Anillin‐related protein Mid1 regulates timely formation of the contractile ring in the fission yeast Schizosaccharomyces japonicus

Abstract

In the fission yeast Schizosaccharomyces pombe (Sp), Mid1/Dmf1 plays an important role in positioning the division site by inducing formation of the contractile ring (CR). Mid1, emanating from the nucleus located in the cell center, forms a dozen of nodes in the middle cell cortex ahead of mitosis, and actin filaments and myosin II accumulated at each node interact and assemble the CR in metaphase. Curiously, in another fission yeast S. japonicus (Sj), CR formation begins after nuclear segregation in late anaphase. Here, we investigated the role of S. japonicus Mid1 during mitosis to compare the molecular mechanisms that determine the cell division site in Schizosaccharomyces. Similar to Sp Mid1, Sj Mid1 often accumulated in the nucleus of interphase cells. Moreover, Sj Mid1 localized to cortical dots with myosin II in the future division site and formed a medial ring in mitotic cells. However, S. japonicus cells without Mid1 function still carried out symmetrical binary division. Therefore, the Mid1 dependency for positional control of the cell division site is possibly different between the two species. Meanwhile, we found that Sj Mid1 enhanced CR formation, in a manner possibly similar to that by Sp Mid1.

Introduction

During the contractile ring (CR)‐dependent cytokinesis of eukaryotic cells, the interactions of filamentous actin (F‐actin) and myosin II in the CR are thought to generate forces to constrict the cell just beneath the cleavage furrow, enabling the mother cell to divide into two daughter cells. However, the molecular mechanism of the mitosis‐specific assembly of F‐actin and myosin II in CR formation at the proper region beneath the plasma membrane is not entirely clear.

The fission yeast Schizosaccharomyces pombe is an ideal organism to study the molecular mechanisms controlling CR formation during cytokinesis. In this organism, it has been showed that the nucleocytoplasmic shuttling protein Mid1/Dmf1 plays a central role in the spatiotemporal control of CR formation (Chang et al. 1996; Sohrmann et al. 1996; Daga & Chang 2005; reviewed by Rincon & Paoletti 2012). Mid1 has a nuclear localization signal (NLS), a nuclear exclusion signal (NES) and an amphipathic sequence followed by a Pleckstrin homology (PH) domain located at the C‐terminus, which are considered important for its functions (Sohrmann et al. 1996; Paoletti & Chang 2000; Lee & Wu 2012). In interphase, most Mid1 is located in the nucleus, but residual amounts of Mid1 form cortical dots called interphase nodes by interacting with the Cdr2 kinase and other proteins associated with the nodes (Almonacid et al. 2009; Rincon et al. 2014). The distribution of interphase nodes is restricted to the middle of the cell by exclusion from cell tips by the Pom1 kinase. As the G2 phase progresses, the amount of Mid1 at the equatorial cortex, a region proximal to the nucleus positioning in the center of interphase cells, increases and other components such as Gef2, Blt1 and Klp8 localize to the interphase nodes (Moseley et al. 2009; Guzman‐Vendrell et al. 2013; Akamatsu et al. 2014). Just after a cell enters mitosis, Cdr2 and other proteins are released from the interphase nodes and Mid1 and its associated proteins form cytokinetic nodes together with an IQGAP‐like protein Rng2, myosin II consisting of the heavy chain Myo2 and light chains Rlc1 and Cdc4, an F‐BAR domain protein Cdc15 and formin Cdc12 (Wu et al. 2003, 2006; Vavylonis et al. 2008). The cytokinetic nodes induce a cortical actin–myosin meshwork encircling the future division site in the cell middle and are merged with each other. As a result, the CR is assembled before the onset of anaphase. In S. pombe, a gene disruption of mid1⁺ is lethal for cell proliferation at high temperatures and causes severe defects in the spatiotemporal control of CR formation. The position and orientation of the CR is frequently disturbed in mid1 null cells (Sohrmann et al. 1996; Bähler et al. 1998a; Paoletti & Chang 2000), and the timing of CR component accumulation is delayed in the mutants than that in the wild‐type cells (Motegi et al. 2004; Hachet & Simanis 2008; Huang et al. 2008). Thus, Mid1 plays a central role in the spatiotemporal control of CR formation in S. pombe.

Mid1 has structural similarity to anillin, which plays an important role in animal cell cytokinesis (reviewed by D'Avino 2009; and Piekny & Maddox 2010). Anillin functionally interacts with F‐actin, myosin II, septin and the plasma membrane under the control of Rho‐type small GTPases to ensure cytokinesis progression (Oegema et al. 2000; Kinoshita et al. 2002; Straight et al. 2005; Hickson & O'Farrell 2008; Piekny & Glotzer 2008; Watanabe et al. 2010). Both Mid1 and anillin share a conserved sequence called an anillin homology domain (AHD) that forms a C2 domain‐like structure with a lipid‐binding amphipathic region followed by a PH domain at their C‐termini (Sun et al. 2015). However, it is still unclear whether anillin and Mid1 share a common ancestral gene. Also, an orthologue of Mid1 has not been found in budding yeasts and filamentous fungi even though many of their genomes have been sequenced and annotated.

