Transient Breakage of the Nucleocytoplasmic Barrier Controls Spore Maturation via Mobilizing the Proteasome Subunit Rpn11 in the Fission Yeast Schizosaccharomyces pombe

Forespore membrane (FSM) closure is a process of specialized cytokinesis in yeast meiosis. FSM closure begins with the contraction of the FSM opening and finishes with the disassembly of the leading-edge proteins (LEPs) from the FSM opening. Here, we show that the FSM opening starts to contract when the event of virtual nuclear envelope breakdown (vNEBD) occurs in anaphase II of the fission yeast Schizosaccharomyces pombe. The occurrence of vNEBD controls the redistribution of the proteasomal subunit Rpn11 from the nucleus to the cytosol. To investigate the importance of Rpn11 re-localization during vNEBD, Rpn11 was sequestered at the inner nuclear membrane by fusion with the transmembrane region of Bqt4 (Rpn11-GFP-INM). Remarkably, in the absence of endogenous rpn11+, the cells carrying Rpn11-GFP-INM had abnormal or no spore formation. Live-cell imaging analysis further reveals that the FSM opening failed to contract when vNEBD occurred, and the LEP Meu14 was persistently present at the FSM in the rpn11-gfp-INM cells. The results suggest that the dynamic localization of Rpn11 during vNEBD is essential for spore development.


Introduction
The fission yeast Schizosaccharomyces pombe enters meiosis to form ascospores upon nitrogen starvation [1]. Meiosis involves two sequential nuclear divisions in meiosis I and meiosis II. S. pombe undergoes meiotic nuclear divisions without nuclear envelope breakdown. Even though the nuclear envelope remains intact, the nucleocytoplasmic barrier is transiently ineffective at the onset of anaphase II, an event called virtual nuclear envelope breakdown (vNEBD) [2,3]. During vNEBD, the nuclear proteins, except for the integral nuclear membrane proteins, diffuse to the cytosol. The occurrence of vNEBD is actively regulated by the RNA binding protein Spo5, a master regulator of meiosis II progression in S. pombe. In the absence of Spo5, the meiotic cells undergo the second nuclear division without vNEBD and fail to form mature spores [2,4,5]. Nonetheless, the causal relationship between vNEBD and spore formation remains undefined. Given that spore development starts in meiosis II, it is tempting to think that the release of the nuclear proteins to the cytosol during vNEBD is involved in the process of spore maturation [6].

S. pombe Cell Growth and Sporulation
S. pombe was grown on yeast extract with supplements of adenine, histidine, leucine, lysine, and uracil (YES) plates or Edinburgh minimal medium (EMM) with the appropriate supplements at 30 • C [21]. For sporulation, cells were freshly grown on YES plates overnight, and resuspension of the overnight culture at a density of 10 9 cells/mL was spotted on a malt extract (ME) plate. The cells on the ME plates were incubated at 26 • C. Between 8 to 15 h after sporulating on the ME plate, the cells undergoing karyogamy were picked for live-cell imaging. For sporulation frequency, the yeast cells were allowed to sporulate on the ME plate for 3-4 days.
The resultant transformants were sporulated and dissected to obtain the haploid rpn11∆ cells that were resistant to aureobasidin A (Takara Bio Inc., Kusatsu, Japan). On the other hand, Rpn11 endogenously tagged with GFP was derived from the strain 9184 (a gift from Kathleen L. Gould, Vanderbilt University School of Medicine, Nashville, TN, USA; [25]). The FSM marker was encoded by mCherry-tagged psy1 + integrated at the lys1 + -locus. Chromosomes were visualized using histone H3 Hht1 tagged with mCherry at the native chromosome locus. A two-step PCR method introducing chromosomal mCherry tag was used to fluorescently label Meu14 and Hht1 [26]. Live-cell imaging analysis was performed using an Olympus inverted microscope IX70 equipped with a CoolSNAP HQ 2 charge-coupled device as previously described (Photometrics, Tucson, AZ, USA; [27]). Briefly, a touch of the sporulating cells on the ME plate was resuspended in the medium of EMM minus nitrogen with the supplements. The solution was added to a lectin-coated (0.2 mg/mL; lectin from Glycine max, Sigma-Aldrich, Tokyo, Japan) glass-bottom dish (MatTek Corp, Ashland, OR, USA). The cells were allowed to adhere to the glass slide via gravity for 5 min. The cells were observed using the Olympus oil-immersion 60× objective lens (PlanApoN60× OSC; NA = 1.4; Olympus, Tokyo, Japan). Optical section images were acquired at 0.5-µm intervals every 6 min using DeltaVision softWoRx 5.5 (Applied Precision, Inc., Seattle, WA, USA). The images were enhanced by three-dimensional constrained iterative deconvolution in softWoRx 5.5.

