Dissecting the first and the second meiotic divisions using a marker-less drug-hypersensitive fission yeast

Abstract

Faithful chromosome segregation during meiosis is indispensable to prevent birth defects and infertility. Canonical genetic manipulations have not been very useful for studying meiosis II, since mutations of genes involved in cell cycle regulation or chromosome segregation may affect meiosis I, making interpretations of any defects observed in meiosis II complicated. Here we present a powerful strategy to dissect meiosis I and meiosis II, using chemical inhibitors in genetically tractable model organism fission yeast (Schizosaccharomyces pombe). As various chemical probes are not active in fission yeast, mainly due to an effective multidrug resistance (MDR) response, we have recently developed a drug-hypersensitive MDR-sup strain by suppression of the key genes responsible for MDR response. We further developed the MDR-supML (marker-less) strain by deleting 7 MDR genes without commonly used antibiotic markers. The new strain makes fluorescent tagging and gene deletion much simpler, which enables effective protein visualization in varied genetic backgrounds. Using the MDR-supML strain with chemical inhibitors and live cell fluorescence microscopy, we established cell cycle arrest at meiosis I and meiosis II and examined Aurora-dependent spindle assembly checkpoint (SAC) regulation during meiosis. We found that Aurora B/Ark1 kinase activity is required for recruitment of Bub1, an essential SAC kinase, to unattached kinetochore in prometaphase I and prometaphase II as in mitosis. Thus, Aurora’s role in SAC activation is likely conserved in mitosis, meiosis I, and meiosis II. Together, our MDR-supML strain will be useful to dissect complex molecular mechanisms in mitosis and 2 successive meiotic divisions.

Introduction

Meiosis consists of 2 rounds of cell division or chromosome segregation following a single round of DNA replication, which leads to generation of haploid gametes from diploid germ cells. Understanding molecular mechanisms during meiosis is important, since missegregation of chromosomes during meiosis links to birth defects, such as Down syndrome and infertility. While homologous chromosomes are bi-oriented and segregate to opposite poles during meiosis I, sister chromatids are bi-oriented during meiosis II, as they are in mitosis. However, whether molecular mechanisms to ensure faithful chromosome segregation, such as the spindle assembly checkpoint (SAC), are conserved or diverged among mitosis, meiosis I, and meiosis II is poorly understood, mainly due to lack of proper approaches to dissect the first and the second meiotic division in genetically tractable model organisms.

The SAC is a surveillance mechanism to ensure that all chromosomes are properly attached by the microtubules emanated from the opposite spindle poles by sensing unattached kinetochores and preventing the metaphase–anaphase transition.¹ The conserved SAC proteins (Mad1, Mad2, Mad3/BubR1, Bub1, Bub3, and Mps1/Mph1) are recruited to unattached kinetochores to activate the SAC. In addition to these factors, Aurora B kinase is another regulatory enzyme of SAC signaling. Aurora B contributes to activation of the SAC in mitosis by at least 2 mechanisms.² First, Aurora B creates unattached kinetochores by destabilizing kinetochore–microtubule attachment, which indirectly activates the SAC. Second, Aurora B recruits all SAC components to kinetochores at least in fission yeast and human cells.³ However, the role of Aurora kinase in SAC activation during meiosis, especially the second meiotic division, has been poorly understood. Canonical temperature-sensitive genetic mutants have not been very useful, since high or low temperature severely affects meiotic progression in yeast.^4,5 In addition, as mechanisms of cell cycle progression are similar between meiosis I and II, mutations of genes involved in cell cycle regulation or chromosome segregation may affect meiosis I, making interpretations of any observed perturbations on meiosis II complicated.