However, the genome of S. pombe contains mid2⁺, which encodes a protein that has an AHD and a PH domain at its C‐terminus and localizes to the cell division site during cytokinesis (Berlin et al. 2003; Tasto et al. 2003). Moreover, Mid2 functionally interacts with the septin cytoskeleton (Berlin et al. 2003; Tasto et al. 2003; Martín‐Cuadrado et al. 2005). Cells lacking mid2⁺ form a normal CR but have defective cell separation after cytokinesis in S. pombe. Importantly, mid2⁺ homologues are well conserved among yeasts and fungi (Sanders & Herskowitz 1996; Gale et al. 2001; Berlin et al. 2003; Si et al. 2012). BUD4 is one of them. This gene product controls bud site selection by interacting with the bud neck filaments consisted of septin in the budding yeast Saccharomyces cerevisiae (Sanders & Herskowitz 1996; Gladfelter et al. 2005; Wu et al. 2015). In the filamentous fungus Neurospora crassa, the Bud4 homologue is required for the process of septum formation by interacting with the Rho family small GTPases (Justa‐Schuch et al. 2010; Delgado‐Álvarez et al. 2014). The primary structure of Bud4 is more similar to Sp Mid2 than Sp Mid1 (Fig. S1 in Supporting Information). It is possible that mid1⁺ may have been derived from a mid2⁺ gene duplication event in Schizosaccharomyces after this group had branched off from other fungi.

Here, we investigated the role of Mid1 through the organization of F‐actin and Rlc1, a light chain of myosin II, during CR formation in Schizosaccharomyces japonicus. This fission yeast is considered as the first to branch among the four Schizosaccharomyces species whose genomes have been sequenced (Rhind et al. 2011). Interestingly, pioneer work carried out by Alfa & Hyams (1990) showed that an accumulation of F‐actin in the equatorial region occurs during anaphase in S. japonicus. However, the CR is formed before the onset of anaphase in S. pombe as mentioned above. The discordance on the timing of CR formation between the two Schizosaccharomyces species may imply that the molecular mechanisms controlling CR formation differ between the two fission yeasts. Moreover, the function of Mid1 in controlling the organization of the CR in S. japonicus has not been explored until recently (Gu et al. 2015). A gene manipulation technique that has recently been established in S. japonicus (reviewed by Niki 2014) enabled us to study the function of Mid1 in this organism. In this manuscript, we report a comparative study of CR formation in Schizosaccharomyces species and address the functional significance of Mid1 during S. japonicus cytokinesis.

Results

Sj Mid1 localizes to the equatorial cell cortex and changes distribution from dots to a ring

To study the cellular localization of S. japonicus Mid1 (Sj Mid1) by fluorescence microscopy, we prepared S. japonicus strains in which the gene for mCherry or gfp was introduced into an upstream position adjacent to the stop codon for mid1⁺ or rlc1⁺. A fluorescence‐protein‐fused myosin regulatory light chain Rlc1 was a conventional marker protein to visualize CR dynamics in S. pombe (Le Goff et al. 2000; Naqvi et al. 2000; Wu et al. 2003). A cylindrically shaped cell of fission yeast grows at the cell tip(s) and symmetrically divides into two daughter cells when the cell size reaches an appropriate length (Mitchison 1957; Mitchison & Nurse 1985). In relatively short cells presumed to be in interphase, we found that a significant fraction of Sj Mid1 localized with Sj Rlc1 in cortical dots at the middle region of cells (Fig. 1Aa, brackets), although only Sj Mid1 was detected in the nucleus. Those dots may correspond to the interphase nodes of S. pombe (Paoletti & Chang 2000). Moreover, Sj Mid1‐mCherry formed a medial ring encircling the cells containing CR (Fig. 1Aa, arrows). Sj Mid1 dots may narrow their distribution in S. japonicus similar to Mid1 behavior during metaphase in S. pombe (Sohrmann et al. 1996; Bähler et al. 1998; Paoletti & Chang 2000; Wu et al. 2003). A strain with a reciprocal combination of the fluorescently tagged genes was also examined. Observing those cells with fluorescence microscopy, discontinuous filament‐like structures in the cytoplasm were frequently observed using fluorescence microscopy with blue‐light excitation for GFP (Fig. 1Ab). The same signal was also detected in wild‐type cells not expressing a GFP construct (Fig. 1Ac), suggesting that the signal was autofluorescence presumably emanating from a cellular structure such as the mitochondria. Such autofluorescence was not detected in the confocal laser microscopy (Fig. 1B). In this strain, Sj Mid1‐GFP and Sj Rlc1‐mCherry colocalized on a medial ring in elongated cells (Fig. 1Ab, arrows,B). However, Sj Mid1‐GFP remained in a lateral region of septating cells, whereas a ring containing Sj Rlc1‐mCherry, namely CR, was closing (Fig.  1Ab, arrowheads). The different behavior of these proteins was confirmed by time‐lapse microscopy. After contraction of the Rlc1‐mCherry ring had occurred, Sj Mid1‐GFP remained on the cell cortex until the signal of Sj Mid1‐GFP reached an undetectable level in late cytokinesis (Fig. 1C).

