nup132Δ cells often wobbled along the spindle axis during the prolonged meiosis I (Fig. 2, B [arrows] and C), possibly reflecting unsatisfied tension at the sister kinetochores. These results also suggest that the sister centromeres of the nup132Δ mutant experienced pulling forces from the opposite spindle poles.

The absence of chiasmata could result in loss of tension at sister kinetochores and activate SAC (Yamamoto et al., 2008). This prompted us to examine whether homologous chromosome recombination occurs in the nup132Δ mutant; this was done using Rhp51-ECFP as a DSB repair marker (Akamatsu et al., 2007). Both the nup132Δ mutant and wild-type cells accumulated bright punctate signals of Rhp51-ECFP in the nuclei during meiotic prophase, indicative of repairing of DSB (Fig. 3, A and B). As a negative control, Rhp51-ECFP appeared to be uniformly distributed in the nucleus of the rec12Δ mutant, which does not initiate meiotic DSBs (Fig. 3 C). The punctate signals were greatly reduced by the time of telomere dispersal (i.e., at the end of meiotic prophase) and totally disappeared at anaphase I (Fig. 3, A and B), suggesting that the DSBs were repaired by homologous chromosome recombination. In agreement with this, recombinant gametes were frequently observed in the nup132Δ mutant, although the recombination rate was slightly decreased compared with that in the wild-type strain (Fig. 3 D). These results demonstrate that activation of SAC in the nup132Δ mutant is not due to a lack of chiasmata.

We then investigated whether the wobbled movement of the sister cen2-GFP in meiosis I resulted from a premature loss of the meiotic cohesin Rec8. Localization of Rec8-GFP in the mutant was examined. In both wild-type and nup132Δ cells, Rec8-GFP localized along the chromosomes during meiotic prophase and metaphase and then suddenly disappeared from the chromosome arms, but it remained at the centromeres at anaphase I (Fig. 4, A and B). The centromeric localization of the Rec8 protector Sgo1 and the Rec8-interacting protein Moa1 was also not affected in the nup132Δ mutants (Fig. S3). To further confirm this result, we performed a cohesin assay in cells with the mes1Δ homothallic strain background expressing cen2-GFP. mes1⁺ was deleted to arrest the cells before the onset of meiosis II (Kitajima et al., 2004). The bub1Δmes1Δ mutant was used as a positive control for the assay because the centromere localization of Rec8 precociously disappears in meiosis I in the absence of Bub1 (Kitajima et al., 2004). Two separate dots were frequently observed in the divided meiosis I nuclei of the bub1Δmes1Δ mutant, indicating precocious splitting of the sister cen2 caused by the loss of Rec8 in meiosis I. The sister cen2-GFP remained close and appeared as a single dot or as a pair of neighboring dots in the divided meiotic nuclei of the mes1Δ or nup132Δmes1Δ cells (Fig. 4, C and D). These results indicate that, in the absence of nup132⁺, Rec8 remains intact in the centromere region in meiosis I.

Reorganization of the meiotic kinetochores has been implicated in the regulation of monopolar spindle attachment. The KMN proteins in the outer kinetochores are disassembled at early meiotic prophase and are reassembled before the onset of metaphase I (Hayashi et al., 2006). Therefore, we examined the meiotic behaviors of the KMN components Mis12 and Spc7 in the nup132Δ mutant. GFP-tagged Mis12 or Spc7 was introduced into cells carrying Cut11-mCherry. Cut11 is a nucleoporin protein that localizes to the nuclear envelope throughout the cell cycle and to the SPB during metaphase (West et al., 1998). Thus, it was used to define the nuclear region and determine the timing of meiotic progression (i.e., karyogamy, meiotic prophase, and prometaphase). As previously reported (Hayashi et al., 2006), in wild-type cells, the Mis12-GFP and Spc7-GFP signals disappeared from the centromeres upon karyogamy and did not reappear until meiosis I onset (Fig. 5, A and B). Remarkably, 53% of nup132Δ cells had Mis12-GFP signals (Fig. 5, C and E), and 77% of nup132Δ cells had Spc7-GFP signals in the nuclei at early meiotic prophase (Fig. 5, D and F). The times of the first appearances of Mis12-GFP and Spc7-GFP signals are plotted in Fig. 5, E and F. These GFP foci of Mis12 or Spc7 localized to one of the centromeres, which were visualized by the mCherry-tagged constitutive inner-kinetochore protein Mis6, but not to all of the centromeres (Fig. 6). These results suggest that Mis12 and Spc7 precociously assembled to, or never disassembled from, some of the centromeres in the absence of Nup132.

