Microtubule-organizing center formation at telomeres induces meiotic telomere clustering

In sexual reproduction, eukaryotic organisms produce haploid gametes through a type of cell division called meiosis. During meiosis, telomeres cluster and promote homologous chromosome pairing, and the paired chromosomes recombine and then segregate (Scherthan, 2001). In many organisms, telomere clustering requires the telomeres to interact with a complex composed of the SUN and KASH nuclear membrane proteins (Hiraoka and Dernburg, 2009; Razafsky and Hodzic, 2009). SUN proteins interact with telomeres, whereas KASH proteins interact with cytoskeletal elements such as cytoplasmic microtubules or actin filaments, linking the telomeres with the cytoskeleton. Cytoskeleton-dependent forces gather the telomeres by as yet unknown mechanisms.

In the fission yeast Schizosaccharomyces pombe, telomeres cluster at the spindle pole body (SPB; a fungal centrosome) upon entering meiosis and remain clustered until the first division (Chikashige et al., 1994). During the period of telomere clustering, the nucleus oscillates between the cell ends, pulled by the SPB. The resultant telomere-led chromosome movements facilitate homologous chromosome pairing (Chikashige et al., 1994; Yamamoto et al., 1999; Ding et al., 2004). Telomere clustering requires telomeric localization of Sad1, an SPB component that belongs to the SUN family of nuclear membrane proteins, and telomere recruitment of Sad1 depends on the meiosis-specific factors Bqt1 and Bqt2 (Chikashige et al., 2006). However, forced telomere recruitment of Sad1 by Bqt1 and Bqt2 fails to induce telomere clustering during mitosis, and the mechanism by which the Sad1-localized telomeres move and gather at the SPB remains unknown.

Cytoplasmic dynein, a protein complex that moves along microtubules by ATP hydrolysis, has been suggested to contribute to telomere clustering. Cytoplasmic dynein drives nuclear oscillation, and the loss of either its motor subunit (dynein heavy chain [DHC]) or its regulatory subunit (Tctex-1–type dynein light chain [DLC]) compromises nuclear oscillation and impairs homologous chromosome pairing (Yamamoto et al., 1999; Miki et al., 2002). However, loss of both subunits reduces homologous recombination more severely than the single loss of either subunit and impairs spatial connection of the telomere-adjacent ribosomal DNA loci with the SPB during the telomere clustering stage (Miki et al., 2002). These data imply that loss of both subunits compromises telomere clustering and impairs homologous chromosome pairing more than the single losses.

Here, we show that telomere clustering in S. pombe depends on cytoplasmic dynein. We also show that telomere clustering depends on different types of microtubule motors, kinesins, and microtubules. Furthermore, a novel, meiosis-specific microtubule-organizing center (MTOC) is formed at the telomere, which we term the “telocentrosome.” This telocentrosome plays a pivotal role in telomere clustering. Based on our findings, we propose a telocentrosome-dependent mechanism for telomere clustering that explains how cytoskeleton-dependent forces gather telomeres.

To investigate the telomere-clustering mechanism, we examined the involvement of cytoplasmic dynein in this process. S. pombe cells normally proliferate in the haploid state. Upon nitrogen starvation, cells of opposite mating types fuse to form a diploid zygote that enters meiosis (zygotic meiosis; Fig. 1 A). Visualization of the telomere-binding protein Taz1 (Cooper et al., 1997) and the SPB component Sid4 (Tomlin et al., 2002) showed that most nuclei contained a single telomere signal adjacent to the SPB signal in wild-type zygotes, both before and after nuclear fusion, confirming telomere clustering at the SPB (Fig. 1, B and C; and Fig. S1, A and B). Likewise, most (but not all) nuclei in dhc1Δ or dlc1Δ single mutant zygotes contained a single telomere signal. In contrast, many nuclei in dhc1Δ dlc1Δ double mutant zygotes contained multiple telomere signals and a single SPB signal (Fig. 1, B and C; and Fig. S1, A and B). These findings confirmed severe telomere clustering defects in the double mutant and indicate that DHC and DLC play independent crucial roles in telomere clustering. There was also an increase in clustering defects upon loss of Ssm4, a p150^Glued subunit of dynactin (Niccoli et al., 2004), in the dhc1Δ background (Fig. 1 C and Fig. S1 B). This indicated that the dynactin complex, which interacts with dynein to aid its functions (Schroer, 2004), also contributes to telomere clustering.

