Mcp6, a meiosis-specific coiled-coil protein of Schizosaccharomyces pombe, localizes to the spindle pole body and is required for horsetail movement and recombination

Sexually reproducing eukaryotic organisms undergo meiosis, a special type of cell division, to generate inheritable haploid gametes from diploid parental cells. This process includes meiosis-specific events that increase genetic diversity, such as synaptonemal complex (SC) formation, homologous pairing and recombination. The fission yeast Schizosaccharomyces pombe proceeds to meiosis when it is nutritionally starved. At this point, two cells with opposite mating types conjugate and the two haploid nuclei fuse, thereby producing a zygote with a diploid nucleus; the meiotic process immediately follows this. Efficient pairing of the homologous chromosomes and the subsequent processing and completion of recombination during the meiotic prophase are pivotal for achieving correct chromosome segregation during meiotic division. An essential event for efficient chromosome pairing in S. pombe is the clustering during prophase of meiosis I of the telomeres of three chromosomes near the spindle-pole body (SPB) (Chikashige et al., 1994; Chikashige et al., 1997). This characteristic arrangement of meiotic chromosomes has been observed in a wide range of organisms and is denoted a `bouquet' arrangement (Loidl, 1990; Scherthan, 2001; Zickler and Kleckner, 1999). This arrangement has been proposed to facilitate homologous chromosome pairing because it generates a polarized chromosome configuration by bundling chromosomes together at their telomeres.

The clustering of telomeres occurs during an event termed horsetail nuclear movement that is characteristic of S. pombe. It occurs at prophase I of meiosis and is characterized by a dynamic oscillation of the nucleus and the adoption by the nucleus of an elongated morphology (Chikashige et al., 1994). This movement has been proposed to facilitate the pairing of homologous chromosomes because it causes the chromosomes, which are aligned in the same direction as a result of bundling at their telomeric ends, to be shuffled around each other (Chikashige et al., 1994; Kohli, 1994; Hiraoka, 1998; Yamamoto et al., 1999; Yamamoto and Hiraoka, 2001). Thus, it enhances the chance that a chromosome encounters its correct partner and thereby promotes the linkage of homologous pairs of chromosomes through homologous recombination. In support of this notion, homologous recombination is reduced in mutants that display impaired telomere clustering owing to the depletion of protein components of the telomere or the SPB, even though recombination machinery is intact in these mutants. For example, elimination of the telomere-binding protein Taz1 (Cooper et al., 1998; Nimmo et al., 1998) or Rap1 (Kanoh and Ishikaw, 2001), or depletion of the SPB component Kms1 (Shimanuki et al., 1997; Niwa et al., 2000) results in a loss of telomere-SPB clustering and reduced meiotic recombination.

It has been proposed that horsetail nuclear movement is predominantly established by pulling the astral microtubules that link the SPB to microtubule-anchoring sites, and that the pulling force is provided by cytoplasmic dynein (Chikashige et al., 1994; Svoboda et al., 1995; Ding et al., 1998; Yamamoto and Hiraoka, 2003). Thus, homologous recombination is also reduced when nuclear oscillation is abolished by disrupting the dhc1⁺ gene, which encodes dynein heavy chain (DHC), a major component of cytoplasmic dynein that is localized to microtubules and the SPB (Yamamoto et al., 1999). It was proposed that dynein drives this nuclear oscillation by mediating the cortical microtubule interactions and regulating the dynamics of microtubule disassembly at the cortex (Yamamoto et al., 2001). Meiotic recombination is also reduced in a null mutant of the dlc1⁺ gene that encodes an SPB protein that belongs to the dynein-light-chain family (Miki et al., 2003). In this mutant, Dhc1-dependent nuclear movement during meiotic prophase is irregular in its duration and direction. This model explains some of the regulatory mechanisms behind nuclear oscillation and chromosome pairing. However, the details of these mechanisms are still mostly unknown. The identification of additional regulatory components is needed fully to elucidate these processes.

In the course of our functional characterization of meiotic-specific proteins that harbour coiled-coil motifs, we found a new SPB-associated protein that is required for meiotic nuclear oscillation and recombination. This gene is expressed specifically during meiosis and thus is referred to as mcp6⁺ (meiotic coiled-coil protein). The coiled-coil motif, which consists of two to five amphipathic α-helices that twist around one another to form a supercoil, is known to be required for protein-protein interaction (Burkhard et al., 2001). In the present study, we report our functional analysis of this protein.

