Inhibition of cell membrane ingression at the division site by cell walls in fission yeast.

Eukaryotic cells assemble actomyosin rings during cytokinesis to function as force-generating machines to drive membrane invagination and to counteract the intracellular pressure and the cell surface tension. How the extracellular matrix affects actomyosin ring contraction has not been fully explored. While studying the Schizosaccharomyces pombe 1,3-β-glucan-synthase mutant cps1-191, which is defective in division septum synthesis and arrests with a stable actomyosin ring, we found that weakening of the extracellular glycan matrix caused the generated spheroplasts to divide under the nonpermissive condition. This nonmedial slow division was dependent on a functional actomyosin ring and vesicular trafficking, but independent of normal septum synthesis. Interestingly, the high intracellular turgor pressure appears to play a minimal role in inhibiting ring contraction in the absence of cell wall remodeling in cps1-191 mutants, as decreasing the turgor pressure alone did not enable spheroplast division. We propose that during cytokinesis, the extracellular glycan matrix restricts actomyosin ring contraction and membrane ingression, and remodeling of the extracellular components through division septum synthesis relieves the inhibition and facilitates actomyosin ring contraction.

INTRODUCTION

Animal cells and fungal cells require assembly and contraction of an actomyosin ring during cytokinesis. In fission yeast, the actomyosin ring contracts to drive membrane ingression and coordinates with the septum assembly machinery to deposit cell wall materials at the division site (Ramos ). The fungal cell wall has been suggested to be a functional equivalent of the extracellular matrix (ECM) in animal cells (Munoz ). The division septum is a special wall structure composed of primary and secondary septa. The primary septum is a structure that must be degraded to permit cell separation and the secondary septum is a structure that forms the cell wall once the cells are separated. The septum assembly machinery consists of α-glucan synthase Ags1 and  β-glucan synthase Bgs1 and Bgs4. The β-glucan synthase Cps1/Bgs1 is essential for primary septum formation. Cps1 synthesizes specifically the linear β-glucan matrix of the primary septum at the division site and couples the extracellular glycan matrix to the actomyosin ring via intermediate protein complexes (Munoz ; Cortes ; Davidson ; Sethi ). The β-glucan synthases Bgs4 and the α-glucan synthase Ags1 are primarily involved in the secondary septum formation and participate in the synthesis of primary septum (Garcia Cortes ). The deposition of extracellular glycan matrix coordinates with actomyosin ring contraction and stabilizes the contracting actomyosin ring at the division site (Munoz ; Arasada and Pollard, 2014).

How the extracellular glycan matrix influences actomyosin ring contraction (apart from its roles in ring stability during cytokinesis) has not been examined closely (Mishra ; Munoz ). In this study, we used the thermosensitive allele of β-1,3-glucan synthase, cps1-191, to address this question. The cps1-191 mutant is defective in β-1,3-glucan and septum synthesis and arrests with a noncontracting actomyosin ring at the nonpermissive temperature (Liu ). Interestingly, we found that weakening of the extracellular glycan matrix in cps1-191 mutant at the nonpermissive temperature enabled actomyosin ring contraction and membrane ingression.

RESULTS AND DISCUSSION

At the restrictive temperature, the β-glucan synthase mutant cps1-191 assembles actomyosin rings that do not contract (Liu ). It has been suggested that β-glucan synthesis at the division site is required to overcome the high intracellular turgor pressure during cytokinesis, and that the actomyosin ring may not be able to overcome the high turgor (Proctor ). To test if the turgor pressure inhibited ring contraction in cps1-191 mutants, we cultured cps1-191 cells in EMMA medium containing 0.8 M sorbitol to decrease the turgor pressure to that inside the cells at the restrictive temperature, and we added 2-deoxyglucose (2-DG) to this culture to prevent further glucan synthesis at the division site and elsewhere in the cells (Megnet, 1965; Svoboda and Smith, 1972; Osumi ). A recent study showed that rings in cps1-191 mutant cells constricted slowly after shifting to the restrictive temperature for ∼2 h before microscopy at the restrictive temperature (Dundon and Pollard, 2020). To ensure a highly penetrant phenotype for cps1-191, we shifted the cps1-191 cells to the restrictive temperature for ∼6 h before microscopy, which was performed at the restrictive temperature. As previously reported, actomyosin rings of cps1-191 cells maintained at normal turgor pressure did not undergo contraction (Figure 1A). We occasionally observed that parts of the cps1-191 cells swelled into bumps and the cells lysed eventually with collapsing rings (Figure 1B). Culturing cps1-191 cells in EMMA medium containing 0.8 M sorbitol did not increase actomyosin ring contraction events, and phenotypically these cells resembled cps1-191 grown under normal growth conditions in EMMA medium, in which a high intracellular turgor pressure is maintained (Figure 1, C and D). Thus, our results showed that a decreased turgor pressure does not allow ring contraction in cps1-191 mutant cells.