The occurrence of telomere clustering in cells lacking dynein suggested that other microtubule motors contribute to telomere clustering. We next investigated the involvement of the microtubule motor kinesin. Depletion or impairment of kinesin motors belonging to the kinesin-5 (Cut7), -7 (Tea2), -8 (Klp5 and Klp6), or -14 (Pkl1 and Klp2) families (Steinberg, 2007) increased clustering defects in dhc1Δ cells, although to a lesser extent than with DLC depletion (Fig. 1 D and Fig. S1, B and C). Thus, these kinesins contribute to telomere clustering in a DHC-independent manner. They may function in the same pathway as DLC and/or play marginal roles because the clustering defects were not increased significantly by kinesin depletion or impairment in the dlc1Δ mutant (Fig. S1 C). Collectively, these results indicate that multiple microtubule motors contribute to telomere clustering.

Because most of the investigated motors are localized on microtubules (Fig. S1 D; Yamamoto et al., 1999; Miki et al., 2002; Niccoli et al., 2004), we examined the dynamics of telomeres with respect to microtubules. In cells undergoing conjugation, telomeres moved with microtubules and form a cluster (Fig. 2 A and Video 1). This observation suggests that telomeres gather via microtubules. To prove the microtubule dependency, we next used the microtubule inhibitor methyl 2-benzimidazole carbamate (MBC). To eliminate microtubules before telomere clustering, we induced meiosis synchronously in haploid cells bearing both mating type genes (Fig. 2 B; Yamamoto and Hiraoka, 2003). Under nitrogen-depleted conditions, these haploid cells underwent a single mitotic division to enter G1 phase and subsequently underwent meiosis without cell conjugation in a fairly synchronous manner (Fig. 2, B and C). MBC or its solvent, DMSO, was added to the medium after mitotic nuclear division. In the presence of DMSO, most telomeres formed a single cluster at the SPB after mitotic division (Fig. 2 C). In the presence of MBC, meiosis proceeded because of incomplete microtubule inhibition (as shown by the occurrence of meiotic divisions); however, most telomeres failed to form a single cluster and only partially clustered at the SPB (Fig. 2 C). These results indicate that telomere clustering depends on microtubules.

To understand the telomere–microtubule interaction and the clustering process, we allowed haploid meiotic cells to reform the microtubules after MBC treatment and examined the dynamics of telomeres and microtubules in detail. After MBC removal, cytoplasmic microtubules extended from the vicinity of telomeres, indicating MTOC formation at the telomeres (Fig. 2 D; Fig. S2, A and B; Video 2; and Video 3). The telomeric MTOC is meiosis-specific, as telomere microtubule nucleation and stable telomere–microtubule interaction were never observed during mitosis (Fig. S2 C and Video 4). We called this unusual, meiosis-specific telomeric MTOC the “telocentrosome.” The telomeres drifted inside the cell with the telocentrosome-nucleated microtubules (Fig. 2 D, 0–344 s) and gathered once they were connected by the microtubules (Fig. 2 D, 367–456 s). These observations show that telomeres gather via telocentrosomal microtubules. Notably, the telomeres have minus end–directed motile activity, as shown by movement on the microtubule toward the nucleation site upon microtubule interaction (Fig. 2 E, arrows; and Video 5). The telocentrosome is also formed during zygotic meiosis. Before cell conjugation, telomeres often partially clustered at the SPB, and Sad1 became localized at the nonclustering telomeres (Sad1-Taz1 colocalization was observed in 90.3% of cells containing multiple Sad1 dots [n = 73]; Fig. 3 A). The telocentrosomes formed at the Sad1-localized telomeres, as shown by microtubule nucleation from the Sad1 dots after MBC treatment (Fig. S2 D and Video 6).