We believe that the use of chemical inhibitors, which will allow acute inhibition in genetically tractable model organisms such as fission yeast (Schizosaccharomyces pombe), should help examine molecular mechanisms of Aurora-dependent SAC regulation in meiosis for at least 4 reasons: (1) tubulin poison, such as nocodazole, can be used to generate unattached kinetochores; (2) proteasome inhibitors, such as Velcade (Bortezomib), can be used to prevent metaphase–anaphase transition even when the SAC is inactivated by Aurora kinase inhibition; (3) off-target effects of Aurora inhibitors can be excluded by comparing inhibitor-sensitive and -resistant isogenic cells; (4) meiotic cell division can be induced by depleting nitrogen source from media. Since fission yeast has robust multidrug resistance (MDR) mechanisms, and use of these chemical inhibitors had been limited, we have recently constructed the drug-hypersensitive (MDR-sup) fission yeast strain in which 2 ABC transporters (Bfr1 and Pmd1), 2 MFS transporters (Mfs1 and Caf5), and a transcription factor (Pap1) were deleted. We have shown that a series of chemical inhibitors, including nocodazole and Velcade, are active in this strain.⁶ In addition, we have identified a specific chemical inhibitor Arkin-1 for fission yeast Aurora kinase Ark1, and characterized a resistance-conferring point mutation (G172D) in Ark1.⁷ Therefore, these tools must be useful to analyze Aurora-dependent SAC regulation in mitosis and meiosis. However, further genetic manipulation, such as fluorescent tagging or gene deletion, in the MDR-sup strain was challenging, since commonly used antibiotic markers, such as kanMX6, hphMX6, and natMX6, were already used.

In this paper, we constructed a new drug-hypersensitive strain without using the commonly used antibiotic markers, which enables efficient fluorescent tagging or gene deletion of SAC proteins selected by these antibiotic marker cassettes. Using this strain and a series of chemical inhibitors, we established M-phase arrest in meiosis I and meiosis II and examined Aurora kinase-dependent regulation of SAC proteins during meiosis.

Results

The erg5 and dnf2 mutations in the MDR-sup strain

In order to develop a novel drug-hypersensitive fission yeast strain in which commonly used antibiotic resistance markers, such as G418-resistant gene (kan), hygromycin B-resistant gene (hph), and nourseothricin-resistant gene (nat), are available, we constructed bfr1, pmd1, pap1, and mfs1 gene deletions without using antibiotic marker cassettes. Each ORF was first deleted by PCR-mediated gene disruption using ura4⁺ auxotrophic marker and then ura4⁺ cassette was deleted by using amplified DNA fragments containing flanking sequences of the ORF. The colonies in which ura4⁺ cassette was deleted were selected on plates containing 5-Fluoroorotic acid (FOA) (Fig. S1A). The caf5 gene was deleted by PCR-mediated gene disruption using less used bsd (Blasticidin S deaminase gene) marker, and bsd marker was not removed for easily selection of the caf5 deletion. Then the 5 genes-deleted strain (5Δ) was constructed (Fig. S1B and C, see “Material and Methods” for the detail of construction). As predicted, the 5Δ strain showed drug sensitivity to 5 μM brefeldin A (BFA) or 2 μM cycloheximide (CHX), while growth of the 5Δ strain is comparable to WT on a plate without compounds (Fig. S1C). Unexpectedly, drug sensitivity of this marker-less 5Δ strain was weaker than that of the original MDR-sup strain (Fig. S1C), suggesting that difference of genetic background between the 5Δ strain and the MDR-sup strain may affect the strength of MDR mechanisms. To address this hypothesis, we crossed the 5Δ strain and the MDR-sup strain, and selected marker-less clones that show sensitivity to G418, hygromycin B (HB), and nourseothricin (NAT), and resistance to Blasticidin S (BS) after random spore analysis. Interestingly, some of colonies that were selected by these criteria showed sensitivity on plates containing 2 μM BFA, and others did not. Therefore, we selected two 2-µM BFA-sensitive clones (clone #1 and #2, Fig. S1D) and compared drug sensitivity with the MDR-sup cells. Both marker-less clone #1 and #2 showed comparable drug sensitivity to the MDR-sup cells on plates containing BFA or CHX (Fig. S1C). These data indicate that genetic mutation(s) unique to either the 5Δ strain or the MDR-sup strain contribute to MDR mechanisms.

In order to identify genetic mutations that affect the MDR mechanisms, we sequenced the whole genome of the 5Δ strain, the MDR-sup strain, clone #1, and clone #2. We found that several missense and 2 frameshift mutations were shared among the MDR-sup strain, clone #1, and clone #2, but not with the 5Δ strain (Table S2). Since frameshift mutation produces a truncated protein that often loses its function, we suspected that the genes with these frameshift mutations are what are involved in MDR mechanisms, and this notion turned out to be the case as described below. The mutations are on the erg5 gene and an uncharacterized SPAC24B11.12c gene (refer as the dnf2 gene below, since it shows high sequence homology to the Dnf2 gene in Saccharomyces cerevisiae). We confirmed these frameshift mutations by direct sequencing. A nucleotide insertion in the erg5 gene produces a truncated Erg5 (1–268) protein that is almost half the size of full-length Erg5 (1–543) (Fig. S1E and F). A nucleotide insertion in the dnf2 gene produces truncated Dnf2 (1–130) protein that is almost 10 times smaller than full-length Dnf2 (1–1402) (Fig. S1E and F).