FIGURE 1:

Lowering turgor pressure prevents cell membrane ingression in cps1 mutant cells. (A) cps1-191 GFP-psy1 rlc1-tdTomato cells were cultured in YEA at the restrictive temperature of 36°C for 6.5 h and were processed similarly using the spheroplasting protocol but omitting lysing enzymes and Zymolyase. Cells in the EMMA medium with 2-DG were imaged at 36°C. Green: GFP-psy1. Red: rlc1-tdTomato. (B) Some cps1-191 GFP-psy1 rlc1-tdTomato cells lysed after more than 6.5 h of incubation at the restrictive temperature. Treatment of cells was the same as in A. Green: rlc1-tdTomato. Red: GFP-psy1. (C) cps1-191 GFP-psy1 rlc1-tdTomato cells were cultured in YEA at the restrictive temperature for 6.5 h and were processed similarly using the spheroplasting protocol but omitting lysing enzymes and Zymolyase. Cells were imaged at 36°C in the EMMA medium containing 2-DG and 0.8 M sorbitol to lower down the turgor pressure. Green: GFP-psy1. Red: rlc1-tdTomato. (D) cps1-191 GFP-psy1 rlc1-tdTomato cells treated as in C were stained with calcofluor dye to reveal the cell wall. Scale bar: 5 μm

Next, we considered the possibility that the extracellular glycan matrix inhibited ring contraction and membrane ingression in cps1-191 mutants in the absence of cell wall remodeling. The Cps1 is a transmembrane protein that (along with other integral membrane proteins, such as Ags1 and Bgs4) links actomyosin rings underneath the cell membrane to the extracellular glycan matrix (Cortes , 2012; Munoz ; Arasada and Pollard, 2015; Davidson ; Sethi ; Martin-Garcia ). It was possible that in the absence of division septum synthesis (and thus cell wall remodeling), the actomyosin rings were stably fixed to the inactive cps1-191 gene product or other integral membrane proteins (such as mok1, sbg1, and bgs4) that link the cell wall to the actomyosin ring. To test if this was the case, we weakened the extracellular glycan matrix by treating the cps1-191 cells with cell wall–lytic enzymes and further blocking new cell wall and septum synthesis by supplementing the culture with 2-DG. Interestingly, upon weakening of the cell wall, myosin rings in cps1-191 mutants expressing the regulatory light chain of myosin tagged with the fluorescent protein tdTomato (Rlc1-tdTomato) underwent contraction coupled with membrane ingression at the restrictive temperature of 36°C (Figure 2A; GFP-tagged Syntaxin-like protein Psy1 was used as a cell membrane marker; 19/29 spheroplasts). Consistently, contracting actin rings labeled with the Lifeact-mCherry were also detected in the cps1-191 mutant upon weakening of the cell wall, suggesting that the actomyosin rings were driving the contraction and membrane ingression (Figure 2B; 5/41 spheroplasts). These mutant spheroplasts with weakened cell walls often divided nonmedially into two, and the rings contracted at much reduced rates (0.061 ± 0.021 μm/min; nspheroplast = 8) compared with wild-type cells (0.299 ± 0.059 μm/min; ncell = 14; Figure 2C). The low rate of ring contraction is comparable to that in the wild-type spheroplasts, in which the rings slide along the cell membrane during ring contraction (0.046 ± 0.031 μm/min; nwild-type spheroplast = 40). We frequently observed that the rings contracted till midphase of division and disassembled before completion of cytokinesis. The spheroplasts, however, went on to divide into two entities (Figure 2, A and B). The segregation of daughter nuclei in the cps1-191 spheroplasts was often not coordinated with the cytokinesis, with some spheroplasts having two nuclei in one of the daughter entities or having cleaved nuclei, presumably due to the nonmedial division (Supplemental Figure 1). The miscoordination of cytofission and nuclear division spatially could arise from the variable dumbbell-shaped morphology of cps1-191 spheroplasts (Mishra ). The functions of Mto1 and Mto2, which are involved in the assembly of postanaphase microtubule arrays, may also contribute to these miscoordination defects (Dundon and Pollard, 2020). Because the cps1-191 mutant spheroplast division was morphologically different from normal fission yeast cell division and was mimicking the morphological changes of some animal cells during division, we have called this type of division cytofission.