Diffusion of Nuclear Rpn11 during Virtual Nuclear Envelope Breakdown (vNEBD)
Localization of the proteasomal subunit Rpn11 was examined during sporulation. Two opposite mating types of the heterothallic haploid cells that respectively expressed Rpn11-GFP and the chromosome marker Hht1-mCherry were mated to enter meiosis and sporulation. Before nuclear fusion, the fluorescence signals of Rpn11-GFP and Hht1-mCherry distinctly labeled each of the two nuclei in the conjugating cell ( Figure 1A, time = 0 min). After karyogamy, the two signals merged into a single nucleus where Rpn11-GFP surrounded the chromosome and outlined the nucleus, indicating that Rpn11 localizes at the nuclear periphery and nucleoplasm ( Figure 1A, time = 0~108 min). During chromosome segregation in meiosis I, the signals of Rpn11-GFP were enriched around the chromosomes, as previously reported [20] ( Figure 1A, time = 180-198 min). On the other hand, the Rpn11-GFP signals shortly disappeared from the nucleus during anaphase II ( Figure 1A, time = 240-246 min) but soon re-localized back to the nuclear periphery of the daughter nuclei after completion of the second nuclear division ( Figure 1A, time = 258 min). The transient disappearance of Rpn11-GFP from the nucleus was also observed in anaphase II of the wild-type homothallic sporulating cells ( Figure 1B, time = 6-12 min). We investigated whether the disappearance of Rpn11-GFP depends on the occurrence of vNEBD that controls the diffusion of nuclear protein to cytosol at the onset of anaphase II [2]. To do so, we used the spo5Δ mutant that undergoes the second nuclear division without vNEBD [2]. In the spo5Δ homothallic mutant, the signals of Rpn11-GFP remained in the nucleus at the onset of anaphase II ( Figure 1C, time = 6-12 min). The results indicated that nuclear Rpn11 diffuses to the cytosol via vNEBD in sporulating cells.

Retention of Rpn11 at the Inner Nuclear Membrane during vNEBD Causes Abnormal Spore Formation
It is hypothesized that subcellular re-localization of the nuclear proteins in vNEBD influences spore formation. In an attempt to sequester Rpn11 in the nucleus, we made a fusion construct in which Rpn11-GFP was additionally tagged with the Bqt4 C-terminus (Rpn11-GFP-INM). The Cterminal region of Bqt4 contains the transmembrane domain and inner nuclear membrane targeting sequences [24]. Meiotic localization of Rpn11-GFP-INM was examined in the cells co-expressing mCherry-labeled Rpn11. Unlike the fluorescence signals of Rpn11-mCherry at the nuclear rim and nucleoplasm, the signals of Rpn11-GFP-INM were restricted at the nuclear periphery before meiosis II ( Figure 2, time = −144-0 min). Moreover, while Rpn11-mCherry diffused away from the nucleus The transient disappearance of Rpn11-GFP from the nucleus was also observed in anaphase II of the wild-type homothallic sporulating cells ( Figure 1B, time = 6-12 min). We investigated whether the disappearance of Rpn11-GFP depends on the occurrence of vNEBD that controls the diffusion of nuclear protein to cytosol at the onset of anaphase II [2]. To do so, we used the spo5∆ mutant that undergoes the second nuclear division without vNEBD [2]. In the spo5∆ homothallic mutant, the signals of Rpn11-GFP remained in the nucleus at the onset of anaphase II ( Figure 1C, time = 6-12 min). The results indicated that nuclear Rpn11 diffuses to the cytosol via vNEBD in sporulating cells.