Erg5 and Dnf2 are required for MDR mechanisms in fission yeast

Erg5 is a C-22 sterol desaturase, one of the enzymes that catalyze a sequence of reactions from zymosterol to ergosterol. As it was reported that the deletion mutant of the erg5 gene showed sensitivity to CHX and staurosporine,⁸ it seems reasonable that the erg5 mutation increases drug sensitivity. Consistent with this, deletion of the erg5 gene by ura4⁺ cassette further increased drug sensitivity in the 5Δ strain (Fig. 1A). It is predicted that Dnf2 is a P4-ATPase, though it has not been characterized in fission yeast. S.cerevisiae has 5 genes encoding P4-ATPases (DNF1, DNF2, DNF3, NEO1, and DRS2),⁹ and all these genes are conserved in S. pombe. While only NEO1 is an essential gene in S. cerevisiae, only Drs2 is an essential gene in Schizosaccharomyces pombe (fission yeast).¹⁰ To examine whether Dnf2 is required for MDR response in fission yeast, we constructed the dnf2 deletion strain. We found that deletion of the dnf2 gene further increased drug sensitivity in the 5Δ strain, while deletion of other non-essential P4-ATPases (dnf1, dnf3, and neo1) did not (Fig. 1A). Taken together, we concluded that both Erg5 and Dnf2 are required for MDR response in fission yeast. During these analyses, we realized that deletion of the erg5 gene by kanMX6 cassette in ura4-D18 background slightly compromised growth even in the absence of chemical inhibitors (Fig. S1G), suggesting that uracil or uridine permeability might be reduced in erg5Δ cells.

Figure 1. Erg5 and Dnf2 are redundantly required for MDR mechanisms in fission yeast. (A) Serial dilutions of the indicated strains were spotted onto YE4S plates, or YE4S plates containing brefeldin A (2 μM) or cycloheximide (2 μM), and incubated at 29 °C. (B) Exponentially growing culture (OD = 0.5) of the indicated strains were diluted 50× in YE4S medium, treated with the indicated compounds at the indicated concentrations (μM), and incubated for 17 h at 29 °C. Growth (%) is presented relative to DMSO-treated cells. See also Figure S1C.

To examine whether Erg5 and Dnf2 regulate MDR mechanisms in the same pathway or not, we constructed the 5Δ dnf2Δ erg5∆ strain and compared sensitivity to CHX, BFA, or Velcade with the 5Δ dnf2Δ or the 5Δ erg5Δ strain. We found that the 5Δ dnf2Δ erg5Δ strain showed higher drug sensitivity to all 3 compounds than the 5Δ dnf2Δ or the 5Δ erg5Δ strain, indicating that Erg5 and Dnf2 cooperatively regulate MDR (Fig. 1B). As drug sensitivity of the 5Δ dnf2Δ erg5Δ strain was completely comparable to clone #1 (5Δ dnf2* erg5*), both frameshift mutations may lead to complete loss of function of these protein (Fig. 1B).

Constructing the MDR-supML strain

Based on above knowledge, we designed a new, marker-less drug-hypersensitive fission yeast, MDR-supML (for marker-less) (Fig. 2A). To construct the MDR-supML strain, first the dnf2 gene was deleted by marker-less method (Fig. S1A) in the 5Δ strain. Second, the erg5 gene was deleted by ura4⁺ cassette, and ura4⁺ cassette was not removed to keep this strain ura⁺ for normal growth (see above). The drug sensitivity of the MDR-supML strain to CHX, BFA, or Velcade was completely comparable to the original MDR-sup strain (Fig. 2B). When we cross the MDR-supML strain and a wild-type-based ura^- strain to isolate a MDR-supML-based strain, we need to select 5 marker-less gene deletions (bfr1-del, pmd1-del, pap1-del, mfs1-del, and dnf2-del) in addition to erg5::ura4⁺ or caf5::bsd, which can be selected on plates lacking uridine or containing Blasticidin S, respectively. As the MDR-supML strain is not able to grow on plate containing 2 μM BFA, we first selected BFA-sensitive, blasticidine S-resistant, and ura⁺ colonies. Then we used PCR-based genotyping to confirm 5 marker-less gene deletions in selected BFA-sensitive colonies. In order to allow a quick genotyping by PCR, we designed PCR primers (MIX1 and MIX2), for which 5 gene deletions can be checked by only 2 PCR reactions (Fig. 2C; Table S3). As the MDR-supML strain showed sensitivity similar to the MDR-sup strain for Velcade, we examined whether Velcade treatment also shows metaphase arrest in the MDR-supML strain. To visualize cell cycle progression, we constructed the MDR-supML strain in which Atb2, α-tubulin, was tagged with GFP at N-terminus with kanMX6 cassette, and Sid4, which constitutively localizes at spindle pole bodies (SPBs), was tagged with mCherry with hphMX6 cassette (scheme of strain construction is summarized in Fig. 2D). Consistent with the previous observation using the MDR-sup strain,⁶ Velcade treatment in the MDR-supML fission yeast cells also showed accumulation of typical metaphase-arrested cells with separated SPBs, short spindles, and condensed chromosomes in a dose- and time-dependent manner (Fig. 2E).