FIGURE 2:

Weakening of cell walls allows ring contraction and cell membrane ingression. (A) Two examples of cps1-191 spheroplasts underwent cytofission at 36°C. The cps1-191 GFP-psy1 rlc1-tdTomato cells were cultured at 36°C for 6.5 h, processed into spheroplasts, and recovered for 1 h at 36°C before imaging. (B) Two examples of cps1-191 spheroplasts expressing Lifeact-mCherry underwent cytofission at 36°C. The cps1-191 GFP-psy1 lifeact-mCherry cells were cultured at 36°C for 6.5 h, processed into spheroplasts, and recovered for 1 h at 36°C before imaging. (C) Quantification of the rate of ring contraction in wild-type cells and cps1-191 spheroplasts undergoing cytofission. (D) Wild-type cells and cps1-191 GFP-psy1 rlc1-tdTomato spheroplasts were stained with the dye calcofluor. The image was pseudo-colored in green to represent calcofluor staining. (E) Electron micrographs of cps1-191 GFP-psy1 rlc1-tdTomato spheroplasts regenerated in medium with or without 2-DG. Scale bar: 5 μm except E, which is 1 μm; error bars: SD.

Analysis of the extracellular glycan matrix using calcofluor staining (a division septum–specific fluorochrome; G. Cortés ) in cells undergoing cytofission in EMMA containing sorbitol and 2-DG medium revealed that the division site of cps1-191 spheroplasts undergoing cytofission with 2-DG medium contained significantly reduced β-glucan materials (Figure 2D). Further study with high-resolution scanning electron microscopy showed that the glucan fibrils regenerated in cps1-191 spheroplasts without 2-DG (Figure 2E, Bottom) while the fibrils were not noticeable in cps1-191 spheroplasts with 2-DG (Figure 2E, Top). The glucan fibrils commonly present at the division site of fission yeast were largely absent in cps1-191 spheroplasts undergoing cytofission (Supplemental Figure 2). Taking these results together, we showed that weakening of cell walls in cps1-191 cells at nonpermissive temperature and ensuing further inhibition of new cell wall and septum synthesis with 2-DG facilitate a novel cytofission event that leads to division of one spheroplast into two in the absence of detectable division-septum growth. Our results also suggested that the extracellular glycan matrix anchored to the actomyosin rings negatively regulates ring contraction and membrane ingression. This is consistent with a previous finding that the absence of Bgs4-synthesized β-glucan in the septum promoted faster ring contraction and membrane ingression than that of normal septa, and at the same time, the synthesis and ingression of septum wall progressed more slowly than that of a normal septum (Munoz ).

A reduction of β-glucan may result in an increased amount of α-glucan in the cell walls of fission yeast. To test if the cytofission of cps1-191 spheroplasts was due to the synthesis of α-glucan at the division site, we prepared, cps1-191 mok1-664 double-mutant spheroplasts containing the thermosensitive alleles of both α- and β-glucan synthases and imaged the myosin rings and cell membranes in these double mutant spheroplasts at the nonpermissive temperature. Similar to the cps1-191 spheroplast, the cps1-191 mok1-664 double-mutant spheroplasts underwent cytofission (Figure 3A, 11/26), suggesting that α-glucan and β-glucan synthesis did not contribute significantly to the cytofission events.

FIGURE 3:

The cps1-191 mutant spheroplasts undergo cytofission independent of the α-glucan synthase and endoglucanases. (A) Cytofission in cps1-191 mok1-664 GFP-psy1 rlc1-tdTomato. (B) Cytofission in cps1-191 agn1Δ GFP-psy1 rlc1-tdTomato. (C) Cytofission in cps1-191 eng1Δ GFP-psy1 rlc1-tdTomato. Scale bar: 5 μm.

