Mei3 and Mei2 independently regulate zygotic growth, cell division, lysis, and mating

We previously reported that in addition to regulating meiotic entry, Mei3 and Mei2 also act to prevent zygotic mating and refertilization [4]. Indeed, deletion of either mei3 or mei2 led to refertilization in approximately 1% and approximately 6% zygotes, respectively (Fig 1B and [4]). We also observed that simultaneous deletion of mei3 and mei2 genes increased the incidence of refertilizations to approximately 13% zygotes (Fig 1B and [4]). These results suggested that Mei3 and Mei2, which initiate meiosis via a linear Mei3-Pat1-Mei2 cascade, may also act independently of each other to repress mating in zygotes. We thus set out to dissect the respective roles of Mei2 and Mei3 in blocking zygotic mating.

We carefully analyzed the phenotypes of wild-type, mei3Δ, mei2Δ, and mei3Δ mei2Δ zygotes (Fig 1C and S1 Movie). After fertilization, wild-type zygotes expanded the fusion neck and underwent sporulation. The mutant zygotes also expanded the fusion necks but never sporulated and instead grew mating projections, refertilized and/or lysed. Surprisingly, we also found that one-third of mei2Δ zygotes went through cell division, a phenotype that was virtually absent from mei3Δ and mei3Δ mei2Δ mutant zygotes (Fig 1B and 1C). Two lines of evidence suggest that these mei2Δ zygotes divide mitotically. First, dividing mei2Δ zygotes formed an actomyosin ring: After mei2Δ gametes fused, shown by the entry of cytosolic mCherry from the P- into the M-partner, the myosin light chain Rlc1-sfGFP formed a contractile ring in the zygotes, which subsequently underwent septation (Fig 1D and S2 Movie). Second, dividing mei2Δ zygotes underwent nuclear division using a microtubule spindle, as shown by the nuclear marker Uch2-mCherry [32] and α-tubulin GFP-Atb2 (Fig 1E). Spindle formation, nuclear division, contractile ring assembly, and septation were not observed in mei3Δ and mei3Δ mei2Δ zygotes. We conclude that (1) in absence of Mei2, a fraction of zygotes progresses through mitosis, and (2) mitotic entry in mei2Δ zygotes depends on Mei3.

The only known molecular function of Mei3 is to inhibit the Pat1 kinase. We thus tested whether Mei3 signals through Pat1 to initiate mitotic entry in mei2Δ zygotes. Since pat1 deletion is inviable due to constitutive Mei2 activation, we analyzed double and triple mei3Δ, pat1Δ, and mei2Δ mutants (S1A and S1B Fig): Approximately 37% pat1Δ mei2Δ double mutant zygotes underwent cell division, similar to mei2Δ. Again, zygotic cell division was virtually absent from mei3Δ pat1Δ mei2Δ zygotes. Thus, Mei3 promotes mitotic entry independently of Pat1 in mei2Δ zygotes.

Even when comparing zygotes that did not undergo mitosis, mei3Δ and mei2Δ single and double mutant zygotes showed clear differences in zygotic growth, morphology, and lysis. To quantify zygotic growth and morphology, we segmented cross sections immediately after fusion necks have expanded (Fig 1C, 150-minute time point and yellow line in the far-right panel) and 10 hours later (Fig 1C, 750-minute time point and green line in the far-right panel). We then quantified the cross section area increase (Fig 1F) and 3 morphological parameters of zygotic growth including (1) the maximum cell perimeter displacement (Fig 1G); (2) the mean coefficient of variation in cell perimeter displacement, which is inversely linked to growth isometry (S1C Fig); and (3) the diameter of the mating projections (S1D Fig). Growth was absent from wild-type zygotes between these 2 time points, whereas the cross section area increased in all mutants (Fig 1F). The growth in mei3Δ mei2Δ zygotes was most prominent (Fig 1F and 1G) and clearly a consequence of cells forming long mating projections, which mostly originated from the fusion neck region (Fig 1C and S1 Movie), leading to a high coefficient of variation (S1C Fig). Nondividing mei2Δ zygotes exhibited a qualitatively similar behavior, but less extensive growth of mating projections. Conversely, mei3Δ zygotes did not form pronounced mating projections, and instead enlarged largely isometrically (Fig 1C, 1F and 1G, S1C and S1D Fig and S1 Movie). Deleting pat1 in mei2Δ and mei3Δ mei2Δ did not change the overall phenotypes of these 2 genetic backgrounds (S1B Fig and S3 Movie).

We further investigated the differences in growth and morphology of mei2 and mei3 mutant zygotes by monitoring the dynamics of the cell growth machinery. Specifically, we imaged Scd2-GFP, a scaffold protein that binds the active form of the key mating and growth regulator Cdc42 [33,34]. We simultaneously monitored Myo52-tdTomato, a type V myosin that delivers secretory vesicles to growth and fusion sites [17,19,35,36]. In wild-type cells, Scd2 and Myo52 focalized at the partner fusion site and then dispersed throughout the zygote after fertilization (Fig 1H, S1E Fig and S4 Movie). In mei3Δ zygotes, Scd2 and Myo52 signal also rapidly dispersed from the fusion site, but then formed relatively short-lived patches throughout the cell cortex (Fig 1H and 1I, S1E Fig). The patches occasionally stabilized at the sites where mating projections formed (S4 Movie). In mei3Δ mei2Δ and nondividing mei2Δ zygotes, Scd2 and Myo52 patches were significantly more stable after fertilization and marked the growing mating projection (Fig 1H and 1I, S1E Fig, and S4 Movie). In these zygotes, Scd2 and Myo52 patches frequently did not disassemble after fertilization, and mating projections grew from the region of the fusion neck. These differences in zygotic growth between mei2Δ and mei3Δ support the conclusion that Mei2 has roles in regulating the zygotic growth machinery that are independent of upstream Mei3 signaling.

Taken together, our analyses of individual and double mei2Δ and mei3Δ mutants show that, in addition to their well-established codependent role in meiotic induction, Mei2 and Mei3 independently regulate the zygotic cell cycle, growth, and mating blocks: (1) Mei3 promotes cell cycle progression in absence of Mei2; (2) Mei2 prevents zygote mating and lysis even in absence Mei3; and (3) mating and lysis phenotypes are exacerbated in the double mutant.

Mating blocks prevent zygote lysis

Lysis is a prominent phenotype of mei3Δ mei2Δ zygotes that we also readily observed in mei2Δ zygotes but not in mutant gametes prior to fertilization (S2A Fig). What causes lysis specifically postfertilization? Cell wall digestion is normally brought about by paracrine pheromone signaling and formation of an actin fusion focus nucleated by the formin Fus1 at the site of partner contact [19,37]. In self-stimulating autocrine cells, high levels of pheromone signaling induce the formation of an ectopic fusion focus, which leads to localized cell wall degradation independently of a partner cell, and consequently lysis [21]. Lysis in autocrine cells is suppressed by deletion of the formin Fus1 [21]. As gamete fusion brings together the machineries for production, secretion, and perception of both pheromones, we hypothesized that zygotes behave as autocrine cells, which would explain their frequent lysis.

To test whether the death of mei3Δ mei2Δ zygotes is a consequence of failed fusion attempts, we aimed to deplete Fus1 specifically postfertilization by developing a zygote-specific artificial degron system (Fig 2A). We first generated DegGreen by fusing the N-terminus of the ubiquitin E3 ligase complex protein Pof1 with the GFP-binding protein (GBP) and expressing it from a strong constitutive promoter. We reasoned that GFP-GBP binding will direct the Pof1 E3 ligase activity toward GFP and lead to its degradation. Indeed, introducing the DegGreen into cells expressing meGFP (monomeric enhanced GFP) lowered the green fluorescent signal by 95% (Fig 2B). We then built DegRed, a fusion between Pof1 N-terminus and mCherry-binding protein, which decreased the red fluorescence in cells expressing mCherry by 63% (Fig 2C). When we introduced DegGreen into gametes that carried the native Fus1 fused with meGFP, the fusion efficiency to fus1Δ partners was reduced by approximately 5-fold as compared to control cells carrying individual proteins (Fig 2D). Likewise, expression of DegRed in cells with native Fus1 fused to mCherry reduced fusion efficiency to fus1Δ partners by approximately 2-fold as compared to cells carrying individual proteins (Fig 2E). Thus, both artificial degrons can decrease the function of Fus1 in gametes.

Download:

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

Fig 2. Zygote lysis is induced by aberrant fusion attempts.