Figure 2. Constructing the “MDR-supML” strain. (A) The list of marker cassettes for deleting 7 genes (bfr1, pmd1, pap1, mfs1, caf5, erg5, and dnf2) in the original MDR-sup strain (MDR-sup) and the MDR-supML strain (ML). Asterisks indicate the frameshift mutations shown in Figure S1E and F. Minus (“−”) indicates gene deletion without using any markers. (B) Exponentially growing culture (OD = 0.5) of the indicated strains were diluted 50× in YE4S medium, treated with the indicated compounds at the indicated concentrations (μM), and incubated for 17 hours at 29°C. Growth (%) is presented relative to DMSO-treated cells. (C) Schematic of strategy for checking marker-less deletions by colony PCR using oligo MIX1 and MIX2 listed in Table S3. The DNA bands amplified from wild-type (WT) cells, and MDR-supML (ML) cells are shown. (D) Construction flow of the GFP-atb2 sid4-mCherry MDR-supML strain. First, h⁹⁰ MDR-supML strain was crossed with h⁹⁰ GFP-atb2-kan strain to construct h⁹⁰ MDR-supML GFP-atb2-kan strain. G418- and BS (Blasticidin S)-resistant, Ura⁺, and BFA (brefeldin A)-sensitive clones were selected after random spore analysis. The marker-less 5-gene deletion was confirmed by colony PCR as shown in (C). Second, h⁹⁰ MDR-supML sid4-mCherry-hph strain was constructed from h⁹⁰ MDR-supML strain by gene targeting. Third, above 2 strains were crossed, and G418- and HB-resistant clones were selected to construct h⁹⁰ MDR-supML sid4-mCherry-hph GFP-atb2-kan strain. (E) The h⁹⁰ MDR-supML sid4-mCherry-hph GFP-atb2-kan cells were synchronized at G₁/S phase by hydroxyurea (HU), and then released from G₁/S by washing HU out (0 min). Velcade (40 μM or 8 μM) or DMSO was added at 30 min after release. The graph shows the percentage of metaphase cells at the indicated time after release. Representative image of metaphase-arrested MDR-supML cells treated by 40 μM Velcade at 120 min (indicated by asterisk in the graph) was shown. Scale bars, 10 μm.

Establishment of cell cycle arrest at meiosis I and meiosis II

We have constructed the homothallic (h⁹⁰) MDR-supML strain that can undergo mating and meiotic divisions in nitrogen-depleted conditions. This strain enables to compare molecular mechanisms of interest in both mitosis and meiosis using the same strain. First, we examined whether cell cycle is arrested at metaphase I or metaphase II in Velcade-treated zygotic MDR-supML cells as observed in mitotic cells. Live imaging of GFP-Atb2 and Sid4-mCherry indicated that most of Velcade-treated zygotes were arrested at metaphase I or metaphase II for at least 100 min, while metaphase–anaphase transition occurred in 30 min after SPB separation in untreated zygotes (Fig. 3A and B; Fig. S2). Second, we examined whether spindle microtubules can be depolymerized by nocodazole treatment in metaphase I- and metaphase II-arrested cells by Velcade. Addition of nocodazole resulted in acute depolymerization of the microtubule cytoskeleton and reduced distance between SPBs as observed in mitotic cells (Figs. 3C and D; Fig. S2).⁶ In contrast, SPBs were separated when meiosis II zygotes were treated with only nocodazole (Fig. S3), which indicates microtubule-independent SPB separation in meiosis II as reported recently.¹¹ These data suggest that inhibition of proteasome by Velcade may suppress mechanisms that regulate microtubule-independent SPB separation in meiosis II. Together, we anticipated that this system is useful to study molecular mechanisms of SAC signaling that respond to unattached kinetochore in meiosis I and meiosis II.