The phenotype of precocious meiotic kinetochore assembly is unique to the nup132Δ mutant. Deletion of any of the five genes encoding other nonessential nucleoporins, ely5⁺, nup37⁺, pom152⁺, nup124⁺, and nup61⁺, which when deleted give rise to abnormal spore numbers (Asakawa et al., 2014), did not affect Mis12-GFP dissociation/association at the centromeres. Moreover, neither rec8Δ nor rec12Δ displays the phenotype of precocious Mis12 reappearance at the centromeres (Fig. S4). Therefore, Nup132 regulates meiotic outer kinetochore reorganization independent of the Rec8-mediated pathway or chiasmata formation.

Another KMN component, Nuf2, which connects the centromeres to the SPB, disappears at the early meiotic prophase and releases the centromeres from the SPB (Asakawa et al., 2005; Hayashi et al., 2006). In wild-type cells, similar to Mis12-GFP and Spc7-GFP, the Nuf2-GFP signals disappeared at early meiotic prophase and reappeared at the later stage of meiotic prophase I (∼40 min before metaphase I onset; Fig. 7 A). In the nup132Δ mutant, the meiotic behavior of Nuf2-GFP was normal, similar to that observed in the wild-type strain (Fig. 7, B and C). Ndc80, the Nuf2 interacting protein, also behaved normally in the mutant (Fig. 7 C). In addition, the Csi1 protein, which is a kinetochore-SPB connector in mitotic interphase (Hou et al., 2012), disappeared at early meiotic prophase (Fig. S5). Csi1-mCherry signals reappeared in late meiotic prophase before telomere dispersal and colocalized with the telomere marker Taz1-GFP (Fig. S5). Because telomeres are clustered at the SPB during meiotic prophase (Chikashige et al., 1997), colocalization of Csi1-mCherry and Taz1-GFP suggests that Csi1 reassembled to the SPB during meiotic prophase. The meiotic behavior of Csi1 in the nup132Δ mutant was the same as that observed in the wild-type strain (Fig. 7 F). Thus, in the mutant, both Nuf2 and Csi1 are properly disassembled early in meiotic prophase and do not reassemble until late in meiotic prophase. Consistent with the fact that Nuf2 and Csi1 are required for the centromere-SPB connection, the centromeric foci of Mis12-GFP or Spc7-GFP did not colocalize with the SPB component Sfi1 or the telomere-bound protein Taz1 (Fig. 7, D and E), indicating that centromeres partially preassembled with Mis12 and Spc7 are not connected to the SPB in early meiotic prophase when Nuf2 and Csi1 are not reassembled at the centromeres.

Spc7 directly interacts with the SAC component Bub1 (Shepperd et al., 2012). Because Spc7 is precociously loaded onto the centromeres in the nup132Δ mutant, we wondered if Bub1 is also precociously recruited. Similar to the KMN proteins, Bub1-GFP signals appeared as foci in the nucleus shortly before the onset of metaphase I in the wild-type cells (Fig. 8 A), suggesting that Bub1 is recruited to the kinetochores when the KMN proteins reassemble. In the nup132Δ mutant, the Bub1-GFP signals appeared at early meiotic prophase (Fig. 8, B and C), indicating that Bub1 is precociously recruited to the centromeres. Nevertheless, Sgo1-GFP signals appeared at the normal time in the nup132Δ mutant (Fig. 8 D). These results suggest that the early recruitment of Bub1 in the nup132Δ mutant is not sufficient to recruit Sgo1 to the preassembled kinetochores and that the defective meiotic kinetochore in nup132Δ forms without affecting Sgo1.