Retention of Rpn11 at the Inner Nuclear Membrane during vNEBD Causes Abnormal Spore Formation
It is hypothesized that subcellular re-localization of the nuclear proteins in vNEBD influences spore formation. In an attempt to sequester Rpn11 in the nucleus, we made a fusion construct in which Rpn11-GFP was additionally tagged with the Bqt4 C-terminus (Rpn11-GFP-INM). The C-terminal region of Bqt4 contains the transmembrane domain and inner nuclear membrane targeting sequences [24]. Meiotic localization of Rpn11-GFP-INM was examined in the cells co-expressing mCherry-labeled Rpn11. Unlike the fluorescence signals of Rpn11-mCherry at the nuclear rim and nucleoplasm, the signals of Rpn11-GFP-INM were restricted at the nuclear periphery before meiosis II (  Figure 3A). Because rpn11 + is an essential gene, the results indicate that Rpn11-GFP-INM is functional for vegetative growth. Next, we examined the ability of cells to undergo sporulation by quantifying the tetrad formation frequency. S. pombe normally forms four spores per ascus (tetrad). In the presence of endogenous rpn11 + , around 90% of the zygotic cells expressing Rpn11-GFP-INM formed the tetrads (rpn11 + rpn11-gfp-INM in Figure 3B,C). In contrast, more than 40% of the rpn11-gfp-INM zygotic cells gave rise to the non-tetrad formation (rpn11Δ rpn11gfp-INM in Figure 3B,C). In particular, over two-thirds of the non-tetrad in the rpn11-gfp-INM mutant showed no spore formation ( Figure 3B,C). We reasoned that the dynamic localization of Rpn11 is vital for its full function in tetrad formation. To test this idea, we performed a rescue experiment. Adding extra copies of genes expressing Rpn11-mCherry in the rpn11-gfp-INM cells fully restored tetrad formation frequency comparable to the wild-type level. By contrast, the mutant cells  Figure 3A). Because rpn11 + is an essential gene, the results indicate that Rpn11-GFP-INM is functional for vegetative growth. Next, we examined the ability of cells to undergo sporulation by quantifying the tetrad formation frequency. S. pombe normally forms four spores per ascus (tetrad). In the presence of endogenous rpn11 + , around 90% of the zygotic cells expressing Rpn11-GFP-INM formed the tetrads (rpn11 + rpn11-gfp-INM in Figure 3B,C). In contrast, more than 40% of the rpn11-gfp-INM zygotic cells gave rise to the non-tetrad formation (rpn11∆ rpn11-gfp-INM in Figure 3B,C). In particular, over two-thirds of the non-tetrad in the rpn11-gfp-INM mutant showed no spore formation ( Figure 3B,C). We reasoned that the dynamic localization of Rpn11 is vital for its full function in tetrad formation. To test this idea, we performed a rescue experiment. Adding extra copies of genes expressing Rpn11-mCherry in the rpn11-gfp-INM cells fully restored tetrad formation frequency comparable to the wild-type level. By contrast, the mutant cells expressing Rpn11-mCherry fused with the Bqt4 C-terminus (Rpn11-mCh-INM) still had a significant ratio of no spore formation. The Bqt4 transmembrane region of the Rpn11-mCh-INM was then removed to allow Rpn11 mobilization (Rpn11-mCh-INM∆TM). Strikingly, the moveable Rpn11 fusion protein expression in the rpn11-gfp-INM cells gave rise to the wild-type level of tetrad formation frequency ( Figure 3C). The results suggest that rpn11-gfp-INM represents a recessive allele for non-tetrad formation and emphasize the important role of moveable Rpn11 in sporulation.