(A) Schematic of the artificial degron system. The DegGreen construct composed of the N-terminal, E3 ligase domain of Pof1 (E3, gray) fused with the GFP-binding nanobody (GBP, dark green) targets GFP (light green; here linked to Fus1) for degradation. Similarly, the DegRed construct, consisting of the E3 ligase domain of Pof1 (E3, gray) fused with the mCherry-binding nanobody (CBP, dark purple), targets mCherry (light purple) for degradation. For zygote-specific protein depletion, 1 partner gamete carries DegGreen and the mCherry-tagged protein of interest, while the other partner brings DegRed and the GFP-tagged protein of interest. Ubiquitination (orange circles) and degradation only happens postfertilization. (B and C) Exponentially growing cells expressing GFP (B) or mCherry (C) from the strong p^act1 promoter without (left) or with DegGreen (B) or DegRed(C) (right). The graph shows the fluorescence normalized to cells without degron systems. Dots show individual measurements, bars mean values, and error bars standard deviation. p-Values for indicated comparisons were obtained from the Kruskal–Wallis test. (D and E) Gamete fusion efficiency between h− strains with incomplete or complete (right) Fus1 degron systems and h+ fus1Δ. Dots show individual replicates, bars mean values, and error bars standard deviation. p-Values for indicated comparisons were obtained from the Welch test. (F) Mating of mei3Δ mei2Δ gametes carrying the indicated halves of the Fus1 artificial degron system. Zygotes with the incomplete Fus1 degron (top and middle panels) undergo cell lysis (examples outlined in yellow), whereas zygotes with a complete Fus1 degron postfertilization (bottom panel) prevent lysis (examples outlined in white). (G) Frequency of lysis in mei3Δ mei2Δ zygotes with incomplete or complete degron systems (genotypes as in (F)). Data are presented as in (D). The data underlying this Figure may be found at https://doi.org/10.6084/m9.figshare.13274837.v1.

https://doi.org/10.1371/journal.pbio.3001067.g002

To specifically degrade Fus1 in zygotes, we used the mei3Δ mei2Δ genetic background and introduced DegGreen into gametes expressing Fus1-mCherry, and DegRed into the other gamete expressing Fus1-meGFP, such that substrates and degrons come in contact only postfertilization (Fig 2F and S5 Movie). Live imaging showed high levels of Fus1-GFP and detectable levels of Fus1-mCherry in control mei3Δ mei2Δ zygotes lacking the cognate artificial degron, with fluorescence concentrated at the shmoo tip, similar to the Scd2 and Myo52 localization described above (S2B Fig, top and middle panels). Remarkably, zygotes with both degrons and Fus1 tags had visibly reduced Fus1 fluorescence postfertilization (S2B Fig, bottom panels), which was also evident from quantification of whole-cell Fus1-GFP signal (S2C Fig), and showed decreased cell lysis by at least 2-fold relative to control zygotes (Fig 2G, S2D Fig). Thus, cell death in mei3Δ mei2Δ mutant zygotes is a consequence of their inability to block mating. We conclude that mating blocks prevent zygote death.

Mei3 promotes G1-S cell cycle transition independently of Pat1 and Mei2

Our findings that mei2Δ, but not mei3Δ mei2Δ, zygotes undergo mitosis strongly suggests that Mei3 drives cell cycle progression postfertilization. To directly assess progress through the cell cycle, we used the Hoechst 3334 dye to stain the DNA of cell mixtures produced by mating of heterothallic h+ (P) and h- (M) strains carrying either cytosolic sfGFP or mCherry expressed from the strong p^tdh1 promoter (Fig 3A). The 2 fluorophores distinguish gametes, which fluoresce in a single channel, from zygotes, which fluoresce in both channels using flow cytometry, and enabled us to determine the cell cycle stage of each cell type in mating mixtures. We performed this experiment in mei3Δ, mei2Δ, mei3Δ mei2Δ, and as control mei4Δ, which arrest with 4C DNA content prior to sporulation [38,39]. As expected, all mutant gametes arrested with 1C content in response to nitrogen starvation and mating (Fig 3A, magenta and green curves). Mutant zygotes however exhibited differences in DNA content (Fig 3A, orange curves). Consistent with previous work, mei4Δ zygotes arrested with 4C content. Approximately half of the mei2Δ zygotes also had 4C DNA content, indicating genome replication. The exact number of mei2Δ zygotes that progress to S-phase could not be determined in this manner since some zygotes pass mitosis, and thus regain a 2C DNA content, prior to our flow cytometry measurements (Fig 1B). By contrast, mei3Δ and mei3Δ mei2Δ zygotes arrested with 2C content, indicating that fertilization brought partner genomes together but the zygotes did not replicate their genomes. We obtained similar results in experiments with homothallic (h90; self-fertile) strains where mCherry expression was driven by the P-cell-specific p^map3 promoter, and sfGFP by the M-cell-specific p^mam1 promoter (S3A Fig). Zygotes produced by h90 mei3Δ and mei3Δ mei2Δ strains arrested the cell cycle in G1 phase, while a large fraction of mei2Δ zygotes progressed through S-phase. In this experiment, we used as control sme2Δ mutants, which lack the Mei2-binding meiRNA and arrest with 4C DNA content prior to sporulation [29]. We conclude that Mei3 promotes S-phase entry in zygotes.

Download:

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

Fig 3. Mei3 promotes G1 exit independently of Pat1 and Mei2.

(A) Strategy used for flow cytometry on mating mixtures of heterothallic strains that constitutively express either sfGFP or mCherry (top). Gametes, which carry a single fluorophore, can be distinguished from zygotes, which carry both fluorophores. Flow cytometry analysis of Hoechst-stained DNA fluorescence (x-axis) in mating mixtures after 24 hours (bottom). The y-axis shows the cell number normalized to mode. Profiles of all cells (gray), gametes (green and magenta), and zygotes (orange) are shown. Gametes largely arrest with unreplicated 1C genomes. mei3Δ and mei3Δmei2Δ zygotes show unreplicated 2C content. mei2Δ and mei4Δ zygotes show a prominent 4C peak. (B) Hoechst-stained DNA fluorescence in zygotes identified through gating for increased size in forward and side scatter measurements (see Materials and methods). Zygotes lacking mei3 arrest with unreplicated genomes (2C), while those with intact mei3 replicate their genomes even in absence of mei2 and pat1. The smaller 2C peak in mei2Δ zygotes compared to panel A likely stems from the gating strategy which excludes zygotes that have undergone mitotic division. (C) Hoechst-stained DNA fluorescence of haploid mei2Δ cells with or without mei3 expressed from the constitutive p^tdh1 promoter grown in presence or absence of nitrogen for 24 hours. Ectopic expression of Mei3 prevents the G₁ arrest in nitrogen starved gametes but has no effect in nitrogen-rich media. (D) Mating efficiency of mei2Δ and mei2Δ pat1Δ cells with or without mei3 expressed from the constitutive p^tdh1 promoter after 24-hour incubation on agar plates. Constitutive mei3 expression reduces mating of gametes independently of mei2 and pat1. Dots show individual replicates, bars mean values, and error bars standard deviation. p-Values for indicated comparisons were obtained from the Welch test. (E) Time lapses show GFP-Pcn1 in haploid wild type and mei2Δ cells with or without constitutively expressed Mei3 upon removal of nitrogen from the media. Nuclei are enlarged in insets below for better visualization of the signal at indicated cell cycle stages with arrowheads pointing to Pcn1 replication foci. (F) Quantification of the number of cells that pass through the first and second S-phases upon nitrogen restriction as shown in (E). Note that analyses include only cells that were in the G2 phase at the beginning of imaging, which represents 66% to 70% of cells (n > 380). The data underlying this Figure may be found at https://doi.org/10.6084/m9.figshare.13274837.v1.

https://doi.org/10.1371/journal.pbio.3001067.g003

We further tested whether Mei3 function in driving the S-phase entry relies on Pat1. To that aim we performed flow cytometry measurements of DNA content in zygotes produced by mutants lacking individual and combinations of mei3, mei2, and pat1 genes. We note that in this instance, we did not label the gametes with fluorescent proteins since we found that we could use the flow cytometry forward and side scatters to separate zygotes based on their larger size (S3B Fig, see Materials and methods). Removal of Pat1 had no effect on the ability of zygotes to enter S-phase as both mei2Δ and mei2Δ pat1Δ zygotes accumulated 4C content, while all zygotes deleted for mei3 arrested in G1-phase (Fig 3B). Mei3-driven DNA replication also occurred in zygotes where Mei2 RNA-binding was abolished due to the F644A mutation in the RRM motif (Fig 3B, [29]). We conclude that Mei3 drives G1-S transition in zygotes independently of Mei2 and Pat1.

Though independent of downstream zygotic signaling, Mei3 may require some zygote-specific factors to promote S-phase entry. To test this hypothesis, we ectopically expressed Mei3 from the strong p^tdh1 promoter in haploid cells that remained alive due to simultaneous deletion of the mei2 gene. During exponential growth in nitrogen-rich media, overexpression of Mei3 did not cause detectable defects in DNA replication based on flow cytometry measurements (Fig 3C). However, upon nitrogen removal, which arrested the mei2Δ haploids in G1-phase, a large fraction of cells overexpressing Mei3 had duplicated genomes (Fig 3C). To directly test whether Mei3 promotes G1/S transition or blocks cells in G2, we imaged the GFP-tagged DNA replication fork component Pcn1 [40] expressed in addition to the native protein. Outside S-phase, GFP-Pcn1 produces a uniform nuclear signal but forms prominent foci during DNA replication [40,41]. While most wild-type and mei2Δ haploid cells responded to nitrogen withdrawal by undergoing division, one S-phase, a second division and then arresting in G1 of the subsequent cell cycle, the majority of cells overexpressing Mei3 underwent a second round of DNA replication (Fig 3E and 3F). Thus, Mei3 does not lead to a premature G2 arrest and instead promotes G1 exit. In agreement with impaired ability to arrest in G1, overexpression of Mei3 resulted in decreased mating efficiency (Fig 3D). Mei3 overexpression impaired mating also in cells lacking Pat1 consistent with our results that Mei3 cell cycle effects are Pat1-independent (Fig 3D). We conclude that Mei3 functions independently of zygote-specific factors to promote the G1-S transition.