Figure 1

Sj Mid1 localizes to the equatorial region of the cell cortex. (A) JK133 (mid1‐mCherry rlc1‐GFP) (a) and JK113 (mid1‐GFP rlc1‐mCherry) (b) were observed. Cells were logarithmically grown in EMM at 25 °C. Bright‐field image (BF; left) and inverted fluorescence images (middle and right) are shown. The bottom panels of (a) show magnified images of the trunk region of the cell indicated by large arrowhead (Bar, 2 μm). Mid1 often localized with Rlc1 in cortical dot(s) in the middle of cells with a relatively short cell length (brackets), whereas both proteins form a medial ring in cells fully elongated. Arrows indicate a cross section of the Mid1 ring. After the onset of CR constriction, the signals deriving from the proteins separated; Sj Mid1 remained in the medial cortex (small arrowheads), whereas Sj Rlc1 associated with the contractile ring (CR). Five dividing cells (cells numbered 1~5) and one interphase cell (cell 6) are shown in (b). The numbers are related to the progression of the cell division stage. The discontinuous filament‐like signal in the cells shown in the middle panel of (b) is autofluorescence possibly from mitochondria, as the same pattern of signal is detected in wild‐type strains not possessing any gfp gene construct (c) by the same exposure condition as in (b). Note that the similar pattern is found in the Rlc1‐GFP panel in (a) although the signal is faint because of the short time exposure. (d) JK92 strain (mid1‐mCherry cut11‐GFP) was observed. Asterisk indicates a septate cell. (B) Three‐dimensional reconstruction image of JK113 cells. Z‐sectioned images (0.3‐μm interval) were reconstructed by maximum projection (upper) and are rotated (70°) around the x‐axis (bottom). Note that Mid1 and Rlc1 form the medial ring. (C) Time‐lapse observation of JK113 cells. Three‐dimensional images reconstructed by maximum projection are shown. Mid1‐GFP and Rlc1‐mCherry localized together to cortical dots before formation of the medial ring (cell 1 and cell 2 at time 0). Within approximately ten minutes after both proteins had formed the medial ring (arrows), contraction of the CR occurred as judged by the behavior of Rlc1‐mCherry. However, Mid1‐GFP remained at the cell cortex after the CR began constricting as indicated by arrowheads, and those signals gradually disappeared from the medial region (cell 1, 33–42 min; cell 2, 42–45 min; cell 3, 6–15 min). Bars, 10 μm.

It has been reported that a large fraction of Sp Mid1 is sequestered in the interphase nucleus (Sohrmann et al. 1996; Bähler et al. 1998a; Paoletti & Chang 2000). Localization of Sj Mid1 in the nucleus was often observed in cells expressing Mid1‐mCherry together with Cut11‐GFP (Aoki et al. 2011) to visualize the nuclear envelope (Fig. 1Ad). In addition, we treated S. japonicus cells with leptomycin B (LMB), an inhibitor of nuclear export of NES‐containing proteins (Nishi et al. 1994), and evaluated the nuclear accumulation of Sj Mid1. We found that cells showed nuclear localization of Sj Mid1 in an LMB‐dependent manner (Fig. 2). Thus, Sj Mid1 may also shuttle between the nucleus and the cytoplasm in S. japonicus cells.

Figure 2

Nuclear localization of Sj Mid1 is enhanced by leptomycin B (LMB) treatment. S. japonicus cells expressing Mid1‐mCherry (JK30) and Rlc1‐mCherry (JK31) were treated with LMB at a final concentration of 100 μg/mL for 2 h. Hoechst 33342 was applied to visualize nuclear DNA and septum. Inverted fluorescence images are shown. Panels in the top row show control cells treated with vehicle (ethanol) for the same period of time. Arrowheads indicate nuclear accumulation of Sj Mid1‐mCherry. Note that LMB treatment induces abnormal condensation of nuclear DNA. No accumulation of Sj Rlc1‐mCherry was found in the presence or absence of LMB. Bar, 10 μm.

Sj Mid1 is not essential for vegetative cell growth

To examine the functional importance of Sj Mid1 for cytokinesis in S. japonicus, the open reading frame of mid1⁺ was replaced with a kanMX6 drug resistance cassette by homologous recombination. The Δmid1‐null strain did not show defective in cell growth or morphology. Δmid1‐null cells were able to form colonies on a YE plate incubated at 16, 25, 30, 36 and 42 °C similar to wild‐type cells (Fig. 3A). Therefore, Sj Mid1 was dispensable for cell growth in S. japonicus. In addition, it appeared that the position of the septum in the cell body and cell separation after cytokinesis was unaffected in Δmid1‐null cells (Fig. 3B). These data suggest that the dependency on Mid1 to position the cell division plane differs between S. japonicus and S. pombe, because a deletion of the mid1⁺ gene caused abnormal septum formation and temperature‐sensitive growth defects in S. pombe (Sohrmann et al. 1996).