FSM Closure is Defective When Rpn11 Is Confined at the Inner Nuclear Membrane during vNEBD
Around one-fourth of the rpn11-gfp-INM zygotic cells failed to form visible spores after three days of sporulation (Figure 3), raising the possibility that the mutant cells are defective in spore development per se. We therefore monitored the initial event of spore development, i.e., FSM assembly, through time-course microscopy of mCherry-tagged Psy1 in the rpn11-gfp-INM cells. The t-SNARE protein Psy1 that mediates membrane vesicle fusion at the FSM serves as the FSM marker [8]. As previously reported [9], the FSM precursors appeared as cytosolic punctate structures that coalesced into four arcs near the two meiosis II nuclei before anaphase II (Figure 4, 0-12 min), indicating fusion of the membrane vesicles into the FSM. After anaphase II, the FSM extended along the nuclear envelope to capture the divided nuclei (Figure 4, 18-42 min). We concluded that FSM initiation and extension is normal in the rpn11-gfp-INM cells.

FSM Closure is Defective When Rpn11 Is Confined at the Inner Nuclear Membrane during vNEBD
Around one-fourth of the rpn11-gfp-INM zygotic cells failed to form visible spores after three days of sporulation (Figure 3), raising the possibility that the mutant cells are defective in spore development per se. We therefore monitored the initial event of spore development, i.e., FSM assembly, through time-course microscopy of mCherry-tagged Psy1 in the rpn11-gfp-INM cells. The t-SNARE protein Psy1 that mediates membrane vesicle fusion at the FSM serves as the FSM marker [8]. As previously reported [9], the FSM precursors appeared as cytosolic punctate structures that coalesced into four arcs near the two meiosis II nuclei before anaphase II (Figure 4, 0-12 min), indicating fusion of the membrane vesicles into the FSM. After anaphase II, the FSM extended along the nuclear envelope to capture the divided nuclei (Figure 4, 18-42 min). We concluded that FSM initiation and extension is normal in the rpn11-gfp-INM cells. assembly, through time-course microscopy of mCherry-tagged Psy1 in the rpn11-gfp-INM cells. The t-SNARE protein Psy1 that mediates membrane vesicle fusion at the FSM serves as the FSM marker [8]. As previously reported [9], the FSM precursors appeared as cytosolic punctate structures that coalesced into four arcs near the two meiosis II nuclei before anaphase II (Figure 4, 0-12 min), indicating fusion of the membrane vesicles into the FSM. After anaphase II, the FSM extended along the nuclear envelope to capture the divided nuclei (Figure 4, 18-42 min). We concluded that FSM initiation and extension is normal in the rpn11-gfp-INM cells.  The final step of FSM assembly is FSM closure, which separates the spore cytoplasm from the ascus cytoplasm. The fusion protein Meu14-mCherry was used to examine FSM closure dynamics because the LEP protein Meu14 decorates the FSM opening. We first monitored the signals of Meu14-mCherry in the cells expressing Rpn11-GFP as the wild-type control. Before anaphase II, Meu14-mCherry appeared as the ring (or line) structures surrounding the nuclei at the diameter of 2 µm ( Figure 5, 0-6 min). Typically, the Meu14 rings' diameter started to shrink as the nuclear Rpn11-GFP diffused to the cytosol during vNEBD at the onset of anaphase II ( Figure 5A upper panel, 12 min). The size of each Meu14 ring continually dropped with an average rate of 0.036 µm per minute until the signals of Meu14-mCherry were undetectable, indicating the closing of the FSM (Figure 5A upper panel, and Figure 5B blue line, 12-54 min). On the contrary, the Meu14 ring size remained constant at the onset of anaphase II, during which Rpn11-GFP-INM stayed at the inner nuclear rim in the rpn11-gfp-INM cells ( Figure 5A lower panel, 12 min). From anaphase II onward, the diameter of the Meu14 ring slightly decreased but never down to wild-type levels ( Figure 5, 12-54 min). Moreover, the signals of Meu14 were persistently present in the ascus cytoplasm of the rpn11-gfp-INM cells. Because Meu14 decorates the FSM opening and disappears from the FSM after FSM closure [12], the results suggest that the FSM fails to close when Rpn11 is forced to stay at the nuclear rim during vNEBD. The final step of FSM assembly is FSM closure, which separates the spore cytoplasm from the ascus cytoplasm. The fusion protein Meu14-mCherry was used to examine FSM closure dynamics because the LEP protein Meu14 decorates the FSM opening. We first monitored the signals of Meu14-mCherry in the cells expressing Rpn11-GFP as the wild-type control. Before anaphase II, Meu14-mCherry appeared as the ring (or line) structures surrounding the nuclei at the diameter of 2 μm ( Figure 5, 0-6 min). Typically, the Meu14 rings' diameter started to shrink as the nuclear Rpn11-GFP diffused to the cytosol during vNEBD at the onset of anaphase II ( Figure 5A  Because Meu14 decorates the FSM opening and disappears from the FSM after FSM closure [12], the results suggest that the FSM fails to close when Rpn11 is forced to stay at the nuclear rim during vNEBD.