Figure 3. Establishment of cell cycle arrest at meiosis I and meiosis II using Velcade and nocodazole in the “MDR-supML” strain. (A–D) Live-cell imaging of the MDR-supML cells expressing GFP-Atb2 (microtubules; green) and Sid4-mCherry (SPBs; red). (A and B) Time-lapse kymographs of spindles during meiosis I (A) and meiosis II (B) in the presence (right) or absence (left) of proteasome inhibitor Velcade (40 µM). To arrest cells at metaphase I or metaphase II, Velcade was added to medium 20 min before SPB separation in meiosis I or meiosis II, respectively. The length of arrows corresponds to 10 min. The shape of cells is outlined in dotted curves and shown above. For meiosis II cells in (B and D), the rectangle regions of zygotes are shown enlarged as kymographs. (C and D) Microtubule depolymerization at metaphase I (C) and metaphase II (D) by nocodazole. Cells arrested at metaphase I or metaphase II by Velcade were subsequently treated with nocodazole (15 µM), and then arrested at prometaphase I (C) or prometaphase II (D) with no microtubules. Arrowheads indicate the timing of nocodazole addition. The length of arrows corresponds to 5 min. Scale bars, 2 µm.

Examining Aurora-dependent SAC regulation during meiosis

It has been shown that SAC-dependent metaphase I delay caused by univalents in recombination-deficient rec12Δ cells was cancelled by Ark1 inhibition during meiosis I in fission yeast,¹² suggesting that Ark1 is involved in SAC activation during meiosis I as in mitosis. However, there remain 2 possibilities to explain this result. First, Ark1 creates unattached kinetochores by destabilizing kinetochore–microtubule attachment on univalents, which indirectly activates the SAC. This function of Ark1 during meiosis has been already reported.¹³ Second, Aurora B recruits essential SAC components to kinetochores as in mitosis,³ although this possibility has not been examined, mainly due to lack of proper experimental approaches. Taking advantages of our experimental system using the MDR-supML cells, chemical inhibitors, and live cell fluorescence microscopy (Fig. 3), we examined whether Ark1’s kinase activity is required for kinetochore localization of Bub1, an essential SAC component, during meiosis I and II. We first treated zygotic cells with Velcade to inhibit metaphase–anaphase transition, then added nocodazole to generate unattached kinetochores. As predicted, strong accumulation of Bub1-GFP signals on unattached kinetochores were observed in response to nocodazole treatment in both metaphase I- and metaphase II-arrested cells (Figs. 4A and B). In several minutes after adding Arkin-1, a specific Ark1 inhibitor,⁷ Bub1-GFP signals were gradually reduced to approximately half, while Plo1-tdTomato signals remained to be strong in both metaphase I- and metaphase II-arrested cells (Figs. 4A, B, E, and F). Delocalization of Bub1-GFP by Arkin-1 treatment was not observed in Arkin-1-treated ark1-G172D zygotes (Figs. 4C–F), excluding the off-target effects of Arkin-1. Together, these data indicate that Ark1-dependent regulation of Bub1 localization on unattached kinetochores is likely conserved among mitosis, meiosis I, and meiosis II.

Figure 4. Examining Ark1-dependent SAC regulation during meiosis I and meiosis II in the “MDR-supML” strain. (A and B) Live-cell imaging of the MDR-supML cells (WT) expressing Bub1-GFP (green) and Plo1–tdTomato (red) in meiosis I (A) and meiosis II (B). Cells arrested at metaphase I or metaphase II by Velcade (40 µM) were subsequently treated with nocodazole (15 µM) to activate the SAC with no microtubules. When Bub1-GFP was strongly accumulated at kinetochores (0 min), cells were treated with Arkin-1 (7.5 µM) to inhibit Ark1 kinase activity. Bub1-GFP levels were dropped in response to Arkin-1. The boxed regions of zygotes are shown enlarged as time-lapse images. The shape of cells is outlined in dotted curves. (C and D) Arkin-1 treatment to the Arkin-1-resistant ark1-GD strain as (A and B). Time-lapse images of zygotic ark1-GD MDR-supML cells in meiosis I (C) and meiosis II (D) are shown. Bub1–GFP levels on kinetochores were maintained after Arkin-1 addition. The boxed regions of zygotes are shown enlarged as time-lapse images. The shape of cells is outlined in dotted curves. (E and F) The signal intensity of Bub1-GFP in Arkin-1 treatment, as shown in (A–D), was quantified in meiosis I (E) and meiosis II (F). Two representative data of Arkin-1 treatment to WT and ark1-GD, and DMSO treatment to WT are shown. Arrowheads indicate the timing of drug addition. Scale bars, 2 µm.