Mei3 promotes S-phase entry and represses mating in diploid fission yeast

Our results contrast earlier work that proposed that Mei2 is required for S-phase entry in nitrogen-starved diploid cells committed to azygotic meiotic differentiation [24,29]. Note that the diploids in these experiments are to be distinguished from zygotes as they do not directly result from the fusion of 2 haploid cells but have been propagated mitotically as diploids before meiotic induction by nitrogen starvation. Diploids are typically isolated from mating mixtures of 2 haploid strains bearing different ade6 mutant alleles that exhibit intermolecular complementation [42]. We decided to revisit these findings and prepared wild-type and diploid cells lacking mei4 or, either individually or combined, mei2 and mei3 (S4A Fig). Since pheromone response is necessary for Mei3 induction and downstream meiotic differentiation [23,25,26], we also monitored the expression of mCherry and sfGFP placed under the pheromone-responsive, mating type-specific p^map3 and p^mam1 promoters, respectively (S4A Fig). Importantly, we observed that even the wild-type diploid population is heterogeneous with (1) cells that do not induce expression from both p^map3 and p^mam1 (S4A Fig, outlined cell); (2) cells that mate (S4A Fig, arrow); and (3) cells that respond to nitrogen removal asynchronously. These issues result in flow cytometry subpopulations (S4C Fig) and complicate the analysis of flow cytometry data.

To overcome these issues, we used 2 complementary approaches. First, we adapted the flow cytometry gating to evaluate the DNA content only in small cells, which excluded mated and clumped cells (see Materials and methods). We also restrained the analysis of nitrogen-starved cells to those with clear p^map3 and p^mam1 promoter activity required for meiotic differentiation (S4B Fig). Before nitrogen starvation, a majority of cells had replicated genomes. Whereas wild-type, mei2Δ, and mei4Δ cells showed clear G1 and G2 cell populations after 24 hours of nitrogen starvation, the mei3Δ and mei3Δ mei2Δ diploids completely arrested in G1 phase within 12 hours after nitrogen removal. In a second approach, we argued that mat1-P and mat1-M loci are not stable in diploids cell produced by standard h+ and h- strains [43], a hypothesis in agreement with our observation of diploid cells with a single active mating type-specific promoter (S4A Fig, outlined cell). To overcome this problem, we produced diploids where mat1 locus switching was prevented by a H1Δ17 mutation [4,44]. These diploids showed activity from both p^map3 and p^mam1 promoters, did not mate, and sporulated within 24 hours after nitrogen removal (Fig 4A). Flow cytometry analysis also showed increased homogeneity and similar activation of both p^map3 and p^mam1 promoters (S4C Fig, note that cells largely fall along the diagonal). Flow cytometry measurements of DNA content on all nonclumped mat1-H1Δ17 diploids (70% to 95% of all cells) clearly showed that mei3Δ and mei3Δ mei2Δ diploids completely arrested in G1, while wild-type, mei2Δ, and mei4Δ cells replicated their genomes (Fig 4B). Furthermore, imaging revealed that approximately 10% mei2Δ diploids, but no other genotype, underwent cell division after cells had activated the 2 pheromone responsive promoters (Fig 4C and 4F and S6 Movie). Thus, nitrogen-starved azygotic diploids, like zygotes, require Mei3 for the G1-S transition.

Download:

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

Fig 4. Mei3 promotes G1 exit in stable diploid cells.

(A) H1Δ17 diploid cells expressing mCherry and sfGFP from P- and M-cell-specific promoters p^map3 and p^mam1 24 hours after removal of nitrogen. Arrows point to shmoo-like projections in mei3Δ and mei2Δ mutants. (B) Flow cytometry analysis of Hoechst-stained DNA fluorescence (x-axis) in diploid cell used in (A) at indicated time points after nitrogen removal. The y-axis shows the cell number normalized to mode. Note that mei3Δ and mei3Δ mei2Δ mutants arrest with unreplicated genomes (2C), while wild-type, mei4Δ, and mei2Δ cells replicate their DNA (4C). The sub-2C population in wild-type sample are likely spores. (C–E) Time lapses show induction of fluorophores in cells as in (A). Note formation of mating projections (arrows) and the septum in mei2Δ mutant (arrowhead). (F) Percentage of H1Δ17 diploids expressing both p^mam1:GFP and p^map3:mCherry that are septating. (G and H) Time lapse (G) and quantification (H) of Mi-sfGFP induction in H1Δ17 diploid cells deleted for mei3. Note that Mi-sfGFP induction (black line in H) coincides with its nuclear accumulation (green arrow in G and green line in H) and precedes the formation of the shmoo (yellow arrow in G). The graph reports mean fluorescence for 8 cells with shaded areas denoting standard deviation. The data underlying this Figure may be found at https://doi.org/10.6084/m9.figshare.13274837.v1.

https://doi.org/10.1371/journal.pbio.3001067.g004

While otherwise wild-type and mei4Δ mat1-H1Δ17 diploids did not grow when nitrogen-starved, mei2Δ and/or mei3Δ mat1-H1Δ17 diploids formed shmoo-like projections (Fig 4A, arrows). Time-lapse microscopy revealed that mating projections in mei2Δ, mei3Δ, and double mutant diploids formed after activating both P- and M- pheromone-responsive promoters (Fig 4C–4E and S6 Movie). These results imply that, like in zygotes, Mei2/Mei3 signaling blocks mating in diploid cells starved for nitrogen.

These observations raise a conundrum about the role of pheromone signaling in diploids. Indeed, expression of the Pi–Mi complex essential for Mei3 and thus meiotic induction relies on pheromone signaling [25], yet our results show that Mei3 actively represses pheromone-induced mating in diploids. We reasoned that induction of the Pi–Mi complex may occur at lower levels of pheromone signaling than needed to form mating projection. While we did not directly measure pheromone levels in diploids, expression of Mi-sfGFP in mei3Δ mutant diploids was observed prior to formation of mating projections (n > 50 diploids forming a shmoo, Fig 4G and S7 Movie). Appearance of the Mi-sfGFP signal though asynchronous in the population was concomitant with its nuclear accumulation (Fig 4G and 4H), which depends on Pi ([4] and S4D and S4E Fig). This suggests that the Pi–Mi complex forms at the time of Mi induction. These results are consistent with the view that in diploids, low levels of pheromone signaling induce the Pi–Mi complex, leading to Mei3 expression, which in turn represses further pheromone-dependent mating and initiates meiosis.

We conclude that in diploid cells, like in zygotes, Mei3 promotes the G1-S transition, and both Mei2 and Mei3 repress mating behaviors.

Mei3 promotes the premeiotic S-phase

We showed above that Mei3 can advance the cell cycle in mitotic cells and is required for premeiotic S-phase in diploids and zygotes. However, interpreting the function of Mei3 in premeiotic S-phase is complicated by the well-established role of Mei3 in relieving the Pat1-mediated inhibition of Mei2. To uncouple possible functions of Mei3 in promoting premeiotic S-phase from Mei2 activation, we built cells with a simplified meiotic signaling that we refer to as SMS cells (Fig 5A). We reasoned that pat1 deletion would be made viable in haploid cells by deletion of mei2 and that Mei2 expression can be induced specifically postfertilization to induce meiosis. Accordingly, we deleted the mei2 and pat1 loci and replaced the native mei3 open reading frame (which is induced postfertilization) with that of mei2. SMS cells did not exhibit any obvious defects during exponential growth (p^{Welch test} = 0.59, n = 5) and produced zygotes that sporulated at wild-type rates (Fig 5B and 5C; p^{Welch test} = 0.28). We note that approximately 3% ± 2% SMS zygotes produced asci with aberrant spore number and that approximately 2% ± 1% SMS zygotes died prior to sporulation. We also compared the levels of meiotic recombination between wild-type and SMS zygotes (Fig 5D). Briefly, strains carrying constructs for expression of cytosolic sfGFP or mCherry integrated at the leu1 and aha1 genomic loci, respectively, were mated, and the proportion of progeny carrying each fluorophore was quantified. The genetic distance between the 2 markers decreased from 17.13 ± 0.14 cM in wild type to 14.70 ± 0.13 cM in SMS mutants. We conclude that SMS zygotes, which lack Mei3, show minor overall defects, but no gross perturbation of the meiotic cycle.