To investigate telocentrosome formation, we looked at telocentrosome-localized factors before cell conjugation. Because of the MTOC activity, we examined SPB components by analyzing their colocalization with multiple Sad1 dots that corresponded to the telocentrosomes and the SPB. The SPB-localized KASH proteins Kms1 and Kms2 (Starr and Fischer, 2005) can interact with Sad1 and/or are required for telomere clustering (Shimanuki et al., 1997; Miki et al., 2004). Kms1 and Kms2 colocalized with the telocentrosome, as shown by colocalization with multiple Sad1 dots (Fig. 3, A and B). Dhc1, Dlc1, and Ssm4 also colocalized with the telocentrosome (Fig. 3, A and C). Furthermore, Alp4, an essential component of the γ-tubulin complex (γ-TuC) that induces microtubule nucleation (Vardy and Toda, 2000), and its regulator, Mto1 (Sawin et al., 2004; Venkatram et al., 2004), colocalized with the telocentrosome (Fig. 3, A and D). Unlike these SPB components, however, Sid4 did not colocalize (Fig. 3, A and D). Thus, a subset of SPB components localized to the telocentrosome.

We next examined the interdependence of telocentrosomal factor localization. Telomere localization of Dhc1 and Dlc1 was mutually independent, and Ssm4 localization to the telomere and the SPB was Dhc1 dependent (Fig. 3 C and Fig. S2 E). Furthermore, Dhc1 localization did not require dynein motor activity, as shown by Sad1 localization of a Dhc1 fragment lacking a motor domain (Fig. 3 E and Fig. S2 F). Dlc1 localization was Kms1 dependent (Fig. 3 C), consistent with their direct interaction (Miki et al., 2004). Importantly, Alp4 telomere localization depended substantially on Kms1 and Mto1, and Mto1 localization depended on Kms1 (Fig. 3 D). Therefore, Kms1 and Mto1 are required for telocentrosome formation, and Kms1 tethers γ-TuC to telomeres, probably via Mto1.

If telomere clustering depends on the telocentrosome, telocentrosome defects should be associated with telomere clustering defects. The kms1 mutant consistently shows severe telomere clustering defects and defective homologous chromosome pairing (Shimanuki et al., 1997; Niwa et al., 2000). The mto1Δ mutant also showed severe telomere clustering defects (Fig. 4, A and C) and phenotypes associated with defective pairing (reduced spore viability and meiotic recombination; Fig. 4 B). These results show that telomere clustering depends on the telocentrosome.

The telomeric localization of Dhc1 and Dlc1 raised the possibility that telomere clustering defects seen in the dhc1Δ dlc1Δ double mutant resulted from defective telocentrosome formation. Indeed, as seen in mto1Δ or kms1Δ zygotes, Sad1 mostly colocalized with telomeres (Fig. 4 C), but Alp4 and Mto1 were frequently absent or barely visible at the Sad1 dots in dhc1Δ dlc1Δ zygotes (Fig. 4, D and E). In addition, microtubules and the SPB-originated MTOC often failed to colocalize with the Sad1 dots (Fig. 4 F). Therefore, the Sad1–γ-TuC interaction appears to be reduced in the double mutant, as in mto1Δ and kms1Δ mutants. However, because strong Sad1 colocalization of Alp4 and Mto1 was observed before cell conjugation as in wild-type cells (Fig. 3 D), the interaction appears to diminish only after cell conjugation.

To evaluate the contribution of DHC and DLC to the telocentrosome, we examined Alp4 telomere localization and the effects of Dhc1 and/or Dlc1 depletion in mat gene-induced haploid meiosis. We inhibited telomere clustering by MBC and simultaneously visualized the SPB (Sid4) to assess Alp4 telomere localization (via Taz1; Fig. 5 A). In wild-type cells treated with MBC, Alp4 telomere localization increased along with the increase in telomere clustering (Fig. 5 B). However, Alp4 localization to the telomere was substantially reduced in telomere clustering–defective mto1Δ cells (Fig. 5 B) despite progression of meiosis (Fig. S3). These results confirm the validity of this assay and the importance of Mto1 in telocentrosome formation. The telomere clustering defects in dynein mutant zygotes were mirrored in haploid meiotic cells (Fig. 5 B, blue lines). Importantly, Alp4 localized to the telomeres in dhc1Δ cells, but Alp4 localization was dramatically reduced in dlc1Δ cells and in dhc1Δ dlc1Δ cells (Fig. 5 B). Therefore, Dlc1 contributes to γ-TuC telomere localization, but Dhc1 does not. Together with the independent telomere/SPB localization of Dhc1 and Dlc1 (Fig. 3 C and Fig. S2 E) and the retention of the Dhc1 function in dlc1Δ cells (Miki et al., 2002), these results indicate that the telomere clustering defects seen in the dhc1Δ dlc1Δ mutant arise from a combination of telocentrosome defects and loss of dynein motor activity.