We reasoned that precocious assembly of kinetochores might affect subsequent monopolar microtubule-kinetochore attachment. To determine whether the deletion of nup132⁺ affected sister chromatid segregation in meiosis I, the segregation pattern of the sister cen2-GFP in meiosis I was determined using live-cell imaging. Coexpression of Cut11-mCherry was used to outline the meiosis I nuclei. Normally, when monopolar spindle attachment of sister chromatids is established in meiosis I, the sister cen2-GFP cosegregates to the same nucleus (Fig. 9 A, top and wild type). In contrast, if bipolar spindle attachment of sister chromatids is established, precocious sister chromatid segregation (i.e., equational segregation) may occur. As shown in Fig. 9 A, the sister cen2-GFP of a nup132Δ or a mad2Δ single mutant cosegregated to the same nucleus with high fidelity. Nevertheless, the frequency of equational segregation in meiosis I increased to 7% in the nup132Δmad2Δ double mutant. This result suggests that erroneous microtubule-kinetochore attachments occurred in the nup132Δ mutant but that the erroneous attachments can be corrected in the presence of SAC.

In addition to SAC, sister centromere coherence prevents equational segregation of the sister chromatids in meiosis I. We therefore examined the segregation pattern of sister cen2-GFP in the sgo1Δ background, in which sister centromere cohesion is compromised (Kitajima et al., 2004). As previously reported (Kitajima et al., 2004), sgo1⁺ deletion alone subtly disturbed the segregation pattern of the sister cen2-GFP. Noticeably, for the deletion of mad2⁺ in the sgo1Δ background, the frequency of equational separation increased to 15% (Fig. 9 A), confirming that both SAC and sister centromere coherence contribute to the cosegregation of sister chromatids in meiosis I. Consistent with this, the depletion of Bub1, which functions upstream of Mad2 and Sgo1, also led to increased frequency of equational segregation, similar to that of the mad2Δsgo1Δ double mutant (Fig. 9 A).

Deletion of nup132⁺ did not exacerbate the level of equational segregation of the sister cen2-GFP in the sgo1Δ or mad2Δsgo1Δ background. However, the deletion of nup132⁺ in the bub1Δ background drastically increased the frequency of equational segregation in meiosis I, and the sister chromatids were randomly segregated (i.e., ∼50% cosegregation and 50% equational segregation; Fig. 9). In addition to functioning in SAC and Sgo1 recruitment, bub1⁺ appears to act synergistically with nup132⁺ to modulate meiotic microtubule-kinetochore attachment (see Discussion).

In this study, we performed live-cell imaging analysis of meiosis in the S. pombe nup132Δ mutant and showed that depletion of Nup132 increased the duration of the meiotic divisions. Meiosis I is prolonged as a result of SAC activation, whereas meiosis II prolongation is not SAC dependent. At least two possible reasons could account for the prolonged meiosis II observed in the nup132Δ mutant: first, the second meiotic divisions are impeded by errors that result from meiosis I chromosome segregation; and second, the untimely breakage of the nucleocytoplasmic barrier during interkinesis interferes with the normal progression of meiosis II (Fig. S2). Currently, we cannot determine whether either of these is the major cause of meiosis II nuclear division failure that leads to nontetrad formation in the nup132Δ mutant.

The SAC-dependent extension of meiosis I observed in the mutant suggests increased erroneous microtubule-kinetochore attachments. In support of this notion, when SAC is absent, as in the bub1Δ or mad2Δ background, the depletion of Nup132 could lead to equational segregation of sister chromatids in meiosis I. This equational segregation was not a result of compromised sister centromere coherence because precocious splitting of cen2-GFP was rarely observed in the nup132Δ mutant.

Strikingly, in the nup132Δbub1Δ mutant, the sister chromatids segregated almost randomly in meiosis I, which was in contrast to the relatively mild phenotype observed in the nup132Δmad2Δ mutant. The differences were not simply caused by defects in Sgo1 recruitment to the centromeres in the bub1Δ mutant because, unlike the nup132Δbub1Δ mutant, the triple mutant of nup132Δmad2Δsgo1Δ did not show random segregation of the sister chromatids. Yet, the meiosis I equational segregation in bub1Δ is thought to be defective in establishing an integrated sister kinetochore for sister chromatids (Bernard et al., 2001). It is possible that, in the nup132Δbub1Δ mutant, the sister kinetochores were not properly formed, leading to random sister chromatid segregation in meiosis I. This might reflect the untimely meiotic kinetochore assembly observed in the nup132Δ mutant.