Download:

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

Fig 5. Mei3 advances premeiotic-S phase entry.

(A) Schematic representation of meiotic signaling in SMS zygotes. (B) Time lapse of SMS cells mating. (C) Quantification of zygotic phenotypes after 48 hours on solid media lacking nitrogen. (D) Results of flow cytometry quantification for nonrecombinant (purple and green) and recombinant (gray and yellow) progeny obtained by crossing strains coding for sfGFP at the leu1 locus with strains coding for mCherry at the linked aha1 locus. (E, G) Time lapses show DIC (gray) and GFP-Pcn1 (green) during mating of cells with indicated genotypes. The fusion time (0 minute) is evident as partner exchange cytosolic mCherry (magenta) expressed in only 1 gamete. Nuclei are enlarged in insets below for better visualization of the GFP-Pcn1 signal with arrowheads pointing to replication foci. Note that SMS zygotes take longer to exit G1 than wild type (E) and that Mei3 promotes S-phase entry in zygotes lacking Pat1 (G). (F, H) Quantification of postfertilization time to appearance of GFP-Pcn1 puncta determined from time lapses as in (E) and (G). Dots show individual measurements, bars mean values, and error bars standard deviation. p-Values for indicated comparisons were obtained from the Kruskal–Wallis test. (I) Quantification of zygotic phenotypes after 24 hours on solid media lacking nitrogen. (J) Flow cytometry analysis of Hoechst-stained DNA fluorescence (x-axis) for gametes (gray) and zygotes (magenta) in mating mixtures after 18 hours. The y-axis shows the cell number normalized to mode. Gametes arrest with unreplicated 1C genomes. Genome duplication occurs in a fraction of cycΔ5 (left panel, arrow points a 4C shoulder) but not SMS cycΔ5 zygotes (right panel). The data underlying this Figure may be found at https://doi.org/10.6084/m9.figshare.13274837.v1.

https://doi.org/10.1371/journal.pbio.3001067.g005

SMS zygotes, however, showed a substantial delay in G1-S transition, with GFP-Pcn1 foci forming before nuclear fusion in wild-type zygotes, but only after karyogamy in SMS zygotes (S8 Movie). The time between partner fusion (monitored by entry of cytosolic mCherry in the P-partner) and S-phase onset was 91 ± 23 minutes in SMS zygotes, but only 35 ± 7 minutes in wild-type zygotes (Fig 5E and 5F). To test whether Mei3 advances S-phase in cells with contracted zygotic signaling, we compared mei3+ and mei3Δ cells, which had pat1 and mei2 loci deleted and mei2 coding sequence under p^mei3 promoter control at the ura4 locus. Mei3 expression reduced the time between fertilization (marked here by entry of cytosolic mCherry in the M-partner) and premeiotic S-phase entry, labeled by Pcn1 foci, by approximately 17 minutes (Fig 5G and 5H). Thus, Mei3 promotes but is not essential for the premeiotic S-phase in zygotes with functional Mei2. This function is also independent of Pat1 kinase.

In wild-type zygotes, previous work showed that the meiotic G1-S transition is largely driven by additive effects of the 5 nonessential cyclins, Cig1, Cig2, Puc1, Crs1, and Rem1 [45]. However, even upon deletion of these 5 cyclins (named here cycΔ5), approximately 23% zygotes sporulate (Fig 5I and [45]). Examination of DNA content in cycΔ5 zygotes by flow cytometry showed a 2C peak with a clear 4C shoulder, indicating that a significant number of zygotes have sufficient residual CDK1 activity for premeiotic S-phase, likely driven by the remaining, essential B-type cyclin Cdc13 (Fig 5J). By contrast, premeiotic S-phase was almost completely blocked when combining the cycΔ5 with the SMS background, which lacks Mei3, as shown by absence of the 4C shoulder in flow cytometry analysis, and only 3% ± 0.5% zygotes sporulated (Fig 5I and 5J). The additive effects between cycΔ5 and SMS genetic backgrounds further support the view that Mei3 serves to elevate cyclin-CDK activity to promote S-phase entry.

Mei3-driven cell cycle progression prevents zygote mating and refertilization

Because Mei3 shows Mei2-independent roles in both promoting S-phase entry and blocking refertilization, we tested whether cell cycle progression per se may block refertilization. This hypothesis is also suggested from the knowledge that haploid cells mate only during the G1-phase of the cell cycle. To address this possibility, we monitored the timing of shmoo formation relative to S-phase entry in mating mixtures of zygotic signaling mutants. Time-lapse imaging of GFP-Pcn1 faithfully reproduced our earlier flow cytometry findings with Pcn1 replication foci forming in virtually all wild-type, sme2Δ, and mei4Δ zygotes but in none of the mei3Δ or mei3Δ mei2Δ zygotes (S5A Fig and S9 Movie). In mei2Δ zygotes, we observed 50% passing S-phase, 37% arresting in G1-phase for the duration of the experiment, and the remaining 13% lysing. Remarkably, all lysing cells were in the G1 phase, and the remaining G1-arrested mei2Δ zygotes formed mating projections (Fig 6A and 6B and S10 Movie). Conversely, only 40% of zygotes that advanced in the cell cycle formed shmoos and did so prior or during S-phase but never after S-phase completion. Thus, cell cycle progression correlates with mating repression in zygotes.

To test whether forced S-phase entry can suppress zygotic mating independently of Mei3 and Mei2, we used mei3Δ mei2Δ cells, which produce G1-arrested zygotes (Fig 3A, S5A Fig), and forced the expression of the G1 cyclin Puc1 under the regulation of the zygote-specific p^mei3 promoter. Imaging of the GFP-Pcn1 showed that forced expression of Puc1 postfertilization drove approximately 15% of zygotes through S-phase (Fig 6C, S5B Fig and S11 Movie) and approximately 7% through division (Fig 6D). Consistent with the hypothesis that cell cycle progression acts as a refertilization block, Puc1 expression decreased the rates of refertilization and lysis in mei3Δ mei2Δ mutant zygotes (Fig 6C and 6D). Puc1 expression also diminished the mating-induced growth in mei3Δ mei2Δ zygotes that did not initiate S-phase (S5C Fig). Our results imply that increased CDK activity and associated cell cycle progression is sufficient to prevent mating and refertilization in zygotes.

We next sought to test whether the G1 exit is necessary to prevent zygotic mating. We used cycΔ5 zygotes, which largely fail to enter S-phase (Fig 5I). Remarkably, mating of cycΔ5 gametes, which are elongated likely due to an extended cell cycle, produced a noticeable fraction of zygotes that grew a shmoo and underwent refertilization (approximately 4% of zygotes; Fig 6E and 6F and S12 Movie). We conclude that rapid exit from the G1 phase is required to prevent zygotic mating and refertilization, even in cells with intact zygotic signaling. These data are consistent with the view that Mei3 blocks zygotic mating and refertilization by promoting the premeiotic S-phase.

Download:

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

Fig 6. Cell cycle progression prevents refertilization.

(A) Time lapse of mating mei2Δ cells. Zygotes that persist in G1 (yellow outlines) form prominent mating projections (yellow arrowhead). Zygotes that progress through S-phase (blue outlines), evident from GFP-Pcn1 puncta (blue arrowhead), largely repress shmoo formation and instead divide. (B) Quantification of phenotypes of mei2Δ zygotes presented in (A). We used GFP-Pcn1 to group zygotes into those that either arrest (left bars) or exit (right bars) G1-phase and scored the mating behaviors of both groups. (C) Mating of control mei3Δ mei2Δ cells (top panel) and mei3Δ mei2Δ cells that express Puc1 from the zygote-specific p^mei3 promoter. GFP-Pcn1 (green) forms puncta (blue arrowhead) in Puc1-expressing zygotes (blue outlines). G1-arrested zygotes (yellow outlines) frequently lyse in control but not Puc1-expressing zygotes. (D) Quantification of zygotic phenotypes observed in time lapses as in (D). (E) Mating of cycΔ5 cells, which undergo refertilization (yellow outlines). (F) Quantification zygotic phenotypes observed in (E). The data underlying this Figure may be found at https://doi.org/10.6084/m9.figshare.13274837.v1.

https://doi.org/10.1371/journal.pbio.3001067.g006

Mei2 blocks refertilization independently of cell cycle progression

While the role of Mei3 in blocking refertilization can be explained by its function in promoting the premeiotic S-phase, it remained unclear whether Mei2 blocks refertilization solely by promoting cell cycle progression or through other means. A role in promoting the meiotic transitions is suggested by our data and previous work [29], as some mei2Δ cells arrest in G1 (Fig 6A). In addition, while mei3Δ pat1Δ mei2Δ triple mutant zygotes arrest in G1, reexpression of mei2 in these cells (in the SMS background) restores a quasi-normal meiotic cycle (Fig 5). However, the higher frequency of lysis and refertilization observed in mei3Δ mei2Δ compared to mei3Δ zygotes, which are both arrested in G1, also indicates a cell cycle-independent function of Mei2 in blocking zygotic mating. To probe this point further, we compared G1-arrested cycΔ5 mei3Δ mei2Δ pat1Δ zygotes with or without Mei2 expression from the p^mei3 promoter. Expression of Mei2, which in these cells is devoid of Pat1 inhibition, resulted in zygotes that still largely arrested in G1 (S6 Fig) but strongly repressed zygotic growth, mating, lysis, and refertilization (Fig 7A–7C and S13 Movie). Consistently, the Scd2-GFP polarity patch was not present after fusion necks expanded in mutant zygotes expressing Mei2 (Fig 7D and S14 Movie). We conclude that Mei2-dependent mating blocks operate largely independently of cell cycle progression.