Our results show that microtubule motor-dependent aggregation of telocentrosomes causes telomere clustering. Studies have shown that acentrosomal spindle poles are formed in a dynein-dependent manner in Xenopus laevis egg extracts and by MTOC aggregation in mouse oocytes (Heald et al., 1996; Schuh and Ellenberg, 2007). Given their similarities, the processes of telomere clustering and acentrosomal spindle pole formation likely share the same mechanism. Considering this possibility, we propose the following telomere clustering mechanism (Fig. 5 C). Upon entering meiosis, Bqt1 and Bqt2 recruit Sad1 to telomeres along with Kms1 and Kms2. This recruitment causes telomere recruitment of the γ-TuC via Mto1, forming the telocentrosome. Simultaneously, cytoplasmic dynein is recruited to the telomere in a motor-independent fashion, whereas Dlc1 is independently tethered to telomeres via Kms1 and may act to maintain the telomere γ-TuC in the later stage. Kinesin motors Pkl1 and Cut7 may also contribute to γ-TuC telomere localization despite the absence of obvious telocentrosomal localization, as they genetically and/or physically interact with γ-tubulin (Pidoux et al., 1996; Paluh et al., 2000; Rodriguez et al., 2008). Subsequently, oligomerized cytoplasmic dynein and minus end–directed kinesin-14 motors gather telomeres at the SPB by cross-linking the SPB- and telocentrosome-nucleated microtubules and by moving along these microtubules toward the nucleation sites. Telomere-tethered dynein probably also facilitates telomere clustering by direct transport of the telomere along the microtubules. Because the investigated kinesins and an Ssm4 homologue p150^Glued are suggested or shown to regulate microtubule polymerization and/or bundling (Hagan and Yanagida, 1992; Pidoux et al., 1996; Browning et al., 2000; Troxell et al., 2001; Ligon et al., 2003; Grissom et al., 2009; Erent et al., 2012), the kinesins and Ssm4 may contribute to telomere clustering by regulating dynamics and/or cross-linking of the telocentrosomal microtubules.

In our model, the telocentrosome is essential for telomere clustering. This is supported by the observation of severe telomere clustering defects in the telocentrosome-defective kms1Δ and mto1Δ mutants. In addition, dynein is a major motor for telomere clustering. This model explains why telomere clustering was severely compromised in the dhc1Δ dlc1Δ double mutant but mildly affected in the dhc1Δ or dlc1Δ single mutant and the dlc1Δ kinesin double mutants. In the absence of dynein, distinct kinesin motors gather telomeres cooperatively. In the absence of DLC, the telocentrosome forms transiently, perhaps at an early stage, and dynein and kinesin motors gather telomeres during this period. This situation is largely unchanged by the concomitant depletion of some kinesins. However, when dynein is depleted simultaneously with DLC, kinesin motors alone fail to gather telomeres.

Our findings have several important implications. First, it is likely that our model applies to telomere clustering in other organisms. In worms, the specialized chromosomal domains called “pairing centers” aggregate to promote homologous chromosome pairing (MacQueen et al., 2005). Pairing centers interact with the SUN and KASH proteins, and their aggregation depends on microtubules and cytoplasmic dynein (Penkner et al., 2007; Sato et al., 2009). These similarities between two evolutionarily distant organisms imply that the telomere clustering mechanism is conserved among eukaryotes. Second, our finding of DLC-dependent regulation of the γ-TuC contributes to understanding of the spatiotemporal regulation of the MTOC and a recently identified link between telomere clustering and meiotic spindle formation (Tomita and Cooper, 2007). In addition, Tctex-1–type DLC regulates neurite outgrowth by interacting directly with a G protein in a dynein motor–independent manner (Sachdev et al., 2007), and our finding sheds light on the dynein motor–independent functions of Tctex-1–type DLC. Finally, because telomere clustering defects cause improper homologous chromosome pairing and thus defective homologous chromosome segregation, our findings may shed light on the mechanisms underlying birth-associated chromosome missegregation that lead to a miscarriage or Down’s syndrome.

The strains used in this study are shown in Table S1. Yeast extract medium with supplements (YES) or Edinburgh minimal medium (EMM) was used for vegetative growth, whereas malt extract (ME) medium or EMM medium lacking ammonium chloride (EMM-N) was used for the induction of meiosis (Moreno et al., 1991).