Discussion

In this study, we constructed a new drug-hypersensitive fission yeast strain (MDR-supML) in which commonly used antibiotic markers are available for further fluorescent tagging or gene deletions. We believe that this strain is useful for efficient chemical screen, target identification, and chemical biology study for dynamic cellular mechanisms.

To increase drug sensitivity at maximum level so far, gene deletion of erg5 and dnf2 is required in addition to that of ABC transporter (bfr1, pmd1), MFS transporter (mfs1, caf5), and transcription factor (pap1). While it has been reported that Erg5 is required for MDR mechanisms,⁸ Dnf2 has not been characterized in fission yeast. We found that Dnf2 and Erg5 were redundantly required for drug resistance to at least 3 distinct chemical inhibitors (cycloheximide, brefeldin A, and Velcade) in fission yeast. Consistent with the role of Dnf2 for MDR response, previous microarray analysis showed that expression levels of the dnf2 gene were induced in response to Purvalanol A, an inhibitor of cyclin-dependent kinases.⁶ Interestingly, deletions of other P4-ATPase genes (dnf1, dnf3, and neo1) did not show drug sensitivity. Therefore Dnf2 may have a specific role for MDR response among P4-ATPases in fission yeast. While P4-ATPase has been originally identified as an aminophospholipid translocase, recent studies of individual P4-ATPase family members have shown that P4-ATPases differ in their substrate specificities and mediate transport of a broader range of lipid substrates.¹⁴ Our data indicate that Dnf2 may recognize and transport not only lipid substrates, but also a wide range of chemical inhibitors, including commonly used chemical probes.

Our MDR-supML strain with chemical inhibitors and live cell fluorescence microscopy is a useful approach to dissect the first and the second meiotic divisions, which occur sequentially after one round of DNA replication. Using Velcade and nocodazole, we established cell cycle arrest at meiosis I and meiosis II, which is an optimal experimental system to study molecular mechanisms of SAC signaling that respond to unattached kinetochore in meiosis I and meiosis II. In addition, the MDR-supML strain could be useful in combination with a series of analog-sensitive alleles of protein kinases,¹⁵ since much lower concentrations of the expensive inhibitors would be required for kinase inhibition. Especially, an ATP analog-sensitive Pat1 protein kinase, which allows synchronization of meiosis at physiological temperature,^4,16 can be combined with the MDR-supML strain to accumulate meiosis I- or meiosis II-arrested zygotes for biochemical analysis.

We treated meiosis I- or meiosis II-arrested cells with Arkin-1 to examine Ark1’s role in SAC activation in both meiotic divisions. We found that Ark1’s kinase activity is required for recruitment of Bub1 to unattached kinetochores in prometaphase I- and prometaphase II-arrested cells. Since Ark1 is also required for Bub1 localization on unattached kinetochore and SAC activation in mitosis,³ our data suggest that Ark1-dependent regulation of SAC activation is likely conserved among mitosis, meiosis I, and meiosis II. In mitosis, kinetochore localization of Bub1 depends on phosphorylation of kinetochore protein Spc7 (KNL1/Blinkin in mammals) by Mps1/Mph1 kinase.^17-19 Therefore, it needs to be addressed whether phosphorylation of Spc7 by Mph1 recruits Bub1 on kinetochores during meiosis. We anticipate that further comprehensive analysis of SAC proteins during meiosis may reveal conserved or diverged aspects of SAC regulatory network between mitosis and meiosis. In meiosis I, when kinetochore geometry is different from mitosis or meiosis II,²⁰ additional kinases (e.g., Plo1 kinase²¹) and phosphatases (e.g., Protein phosphatase 2A²²) are specifically enriched at kinetochores. Thus, it will be interesting to address whether phospho-regulation of SAC signaling at the first meiotic division may be fine-tuned by a different set of kinases and phosphatases from those in mitosis or the second meiotic division.