The S. pombe strains used in this study are listed in Table 1. Complete media YPD or YE, the synthetic minimal medium EMM2 and the sporulation media ME or EMM2-N (1% glucose) were used (Alfa et al., 1993). Homozygous diploid strains were constructed by cell fusion. Cells were converted to protoplasts by treatment with lysing enzyme. Then, cells were fused using CaCl₂ and polyethylene glycol (Sipiczki and Ferenczy, 1977). Plates with EMM2 containing 1 M sorbitol were used in the cell fusion experiments. Induction of meiosis in the genetic background of the pat1-114 mutant (Shimada et al., 2002). Northern (Watanabe et al., 2001) and western blot (Okuzaki et al., 2003) analyses were performed as described previously.

To disrupt the mcp6⁺ gene by replacing it with the ura4⁺ gene, we used the polymerase chain reaction (PCR) to obtain a DNA fragment carrying the 5′ upstream region and 3′ downstream region of the mcp6⁺ gene. For this purpose, we synthesized the following four oligonucleotides and used them as primers: mcp6-5F, 5′-GGTACCTTCTGGTGCCGCCGACCTTC-3′; mcp6-5R, 5′-CTCGAGATTAAATCAATCTGTTAATC-3′; mcp6-3F, 5′-CCCGGGGGATAGCTATGAAACCCTGA-3′; mcp6-3R, 5′-GAGCTCTCATTTTTTTTATAAGAAGG-3′. (The underlined sequences denote the artificially introduced restriction enzyme sites for KpnI, XhoI, SmaI and SacI, respectively.) These PCR products and the 1.8 kb HindIII fragment containing the ura4⁺ gene (Grimm et al., 1988) were inserted into the pBluescriptII KS (+) vector via the KpnI-XhoI, SmaI-SacI or HindIII sites. This plasmid construct was digested with KpnI and SacI, and the resulting construct was introduced into the diploid strain TP4-5A/TP4-1D. The Ura⁺ transformants were then screened by Southern blot analysis to identify the disrupted strain.

To construct green fluorescent protein (GFP)-tagged mcp6⁺ strains, we performed PCR using the wild-type (TP4-5A) genome as a template and obtained a DNA fragment carrying the open reading frame (ORF) region and the 3′ downstream region of the mcp6⁺ gene. For this purpose, we synthesized the following two oligonucleotides and used them as primers: mcp6-ORF-F, 5′-CGGCGCGCCGCATATGGAATATCAAGAAGAGGC-3′; mcp6-ORF-R, 5′-GTACTCGAGGCGGCCGCGGGCTCAGATCGTGATTGACAG3′. The underlined sequences denote the artificially introduced restriction enzyme sites for AscI and NdeI, and XhoI and NotI, respectively. To obtain the 3′ downstream region, we used the same primers as described above. These PCR products were inserted into the pTT(GFP)-Lys3 vector (T.T., unpublished) (see supplementary material Fig. S1), which is designed to allow one-step integration via NdeI-NotI and SmaI-SacI sites. This plasmid construct was digested with PmeI. The resulting construct was introduced into strain HM105 (h^- lys3). We then screened the Lys⁺ transformants and confirmed the precise integration of the constructs by PCR.

The crossing-over rate was determined as described previously (Fukushima et al., 2000). Briefly, haploid parental strains were grown on YPD plates at 30°C and cells were mated and sporulated on ME plates at 28°C (zygotic meiosis). After 1 day of incubation, the spores were separated by a micromanipulator (Singer Instruments, UK). To examine the frequency of crossing over, we measured the genetic distance (in centiMorgans) between leu1⁺ and his2⁺, and between lys3⁺ and cdc12⁺. Genetic distance was calculated according to the formula 50×[T+(6NPD)]÷(PD+T+NPD) (Perkins, 1949), where T, NPD and PD indicate the number of tetratypes, nonparental ditypes and parental ditypes, respectively.

Intragenic recombination rate and spore viability were determined as described previously (Shimada et al., 2002). Briefly, haploid parental strains were grown on YPD plates at 33°C. Cells were mated and sporulated on ME plates at 28°C (zygotic meiosis). After 3-4 days of incubation, spores were treated with 1% glusulase (NEN Life Science Products) for 2-3 hours at room temperature and checked under a microscope for complete digestion of contaminating vegetative cells. The glusulase-treated spores were then washed with water and used to measure the intragenic recombination rate and in the spore-viability assays. To examine the frequency of intragenic or ectopic recombination (or prototroph frequency), we used two ade6 alleles -ade6-M26 and ade6-469 (Gutz, 1971) or ade6-M26 and z7 (ade6-469) (Virgin and Bailey, 1998), because the reciprocal recombination between these alleles produces the ade6⁺ allele.