To visualize telomeres, we constructed a plasmid, pMY23, encoding an mCherry and taz1⁺ fusion as follows: An ars1-coding region was removed from pAUR224 (Takara Bio Inc.) by digesting pAUR224 with MluI and AflII and self-ligating it after blunting the digested DNA ends with a Klenow fragment. A critical portion of the cytomegalovirus promoter was removed from the resulting plasmid by digesting the plasmid with NcoI and XhoI and then self-ligating it after blunting the digested DNA ends with a Klenow fragment. This resulted in pTO2. The mCherry (Shaner et al., 2004) coding region was excised from an mCherry-bearing plasmid (a gift from K. Tanaka, University of Leicester, Leicester, England, UK) with BamHI and MluI and inserted in the corresponding sites of pTO2, producing pHM4. Finally, a DNA region encoding the taz1⁺ promoter and ORF was amplified by PCR using two synthetic oligonucleotide primers, 5′-CCGCTCGAGCATCGACAAGGCATGCGAAGCT-3′ and 5′-GCGGATCCTAGATTGATAATTAACAAGCTCTTCC-3′, and fission yeast genomic DNA as a template. The amplified DNA fragment was digested with BamHI and XhoI and inserted in the BamHI and SalI sites of pHM4. The resulting plasmid, pMY23, was transformed into cells, and integrants were selected by resistance to the antibiotic aureobasidin A (Takara Bio Inc.).

Alternatively, we generated C-terminal mCherry-tagged Taz1 by the two-step PCR-based method (Krawchuk and Wahls, 1999). We first constructed an mCherry module plasmid (pHM22) by inserting the mCherry-encoding DNA region of the mCherry-bearing plasmid between the SalI and BglII sites of pFA6a-natMX6 (obtained from T. Carr, University of Sussex, Sussex, England, UK; Hentges et al., 2005). We amplified DNA fragments encoding the Taz1 C terminus–coding region or the taz1⁺ terminator by PCR using two sets of synthetic oligonucleotide primers (5′-CTGGTGGTAGAAGTTCGGCAA-3′ and 5′-CCTTAATTAACCCGGGGATCCGAGATTGATAATTAACAAGCTCTTCC-3′; and 5′-TTTTCGCCTCGACATCATCTTTTTTCGTCTCCGGATAGAGTTT-3′ and 5′-CGCAATTTGGGATTGAAAACCTA-3′) and genomic DNA as a template. We then amplified the mCherry module by PCR using the two PCR products as primers and pHM22 as a template. The resulting PCR product was introduced into cells, and integrants were selected by resistance to the antibiotic nourseothricin (WERNER BioAgents) and confirmed by PCR and microscopic observation.

To visualize microtubules, we constructed plasmids pMY53 and pMY56, which encoded an mCherry and atb2⁺ fusion gene as follows: A coding region of the nda3 promoter in a plasmid containing the mDsRed and atb2⁺ fusion gene (Yamamoto et al., 2008) was removed to create a BamHI site at the removal site using two synthetic oligonucleotide primers, 5′-CTAGACTAATTTATGGTGGTTTGTTTTG-3′ and 5′-GGATCCATGAGAGAGATCATTTCCATTCACGTT-3′, following the instructions for the KOD-Plus-Mutagenesis kit (Toyobo Co., Ltd.). A DNA fragment encoding mCherry was amplified by PCR using two synthetic oligonucleotide primers, 5′-TCAGATCTCTTGTACAGCTCGTCCATGCC-3′ and 5′-GGGGGCGAGCTCAGATCTATGGTGAGCAAGGGCGAGGA-3′, and the mCherry-bearing plasmid as a template. The amplified DNA fragment was digested with BglII and inserted at the BamHI site of the atb2⁺ gene-bearing plasmid to yield the mCherry-atb2⁺ integration plasmid pMY53. For constructing pMY56, the mCherry-atb2⁺ fusion gene along with the nda3 promoter was amplified by PCR using two synthetic oligonucleotides, 5′-CAGGAAACAGCTATGACCATGA-3′ and 5′-CGGGATCCCTATAGGGCGAATTGGAGCTC-3′, and pMY53 as a template. The amplified DNA fragment was digested with XhoI and BamHI and inserted between the SalI and BamHI sites of pTO2 to yield pMY56. To construct strains expressing the mCherry-atb2⁺ fusion gene, pMY53 or pMY56 was transformed into lys1-131 or aureobasidin A–sensitive cells, respectively. The integrants were selected by the lys⁺ or aur^r phenotype and further confirmed by microscopic observation.