Normal fission yeast cells that just complete ring contraction and membrane ingression are not entirely separated until the primary septum digestion of the division septum connecting the two newly divided cells (Sipiczki, 2007). This process is achieved in fission yeast through the action of endoglucanases (Martin-Cuadrado ; Dekker ; Garcia ). We tested if proteins involved in the separation of fission yeast cells were also involved in the cytofission, which would be expected if trace amounts of division septum had been deposited during ring contraction. To this end, we constructed double-mutant spheroplasts of cps1-191 lacking the endoglucanases eng1 (β-glucanase) and agn1 (α-glucanase), respectively. Similarly to the single mutant cps1-191, the double mutants lacking either of the two endoglucanases underwent cytofission upon weakening of the cell wall (Figure 3B, 8/23; Figure 3C, 11/38). The results indicated that the cytofission events of cps1-191 mutants do not require the breakdown of cell wall materials by endoglucanases, even though cytofission leads to the complete separation of spheroplasts.

In ∼53% of the cps1-191 spheroplasts that underwent cytofission (53 out of 99 spheroplasts), the rings contracted till midphase of division and disassembled before division into two entities. We tested if the ESCRT abscission complex was involved in the cytofission by removing two of the ESCRT proteins Vps4 and Vps20 in the cps1-191 spheroplasts. The cps1-191 vps4Δ and cps1-191 vps20Δ double-mutant spheroplasts underwent cytofission as in the single cps1-191 mutant spheroplast (Supplemental Figure 3A, 14/14; Supplemental Figure 3B, 8/8). It is possible that the completion of cytofission without the actomyosin rings was achieved via an unknown cell abscission mechanism.

Previous studies suggested that in certain circumstances, some eukaryotic cells are able to divide without an actomyosin ring (Proctor ; Choudhary ; Flor-Parra ; Dix ; Ramos ). To see if the cytofission was driven by contraction of the actomyosin ring, we first perturbed the functions of rings using Latrunculin-A (LatA) to inhibit actin polymerization (Morton ; Fujiwara ). cps1-191 spheroplasts treated with DMSO underwent ring contraction and membrane ingression and completed cytofission (Figure 4A; 11/16 spheroplasts). In contrast, cps1-191 spheroplasts treated with LatA underwent ring disassembly and failed to divide into two entities or ingressed very slowly, resembling dividing cells after long incubation with LatA (Ramos ). Interestingly, the smaller entity retracted into the bigger entity, probably due to the imbalance of intracellular pressures (Figure 4B; 27/33 spheroplasts).

FIGURE 4:

The function of actomyosin rings is required for the cps1-191 mutant spheroplasts to undergo cytofission. (A) cps1-191 GFP-psy1 rlc1-tdTomato spheroplasts underwent cytofission in the presence of DMSO. Left panel shows the DIC images; Right panel shows the fluorescence micrographs. (B) cps1-191 GFP-psy1 rlc1-tdTomato spheroplasts were incubated with 150 μm Lat-A. Left panel shows the DIC images; Right panel shows the fluorescence micrographs. (C) rlc1Δ cyk3-GFP spheroplasts failed to undergo cytofission at 36°C. The rlc1Δ cyk3-GFP cells were cultured at 36°C for 6.5 h, processed into spheroplasts, and then recovered in minimal medium containing sorbitol before imaging at 36°C. Top panel shows the DIC images; Bottom panel shows the fluorescence micrographs. (D) The cps1-191 GFP-psy1 rlc1-tdTomato spheroplasts failed to undergo cytofission in the presence of 75 μM brefeldin A. Scale bar: 5 μm.

Next, we perturbed the myosin component of actomyosin rings by deleting rlc1, the regulatory light chain of myosin II (Le Goff ; Naqvi ; Pollard ), in cps1-191 mutants. It has been shown that the cells lacking rlc1 (rlc1Δ) are cold-sensitive and fail to undergo cytokinesis at low temperatures, but at high temperatures, the rlc1Δ cells assemble an intact actomyosin ring that contracts normally (Naqvi ; Supplemental Figure 4; 31/31 cells). We used this differential temperature requirement to test the essentiality of actomyosin ring functions in cytofission. If the actomyosin ring was essential in driving cytofission, the absence of rlc1 might render the cells with weakened cell walls unable to undergo cytofission at high temperatures, which are normally permissive for cell division in rlc1Δ cells alone (Naqvi ). Consistent with the LatA experimental findings, the double mutant cps1-191 rlc1Δ with weakened cell walls did not undergo ring contraction at a high temperature (Figure 4C; 23/24 spheroplasts), whereas the single mutant of cps1-191 could undergo cytofission.