Cells from a single colony were cultured at 28°C in 10 ml EMM2 plus supplements [adenine (75 μg/ml), histidine (75 μ/ml), leucine (250 μg/ml), lysine (75 μg/ml) and uracil (75 μ/ml)] until they reached mid-log phase. The cells were collected by centrifugation, washed three times with 1 ml EMM2-N and then induced to enter meiosis by incubation in EMM2-N at 28°C for 6 hours. For live observations, we added 0.5 μg ml^-1 Hoechst 33342 to 200 μl of the cells and an aliquot was observed under a fluorescence microscope (Olympus BX51).

For methanol fixation, cells were collected by aspiration through a glass filter (particle retention 1.2 μm; Whatman, Brentford, UK) that traps cells. The cells were then immediately immersed into methanol at -80°C and left overnight to fix the cells. The cells were then washed off the glass filter with distilled water and collected by centrifugation (2000 g, 5 minutes), and the pellet was washed three times with PBS. 0.5 μg ml^-1 Hoechst 33342 was added and the cells were observed under a fluorescence microscope.

For time-lapse observations, cells expressing GFP-tagged DNA polymerase α (Polα) (ST193 and CRL026-1) or cells expressing LacI-NLS-GFP and integrated LacO repeat at lys1 locus (ST197 and AY174-7B) were cultured in 10 ml EMM2 plus supplements until they reached mid-log phase at 28°C. They were then induced to enter meiosis by incubation in EMM2-N at 28°C. After 5 hours of nitrogen starvation, the cells were put on a glass-bottomed dish whose surface was coated with 0.2% concanavalin A and images under a fluorescence microscope (Olympus IX71) were recorded every 2.5 minutes (1 second of exposure time) after the initiation of karyogamy. For observation of LacI-GFP dots, images were taken with a 0.3 second exposure at 5 minute intervals, with ten optical sections made at 0.5 μm intervals for each time point. Projected images obtained with Meta Morph software were analysed.

Cells were fixed following the procedure of Hagan and Hyams (Hagan and Hyams, 1988) using glutaraldehyde and paraformaldehyde. In indirect immunofluorescence microscopy (Hagan and Yanagida, 1995), the SPB was stained with the anti-Sad1 antibody (a gift from O. Niwa, Kazusa DNA Research Institute, Kisatazu, Japan), microtubules were stained with the anti-α-tubulin antibody (TAT1). Subsequently, we added an Alexa-488-conjugated goat anti-mouse antibody (Molecular Probes, Eugene, OR) to label TAT1 or an Alexa-594-conjugated goat anti-rabbit antibody (Molecular Probes, Eugene, OR) to label anti-Sad1 antibody. The samples were then stained with 0.2 mg ml^-1 Hoechst 33342 in PBS for 1 minute and mounted with antifade mounting medium containing 10 mg/ml p-phenylendiamine in 100 mM Tris-HCl (pH 8.8). Fluorescence images of these cells were observed using a fluorescence microscope (Olympus BX51) with a Cool SNAP CCD camera (Roper Scientific, San Diego, CA). Immunofluorescence images were acquired using Adobe Photoshop 7.0.

Meu13 harbours a coiled-coil motif and plays a pivotal role in homologous pairing, meiotic recombination at meiosis I (Nabeshima et al., 2001) and the meiotic recombination checkpoint (Shimada et al., 2002). To identify a meiosis-specific coiled-coil protein of S. pombe that might interact with Meu13, we searched the genome database for uncharacterized genes that harbour coiled-coil motifs (http://www.sanger.ac.uk/Projects/S_pombe/) and found 60 genes. We then obtained DNA fragments from each of these genes and used them as probes for time-course northern blot analysis to examine their meiosis-specific transcription. This analysis identified seven novel mcp genes (Saito et al., 2004): mcp1⁺ (AB189991); mcp2⁺ (AB189990); mcp3⁺ (AB189989); mcp4⁺ (AB189988); mcp5⁺ (AB189987); mcp6⁺ (AB189986); mcp7⁺ (AB189985).

Mcp6 consists of 327 amino acids and harbours two putative coiled-coil motifs, a leucine zipper (LZ), a nuclear localization signal (NLS), a peroxisomal targeting signal (PTS) and four potential Rad3-kinase phosphorylation target sites (SQ/TQ motifs) (Fig. 1A). Homology searches using the BLAST algorithm (at http://www.ncbi.nlm.nih.gov/BLAST/) indicate that Mcp6 is specific to S. pombe, because orthologues were not found in other organisms. Homology searching using the Block Maker program (http://bioinformatics.weizmann.ac.il/blocks/blockmkr/www/make_blocks.html) revealed that not only the coiled-coil domains but also other regions (depicted by filled vertical arrowheads) of Mcp6 are partly homologous to the myosin heavy chain (MHC), which is essential for cytokinesis (Rajagopalan et al., 2003) (Fig. 1B). Partial homology was also found with the SMC family of proteins, which are core components of the cohesin and condensin complexes that are required for chromosome movement (Jessberger, 2002). Moreover, we detected homology to Uso1, a protein required for endoplasmic-reticulum-to-Golgi vesicular transport in Saccharomyces cerevisiae (Sapperstein et al., 1996). A common functional feature of these proteins is their involvement in the dynamic movement of subcellular components. This suggests that Mcp6 might also be involved in subcellular dynamics.