Failure to disassemble the KMN complex at the kinetochores could result in precocious sister chromatid segregation (Chikashige et al., 2004; Hayashi et al., 2006). Consistently, we observed precocious assembly of the KMN proteins Mis12 and Spc7 at the meiotic centromeres in the nup132Δ mutant. Interestingly, our results showed that Rec8 and Rec12 are not involved in outer kinetochore reorganization during meiotic prophase and that depletion of Nup132 neither disturbs the centromere localization of Rec8 nor affects chiasmata formation. Therefore, we propose that independent of centromeric cohesin and the recombination pathway, Nup132 regulates the assembly of the KMN complex at early meiotic prophase to modulate monopolar spindle attachment in meiosis I.

Upon mating pheromone signaling, Nuf2, Mis12, and Spc7 disassemble from the centromeres and then return to the centromeres late in meiotic prophase (Hayashi et al., 2006). In the absence of Nup132, whereas Nuf2 remains dissociated from the centromeres until late meiotic prophase, Mis12 and Spc7 often appear at the centromeres early in meiotic prophase, and these partially assembled centromeres remain separated from the SPB. In this scenario, Nuf2 dissociation releases the centromeres from the SPB, and Nup132 prevents Mis12 and Spc7 from precociously assembling to the centromeres in early meiotic prophase. Bub1 is recruited early to the centromeres partially assembled with Mis12 and Spc7. Thus, Mis12 and Spc7 likely provide an interaction hub for the recruitment of SAC components. However, the timing of Sgo1 appearance was not affected by the absence of Nup132. Hence, it is likely that reorganization of the meiotic outer kinetochore modulates monopolar spindle attachment independent of Sgo1.

The metazoan homologues of Nup132 localize to the NPCs during interphase and were enriched at the kinetochores during mitosis (Belgareh et al., 2001; Mishra et al., 2010; Ródenas et al., 2012). In contrast, S. pombe Nup132 localizes to the NPCs throughout mitosis and meiosis, and there is no evidence of kinetochore localization (Asakawa et al., 2010, 2014). How Nup132 regulates outer kinetochore disassembly without localizing at the kinetochores is not yet known. One possible explanation is that Nup132 mediates the localization of the kinetochore proteins by sequestering them at the NPCs. Interestingly, it has been reported that the human homologue of Nup132, hNup133, interacts with the outer kinetochore protein CENP-F at the nuclear envelope (Zuccolo et al., 2007; Bolhy et al., 2011). Although there is no obvious CENP-F homologue in S. pombe, we noted a distant homology between human CENP-F and S. pombe Spo15 using a DELTA-BLAST search (Boratyn et al., 2012). Spo15 is required for spore formation and localizes to the SPB throughout mitosis and most of meiosis (Ohta et al., 2012). Thus, Spo15 can be a candidate for a functional homologue of CENP-F that participates in outer kinetochore reorganization during meiotic prophase in S. pombe. However, this remains to be tested.

The nup132Δ mutant exhibits different effects in mitosis and meiosis (i.e., normal progression of mitosis but delays in the metaphase-anaphase transition in meiosis). This may result from differences in the dynamic nature of the outer kinetochores between mitosis and meiosis in S. pombe: Although the outer kinetochore dynamically disassembles and reassembles during meiotic prophase, it is a stable structure that constantly associates with the SPB during the mitotic cell cycle (Asakawa et al., 2007). The steady assembly of the outer kinetochore during the mitotic cell cycle diminishes the need for Nup132. This idea is also supported by the fact that mitotic delays were observed when hNup133 was depleted from the kinetochores in human cells (Zuccolo et al., 2007) as the human outer kinetochore is dynamically assembled and disassembled during mitosis (Gascoigne and Cheeseman, 2011). Because delays in the metaphase-anaphase transition occur in both human mitosis and S. pombe meiosis in the absence of hNup133/Nup132, human hNup133 and S. pombe Nup132 may play a similar, conserved role.