Materials and Methods

Strains, media, and chemical inhibitors

All strains used are listed in Table S1. Standard growth conditions and methods were used.²³ All experiments except live imaging for meiotic cells were performed in yeast extract (YE) medium containing adenine, leucine, uridine, and histidine. The PCR-based gene targeting²⁴ was used to construct gene disruptants and fluorescent protein-tagged strains with selection marker gene cassettes. Information of chemical inhibitors were described previously.⁷

Construction of the bfr1-del pmd1-del pap1-del mfs1-del caf5::bsd (5Δ) strain

See Figure S1A for a schematic. The bfr1⁺ gene of the S. pombe strain JY878 (h⁹⁰ ade6-M216 leu1 ura4-D18) was deleted with the ura4⁺ gene cassette to yield the bfr1::ura4⁺ strain. Sewing PCR was done to amplify the DNA fragment connecting upstream and downstream sequences outside the bfr1⁺ coding region, and the amplified fragment was then introduced into the bfr1::ura4⁺ strain to excise out the ura4⁺ cassette gene. Colonies that conferred resistance against 5-FOA (5-fluoroorotic acid) were then selected, and colony PCR was performed to confirm the correct gene replacement. The resultant strain is denoted bfr1-del (Fig. S1A). The pmd1⁺ gene of the bfr1-del strain was next deleted with the ura4⁺ cassette, and the resulting pmd1::ura4⁺ strain was remade into pmd1-del in similar ways. Repeating this with pap1⁺ and mfs1⁺ genes led to generation of the 4-deletion (4∆) marker-less strain: bfr1-del pmd1-del pap1-del mfs1-del. The caf5⁺ gene of the strain was then disrupted with the blasticidin S-resistant gene (bsd), to produce the 5∆ strain MJ1682: bfr1-del pmd1-del pap1-del mfs1-del caf5::bsd.

Deep sequencing and analysis

Genome DNA isolated from SAK84 (MDR-sup), MJ1682 (5Δ), SAK550, and SAK551 (clone #1 and #2 in Fig. S1C, respectively) strains were sequenced by HiSeq 2500 (Illumina) following manufacturer's instruction. More than 90 million of single-end reads were obtained for each strain and aligned to Schizosaccharomyces pombe reference genome sequence using bowtie 2 aligner²⁵ with default parameters. More than 94% of the obtained reads were mapped on the reference, and more than 96% of the genome was covered with 3 or more reads, with the average read depths of ~50. The outputs of bowtie 2 aligner were further processed to extract potential SNV (single nucleotide variation) information. First, clonal reads, which mapped at the same genomic position on the same strand, were collapsed and replaced by the read with highest alignment score. Then, reads with low mapping quality score (less than 10) were discarded. Finally, from the alignment result of the retained sequences, difference between the sequenced and reference genome, or potential SNV information, was extracted. When the same SNV candidate was detected at the same genome position twice or more, and the frequency of the candidate detection among the reads covering this locus is above a defined threshold, the detected SNV candidate was identified as a genuine SNV and reported. The threshold used was 0.8 for mismatch detection, while 0.27 for indel to allow sensitive detection of indels in homopolymeric repeats. The SNVs, which (1) are located within protein-coding genes and (2) appear in strains SAK84, SAK550, and SAK551 but not in MJ1682, or appear exclusively in MJ1682, were found at 8 genome positions and listed in Table S2. All scripts used in this analysis are available upon request.

Microscopy for fixed cells

Cells were fixed using methanol, stained with DAPI, imaged by a microscope (Axio Imager 2; Carl Zeiss), and processed with Axio Vision 4.8 software (Carl Zeiss), as described previously.⁶

Live imaging

For live-cell imaging of cells in meiosis, homothallic h⁹⁰ cells were spotted on sporulation agar (SPA) and incubated for 8–12 h at 25 °C. Zygotes were then mounted onto multi-well glass-bottom dishes (Matsunami) precoated with lectin, and the dishes were filled with liquid Edinburgh minimal medium without a nitrogen source. Live-cell imaging was performed on the AF6000 system (Leica). Images were acquired as serial sections along the z-axis at 0.5-µm intervals. 3D deconvolution, z-stack projection, and quantification of signal intensity were performed using LAS AF software (Leica).