Download:

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

Fig 7. Mei2 blocks zygotic mating independently of cell cycle progression.

(A) Mating mixtures of SMS cycΔ5 cells (bottom), which induce Mei2 postfertilization and control cells that lack Mei2 (top). Mei2 expression prevents zygote lysis (yellow outlines) and refertilization (white outlines). (B) Quantification of phenotypes in zygotes as in (A). (C) Quantification of growth in zygotes as in (A) in the 10 hours after concave neck expansion was completed. Dots show individual measurements, bars mean values, and error bars standard deviation. p-Values for indicated comparisons were obtained from the Kruskal–Wallis test. (D) Projection of 60 time points during 10 hours from the time of fertilization of polarity and growth marker Scd2-GFP. The yellow arrowheads point to the signal at the neck. Scd2 is restricted to the neck in the zygote expressing Mei2 (right) but continuously trails the growth projection in the zygotes lacking Mei2 (blue arrow). The data underlying this Figure may be found at https://doi.org/10.6084/m9.figshare.13274837.v1.

https://doi.org/10.1371/journal.pbio.3001067.g007

Refertilization blocks protect zygotes from death

Both Mei3 and Mei2 strongly contribute to preventing zygote refertilization, with 10% to 15% of mei3Δ mei2Δ double mutant zygotes fusing with an additional gamete, consistent with our previous observations [4]. However, an even more prevalent phenotype of mei3Δ mei2Δ zygotes is cell lysis. To test whether lysis may be due to failed zygotic fusion attempts, we developed a tool to target proteins for degradation specifically in zygotes. By linking an E3 ligase domain to the GFP- or mCherry-binding proteins, which bind their target with nanomolar affinities, this tool can be used to target any GFP- or mCherry-tagged protein for degradation. While combination of degron and target in the same cell leads to constitutive target degradation, as we have shown for cytosolic GFP and mCherry, the mating of gametes carrying the degron for the target expressed in the partner cell leads to target degradation specifically in zygotes. This is particularly interesting to test the zygotic role of essential proteins or proteins necessary for the earlier mating steps, and thus we expect it will be useful beyond the specific question asked here.

Using this zygotic degradation tool, we now show that the cell death of mei3Δ mei2Δ zygotes depends on the formin Fus1, the critical mating-specific organizer of the actin fusion focus. Indeed, degron-induced Fus1 depletion reduced fusion in gametes and lysis of mei3Δ mei2Δ zygote to similar extent, suggesting that most, if not all, lysis is due to erroneous activity of Fus1. This shows that zygotic death is due to ectopic formation of the fusion focus, which promotes local cell wall digestion. While we do not formally show it, these ectopic fusion attempts are most likely due to zygotes becoming autocrine cells, which exhibit similar lysis phenotypes [21], as fertilization brings together the production and perception machineries for both pheromones. We conclude that postfertilization mating blocks not only prevent polyploid formation to ensure genome maintenance across generations but also are essential to prevent zygotic death.

The high rates of refertilization and cell death in mei3Δ mei2Δ zygotes point to strong selective pressures to evolve mating blocks. This finding is likely to have relevance well beyond the fission yeast. In principle, species where sex relies on a diffusible signal produced by one and perceived by the other partner will lead to autocrine signaling postfertilization. In walled cells, sexual signaling normally promotes gamete cell wall digestion for fertilization and is likely to similarly induce cell wall breakage and thus cell lysis when activated in an autocrine manner in zygotes. Thus, we expect that mechanisms to down-regulate autocrine signaling and prevent cell death postfertilization may be a common feature of sex in fungi [46] and likely other walled organisms that employ diffusible signals [47,48].

Mei3 promotes S-phase entry independently of Mei2

The strong phenotype of mei3Δ mei2Δ zygotes is a combination of the contributions of Mei3 and Mei2 functions, whose individual deletions produce much milder phenotypes. Multiple lines of evidence indicate that, in addition to its role in relieving Pat1 inhibition of Mei2 [49], Mei3 drives S-phase entry. First, mei3Δ and mei3Δ mei2Δ but not mei2Δ zygotes are blocked before S-phase, consistent with previous work on mei3 [31]. Second, overexpression of Mei3 prevents the G1 arrest in response to nitrogen starvation. Third, Mei3 expression advances the premeiotic S-phase entry in SMS zygotes, which lack pat1 and induce Mei2 expression only postfertilization. Indeed, in the aforementioned experiments, the pat1 gene is dispensable for the Mei3-driven cell cycle progression, and thus Mei3 must operate independently of Mei2 activation. In the SMS background, the effect of mei3 deletion is relatively weak, with only a delay in S-phase entry. This reflects the direct, Mei2-independent role of Mei3 in promoting S-phase. In otherwise wild-type zygotes, the deletion of mei3 has a much more severe outcome, completely blocking S-phase. This is because in these zygotes Mei2 also cannot be activated, and thus the phenotype is identical to the deletion of both mei3 and mei2. We conclude that Mei3 promotes S-phase entry both by relieving Pat1 inhibition of Mei2 and by means independent of its activation of Mei2.

While the initial studies on meiosis suggested that Mei2 is essential for premeiotic S-phase [29], these have relied on h+/h− diploids, which are unstable. We have revisited this work and introduced controls that ensure we are analyzing the h+/h− diploids stimulated by pheromone signaling. Our results mirror those obtained from zygotes and find that S-phase entry completely relies on Mei3 and only partially on Mei2. A direct implication of this result is that initiation of meiosis upon Pat1-inactivation, a widely used tool to initiate azygotic meiosis [27], does not fully reproduce the physiological process. Instead, azygotic diploids with inactivated Pat1 are analogous to SMS zygotes as both lack Mei3 and Pat1 functions and start S-phase with a delay. This conclusion agrees with other notable differences between Pat1 inactivation-induced meiosis and normal zygotic meiosis [27,50] and suggests that Mei3 (and possibly Pat1) may have additional roles to that of regulating Mei2 activity.

One important question for the future is how Mei3 promotes premeiotic S-phase. A detailed recent study on the meiotic roles of cyclins found a number of differences between zygotic and azygotic meiosis [45]. In particular, removal of all nonessential cyclins almost fully impaired S-phase entry during azygotic meiosis, but only partially prevented zygotic sporulation. Consistent with this work, we found that the deletion of nonessential cyclins impairs the progression through premeiotic S-phase. As the G1-block is exacerbated in cycΔ5 zygotes lacking Mei3 (but expressing active Mei2), we hypothesize that Mei3 at least in part promotes the activity of the only remaining cyclin Cdc13-CDK complex.

Cell cycle progression acts as zygotic mating block

Irrespective of the precise molecular mechanisms by which Mei3 promotes premeiotic S-phase, this role is critical to prevent zygotic mating. Indeed, 3 lines of evidence show that cell cycle progression prevents mating postfertilization. First, mating of mei2Δ zygotes correlates with their cell cycle stage, with only G1 and early S phase zygotes displaying mating behaviors. Second, G1 arrest upon removal of nonessential cyclins leads to refertilization, even in presence of functional Mei3-Pat1-Mei2 signaling. Third, forced cell cycle progression by zygotic expression of the G1 cyclin Puc1 substantially decreases the mating behaviors of mei3Δ mei2Δ zygotes. The observation that Puc1 reduces mating even in cells that fail to replicate their genomes suggests that CDK represses mating independently of the cell cycle transition. These data are in line with transcriptional genome-wide studies of meiosis showing that the expression of key mating regulators rapidly declines as cells progress through the meiotic stages [51]. They also agree with the well-established knowledge that gamete mating strictly requires G1 arrest, as CDK activity inhibits Ste11 at both pre- and posttranscriptional levels [15,16]. Thus, we propose that one important way Mei3 blocks refertilization is by promoting CDK activity.

Similar to the case of S. pombe, only the G1 phase is permissive for mating in Saccharomyces cerevisiae, and zygotes reenter the cell division cycle. Antagonism between CDK activity and pheromone signaling is very well established in this organism and leads to arrest of gametes in G1 [52–54]. Pheromone signaling also arrests the cell cycle in other fungi such as C. albicans, F. oxysporum or U. maydis [55–57]. Beyond fungi, gametes generally arrest the cell cycle at various stages and zygotes resume cell cycle progression. We can thus speculate that zygotic reentry in the cell cycle may have conserved roles in bringing the zygote out of a fertilization-permissive state.