To visualize Kms2, we constructed plasmid pMY16, which encoded a GFP and kms2⁺ fusion gene as follows: A DNA fragment coding for GFP was amplified by PCR using two synthetic oligonucleotides, 5′-CCGCTCGAGATCTATGTCTTTAATTAACAGTAAAGGAG-3′ and 5′-CCGCTCGAGGGATCCTTTGTATAGTTCATCCATGCCAT-3′, and pFA6a-GFP(S65T)-kanMX6 (Bähler et al., 1998) as a template. The amplified DNA fragment was digested by XhoI and inserted at the SalI site of pYC36 (obtained from Y. Chikashige, Advanced Information and Communications Technology Research Center, Kobe, Japan; Chikashige et al., 2004) to produce pMY12. A DNA fragment encoding the kms2⁺ promoter was amplified by PCR using two synthetic oligonucleotide primers, 5′-CGGGATCCTTTGATTGTTTATTGACATGTGCT-3′ and 5′-CGGGATCCGAGAAGTAGGAGATATAATCGTC-3′, and fission yeast genomic DNA as a template. The fragment was inserted at the BglII site of pMY12 after digestion with BamHI. Subsequently, a Kms2-coding region and its terminator were amplified by PCR using two synthetic oligonucleotide primers, 5′-GAAGATCTATGGATGAATACATCCCTTTTGA-3′ and 5′-GAAGATCTAGTTATAAAAGAGAAATTATGAGAAA-3′, and fission yeast genomic DNA as a template. The amplified gene was then inserted in frame after the GFP gene in the constructed plasmid by inserting the fragment at the BamHI site after BglII digestion. The resulting plasmid, pMY16, was introduced into cells as described for pMY53.

To visualize Dlc1, we constructed a plasmid, pMY21, encoding the dlc1⁺ and GFP fusion gene as follows: A DNA fragment coding GFP with the Saccharomyces cerevisiae ADH1 terminator was amplified by PCR using two synthetic oligonucleotides, 5′-ACGCGTCGACGAAGATCTTCGGATCCCCGGGTTAATTAAC-3′ and 5′-GGGGTACCATATTACCCTGTTATCCCTAGCG-3′, and pFA6a-GFP(S65T)-kanMX6 as a template. The amplified DNA fragment was digested with SalI and KpnI and inserted between the corresponding sites of pYC36 to yield pTO7. Subsequently, a DNA fragment encoding the dlc1⁺ promoter and ORF was amplified by PCR using two synthetic oligonucleotides, 5′-ACGCGTCGACTAAAGATTTGGATGATTTCGCTAC-3′ and 5′-GAAGATCTATAATGGAGATCCACATAATGCTC-3′, and fission yeast genomic DNA as a template. The fragment was then digested with BglII and SalI and inserted between the corresponding sites of pTO7 to yield pMY21. pMY21 was introduced into cells as described for pMY53.

To visualize Tea2, we constructed a plasmid, pMY36, encoding a tea2⁺ and GFP fusion gene as follows: A DNA fragment encoding the promoter and ORF of the tea2⁺ gene was amplified by PCR using two synthetic oligonucleotides, 5′-AAACTGCAGACGAAAGCCAGCTACCTGTTG-3′ and 5′-CGCGGATCCAAAGAAAGAAGTTGCGTTTCCTGAG-3′, and fission yeast genome DNA as a template. The fragment was then digested with PstI and BamHI and inserted between the PstI and BglII sites of pTO7 to yield pMY36. pMY36 was introduced into cells as described for pMY53.

To visualize the SPB, we constructed a plasmid, pMY2, that encoded the mRFP and sid4⁺ fusion gene as follows: Using an epitope-tagging cassette (Sato et al., 2005), the mRFP gene together with the TEF terminator was integrated into the yeast chromosome so that the mRFP ORF was fused in frame with the sid4 ORF at the Sid4 C terminus. For alternative visualization, a DNA fragment encoding the sid4⁺-mRFP fusion gene together with its promoter and terminator was amplified by PCR using two synthetic oligonucleotide primers, 5′-CCGCTCGAGCGTTAAAGCTCAATGATCAATCTT-3′ and 5′-CCGGCGGGGACGAGGCAA-3′, and genomic DNA from cells containing the sid4⁺-mRFP fusion gene as a template. The fragment was digested with BglII and XhoI and inserted between the BamHI and SalI sites of pTO2. The resulting plasmid, pMY2, was introduced into cells as described for pMY53.