To examine the meiosis-specific transcription of mcp6⁺, we performed northern blot analyses of RNA obtained from CD16-1 (h⁺/h^-) and CD16-5 (h^-/h^-) cells harvested at various times after the induction of meiosis by nitrogen starvation. In this experiment, we took advantage of the fact that the heterozygous CD16-1 strain initiates meiosis upon nitrogen starvation, whereas the homozygous CD16-5 strain does not. This analysis revealed that mcp6⁺ displays meiosis-specific transcription that peaks at the horsetail phase (6 hours after induction), which is when homologous chromosome pairing and recombination occur (Fig. 1C).

When the Mcp6 protein appears during meiosis was then assessed by western blot analysis. To attain synchronous meiosis, we used the pat1-114 temperature-sensitive strain, which enters meiosis in a highly synchronous manner when it is shifted to the restrictive temperature (Iino and Yamamoto, 1985). Thus, pat1-114 mcp6⁺-gfp diploid cells that express Mcp6 protein tagged with GFP were induced to enter synchronized meiosis and their lysates were subjected to western blot analysis with an anti-GFP antibody. As shown in Fig. 1D, the frequency of horsetail nuclei is at a normal level, like the pat1 control (Fig. 3A). Thus, we judged that the function of Mcp6-GFP is intact in this strain. Mcp6-GFP first appeared at the horsetail phase and its expression peaked at 3.5 hours after nitrogen starvation. This is similar to the timing of the production of Meu13 (Nabeshima et al., 2001). These results indicate that Mcp6 is a meiosis-specific protein that is exclusively expressed at the horsetail phase.

We first examined the subcellular localization of Mcp6 by constructing a Mcp6-GFP-expressing strain in the h⁹⁰ genetic background and inducing it to undergo meiosis by nitrogen starvation. As shown in the fluorescent microscope images in Fig. 2A (top), no GFP signal was detected during mitosis. Upon mating, however, the Mcp6-GFP fusion protein appeared as a dot near the edge of the nucleus during the horsetail period of meiosis (Fig. 2A, rows 2, 3). The dot disappeared after the horsetail period and the signal was not detected at meiosis I (MI) or meiosis II (MII). The dot localized with the fluorescence signal of Sad1-DsRed, which is known to localize to the SPB (Fig. 2A). Thus, Mcp6 is expressed only during the horsetail period of meiosis and might localize to the SPB.

Because the SPB and telomeres colocalize at this stage of meiosis, it was not clear whether Mcp6 localizes to the SPB or the telomeres. Thus, we expressed Mcp6-GFP (from the nmt1 promoter in an expression vector pRGT81) and Sad1-DsRed (from the native promoter of sad1⁺) during mitotic growth, which is when the SPB and telomeres localize to distinct subcellular loci. We confirmed that Sad1-DsRed does not localize with GFP-tagged Taz1, a component of telomeres, during mitosis (Fig. 2B). However, all of the Mcp6-GFP and Sad1-DsRed colocalized to the edge of the nucleus of the mitotic cells (Fig. 2B), which is where the SPB is known to be located. These results indicate that Mcp6 localizes to the SPB.

To determine the role that Mcp6 plays in meiosis, we first examined the meiotic progression of mcp6Δ cells. To attain synchronous meiosis, we used the pat1-114 temperature-sensitive strain again. Thus, homozygous diploid pat1-114 cells were arrested at the G₁ phase by nitrogen starvation and then shifted to the restrictive temperature to induce synchronous meiosis. We then observed over time the number of nuclei in the cells of the pat1-114 strain, whose mcp6⁺ gene is intact, and in the cells of the pat1-114 mcp6Δ mutant. We found that the times at which cells with two or four nuclei peaked were similar for both strains (Fig. 3A).

Notably, almost no pat1-114 mcp6Δ cells (<0.5%) displayed the flat nuclear shape or the non-central position of the nucleus that are characteristic of the horsetail period. This is reminiscent of the description of the cells that bear a mutation in the SBP component Kms1 - that the nuclear shapes of these cells at prophase I of meiosis were aberrant (Shimanuki et al., 1997). Because the oscillatory nuclear movement that normally occurs during the prophase in wild-type meiosis (Chikashige et al., 1994) is impaired in the kms1-null mutant (Niwa et al., 2000), we surmised that deletion of mcp6⁺ would also abolish nuclear migration. Thus, we examined over time the movement of chromosomes in pat1-114 mcp6Δ cells under a microscope. We found that horsetail oscillation at the prophase of meiosis I was indeed abrogated in these cells (Fig. 3B, top). In fact, almost no nuclear movement was observed throughout the prophase of meiosis in any of the pat1-114 mcp6Δ cells that we examined. By contrast, pat1-114 cells displayed the marked nuclear oscillations that characterize this meiotic period (Fig. 3B, bottom).