Mei2 blocks zygotic mating independently of cell cycle progression

The role of Mei2 is more complex, with at least 3 different functions. First, Mei2 is absolutely critical to induce meiosis. Indeed, we found that a fraction of mei2Δ zygotes enter a cell cycle akin to mitosis, with assembly of a single microtubule spindle followed by formation of a cytokinetic ring and septation. This indicates that upon cell fusion, zygotes lacking mei2 can reenter the cell cycle (driven by Mei3), but that this cycle is mitotic. Thus, Mei2 is the master specifier of meiosis. Second, Mei2 contributes in driving the meiotic cycle forward. Indeed, a fraction of mei2Δ zygotes and azygotic diploids are blocked or delayed in G1-phase. The ability of Mei2 to drive meiosis in SMS zygotes lacking Mei3 and Pat1 also shows that active Mei2 is in principle sufficient to allow a whole meiotic cycle, albeit with a G1 delay. A third function of Mei2, independent of promoting the meiotic cycle, is to block mating in zygotes. Indeed, mei3Δ mei2Δ double mutants show substantially more refertilization, growth, and lysis than mei3Δ zygotes, even though both are blocked in G1. This is also evident form the difference in dynamics of the polarity regulators Scd2 and Myo52, which were relatively short-lived in mei3Δ but stable in mei2Δ zygotes. These dynamics differences are reminiscent of dynamics observed in gametes where low levels of pheromone lead to unstable polarity zones which stabilize at elevated pheromone levels [58], suggesting that Mei2 suppresses pheromone signaling or downstream response. Expression of Mei2 in SMS zygotes lacking all nonessential cyclins was also sufficient to block most mating behaviors including Scd2 patch formation. We conclude that Mei2 imposes the zygotic fate, characterized by meiotic competence and mating blocks.

Our results challenge the current understanding of Mei2 regulation. The canonical view is that Mei2 is kept inactive by Pat1 until Mei3 relieves this inhibition. However, the observation that mei3Δ mutants do not lyse postfertilization like mei3Δ mei2Δ zygotes indicates that Mei2 must prevent the pheromone from stabilizing mating and fusion machineries even prior to Mei3-dependent activation. This suggests that Pat1 inhibition of Mei2 is not complete, a view in agreement with a previous report that Mei2 promotes efficient mating when Mei3 is not expressed [59]. These findings raise the question of how Mei2 allows mating prior to fertilization but prevents mating-induced lysis postfertilization. The simplest explanation may be that there are additional differential regulations of Mei2 between gametes and zygotes, for instance, inhibition by TOR kinase [59] or other unidentified factors.

Synergistic action of 2 mating blocks in zygotes

We propose that the dramatic mating phenotypes of mei3Δ mei2Δ double mutant zygotes are due to synergistic actions of Mei3 and Mei2, which act in parallel to block mating. Mei2 is the key inducer of the zygotic fate, which represses mating and promotes meiosis. Mei3 involvement is 2-fold: On the one hand, Mei3 promotes Mei2 activation and in this way acts upstream of Mei2 to promote meiosis and inhibition of the pheromone response. On the other hand, Mei3 promotes the exit from the mating permissive G1 phase, independently of Mei2, and in this way acts synergistically with Mei2. Indeed, we found that combining mutations in nonessential cyclins, which block zygotes in G1, with mei2Δ also show synergistic mating block defects. Prolonging the G1 arrest exacerbates the mating block defects of mei2Δ zygotes by allowing extra time to grow mating projection and thus increasing chances of refertilization and lysis. Collectively, our work shows that in fission yeast, distinct zygotic signaling branches initiate cell cycle-dependent and independent refertilization blocks that ensure survival and ploidy maintenance.

Robustness of biological systems often relies on parallel mechanisms regulating a single process [60]. This is also the case for refertilization blocks in several organisms [61]. For example, in amphibians, changes in the plasma membrane potential provide the so-called “fast” block to refertilization, while secretion of an outer cell layer presents a physical barrier to further sperm entry and a “slow” block to refertilization [62]. Subsequent development in all instances requires cell cycle progression, but its relevance as a refertilization block has not been studied. Fungal sexual reproduction might have experienced similar evolutionary pressures for rapid and long-term blocks to refertilization and lysis. Thus, Mei2- and Mei3-dependent mating blocks may act analogously, as rapid and delayed blocks, respectively. The fertilization-induced exit from G1 occurs only approximately 35 minutes after cell fusion, leading to the speculation that evolution could have selected for additional rapid refertilization blocks. Furthermore, the G1 exit alone may not be sufficient to fully block mating since gametes arrested in S-phase do mate, albeit at a very low frequency [16]. As the Mei2-dependent mating block is already operational independently of Mei3 at the time of fertilization, it may act as a “fast” block preventing refertilization during this permissive time frame.

Strains and genetic markers

All strains used in the study are shown in S1 Table.

All genetic markers with detailed description are available in S2 Table. As indicated, previously reported markers were either present in the lab stock, received as a gift, or obtained from the National BioResource Project (yeast.nig.ac.jp/yeast/top.jsf) or fission yeast deletion library produced by Bioneer (Daejeon, Republic of Korea).

For markers generated in this study, we provide sequences of plasmids and primers used to make them in the Supplemental Sequences on Figshare (https://doi.org/10.6084/m9.figshare.13274837.v1). Briefly, we used SIV vectors [41] to integrate artificial constructs into the ade6, his5, lys3, or ura4 genomic loci. The pRIP vector [63] was used to integrate GFP under the p^mam2 promoter at the mam2 locus. Recombination assay markers were integrated into either leu1-apc10 or aha1-SPBC1711.09c intergenic regions. Artificial degrons were integrated at either ura4 or fus1 loci. Fluorophore tagging of fission yeast proteins was performed at their native genomic loci with the exception of the S-phase marker Pcn1 which was expressed as second copy. To achieve constitutive gene expression, we used the strong promoters of act1 and tdh1 genes or moderate promoters of the pcn1 gene or the SV40 viral promoter. To induce mating type-specific expression of genes, we used either the M-cell-specific p^mam1 and p^mam2 promoters or the P-cell-specific p^map3 promoter. Zygote-specific gene expression was achieved by using the p^mei3 promoter. All constructs were sequenced and their integration into the genome confirmed by genotyping.

To prevent the mat1 locus switching, wild-type H1-homology box was replaced with the one amplified from the PB9 strain carrying the H1Δ17 mutation [4,44].

The DegRed artificial degron was comprised of the N-terminus of E3 ligase Pof1 (a.a. 1–261) fused to the mCherry-binding protein (ChBP, [64]). Similarly, the DegGreen construct consisted of the Pof1 N-terminus (a.a. 1–261) fused to the GBP [65]. Both constructs were placed under the regulation of the strong p^tdh1 promoter (1,000 bp upstream the START codon) and the budding yeast ADH1 transcriptional terminator and integrated at either the ura4 or fus1 locus. We note that the 2 artificial degrons differ significantly in the ability to deplete target proteins (Fig 2B and 2C) for reasons that are not immediately clear. It is possible that the degron activities differ due to different expression levels, strength of nanobody-target binding, or availability of residues for ubiquitination. We highlight that DegRed and DegGreen lack the WD40 repeats that target Pof1 to native target proteins and that neither caused any obvious defects in cells.

Growth conditions

Note that all strains used in the study are prototrophs. The growth conditions used for experiments with haploid cells are detailed in [66]. Briefly, freshly streaked cells were inoculated into MSL +N medium and incubated overnight at 25°C with 200 rpm shaking to exponential phase. The next day, cultures were diluted to OD₆₀₀ = 0.025 in 20 ml of MSL +N medium and incubated overnight at 30°C with 200 rpm agitation to exponential phase. Experiments on exponentially growing cultures were performed at this point. For analysis of mating cells, homothallic cells or 1:1 mixtures of heterothallic cells were pelleted for 1 minute at 1,000g, shifted to microcentrifuge tubes and washed 3 times with 1 ml of MSL −N medium. To prepare the mating mixtures for flow cytometry and quantifications of sporulation, fusion, and mating efficiencies, we resuspended 1 OD₆₀₀ (approximately 1.4 × 10⁷ cells for our spectrophotometer) in 20 μl of MSL −N media, dropped them onto solid MSL −N media and incubated at 25°C for 24 hours unless otherwise indicated. For time-lapse imaging and related quantifications, we diluted cells in 3 ml of MSL −N medium to final OD₆₀₀ = 1 to 1.5 and incubated them at 30°C with 200 rpm agitation for 4 to 6 hours. The exception is the cycΔ5 mutants, which once nitrogen starved rapidly formed clumps difficult to image. Thus, we either shortened the incubation in liquid MSL −N to 2 to 3 hours, or alternatively, starved the h+ and h− cells separately for 5 hours before mixing them immediately prior to imaging. Finally, cells were mounted onto MSL −N agarose pads and covered with a coverslip, and the chamber was sealed using VALAP (vaseline, lanolin, and paraffin; 1∶1∶1). We directly proceeded with imaging, or incubated slides until cells were 24 hours without nitrogen before scoring for phenotypes. The same growth conditions were used in nitrogen starvation experiments with a single heterothallic strain.