Conclusions and Discussion
The role of vNEBD in spore formation has not previously been defined. The current study has identified proteasomal Rpn11 as the vNEBD substrate of which nuclear localization changes to the cytoplasm in anaphase II (Figure 1). Through fusing Rpn11-GFP with the transmembrane domain of the inner nuclear membrane protein Bqt4, we were able to constrain Rpn11 at the nuclear periphery

Conclusions and Discussion
The role of vNEBD in spore formation has not previously been defined. The current study has identified proteasomal Rpn11 as the vNEBD substrate of which nuclear localization changes to the cytoplasm in anaphase II (Figure 1). Through fusing Rpn11-GFP with the transmembrane domain of the inner nuclear membrane protein Bqt4, we were able to constrain Rpn11 at the nuclear periphery during vNEBD (Figure 2). Many of the cells with sequestered Rpn11 at the nuclear periphery barely had ascospore formation (Figure 3). The FSM closure, but not the FSM growth, was defective in the rpn11-gfp-INM mutant (Figures 4 and 5), consistent with the fact that the occurrence of vNEBD is after FSM growth and followed by FSM closure. Therefore, Rpn11 redistribution to the cytosol via vNEBD is important for spore maturation. At this moment, we cannot rule out that a small fraction of Rpn11 goes to the cytoplasm in meiotic prophase and meiosis I. However, we argue that the spo5 mutants whose cytoplasmic proteasome availability is normal before vNEBD still fail to form mature spores [5], suggesting that cytoplasmic proteasome availability in prophase and meiosis I plays little role in spore development.
FSM closure, a unique form of cytokinesis, starts with constriction of the LEP ring after anaphase II. The septation initiation network (SIN) coordinates the anaphase exit and cytokinesis onset [28]. Intriguingly, the sin mutant displays belated LEP ring contraction similar to the rpn11-gfp-INM mutant [11]. Moreover, several SIN components undergo proteasomal degradation after anaphase II [29]. Proteolysis of the SIN components is inhibited in the spo6 mutant of which vNEBD fails to occur [29]. Hence, it is possible that the proteasome subunit Rpn11 regulates FSM closure through the SIN signaling pathway.
It is tempting to think that not only Rpn11 but the 26S holoenzyme disperses from the nucleus to the cytosol during vNEBD, given that the proteasome anchor Cut8 [30,31] also diffuses at anaphase II onset (data not shown). Cytoplasmic relocation of the proteasome occurs possibly to obtain access to the cytosolic substrates, such as the LEP at the FSM leading edges. In S. cerevisiae, the LEP Ssp1 is targeted towards proteasomal degradation by the Ama1-activated APC/C. The ama1∆ mutant whose Ssp1 fails to be degraded has an unsealed FSM and no spore wall [16,32]. FSM closure is a prerequisite for spore wall development [33,34]. Intriguingly, S. pombe cells lacking Fzr1 (homolog of S. cerevisiae Ama1) barely assemble the spore wall [35]. Moreover, S. pombe LEP Meu14 undergoes proteolysis after anaphase II [12]. It remains to be determined whether Meu14 is subjected to proteasomal degradation by Fzr1-activated APC/C. On the other hand, we observed that Meu14 is persistently present when the proteasome subunit Rpn11 is sequestered to the inner nuclear membrane. Thus, proteasomal degradation of the LEP is likely the conserved mechanism underlying FSM closure in both budding and fission yeasts.