Because the effects of mating and karyogamy cannot be observed at the restrictive temperature in the azygotic meiosis of h^-/h^-pat1-114 homozygous diploid cells, we also assessed the effect on nuclear oscillation of deleting the mcp6⁺ gene in the h⁹⁰ genetic background. Thus, we subjected these cells expressing Polα-GFP to time-lapse observation under a microscope. As shown in Fig. 3C (top), it is apparent that nuclear oscillation at prophase of meiosis I is also impaired in mcp6Δ cells. Notably, in contrast to pat1-114 mcp6Δ cells, slight movement was observed just after karyogamy, although no apparent movement was detected thereafter. The nucleus does seem to be moving a little in mcp6Δ cells because a trace of the trailing edge of the moving nucleus can be observed (Fig. 3B, white arrowheads). Nonetheless, it is evident that chromosomal movement is largely hampered in mcp6Δ cells compared with the vigorous nuclear movements in wild-type cells that occur several times after karyogamy (Fig. 3C, bottom).

Although it seems that the horsetail period (117 minutes) and the periods from karyogamy to meiosis I (154 minutes) and from karyogamy to meiosis II (195 minutes) in mcp6Δ cells were slightly shorter than those of wild-type cells (136 minutes, 166 minutes and 209 minutes, respectively) (Fig. 3D), these changes were not statistically significant (P>0.05 in Student's t-test). Thus, we conclude that the durations of the horsetail period, meiosis I and meiosis II are almost normal in h⁹⁰ mcp6Δ cells (zygotic meiosis), as they are in the pat1-114 genetic background (azygotic meiosis).

To investigate the requirement of Mcp6 in the process of chromosome pairing, we observed the homologous chromosomal regions in living zygotes during horsetail phase. By integrating a tandem repeat of the Escherichia coli lac operator sequence into the S. pombe genome at the lys1⁺ locus (near the centromere of chromosome I), the fusion gene encoding GFP-LacI-NLS (which binds to the lac operator) was expressed in this strain. Consequently, two homologous loci on the chromosomes were visualized with GFP fluorescence (Nabeshima et al., 2001). In wild-type background, these homologous loci repeatedly associated and dissociated in the horsetail nucleus, oscillating back and forth between the cell poles (Fig. 4A, bottom). By contrast, the paired GFP signals were less frequent in mcp6Δ cells than in wild-type cells (Fig. 4A, top).

In order to quantify the pairing activity during the horsetail phase, we scored the occurrence of paired signals every 5 minutes from karyogamy to the first meiotic nuclear division in 20 live individuals of each strain. In the wild-type strain, the population of cells with paired signals during the initial 45 minutes was about 30% and then increased to about 80%. This level was maintained until 120 minutes (Fig. 4B, blue line). In mcp6Δ cells, there was no significant difference from the wild type in the initial 45 minutes but the subsequent increase in pairing observed in the wild type was completely absent (Fig. 4B, red line). These results indicate that Mcp6 is required for promoting homologous pairing with horsetail movement.

To determine whether, like the other SPB component Kms1, Mcp6 plays a role in meiotic recombination (Niwa et al., 2000), we compared the rates of intergenic and intragenic recombination in mcp6Δ and wild-type strains. We first investigated the crossover recombination of zygotic meiosis by tetrad analysis, which allowed us to measure the genetic distance between leu1 and his2 (Fig. 5A, insets). When the mcp6Δ strain was crossed, the genetic distance between leu1 and his2 was only 12% of the distance obtained when the wild-type strain was crossed (Fig. 5A, left). The genetic distances between lys3 and cdc12 in the mcp6Δ strain was also reduced to 44% of that observed in the wild-type strain (Fig. 5A, right). We also characterized the intragenic recombination of the mcp6Δ strain between two different mutant alleles of ade6 (M26 and 469). When the mcp6 strain was crossed, the frequency of Ade⁺Δ recombinant spores was 19% of the wild-type frequency (Fig. 5B). Thus, it appears that Mcp6 plays a significant role in meiotic recombination.