To summarize, we monitored exponentially growing cells in MSL +N liquid media and mating on MSL −N solid agar media plates or agarose pad chambers. We note that we observed quantitative differences between matings performed in different conditions (e.g., mating efficiency is higher on MSL −N solid agar plates) possibly due to limitation of nutrients and lower cell density in agarose chambers. Therefore, in all experiments, control strains were always run in parallel with identical conditions.

To obtain diploid cells through ade6 complementation, we mated ade6-M210 and ade6-M216 strains on ME media and the next day streaked cells onto EMM media lacking adenine. As soon as visible, colonies of adenine prototrophs were restreaked and inoculated for experiments the following day. We paid attention that cells never overgrow as this may lead to meiotic activation. To obtain H1Δ17 diploids, we mated on ME heterothallic strains that have the H1Δ17 allele linked with different antibiotic resistance markers [4]. The next day, we streaked cells onto YE media selecting for both antibiotics and waited for double resistant clones to form. We cultured both types of diploids starting with fresh cells OD₆₀₀ = 0.025 in 20 ml of MSL +N media lacking amino acids. Cells were incubated overnight to OD₆₀₀ = 0.3 to 0.5, collected by centrifugation at 1,000g for 1 minute, and washed in MSL −N media 3 times. For time-lapse imaging, we imaged cells in the CellASIC ONIX microfluidic device (Merck Group, Darmstadt, Germany). For flow cytometry and snapshot images, we resuspended cells in 25 ml of MSL −N media to final O.D.₆₀₀ = 0.5. A volume of 5 ml aliquots were collected at indicated time points, cells pelleted by centrifugation at 1,000g for 1 minute, supernatant discarded, and cell pellet thoroughly resuspended in 50 μl of MSL −N on ice before ethanol fixation.

To evaluate the meiotic recombination frequency, we prepared mating mixtures on solid MSL −N media as described above. After 48-hour incubation at 25°C, we collected cells using inoculation loops, resuspended them in 1 ml of MSL −N media containing 10 μl of glusulase (NEE154001EA, Perkin-Elmer, Massachusetts, United States of America), and incubated the samples for 24 hours at 30°C. Samples were then washed 3 times in water, and we visually confirmed that glusulase treatment killed all cells except spores, thus ensuring that subsequent analyses evaluated only the filial generation. Samples were then resuspended in YE media and incubated at 30°C for approximately 8 hours. We diluted samples to final O.D.₆₀₀ = 0.025 and grew them to exponential phase for additional approximately 12 hours when samples were subjected to flow cytometry.

Genetic crosses were performed by mixing freshly streaked strains on either ME or MSL −N media. We note that MSL −N media enhanced mating and sporulation of mutants lacking multiple cyclins and zygotic signaling components.

We used standard conditions and protocols for genetic manipulations [67]. Briefly, yeast cells were grown in standard fission yeast YES media at either 25°C or 30°C and using 200 rpm rotators for liquid media. Cells were transformed with the lithium-acetate protocol [67]. For selection, we supplemented YES with 100 μg/ml G418/kanamycin (CatNo.G4185, Formedium, Norfolk, United Kingdom), 100 μg/ml nourseothricin (HKI, Jena, Germany), 50 μg/ml hygromycinB (CatNo.10687010, Invitrogen, California, USA), 100 μg/ml zeocin (CatNo.R25001, ThermoFischer, Massachusetts, USA), and 15 μg/ml blasticidin-S (CatNo.R21001, ThermoFischer). To our knowledge, this work is the first fission yeast report that employs patMX dominant marker to select for growth on glufosinate-ammonium. The patMX sequence was obtained from pAG31 plasmid [68] and introduced into the pAV0892 (see Supplemental Sequences). After plasmid transformation into the prototroph yeast, cells were plated onto nonselective YES media and incubated overnight at 30°C. The next day, we replicated the cells onto the EMM media without amino acids and supplemented with approximately 400 μg/ml of glufosinate-ammonium (CatNo. G002P01G, Cluzeau Info Labo, Sainte-Foy-La-Grande, France) and incubated them at 30°C. If background growth was observed after a couple of days, we replicated cells once more onto the fresh selective media. Individual colonies developed within 5 days.

To determine growth rates, strains were cultured in YE media overnight, diluted to O.D.₆₀₀ = 0.01, and arrayed in the 96-well plate. After a 2-hour incubation at 30°C with constant shaking to relaunch cultures out of stationary phase, the plate was placed in plate reader Synergy H1M (Biotek, Winooski, USA), which measured to O.D.₆₀₀ every 2.5 minutes for a total of 5 hours at 30°C without shaking. A total of 5 replicates of each sample in each condition were measured. Growth rates were calculated by first subtracting the blank from the O.D.₆₀₀ reading and then performing log₂ transformation. The slope of a linear regression line fitted on a plot of the log₂-transformed O.D.₆₀₀ values over time, within a selected time window, which was manually determined to find a slope associated with the best fit reported through R². We report the Welch test p^value in the text and provide raw data and calculations as Source Data.

Flow cytometry

To prepare nitrogen-starved and mated cells for DNA content measurements, we collected cells from solid media using inoculation loops and resuspended them in 1 ml of MSL −N media precooled to 4°C. Cells were pelleted by centrifugation at 1,000g for 1 minute at 4°C, supernatant discarded, and cells thoroughly resuspended in 50 μl of MSL −N on ice before ethanol fixation.

For cell fixation, we added 1 ml of 70% EtOH (v/v in H₂O) prechilled to −20°C and rapidly mixed by pipetting and vortexing for 5 seconds, followed by incubation at −20°C for 15 minutes. Samples were once more vortexed for 5 seconds, centrifuged at 4°C for 2 minutes at 2,000g, supernatant discarded, and cells gently resuspend in 1 ml of room-temperature PBS (NaCl 8 g/l, KCl 0.2 g/l, Na₂HPO₄ 1.44 g/l, KH₂PO₄ 0.24 g/l, HCl adjusted (pH 7.4); filtered using 0.2 μM filter). Samples were pelleted by centrifugation at room temperature for 1 minute at 1,000g, supernatant discarded, and cells resuspended in 1 ml of PBS. We did not store the samples as we noticed that their quality degraded considerably already after a day at 4°C. Instead, we immediately stained samples with Hoechst-33342 (Sigma-Aldrich B2261-25MG) at final concentration 25 ng/μl. We note that increasing the dye concentration produces a strong background signal. We resuspend 0.1 O.D.₆₀₀ of cells in 0.2 ml of PBS in 5 ml flow cytometry Falcon tubes, added 1.8 ml of Hoechst working solution (27.7 ng/ml Hoechst-33342 in PBS) leading to final 25 ng/μl Hoechst-33342 concentration, and briefly vortexed. Tubes were closed and incubated in the dark ensuring thorough mixing using a rocking platform for 15 to 60 minutes before flow cytometry. We note that the signal produced by staining showed variation between experimental replicas, possibly due to stability of Hoechst-33342 stock (25 mg/ml in DMSO, −20°C) and slightly different concentrations of the Hoechst working solution, which was freshly prepared for each experiment. Furthermore, experimental variation arose due to laser intensities and recording settings. Thus, we refrain from comparing samples from different experimental replicates and show in the same figure panel the data obtained from samples processed simultaneously.

Flow cytometry was performed on Fortessa instrument (Becton Dickinson, Franklin Lakes, USA) with proprietor software platform. The analysis was performed using the FlowJo (Becton Dickinson, Franklin Lakes, USA) software package. The 488-nm laser was used to determine the forward scatter (FSC) and side scatter (SSC). We used the 355-nm laser with 450/50 filter to detect Hoechst 33342 stained DNA, 488-nm laser with 505LP mirror and 550/50 filter to detect sfGFP, and 561-nm laser line with 610LP mirror and 610/20 filter to detect mCherry.

Raw data (FCS format files) with sample legends and Source Data figures detailing the gating strategies for each analysis are available from Figshare (https://doi.org/10.6084/m9.figshare.13274837.v1). Importantly, data presented in the same panel use identical gates. We used FSC and SSC signal gating to exclude dust and clumped cells. Mating mixtures of P- and M-cells differentially labeled with sfGFP and mCherry provided fluorescent signals that were gated to distinguish gamete and zygote populations (S3B Fig), since gametes possess only one and zygotes both fluorophores. Importantly, these experiments also showed that gametes and zygotes produce populations with distinct FSC-SSC profiles (S3B Fig). We exploited this observation and used the FSC-SSC profiles to gate the zygote populations in mating mixtures produced by unlabeled gametes. When evaluating the frequency of meiotic recombinants, we used the red and green fluorescence signal to gate the populations of cells expressing neither, individual, or both fluorophores.