To visualize the nuclear membrane, we generated C-terminal mCherry-tagged Cut11 (West et al., 1998) by the two-step PCR-based method, as described for telomere zvisualization. We used two sets of synthetic oligonucleotide primers (5′-CTACAACCCCGGTTAATGAAGG-3′ and 5′-CCTTAATTAACCCGGGGATCCGGGAGTTGCTTTTGTATTCTCGAA-3′; and 5′-TTTTCGCCTCGACATCATCTGTCATCCTATAAGGTAGGTTTGT-3′ and 5′-CAAGGATCTTAAACCTCTTGGG-3′) and pHM22 as a module plasmid.

Cells grown on YES solid medium at 30°C were transferred to ME solid medium and induced to enter meiosis by incubation at 25°C for 16–18 h. Nuclear DNA in meiotic zygotes was stained with DNA-specific Hoechst 33342 dye (Invitrogen) by incubating the zygotes in distilled water containing 5 µg/ml Hoechst 33342 for 5 min. Telomeres visualized using GFP-tagged Taz1 were observed using an epifluorescence microscope (IX51; Olympus) or a fluorescence microscope (IX71; Olympus) equipped with a 60×/1.42 NA or 100×/1.40 NA Plan Apochromat oil immersion objective lens and a charge-coupled device (CCD) camera (DP30BW; Olympus). All images were analyzed by MetaView or MetaMorph software (Molecular Devices).

Haploid cells bearing both types of the mating type genes were constructed by integrating the mat-P gene at the lys1⁺ or the aur1⁺ locus of h⁻ haploid cells. The mat-P gene was integrated at the lys1⁺ locus using the mat-P integration plasmid (pAY153) as described previously (Yamamoto and Hiraoka, 2003). For mat-P integration at the aur1⁺ locus, we constructed an integration plasmid, pKT3, by amplifying a DNA fragment encoding the mat-P gene and inserting it at the SalI site of pTO2. The mat-P gene was integrated at the aur1⁺ locus as described for pAY153, except that the integrants were selected by a resistant phenotype to the antibiotic aureobasidin A.

Haploid cells bearing both types of the mating type genes were grown to the early stationary phase in YES medium at 30°C and transferred to EMM-N medium. Cells were induced to enter meiosis by incubation at 30°C with vigorous shaking. To monitor meiotic progression, a portion of the cell culture was removed every hour, and the chromosomal morphology of the cells was analyzed by staining DNA in 0.1 M Tris-Cl, pH 9.0, containing 1 µg/ml DAPI. To disrupt microtubules, 1/200 volume of 10-mg/ml MBC (Sigma-Aldrich) in DMSO was added to the culture medium. For the control experiment, the same amount of DMSO without MBC was added instead.

To follow dynamics of telomeres and microtubules in normal zygotic meiosis, cells expressing mCherry-tagged Taz1 and GFP-tagged Atb2 were induced to enter meiosis on the solid ME medium, as described for telomere clustering analysis. The cells were suspended in EMM-N medium, and a total of 20 µl of the cell suspension was placed on a 35-mm glass-bottomed culture dish (Matsunami Glass Ind., Ltd.) precoated with Bandeiraea simplicifolia lectin (Sigma-Aldrich). To follow telomere and microtubule dynamics in haploid meiotic cells, cells expressing mCherry-tagged Taz1 and GFP-tagged Atb2 or those expressing GFP-tagged Taz1, mCherry-tagged Atb2, and RFP-tagged Sid4 were induced to enter meiosis synchronously and treated with MBC as described for the analysis of haploid meiotic cells. A total of 30 µl of the MBC-containing cell suspension was placed on a 24 × 60–mm coverslip (Matsunami Glass Ind., Ltd.) precoated with lectin. After the cells were immobilized on the surface of the coverslip, the MBC-containing EMM-N medium was removed using a micropipette, and 100 µl MBC-free EMM-N medium was immediately placed on the coverslip to allow the cells to reform microtubules. Cells were kept at 25–27°C during live-cell observation.