The S. pombe strains used in this study are listed in Table S1. The culture media used here have been described previously (Moreno et al., 1991). All strains were grown on yeast extract with supplements (YES) plates or Edinburgh minimal medium (EMM) with the appropriate supplements at 30°C. Sporulation was induced on malt extract (ME) plates at 26°C. To induce meiosis entry, cells were freshly inoculated on a YES plate and were resuspended in nitrogen-free minimal medium supplemented with adenine, uracil, histidine, lysine, and leucine (EMM-N+5S) at a density of 10⁹ cells/ml. The cell suspension was then spotted onto an ME plate. Heterothallic strains were premixed with EMM-N+5S medium before plating on the ME plate. After 8–10 h of incubation on the ME plates, the cells that had undergone karyogamy were selected for live-cell imaging. The sporulation frequency was determined after the cells were sporulated for 2 d on the ME plates.

The nup132⁺ gene was disrupted with the ura4⁺ gene by using the template plasmid of pCSU3 and the primers NM4-UPF (5′-ATGAAAAATAGCTTTCCGATTCGGC-3′), NM4-UPR (5′-CCCACAGTTCTAGAGGATCCGGTCAAGCTTAAACTACTTT-3′), NM4-DWF (5′-GCCTTAACGACGTAGTCGACTTTATCTTAATCATACAAAC-3′), and NM4-DWR (5′-CCGAGGCAGCCAACACTGTACTTGG-3′; Chikashige et al., 2006; Asakawa et al., 2014). The underlined portions are the sequences shared by the pCSU3 plasmid. In cases where the ura4⁺ marker was not available, nup132⁺ was deleted with a kanMX6 marker or a natMX6 marker (Krawchuk and Wahls, 1999; Hentges et al., 2005). The nup132⁺ gene was disrupted with a drug resistance gene cassette (kanMX6 or natMX6) by using pFA6a derivatives and the primers nup132D-1 (5′-GAAATCTGATGTTTCCAACC-3′), nup132D-R1 (5′-GAGGCAAGCTAAACAGATCTGACTATTTGACGATATCAGT-3′), nup132D-F2 (5′-GTTTAAACGAGCTCGAATTCTAACCTTTATCTTAATCATA-3′), and nup132D-2 (5′-GTTCATTACCGCGTTGG-3′). The underlined portions are the sequences shared by the pFA6a plasmid. The nup132⁺ gene disruption was confirmed by PCR, and the phenotype of nontetrad formation or unevenly distributed nucleoporins (Baï et al., 2004; Asakawa et al., 2014). The mad2Δ background is derived from the HR105 strain (a gift from T. Matsumoto, Kyoto University, Kyoto, Japan; Kim et al., 1998), whereas the bub1Δ background is originated from strain 393 (a gift from J.-P. Javerzat, Institut de Biochimie et Génétique Cellulaires, Bordeaux, France; Bernard et al., 1998). The rec12Δ background was derived from the YY290-7B strain, in which the whole ORF of rec12⁺ was replaced with the kanMX6 marker (Ding et al., 2004). The sgo1Δ background originated from strain PZ856 (obtained from the Yeast Genetic Resource Center of Japan supported by the National BioResource Project; Hauf et al., 2007).