Microscopy and image processing

Images were obtained by wide-field microscopy performed on a DeltaVision platform (Cytiva, Massachusetts, USA) composed of a customized inverted microscope (IX-71; Olympus, Japan), a UPlan Apochromat 100x/1.4 NA or 60x/1.4 NA oil objective, either camera (CoolSNAP HQ2; Photometrics, Arizona, USA) or camera (4.2Mpx PrimeBSI sCMOS camera; Photometrics, Arizona, USA), and a color combined unit illuminator (Insight SSI 7). Images were acquired using softWoRx v4.1.2 software (Cytiva, Massachusetts, USA). We used the GFP-mCherry filterset to detect the green (Ex: 475/28, Em: 525/50) and red (Ex: 575/25; Em: 632/60) fluorescent proteins. We imaged a different number of z-sections to best capture the structure of interest. We present either a single z-plane or a projection image.

All image analysis and processing was performed using Fiji software package [69] using either built-in modules or additional modules as specified. We used the MultiStackReg v1.45 module to adjust drifting time lapses.

Time-projection images were prepared from 60 time points during 10 hours. Fluorescence channels show the maximum projection image. DIC images show either maximum or average projection image to allow for best visualization of the discussed phenotype.

Quantifications and statistics

Sample sizes were not predetermined. No randomization was used. No blinding was used. In analysis of imaging data, we excluded regions with multiple cell layers as we could not obtain reliable data. In morphometric analysis, we excluded cells where we could not follow the shmoo growth because it went out of the imaging focal plane. We report p-values from Kruskal–Wallis test generated using Matlab function ranksum() and p-values from Welch t test generated using Excel function TTEST(). Flow cytometry gating strategy is explained in detail, and identical gates are applied to samples under comparison. All quantifications except those requiring long-term time lapses were obtained from experimental triplicates. Variability between imaging chambers and their deterioration between different time points makes direct comparison between different imaging experiments difficult. For example, refertilization and lysis in meiotic signaling mutants increase over time and reach different incidence at 18 hours and 24 hours, and thus results depend on when the chambers fail. Thus, we refrain from statistical comparison between different replicates and make comparisons only between samples processed and imaged simultaneously. We stress that we reached the same qualitative results and conclusions from experimental replicates. In panels 1B, 1C, 1F, 1G, S1C, S1D, 6E, 6F, and S2A, the wild type was imaged separately from other samples.

Quantifications of sporulation efficiencies were performed on mating mixtures spotted on solid MSL −N media for 24 hours unless indicated otherwise. Samples from the edge of the mating mixture were scored for zygotic phenotypes by microscopy. We report percentages of each phenotype and p-values from Welch t test.

Quantification of zygotic phenotypes was performed from DIC time lapses where we could follow zygotes for 10 or more hours. We scored zygotes that underwent cell lysis, cell division, refertilization, or neither, and scored only the first observed event. Such scoring was adapted zygotes could carry out several events consecutively. For example, a zygote may undergo division and the daughter cells may then fuse with another partner. In quantifications that include of mei3Δ zygotes, we classified the remaining zygotes as nondividing and refrain from denoting the percentage of shmooing since zygotes form subtle shmoos which we may underreport. Since mei2Δ and mei3Δ mei2Δ zygotes form prominent mating projections, we additionally distinguish shmooing zygotes from those that abolished mating.

Quantification of diploid phenotypes (septation) was performed from DIC time lapses.

In experiments targeting Fus1 with the synthetic degron, quantification of fusion efficiencies and zygotic lysis frequencies were performed from cells mating in agarose pad chambers 24 hours after nitrogen removal. Fusion efficiency was calculated as the percentage of fused gametes compared to the total gamete pairs formed. The detailed analysis of zygotic phenotypes was performed from long-term time lapses.

DIC images were used for analyses of zygotic growth and morphology. We measured the mating projection diameter by determining the diameter of the half circle that fitted in the shmoo tip. Our measurements correlated well with the diameter across the mating projections in mei2Δ and mei3Δmei2Δ zygotes, which could not be reliably measured for the subtle mating projections in mei3Δ zygotes. We compared the cell outlines immediately after the concave neck expansion ceases (initial perimeter) and 10 hours later (final perimeter). Subtracting the area of 2 outlines determined the growth of the cross section area. To quantify the displacement of the cell perimeter during this period, we used the MorphoLibJ library [70]. Briefly, for each zygote, the initial perimeter was used to generate a binary image. We then used the MorphoLibJ function “Chamfer Distance Map” on the image to convert each pixel’s minimal distance from the initial outline into intensity of that pixel. We could then obtain intensities, which correspond to minimal distances from the initial perimeter, for each pixel on the final perimeter using built-in Fiji function “Plot Profile.” For each zygote, we then determined the maximum perimeter displacement and the coefficient of variation in perimeter displacement (for perfect isometric growth, where all measured values are identical, coefV = 0). A minimum of 20 cells was analyzed, and we report the mean measurement values and its standard deviation.

Polarity patch lifetime was determined from time lapses of cells co-expressing Scd2-GFP and Myo52-tdTomato imaged at a single plane with 10-minute intervals. The patch was visually assessed and considered 1 entity when at consecutive time points the signal, irrespective of intensity, maintained position or moved to an adjacent location. Thus, we considered as a single patch those that trail the growth site, partially lose focus, transiently dim or slide across the cortex. We evaluated 10 or more cells per genotype and compared samples using the Kruskal–Wallis test.

Fluorescence quantifications to assess efficiency of the artificial degron was performed on average projections of images containing 12 z-sections with 0.5 μm spacing. We measured the fluorescence in unlabeled wild-type cells, and the cells expressing either the fluorophore alone or together with the degron. We used the DIC channel to outline the cells, measured the mean fluorescence of 30 cells per sample and the slide background florescence. We subtracted the slide background fluorescence and compared samples using the Kruskal–Wallis test. We then calculated the average fluorescence for cells from each sample and its standard deviation. We set the average wild-type cell fluorescence to 0, the average fluorescence of cells expressing the fluorophore alone to 100, and used error propagation formula to calculate standard deviations.

To quantify Fus1-GFP levels, we first determined the average cell background fluorescence by measuring fluorescence in cells not encoding GFP and subtracting the media fluorescence measured in the neighboring regions. Similarly, we determined fluorescence in gametes expressing Fus1-GFP from 1 hour before fusion for 5 hours with 10-minute intervals and subtracted the average cell background. Since Fus1-GFP signal varied between cells, we normalized the signal at the time of gamete fusion to 1 and calculated average and standard deviation for 8 different mating pairs per sample. We note that the whole-cell Fus1-GFP fluorescence is weak, which makes it difficult to determine how much of the signal remains in zygotes co-expressing DegGreen.

Quantifications of Mi induction were performed on single z-plane time lapses of nitrogen-starved H1Δ17 diploids. We used the DIC channel to outline the cells over time and measure mean cell fluorescence. We used these measurements to visually determine the time of Mi induction and align time courses from different cells. The Mi signal also allowed us to visualize and outline the circular nuclei and measure the mean fluorescence. We then subtracted the slide background and report the mean fluorescence (line) and standard deviation from measurements of 8 cells.

To quantify cell cycle progression upon nitrogen starvation, we performed 24-hour time-lapse imaging of eGFP-Pcn1 immediately after cells were removed from MSL+N media, washed and mounted onto MSL −N agarose pads. We tracked cells which were in G2 phase at the beginning of imaging and scored their passage through S-phase. All cells went through at least 1 cell division, and we report the frequency of cells passing the subsequent S-phase. For cells that progressed to divide for a second time, we also report the frequency of their entry into the next S-phase.

Quantifications of the S-phase timing used the exchange of cytosolic mCherry between partner gametes to determine the fusion time (T_fusion). Time of S-phase was recorded as the first time point with punctate GFP-Pcn1 localization. We analyzed only cells that fused within the first 5 hours since the onset of imaging since we noticed that zygotes formed at late time points took longer time to exit G1 in all crosses investigated, possibly due to depletion of nutrients from the agarose pads used for imaging. Additionally, we ensured that the populations we compare have similar average T_fusion and do not show statistically significant T_fusion differences (p^{Kruskal–Wallis} > 0.1).

To determine the meiotic recombination frequencies, in 1 parental strain, we introduced at the leu1 genomic locus sequences driving sfGFP expression from the strong p^tdh1 promoter, and in the other parental strain at the aha1 genomic locus sequences driving mCherry expression also from the strong p^tdh1 promoter. We chose to integrate the constructs proximal to leu1 and aha1 loci in regions with no known function. Mating between these 2 parental strains yielded nonrecombinant progeny that expresses 1 fluorophore and recombinant progeny that expresses either both or neither fluorophore. We used flow cytometry to determine the frequencies of recombinant and nonrecombinant progeny and the genetic distance of the 2 loci in cM (centimorgan) was determined according to the following formula:

The experiment was performed in triplicate, and we report mean values, standard deviation, and p-values obtained using the Welch t test. We obtained that the genetic distance between the leu1 and aha1 loci is 17.13 ± 0.14 cM in wild-type crosses, which is consistent with previous report of approximately 18 cM separating the leu1 and aha1-proximal his2 gene [71,72].