Figure 3

Sj Mid1 is not essential for cell proliferation. (A) Wild‐type (WT; NIG2028) and Δmid1 (JK105) cells were spread on YE plates and incubated at the indicated temperatures for 3 days. Incubation of a plate at 16 °C was for 1 month. No significant difference in cell growth was found between the strains under any condition. (B) Bright‐field images of WT and Δmid1 cells incubated overnight in YE liquid medium at the indicated temperatures. Chain‐like cells not separated after cytokinesis were occasionally found in WT and Δmid1 cells (arrows), but their frequency was not significantly different between the two strains. Bar, 10 μm.

Sj Mid1 promotes myosin II and F‐actin accumulation for CR formation

To examine CR formation in Δmid1‐null cells, Sj Cut11‐GFP and Sj Rlc1‐mCherry were simultaneously introduced into the deletion strain. We noticed that the percentage of anaphase cells with Sj Rlc1‐mCherry localization to the middle of the cell was lower in the Δmid1 strain compared to wild type (Fig. 4A,B). Sj Rlc1‐mCherry localized to the middle of the cell just before telophase and formed a ring, but its distribution was somewhat faint and uneven in Δmid1‐null cells (compare ‘cell 3’ in Δmid1 with WT in Fig. 4A). Thus, these data suggest that Sj Mid1 might facilitate myosin II localization to the middle of the cell before cytokinesis and might be involved in the homogeneous distribution of myosin II in the CR during CR formation.

Figure 4

Effects of gene disruption of mid1⁺ on contractile ring (CR) formation. (A) Cortical accumulation of Rlc1 at the cell center is affected in Δmid1 cells. WT (JK94) and Δmid1 (JK148) cells expressing Rlc1‐mCherry and Cut11‐GFP, as a nuclear marker, were grown in EMM at 25 °C. ‘Cell 1’ is in interphase. ‘Cell 2’ is at the onset of nuclear division in early anaphase. ‘Cell 3’ is probably just after anaphase or telophase onset. In WT cells, Rlc1 accumulates as cortical dots in the middle region in early anaphase, and Rlc1 is merged into the CR before the onset of cytokinesis. However, Rlc1 distribution is considerably affected in Δmid1 cells. Brackets indicate Rlc1 accumulation in the middle region. Asterisk indicates a cell that has almost completed septum formation. (B) Quantification of Rlc1 localization patterns in mitotic cells. Cells incubated as in A were classified. At least 100 cells were counted at each stage. Average percentage from two independent experiments was shown. (C) F‐actin distribution was compared in WT (NIG2028) and Δmid1 (JK105) cells. Cells logarithmically growing at 25 °C in YE were fixed and processed for visualizing DNA and F‐actin. ‘Cell 1’ is before mitosis. ‘Cell 2’ is in late anaphase judging from condensed nuclear DNA. Single asterisks indicate Δmid1 cells with no F‐actin distribution in the cell center even at this stage. ‘Cell 3’ is probably just after anaphase or in early telophase. Although ‘cell 3’ of both strains form an F‐actin ring in the cell center, the ring appeared fragile in Δmid1 cells compared to WT (arrowheads). As telophase progressed, the F‐actin ring tightly formed even in Δmid1 cells (arrows) as well as in WT cells. Double asterisk indicates Δmid1 cells having a fully formed septum in position. Actin dots were distributed on both sides of the septum in this cell. (D) Quantification of the number of cells showing F‐actin distribution in the middle region in anaphase and telophase. Cells processed as in C were examined. At least 100 cells were counted at each stage. Average percentage with standard deviation from three independent experiments was shown. *T‐test was carried out.

We further investigated the distribution of F‐actin in Δmid1‐null cells. Previously, it has been shown that F‐actin appears at the future division site after the onset of nuclear division in S. japonicus (Alfa & Hyams 1990). As shown in Fig. 4C, the CR tightly forms just before the completion of the segregation of daughter nuclear DNA in this organism. Remarkably, in a significant population of Δmid1‐null cells, F‐actin accumulation was not detected in the middle of cells whose nuclear DNA had almost completely divided (‘cell 2′ indicated by a single asterisk in Fig. 4C,D). Moreover, F‐actin was likely to have been loosely assembled into the CR in Δmid1‐null cells compared to wild‐type cells (arrowheads in Fig. 4C). However, eventually, the F‐actin seemed to be tightly arranged in the CR as telophase progressed (arrows in Fig. 4C), and CR constriction occurred (double asterisk, Fig. 4C). We therefore conclude that the timing of myosin II and F‐actin recruitment to form the CR in the future division site was significantly delayed in Δmid1‐null cells. Moreover, our observation of Rlc1‐mCherry localization showed that the distribution of myosin II in the early stages of CR formation appears to be quantitatively and qualitatively disturbed in the absence of Sj Mid1. Furthermore, we found that Sj Mid1 appears to facilitate the formation of a tightly packed F‐actin ring before the onset of telophase. It is possible that Sj Mid1 may organize F‐actin in the middle of the cell by controlling the distribution of myosin II. Alternatively, Sj Mid1 may induce actin polymerization by interacting with actin‐regulatory proteins other than myosin II, as discussed in detail below.