GFP and RFP images of the cells at seven focal planes spaced at 0.5-µm intervals were taken through a 60×/1.42 NA Plan Apochromat oil immersion objective lens using a fluorescence microscope (IX71) equipped with a cooled CCD camera (CoolSNAP-HQ2; Nippon Roper Co. Ltd.) and controlled by MetaMorph software. The images taken at each time point were processed by deconvolution and combined to form a 2D projection in MetaMorph software.

The gene encoding GFP was amplified by PCR using two synthetic oligonucleotide primers, 5′-CGCGGATCCATGAGTAAAGGAGAAGAACTT-3′ and 5′-GAAGGCCTATTTGTATAGTTCATCCATGCC-3′, and digested with BamHI and StuI. It was then fused with the nmt1 terminator by inserting it between the BamHI and SmaI sites of the pREP1 plasmid (Maundrell, 1993). A DNA fragment encoding GFP and the nmt1 terminator was excised from the resulting plasmid by digestion with BamHI and SacI and inserted between the corresponding sites of the pRS405 plasmid bearing the S. cerevisiae LEU2 gene (Sikorski and Hieter, 1989) to produce pAY147. A DNA fragment encoding a portion of the dhc1⁺ ORF was amplified by PCR using two synthetic oligonucleotide primers, 5′-AGAGATTCCTGCTTCAGG-3′ and 5′-GAAGATCTACTTGAAGCACACCATTCAGA-3′, and pDHC1-Cla (Yamamoto et al., 1999) as a template. The PCR product was digested with EcoRV and BglII and inserted between the BamHI and SmaI sites of pAY147 to generate a fusion gene that encoded part of the dhc1⁺ ORF and the GFP ORF. The resulting plasmid, pAY182, was integrated at the dhc1⁺ locus by transforming it into fission yeast cells bearing the leu1-32 mutation. When integrated at the dhc1⁺ locus by homologous recombination, the GFP gene was fused to the dhc1⁺ gene so that an N-terminal portion of Dhc1 that lacked the motor domain was tagged with GFP. The integrants were selected by the leu⁺ phenotype and confirmed by Southern blot analysis.

For spore viability analysis, cells were grown on solid YES medium at 30°C and induced to enter meiosis on ME solid medium at 25°C. The spores were dissected randomly under a microscope (BX40; Olympus), and their viability was determined by colony formation on the YES solid medium. To determine the recombination frequency, two strains bearing appropriate genetic markers were induced to enter meiosis as for the spore viability analysis. Crossover recombination was examined by tetrad analysis. Gene conversion at the ade6 locus was examined as described previously (Ponticelli and Smith, 1989) except that the spore suspension was heated at 55°C for 30 min after glusulase treatment to eliminate vegetative cells.

Fig. S1 shows telomere clustering defects in dynein and kinesin mutants and the intracellular localization of kinesin motors. Fig. S2 shows microtubule and telomere dynamics and Sad1 localization of Dhc1 and Dlc1. Fig. S3 shows meiotic progression in haploid wild-type, mto1Δ, and dynein mutant cells in the presence of DMSO or MBC. Table S1 is a list of strains used in this study. Video 1 shows dynamics of microtubules and telomeres in cells undergoing cell conjugation. Video 2 shows dynamics of microtubules and telomeres in haploid meiotic cells after MBC removal. Video 3 shows microtubule nucleation from the vicinity of the telomere in haploid meiotic cells. Video 4 shows telomere and microtubule dynamics in a mitotic cell. Video 5 shows dynamics of microtubules and telomeres in haploid meiotic cells after MBC removal. Video 6 shows microtubule nucleation from telomere-localized Sad1 before cell conjugation.

We thank Y. Chikashige, T. Carr, J.R. McIntosh, J. Paluh, K. Sawin, K. Tanaka, T. Toda, A. Yamashita, M. Yamamoto, and the Yeast Genetic Resource Center for strains or reagents and J. Cooper, A. Shinohara, K. Tanaka, M. Uritani, and T. Ushimaru for critical reading of the manuscript and many helpful suggestions. We also thank K. Ichikawa for assisting live-cell observation.

This work was supported by Grants-in-aid for Scientific Research (C) and on Innovative Areas to A. Yamamoto and was performed in part under the Cooperative Research Program of the Institute for Protein Research, Osaka University.