To characterize the meiotic recombination competency of the mcp6Δ strain further, we examined the ectopic recombination rate. We used a pair of ade6 alleles from different chromosomal loci, one at the natural position of ade6 on chromosome III and the other at the pac1 locus on chromosome II (z7) (Virgin and Bailey, 1998) (Fig. 5C). Notably, the frequency of ectopic recombination between these two loci in mcp6Δ cells was twice (202%) (in the M26 and 469 pair) that of wild-type cells (Fig. 5C). A similar increase in the ectopic recombination rate between M26 and z15 (telomere of chromosome I) has been reported for the kms1Δ strain (Niwa et al., 2000). However, mcp6Δ cells are distinct from kms1Δ cells in that they do not display the abnormal telomere clustering that characterizes kms1Δ cells.

When we observed the spore morphology of mcp6Δ cells, we found that almost all of the spores looked normal (Fig. 5D). Indeed, depletion of Mcp6 caused only about 4% of the asci to display abnormal numbers of ascospores. The spore viability is also similar to that of wild-type cells (Fig. 5E). Thus, we conclude that Mcp6 does not play an important role in spore formation.

In addition to chromosome oscillation at the horsetail phase, telomere clustering is required for the alignment and subsequent association of homologous chromosome arms (Ding et al., 2004). The telomere protein Taz1, whose deletion causes G₂/M-phase DNA-damage-checkpoint delay, chromosome mis-segregation and double-stranded DNA breaks, plays a role in preventing and repairing DNA breaks (Miller and Cooper, 2003) and telomere clustering (Cooper et al., 1998). To determine whether the SPB localization of Mcp6 depends on proper telomere clustering, we prepared taz1 null mutant cells harbouring the integrated mcp6⁺-gfp gene driven by its own promoter. These cells were then induced to enter meiosis by nitrogen starvation and the Mcp6-GFP signal was observed by immunofluorescence. As shown in Fig. 6A,C, the subcellular localization of Mcp6-GFP was almost normal in taz1Δ cells (71 cells were counted in both wild-type and mcp6Δ cells).

We next examined whether SPB localization of Mcp6 is dependent on dynein. To do this, we prepared a dhc1-d3 mutant harbouring the integrated mcp6⁺-gfp gene driven by its own promoter and induced it to enter meiosis by nitrogen starvation. Mcp6-GFP colocalized normally with DsRed-Sad1 to the SPB in dhc1-d3 cells, which suggests that Dhc1 is not required for the proper localization of Mcp6 at the SPB (Fig. 6B). The frequency of cells in which Sad1-DsRed and Mcp6-GFP colocalized to the leading edge of the horsetail nucleus was 94% (15/18) in dhc1-d3 cells, which is almost equal to the frequency in wild-type cells (93%; 32/36) (Fig. 6D).

For proper horsetail movement, it is essential that the telomeres cluster at the SPB after karyogamy and that the telomere/SPB complex migrates on the microtubule that radially extends from the SPB to the cell cortex on the opposite site of the cell. To understand the role of Mcp6 in horsetail movement, we examined the subcellular localization of the SPB components Sad1, Spo15 and Kms1 in mcp6Δ cells. Sad1 is a constitutive component of SPB that is essential for normal bipolar spindle formation (Hagan and Yanagida, 1995), Spo15 is associated with SPBs throughout the life cycle and plays an indispensable role in the initiation of spore membrane formation (Ikemoto et al., 2000), and Kms1 is required for the formation of meiotic prophase-specific nuclear architecture (Shimanuki et al., 1997). To do this, we prepared strains that express the Sad1-dsRed protein and the other GFP-fused components from their own promoters in the mcp6-null genetic background. The cells were induced to enter meiosis by nitrogen starvation and then observed under a fluorescence microscope.

We first investigated the localization of Sad1-dsRed during the horsetail phase and found that, in mcp6Δ cells, 68% (17/25) of the Sad1 signal constituted a single dot at the leading edge of the nucleus (Fig. 7A,D). This is similar to what is observed in wild-type cells (59/82, or 72%, of the Sad1 signal is present as a single dot). These results indicate that Mcp6 is not required for Sad1 localization to the SPB. We next examined the localization of Spo15-GFP and found that most of the Spo15 signal localized with Sad1-DsRed to the SPB in both mcp6Δ and wild-type cells (Fig. 7B,E). We also examined the localization of Kms1-GFP and found that most Kms1-GFP and Sad1-DsRed colocalized to either the SPB or the Sad1 body in both mcp6Δ and wild-type cells (Fig. 7C,F). These results indicate that Mcp6 is not required for the proper organization of SPB architecture.