S. japonicus Mid1 fails to rescue the defects of an S. pombe Mid1 mutant

Although a gene disruption of Sp mid1⁺ dramatically affects CR formation in S. pombe (Sohrmann et al. 1996), fatal defects of CR formation were not occurred in Sj mid1‐null cells of S. japonicus. The difference of Mid1 dependency on CR formation in the fission yeast species might result from poor conservation of Mid1 functional domains during evolution. We thus examined whether artificial expression of Sj Mid1 from the plasmid pREP1 (Maundrell 1993) could rescue the lethality of S. pombe Δmid1‐null cells at 36 °C. We found that Δmid1 cells transformed with either pREP1 (empty vector) and pREP1‐HA Sj Mid1 did not grow at 36 °C, whereas Δmid1 cells transformed with the control plasmid pREP1‐HA‐Sp Mid1 grew at 36 °C (Fig. 5A). Thus, artificial expression of Sj mid1⁺ failed to complement the temperature‐sensitive growth defect of S. pombe cells lacking endogenous mid1⁺. Meanwhile, we found that Sp mid1⁺ could not significantly complement a cellular function of Sj mid1⁺ (Fig. S2 in Supporting Information). Therefore, a gene function of mid1⁺ was not highly conserved in these fission yeast species.

Figure 5

Examination of the functional conservation of Mid1 in fission yeast species. (A) Expression of Sj Mid1 was not able to complement the temperature‐sensitive growth defect of an S. pombe Δmid1 strain. S. pombe Δmid1 cells transformed with pREP1 expressing HA‐tagged proteins were streaked on SD plates, respectively. Cells were incubated at the indicated temperatures for 3 days. (B) Localization of YFP‐fused Sj Mid1 and the truncated proteins in S. pombe cells. Wild‐type S. pombe cells transformed with pREP1 containing YFP‐Sj mid1⁺ gene and deletion mutants were incubated in EMM without thiamine for 20 h at 25 °C to induce expression of the inserted gene. Nuclear localization and vacuolar localization of YFP‐fused proteins were indicated by arrowheads and arrows, respectively.

We next investigated the cellular localization of Sj Mid1 in S. pombe. The protein motifs and domains required for the cellular localization of Sp Mid1 have been studied in detail (Paoletti & Chang 2000; Celton‐Morizur et al. 2004; Almonacid et al. 2009, 2011; Lee & Wu 2012; Saha & Pollard 2012b; Ye et al. 2012; Guzman‐Vendrell et al. 2013; Rincon et al. 2014). Opposing activities of the NLS and NES motifs control nucleocytoplasmic shuttling of Sp Mid1 (Paoletti & Chang 2000), and cytoplasmic Mid1 can relocate to the plasma membrane and interact with Sp Cdr2 via its central region to form nodes on the cell cortex around the nucleus positioned in the cell center (Almonacid et al. 2009; Lee & Wu 2012; Guzman‐Vendrell et al. 2013; Rincon et al. 2014). The amphipathic region in AHD and the following C‐terminal PH domain are important for the membrane association of Sp Mid1 (Celton‐Morizur et al. 2004; Almonacid et al. 2009; Lee & Wu 2012; Saha & Pollard 2012b; Guzman‐Vendrell et al. 2013). Immediately after the G2/M transition, Sp Mid1 at the nodes recruits myosin II and CR‐assembling proteins such as Sp Rng2 and Sp Cdc12, and the CR is formed by the interaction of F‐actin with myosin II before the onset of chromosome segregation (Wu et al. 2003, 2006; Motegi et al. 2004). It has been showed that the N‐terminal half of Sp Mid1 can associate with the CR (Celton‐Mozier et al. 2004; Almonacid et al. 2009; Saha & Pollard 2012b; Lee & Wu 2012; Guzman‐Vendrell et al. 2013).

To explore which domain is less functionally conserved between Sj Mid1 and Sp Mid1, we transformed wild‐type S. pombe cells with expression plasmids for YFP‐fused Sj Mid1 or truncated mutant proteins. We found that YFP‐Sj Mid1 did not localize to nodes or the CR in S. pombe cells (Fig. 5B). Instead, localization to membranous structures was seen. The membrane association of Sj Mid1 was probably exerted via its C‐terminal region, as the N‐terminal deletion mutant Sj Mid1ΔN2 (592–954 a. a.) retained its membrane association, whereas the C‐terminal half‐deletion mutant Sj Mid1ΔC1 (1–594 a. a.) failed to localize to membranes. A significant portion of Sj Mid1ΔC1 accumulated in a vacuole‐like compartment. This truncated protein may be structurally disordered and brought in a vacuole. Meanwhile, Sj Mid1ΔC2 (1–415 a. a.) did not associate with the CR, whereas the corresponding domain of Sp Mid1 (1–420 a. a.) did (Lee & Wu 2012). Therefore, it was suggested that Sj Mid1 was unlikely to interact stably with CR components of S. pombe. However, Sj Mid1ΔC2 showed a nuclear localization which was especially evident in interphase cells, suggesting that this region of Mid1 may be responsible for the nuclear localization of Sj Mid1. Coincidently, N‐terminal fragment of Sp Mid1 (150–308 a. a.) also localizes in the nucleus (Saha & Pollard 2012b), although the canonical NLS of Sp Mid1 was found in its C‐terminal region (Paoletti & Chang 2000; Celton‐Morizur et al. 2004; Almonacid et al. 2009). Therefore, an uncharacterized NLS may be located in the N‐terminal half of Mid1 proteins (Fig. S3 in Supporting Information). Alternatively, Mid1 may also have NLS‐independent nuclear localization activity.