Unless stated otherwise, a two-step PCR method was used to introduce a chromosomal CFP, GFP, or mCherry tag to produce fluorescently labeled proteins (Hayashi et al., 2009). To visualize tubulin, integrating plasmids carrying GFP-atb2⁺or mCherry-atb2⁺ were introduced into the cells (Masuda et al., 2006; Unsworth et al., 2008). The telomere marker and nuclear reporter were encoded by plasmid-borne taz1⁺-GFP and 3GFP-NLS integrated at the lys1⁺-locus, respectively (Chikashige et al., 2006; Asakawa et al., 2010). mCherry-tagged histone H2B is expressed from an aur1^r-integrating plasmid that carries the mCherry-htb1⁺ gene (Ruan et al., 2015). The aur1^r gene confers resistance to the toxic drug aureobasidin A (Takara Bio Inc.). To visualize the centromere of chromosome II, cells were crossed to strains carrying lacI-GFP and tandem repeats of the lacO sequence integrated at the cen2-proximal locus (Yamamoto and Hiraoka, 2003). Cells carrying GFP-3pk-moa1⁺ are cross products derived from strain PZ425 (GFP-3pk-moa1+-kanr, ade6-M216, leu1; a gift from Y. Watanabe, University of Tokyo, Tokyo, Japan). The strains carrying sgo1⁺-flag-GFP originated from strain FY13800 (leu1 sgo1⁺-flag-GFP ade6-M210 pREP81 [CFP-atb2+]), which was obtained from the Yeast Genetic Resource Center of Japan supported by the National BioResource Project. The Rhp51-ECFP strains are derived from strain YA1083 (TH805 rhp51-ECFP::ura4+::rhp51; a gift from H. Iwasaki, Tokyo Institute of Technology, Tokyo, Japan; Akamatsu et al., 2007).

For the live-cell imaging, induced meiotic cells in the liquid medium of EMM-N+5S were immobilized on lectin (0.2 mg/ml; Sigma)–coated 35-mm glass-bottomed culture dishes (MatTek Corp.) and observed at 26°C (Asakawa and Hiraoka, 2009). Images were obtained with a DeltaVision deconvolution microscope system (Applied Precision, Inc.), in which an Olympus inverted microscope IX70 is equipped with an interline CoolSNAP HQ² charge-coupled device (Photometrics). The acquisition software was DeltaVision softWoRx 5.5. At each time point, optical section images were acquired at 0.3-µm or 0.5-µm intervals by using an Olympus oil-immersion 60× objective lens (PlanApoN60x OSC; NA = 1.4). Three-dimensional constrained iterative deconvolution of the images was done by the “enhanced ratio” method in softWoRx 5.5, and the images of Figs. 5, 6, 7, and 8 were processed using the denoising algorithm (Boulanger et al., 2009) before deconvolution.

Cells were sporulated on ME plates for 3 d at 26°C. To release the spores, the asci were collected from the plate, suspended in a solution containing a 1:10 dilution of β-glucuronidase (Sigma), and incubated for several hours at 30°C. Breakdown of the ascus cell wall by the enzyme and spore release were verified under the microscope. Spore numbers were determined with a CDA-1000 particle analyzer (Sysmex Corp.), and the spores were properly diluted to spread on YES plates. The YES plates were incubated at 30°C for 3–4 d to allow for colony formation. Spore viability = number of colonies formed × dilution factor/total number of spores applied.

Fig. S1 shows additional evidence of the delays in the metaphase-anaphase transition in the nup132Δ mutant by tracking Cut2-GFP dynamics. Fig. S2 shows localization of 3GFP-NLS during meiosis. Fig. S3 shows centromere localization of Moa1-GFP and Sgo1-Flag-GFP during metaphase I in the nup132Δ mutant. Fig. S4 summarizes the timing at which the Mis12-GFP first appeared during meiotic prophase in the various mutants. Fig. S5 shows live-cell imaging of a wild-type cell expressing Csi1-mCherry during meiotic prophase. Table S1 lists the strain used in this study.

We are indebted to Yuji Chikashige and Atsushi Matsuda for valuable suggestions on image acquisition and processing. We thank Chizuru Ohtsuki, Kun Ruan, and Xu Zhang for technical assistance. We also thank Hiroshi Iwasaki, Jean-Paul Javerzat, Tomohiro Matsumoto, Yoshinori Watanabe, and the National BioResource Project for kindly providing yeast strains.

This work was supported by a Japan Society for the Promotion of Science postdoctoral fellowship to H.-J. Yang and by Ministry of Education, Cultutre, Sports, Science and Technology/Japan Society for the Promotion of Science KAKENHI grants to H. Asakawa, T. Haraguchi, and Y. Hiraoka.

The authors declare no competing financial interests.