Telomere clustering near the SPB, which occurs during the prophase of meiosis I, is an essential event for efficient chromosome pairing and cells deficient in the telomere-associated protein Taz1 (taz1Δ) have been reported to have defective telomere clustering, reduced recombination rates, abnormal spore formation and reduced spore viability (Nimmo et al., 1998; Cooper et al., 1998). To determine whether telomere clustering is normal in mcp6Δ cells, we examined the subcellular localization of the telomere proteins Taz1 and Swi6 by preparing h⁹⁰ taz1⁺-gfp sad1⁺-dsred and h⁹⁰ swi6⁺-gfp sad1⁺-dsred strains, which express Sad1-dsRed together with GFP-tagged Taz1 or Swi6 from their own promoters. These strains were induced to enter meiosis by nitrogen starvation and then observed under a fluorescence microscope. The Taz1-GFP signal of the telomere at the leading edge of the nucleus localized with the Sad1-DsRed signal to the SPB in 62% (23/37) and 58% (14/24) of the mcp6Δ and wild-type cells, respectively (Fig. 8A,D). The Swi6-GFP signal also localized with the Sad1-DsRed signal to the SPB at similar levels in mcp6Δ (76%; 13/17) and wild-type cells (85%; 35/41) (Fig. 8B). These results indicate that the subcellular localization of Taz1 and Swi6 is normal in mcp6Δ cells - namely, Mcp6 is required for neither SPB organization nor telomere clustering.

The oscillatory nuclear movement is mediated by dynamic reorganization of astral microtubules originating from the SPB. To observe the organization of microtubules in mcp6Δ cells, we prepared the h⁹⁰ strain that expresses GFP-fused α-tubulin from the nmt1 promoter. After 6 hours' induction of meiosis in EMM2-N medium, the strain was fixed with glutaraldehyde and paraformaldehyde for immunostaining. In wild-type cells, 95% (156/165) of astral microtubules originated from the SPB during the horsetail phase. In mcp6Δ cells, however, collapse of astral microtubule organization was observed (Fig. 8C,F): 21% (21/100) of mcp6Δ cells displayed the microtubules not associated with SPB [Fig. 8F, abnormal (i)]. We also found that 60% (60/100) of mcp6Δ cells showed abnormal astral microtubules that originated from the SBP but branched elsewhere [Fig. 8F; abnormal (ii)]. These results indicate that Mcp6 is required for proper astral microtubule positioning during horsetail phase.

In the present study, we show that Mcp6 is a novel coiled-coil protein that is only expressed during the horsetail period of meiosis (Fig. 1) and localizes to the SPB (Fig. 2). We found that the deletion of mcp6⁺ almost abolished horsetail movement of chromosomes (Fig. 3) and reduced recombination rates (Fig. 5). Notably, we observed that, whereas deletion of mcp6⁺ from the homozygous diploid pat1 genetic background completely abolished horsetail movement during azygotic meiosis of these cells (Fig. 3B), a slight chromosome movement after karyogamy was observed during the zygotic meiosis of mcp6Δ cells in the h⁹⁰ genetic background (Fig. 3C). This indicates that karyogamy is important for initiating horsetail movement and suggests that Mcp6 plays a role during or just after karyogamy. BLAST-based homology searches failed to identify proteins in other species that are homologous to Mcp6. Thus, Mcp6 is an S.-pombe-specific protein. This is reasonable because S. pombe is the only organism examined so far that displays horsetail movement of nucleus.

Two S. pombe mutants [kms1-1 (Shimanuki et al., 1997; Niwa et al., 2000) and dhc1 mutants (Yamamoto et al., 1999)] have been reported to lack horsetail movement. Another mutant (dlc1Δ) also shows abnormal horsetail movement (Miki et al., 2002). Although Kms1, which is also an S.-pombe-specific protein, localizes to the SPB throughout mitotic and meiotic phases, abnormal phenotypes of kms1-1 cells are only detected in meiosis (Niwa et al., 2000). Dhc1, the dynein heavy chain that is conserved among various species, localizes to the SPB, microtubules and cell cortex, and is predominantly expressed from karyogamy through to meiosis I (Yamamoto et al., 1999). Thus, Mcp6 is a new member of this group of SPB- and horsetail-movement-associated proteins. In support of this, mcp6Δ cells show similar phenotypes to the kms1-1, kms1Δ, dhc1 and dlc1Δ mutant cells as follows. First, like kms1, dhc1 and dlc1Δ mutants, the rates of intergenic or intragenic recombination of the mcp6Δ mutant are markedly reduced (Fig. 5A,B). Second, the ectopic recombination rate of mcp6Δ cells is increased about twofold (Fig. 5C), as has also been observed for the ectopic recombination of kms1Δ cells between ade6-M26 and ade6-469 (z15) loci, although the rate of ectopic recombination between the ade6-M26 and ade6-469 (z7) loci in kms1Δ cells was almost the same as that in wild-type cells (Niwa et al., 2000).