Discussion

In this study, we showed a functional role of Mid1 in S. japonicus cytokinesis. Sj Mid1 appeared as cortical dots with Sj Rlc1 in the middle region of interphase cells, and these proteins formed the medial ring in late anaphase (Fig. 1). As cytokinesis progressed, Sj Rlc1 localized to the contracting CR whereas Sj Mid1 remained at the cell periphery for a while and then disappeared from that location. This behavior of Sj Mid1 is reminiscent of Mid1 in S. pombe (Sohrmann et al. 1996; Paoletti & Chang 2000). Sp Mid1 plays a central role for determining the division plane before cytokinesis (Sohrmann et al. 1996; Paoletti & Chang 2000; Wu et al. 2003; Daga & Chang 2005). Nevertheless, we found that mitotic S. japonicus cells were able to divide into two cells with nearly equal sizes without mid1⁺ gene function (Fig. 3B), suggesting that the importance of Mid1 to spatiotemporally control cytokinesis is not conserved in these fission yeast species. Coincidently, artificial expression of Sj Mid1 did not improve the temperature‐sensitive cell growth of an S. pombe Δmid1 strain (Fig. 5A). The roles of Mid1 proteins in cytokinesis progression may have functionally diversified after the two species had evolutionally separated. Although our manuscript was in preparation for submission, Gu and colleagues had independently reported the functional diversity of Mid1 proteins between S. pombe and S. japonicus (Gu et al. 2015). Their experimental results and ours agree with a few exceptions. In our study, it was originally showed that Sj Mid1 enhanced the assembly of F‐actin into the CR (Fig. 4D) and possessed a nuclear‐localizing activity similarly to Sp Mid1 (Figs 1, 2 and 5B). It is possible that the functional significance of Mid1 proteins in cytokinesis may have diverged in fission yeast species by using other proteins to control the spatiotemporal formation of the CR.

Sj Mid1 is required for the prompt CR formation

It has been reported that CR formation is delayed in S. pombe when mid1⁺ gene function is removed (Motegi et al. 2004; Hachet & Simanis 2008; Huang et al. 2008). In those cells, CR components are assembled into a ring without intermediation of cortical nodes dependent on the SIN (septation initiation network)‐signaling pathway (Hachet & Simanis 2008). However, CR formation occurs in anaphase in S. japonicus (Alfa & Hyams 1990). We found that a gene deletion of Sj mid1⁺ caused a delay in CR formation in S. japonicus. In those cells, the cortical localization of Rlc1 in the cell center is affected (Fig. 4A). It is possible that the timing of CR formation may be retarded due to a reduction in myosin II levels in the medial region. It has been showed that Mid1 plays an important role for anchoring myosin II in the medial cortex together with Rng2 before CR formation in S. pombe (Motegi et al. 2004; Almonacid et al. 2011; Padmanabhan et al. 2011; Takaine et al. 2014). Similarly, Sj Mid1 may also contribute to CR formation by interacting with myosin II in S. japonicus. In addition, it is possible that the node‐mediated Mid1 function may be exerted at nearly the same time as the activation of CR formation by SIN‐dependent signaling in S. japonicus.

The process of CR formation in S. pombe has been well studied. Several minutes after the G2/M transition, Mid1 in nodes localizes myosin II to the cell center via interacting with Rng2 and induces actin polymerization from the nodes by recruiting a formin Cdc12 (Chang et al. 1997; Padmanabhan et al. 2011; Takaine et al. 2009, 2014; Laporte et al. 2011; Saha & Pollard 2012a). Myosin II and F‐actin then interact with each other and CR formation rapidly progresses in metaphase (Motegi et al., 2000; Wu et al. 2003). However, Sj Mid1 organized cortical dots containing Rlc1 before mitosis, but they stayed in the equatorial region without F‐actin appearance until late anaphase. This is an interesting point that may possibly explain the difference in timing of CR formation between the two species. We consider two possibilities regarding the late CR formation in S. japonicus. First, Sj Mid1 is possibly maintained in an inactive state and cannot promote actin polymerization until late anaphase. The other is that node‐associated proteins, which induce CR formation such as Cdc12, may not be fully activated until late anaphase. Interestingly, it has been showed that SIN‐dependent phosphorylation releases Cdc12 from a semi‐functional oligomer state to a fully functional monomer state specifically in anaphase in S. pombe (Bohnert et al. 2013). Clarifying the function and regulation of the S. japonicus Cdc12 homologue may be important in determining the reason for the difference in CR formation timing in these fission yeast species.

Multiple molecular mechanisms exist to position the cell division site in fission yeast

In S. pombe, Mid1 is released from the nucleus, which is positioned at the cell center to the proximal cell cortex during G2 phase and forms nodes by interacting with Cdr2 and other functionally related proteins (Paoletti & Chang 2000; Wu et al. 2003, 2006; Almonacid et al. 2009; Guzman‐Vendrell et al. 2013; Akamatsu et al. 2014; Rincon et al. 2014). Mid1‐anchored nodes play critical roles for CR positioning in the cell center. Several mechanisms function cooperatively to ensure Mid1 localization to the middle cortex in S. pombe. The tip complex consisting of Tea1 with its associated proteins and Pom1 occludes Mid1 nodes from the cell tips (Celton‐Morizur et al. 2006; Padte et al. 2006; Huang et al. 2007). Also, the cortical ER network retains Mid1 in the center of cells by restricting its diffusion (Zhang et al. 2010).

However, Mid1‐independent mechanisms that position the CR in the cell middle or occlude CR formation at the cell tip may exist in S. pombe, although the molecular basis underlying these activities has not been fully uncovered (Huang et al. 2007; Padte et al. 2006; Rincon et al. 2014). It is possible that S. japonicus may mainly depend on Mid1‐independent mechanisms for CR positioning. It has very recently been reported that the Pom1 homologue in S. japonicus plays an important role for determining the division site via controlling the CR anchoring protein Cdc15 (Gu et al. 2015). In their paper, they showed that a gene deletion of mid1⁺ does not exacerbate the defect in CR positioning of Δpom1‐null cells, suggesting that Mid1 possibly plays a minor role in positioning the cell division site as compared to Pom1 in S. japonicus. However, Pom1 is not essential for cell growth in S. japonicus (Gu et al. 2015). Therefore, it is possible that another pathway may also function to control the cell division site in S. japonicus.

The mid1⁺ gene might have been derived from mid2⁺ by gene duplication after the ancestor of Schizosaccharomyces branched from other taxa (see Introduction). Considering the phylogenic relationship of fungi and yeasts, the Mid1‐dependent system for positioning the division site may not be the most prominent mechanism. Molecular evolution studies showed that S. japonicus is probably the earliest branched species of the fission yeasts (Rhind et al. 2011). It is therefore possible that S. japonicus may have some ancestral traits. As mentioned above, a recent study (Gu et al. 2015) is in agreement with our studies, which showed that deletion of mid1⁺ did not cause a fatal defect in S. japonicus cytokinesis. It is possible that a dependency on Mid1 for cytokinesis may not have been required in this organism compared to S. pombe. Instead, ancestral spatiotemporal controlling machinery may have a predominant function during cytokinesis in S. japonicus. Future studies comparing the molecular mechanism of cytokinesis between S. pombe and S. japonicus will shed light on how important mechanisms including nucleocytoplasmic shuttling of Mid1 for cell division site position‐ing and the node‐dependent spatiotemporal control system of CR formation have been established in the Schizosaccharomyces lineage.

Experimental procedures

Strains and handling of cells

Schizosaccharomyces pombe and S. japonicus strains used in this study are listed in Tables S1 and S2 in Supporting Information, respectively. Fundamental strains for launch of our work in cytokinesis of S. japonicus were kindly provided by Prof. H. Niki (National Institute of Genetics, Japan). We used the NIG2028 strain as wild type in this study according to pioneer work (Furuya & Niki 2009).

YE [0.5% Bacto yeast extract (Becton & Dickinson; BD), 3% d‐glucose (Wako), 2% Bacto agar (BD) for a plate] was used as the standard medium. Geneticin (G418; Sigma Aldrich or Alexis) was supplemented in YE plates at a final concentration of 100 μg/mL to select for KanMX6‐positive transformants. For microscopic observation, cells were cultured in YE liquid medium containing supplements including adenine, uracil, leucine, histidine and lysine at final concentrations of 225 mg/L for each. For the same purpose, Edinburgh minimal medium (EMM) (Forsburg & Rhind 2006) with supplements as mentioned above and SD (BD) was also used.

To establish an S. japonicus strain by making a genetic cross of parental strains, fresh colonies of parental strains grown on YE plates were mixed with a toothpick in a drop of sterilize water on an MEA plate [3% Bacto malt extract (BD), 2% Bacto agar (BD)]. After incubation at 25 °C for several days, spores derived from a mating parental strains were treated with β‐glucuronidase (Sigma) for separating spores from each other by digesting asci. Then, spores were spread on a plate after a brief treatment with 30% ethanol to kill the parental cells. The descendant colonies were picked up and their genetic markers were tested.

Gene manipulation

All strains prepared in this study were constructed by homologous recombination of a PCR‐based altered gene with a target chromosomal gene using a set of pFA6a‐kanMX6 series as a template (Bähler et al. 1998b).