However, Mcp6 does have some phenotypes that differ from those of kms1, dhc1 and dlc1Δ mutants, and thus it seems to play a distinct role in maintaining the horsetail movement. One such phenotype is the almost normal spore formation and spore viability in mcp6Δ cells (Fig. 5D,E). By contrast, spore formation and spore viability are abnormal in kms1Δ, dhc1Δ and dlc1Δ cells. Second, unlike kms1Δ cells (Shimanuki et al., 1997), telomere clustering, as determined by examining the subcellular localization of Taz1-GFP and Swi6-GFP, appears almost normal in mcp6Δ cells (Fig. 8). Third, although kms1Δ and dhc1-d3 cells showed reduced SPB integrity and abnormal chromosome segregation, respectively, the mcp6Δ mutant does not show such abnormalities (Figs 6, 7). Fourth, Mcp6 is the only meiosis-specific protein of this group, because Kms1, Dhc1 and Dlc1 are also detected in both the mitotic and meiotic phases (Goto et al., 2001; Miki et al., 2002).

We report here that recombination rates are greatly reduced in mcp6Δ cells compared with wild-type cells (Fig. 5), which is due primarily to the inefficient homologous pairing of chromosomes in mcp6Δ cells (Fig. 4). This inefficient homologous pairing is mainly derived from the impaired horsetail movement of nucleus (Fig. 3). We previously showed that meu13Δ and mcp7Δ cells show a delay in entering meiosis I, and that this is due to the meiotic recombination checkpoint that provides cells with enough time to repair double-strand breaks (DSBs) in a manner dependent on checkpoint rad⁺ genes (Shimada et al., 2002; Saito et al., 2004). In the case of mcp6Δ cells, however, even though recombination rates were reduced, a delay in entering meiosis I was not observed (Fig. 3A). Thus, we surmise that Mcp6 is not involved in DSB repair like other SPB components, probably because Mcp6 is located at the SPB, a location that is too remote directly to regulate DSBs on chromatin. Similarly, Mcp6 might not directly regulate the recombination machinery. Thus, it is most probable that Mcp6 is involved in regulating the pairing of homologous chromosomes by inducing proper horsetail movement.

Some mutants (lot2-s17, lot3-uv3/taz1, dot1 and dot2) show defective horsetail movement and reduced meiotic recombinations, but their phenotypes are distinct from mcp6Δ. For example, lot2-s17 and lot3-uv3/taz1 mutants display a dramatic lengthening of telomeric repeats, low spore viability and chromosome mis-segregation through meiosis (Nimmo et al., 1998; Cooper et al., 1998; Hiraoka et al., 2000). Two mutants (dot1 and dot2) do not sporulate and show defective SPB integrity and impaired telomere clustering (Jin et al., 2002). By contrast, we show here that mcp6Δ differs from these mutants because sporulation (Fig. 5D,E), SPB integrity (Fig. 7) and telomere clustering (Fig. 6) are almost normal.

Some mutant strains of taz1 that show abnormal horsetail movement (Hiraoka et al., 2000) are also distinct from mcp6Δ cells. First, telomere clustering is abnormal in taz1Δ cells. Second, taz1Δ cells display abnormal spore formation and reduced spore viability. Third, although the subcellular localization of Taz1-GFP is normal in mcp6Δ cells (Fig. 8A), Mcp6-GFP localization became slightly abnormal in taz1Δ cells (Fig. 6A,C). Thus, Mcp6 might interact with Taz1 but its role in chromosome maintenance during meiosis is quite distinct.

Finally, microtubules are nucleated exclusively from SPBs immediately after karyogamy and form typical X-shaped configurations during nuclear fusion of meiotic cells (Svoboda et al., 1995; Yamamoto et al., 1999). However, we report here that the astral microtubule organization was largely abnormal in mcp6Δ cells (Fig. 8). Thus, in the absence of Mcp6, many cells fail to organize a long, curved microtubule array that extends from the cell ends to provide the tracks for nuclear horsetail movement, resulting in the abolished nuclear oscillation and reduced chromosome pairing. Taken together, we conclude that Mcp6 controls horsetail movement by regulating the astral microtubule organization during meiosis.

We thank O. Niwa, Y. Hiraoka, A. Yamamoto, C. Shimoda, M. Yamamoto, G. R. Smith, K. M. Miller and W. Z. Cande for providing S. pombe strains, and O. Niwa for the anti Sad1 antibody. We also thank K. Nabeshima for technical advice in the experiments of chromosome pairing, E. Okamura for technical assistance, and P. Hughes for critically reading the manuscript. This work was supported by a Grant-in-aid for Scientific Research on Priority Areas from the Ministry of Education, Science, Sports and Culture of Japan and a grant from the Uehara Memorial Foundation to H.N.