To transform S. japonicus cells with the altered genes, an electroporation method was applied according to Aoki et al. (2010) as follows. NIG2028 strain inoculated in 50 mL of YE liquid medium was incubated overnight at 30 °C in a shaking bath to a 3 × 10⁶ cells/mL cell density. Cells were isolated by centrifugation, and the resultant pellets were washed with 40 mL of ice‐chilled Milli‐Q water three times. Cell pellets were suspended in 10 mL of 1 M sorbitol aqueous solution containing 50 mm dithiothreitol (DTT), and the suspension was shaken at 30 °C for 15 min. After removing the supernatant by centrifugation, cells were suspended in 10 mL of 1 m sorbitol aqueous solution. Cells were centrifuged again and resuspended in 5 mL of 1 m sorbitol aqueous solution. This manipulation was repeated one more time, and cells were finally split into two 1.5‐mL microtubes. After gently pipetting cells with 5 μL of 1 m sorbitol and 1 μg transforming DNA, the microtubes were left on ice for 30 min. Electroporation was carried out in a dedicated cuvette (gap distance, 2‐mm) containing 40 μL of the cell suspension with transforming DNA with the Genepulser II or Genepulser Xcell, the electroporation instruments of Bio‐Rad. Cells were perforated with a single electric pulse (200 Ω, 25 μF, 2.3 kV, exponential) which allowed the DNA to be introduced into the cells. At once, a small amount of 1 M sorbitol aqueous solution was added to the cuvette and the suspension of cells was transferred into 5 mL YE medium supplemented with 40 μg/mL adenine and 20 μg/mL uracil (final concentration). After an overnight incubation at 30 °C, the transformed cells were spread onto YE plates containing G418 (100 μg/mL), adenine (40 μg/mL) and uracil (20 μg/mL) (YEAUG), and the plates were further incubated at the same temperature. After several days, colonies growing on the plate were selected and streaked onto a new YEAUG plate.

Transformation of S. pombe was carried out mainly by the lithium acetate method (Okazaki et al. 1990). An electroporation method (Suga & Hatakeyama 2001) was also used in some cases.

Localization domain analysis

To observe the cellular localization of Sj Mid1 in S. pombe, yfp gene‐fused Sj mid1⁺ cDNA was inserted between NdeI and SalI in the multiple cloning site of the pREP1 expression vector (Maundrell 1993). Subsequently, vectors for expression of various lengths of truncated Sj Mid1 were constructed using a PCR‐based mutagenesis kit (Takara) using the following sets of oligonucleotides:

• 5′‐TCTCGAGCGGCCGCCCTTGTACAGCTCGTCCAT‐3′ and 5′‐GGCGGCCGCTCGAGAGACGAAGATGATAGC‐3′ for ΔN1; 5′‐TCTCGAGCGGCCGCCCTTGTACAGCTCGTCCAT‐3′ and 5′‐GGCGGCCGCTCGAGAGGCAAGCTTTATCTT‐3′ for ΔN2; 5′‐TGACTGACTGACGATGTAAAAGGAATGTC‐3′ and 5′‐ATCGTCAGTCAGTCAGCAAGGTGAAGATAA‐3′ for ΔC1; and 5′‐TGACTGACTGACGATGTAAAAGGAATGTC‐3′ and 5′‐ATCGTCAGTCAGTCATGTTACCGGCAGACA‐3′ for ΔC2. Wild‐type cells transformed with the generated vectors were cultured in EMM at 25 °C and observed under fluorescence microscopy.

Microscopy

Cells were photographed with the upright microscope BX51 (Olympus) equipped with the cooled charge‐coupled device (CCD) camera ORCAII‐ER‐1394 (Hamamatsu photonics). Digital images were acquired with Simple PCI software (Compix Inc.).

For time‐lapse observation of fluorescent protein‐expressing cells, the inverted microscope BX71 (Olympus) equipped with the spinning disc confocal scanner unit CSU22 (Yokogawa) with 488 and 568 nm laser units, a piezo‐actuator for Z‐scanning (Physik), and the EMCCD camera iXon3 885 (Andor) was used. The system was controlled by MetaMorph software (Molecular Devices). Cell cultures were mounted on glass slides, and the prepared slides were sealed with vacuum grease to avoid evaporation of water. Z‐sectioned images were collected at 0.3‐μm intervals. The acquired images were processed by ImageJ 1.43 software (Wayne Rasband, National Institutes of Health, USA).

Cell staining

Cell cultures were fixed with one‐tenth volume of 30% formaldehyde‐containing PEM buffer [0.1 M Pipes (pH 6.8), 1 mM EGTA, 1 mm MgCl₂]. After 1 h, cells were washed with PEM buffer three times and were suspended in PEM buffer with 0.6 mg/mL of Zymolyase 100T (Seikagaku Co. Ltd) for partial cell wall lysing. After 5 min, cells were gently resuspended in PEM buffer containing 1% Triton X‐100 and were washed with PEM buffer three times. To stain for F‐actin, the perforated cells were treated with 0.15 U BODIPY‐phallacidin (Invitrogen) for 1 h. After replacing the staining solution with PEM containing 4′, 6‐diamidino‐2‐phenylindole (DAPI), cells were observed with fluorescence microscopy.

Acknowledgements

We are deeply grateful to Prof. H. Niki & Dr K. Aoki (National Institute of Genetics, Japan) for providing us S. japonicus strains and genetic materials and for valuable advice. This work was supported by a Grant‐in‐Aid for Scientific Research to K. N. from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) (No. 24570207).