Targeted membrane deposition is required in the cytokinesis of fission yeast (Wang ; Onwubiko ). Next, we tested if targeted membrane deposition at the division site facilitates actomyosin ring contraction in cytofission. When the vesicular trafficking across the endomembrane system was inhibited using brefeldin A in the cps1-191 spheroplasts, the myosin rings were not able to contract to drive cytofission events (Figure 4D, 27/27). This result suggested that addition of cell membrane via targeted membrane trafficking at the division site is necessary to enable cytofission.

Our study reveals that the extracellular glycan matrix inhibits actomyosin ring contraction in the absence of cell wall remodeling and division septum synthesis. When the inhibition is relieved by experimental treatments such as ones reported in this study, or by septum synthesis, the actomyosin ring contracts to drive the membrane ingression. A previous study by Proctor analyzed cps1-191 mutants and explained that the failure of membrane ingression in the mutant was due to a defect in division–septum assembly. The authors also proposed that the high intracellular turgor pressure prevents actomyosin ring contraction in fission yeast. We tested this model by lowering the turgor pressure in cps1-191 mutant cells and found that it was not sufficient to enable membrane ingression in the absence of cell wall remodeling in the cps1-191 cells. However, the ability of cps1-191 mutant cells to divide upon weakening of the cell wall indicates that the actomyosin ring in cps1-191 mutant cells is capable of driving membrane ingression even when the division-septum assembly is defective. When cell wall remodeling is normal, as in wild-type cells, ring contraction and membrane ingression coordinate with cell wall and septum growth. The lowering of turgor pressure by sorbitol addition in wild-type cells with normal cell wall remodeling may facilitate ring contraction, explaining the findings of Proctor et al (Proctor ). The fact that ring contraction is slower during cytofission however, agrees better with the work of O’Shaughnessy and colleagues, who have proposed that the rate of septum synthesis sets the rate of cytokinesis (Stachowiak ). It is possible that our work reveals the highest rate of actomyosin ring contraction when confronted with membrane drag and viscous drag of the cytosol. The low ring contraction rate could be as a result of a reduced amount of Cps1-191 or cytokinetic proteins at the division site (Cortes ). In fission yeast spheroplasts, the actomyosin ring is probably required at the early phase of cytofission to drive spheroplasts into a dumbbell shape with high curvature. Although we cannot exclude the possibility that residual and undetectable actomyosin structures may facilitate division after seeming actomyosin ring disassembly, recent work suggests other potential mechanisms not involving actomyosin rings or ESCRT in division of dumbbell-shaped vesicles. It has been proposed that the spontaneous curvature in dumbbell-shaped lipid vesicles generates constriction forces to induce membrane fission. This leads to the division of a dumbbell-shaped lipid vesicle into two with increased curvature (Steinkuhler ).

The yeast cell wall consists mainly of glycan matrix and glycosylated proteins and has been suggested as a functional equivalent of the extracellular matrix (ECM) in animal cells (Munoz ). The mechanical interaction between the cytokinetic actomyosin ring and the ECM is not well understood. A recent study of zebrafish epicardial cells in the heart explants shows the cell–ECM adhesions at the division site. The cell–ECM adhesions lead to the traction forces at the cytokinetic ring that inhibit cytokinesis (Uroz ). An early biophysical study also detected a large traction force at the cleavage furrow of fibroblast cells cultured on an elastic substrate, suggesting an interaction of cytokinetic machinery and ECM (Burton and Taylor, 1997). When cell–ECM adhesion is enhanced during mitosis, cleavage furrow ingression is inhibited in the epithelial cells (Taneja ). Consistently, our study shows that the anchoring to the extracellular glycan matrix of actomyosin rings that do not undergo remodeling (due to a defective Bgs1) prevents actomyosin ring contraction and cell membrane ingression. Weakening of the extracellular glycan matrix, presumably mimicking decreased cell–ECM adhesion, has enabled cytofission events.

MATERIALS AND METHODS

Yeast strains, media, and culture conditions

Table S1 lists the S. pombe strains used in our study. Standard fission yeast genetic techniques were used to prepare the strains. The rich medium YEA (5 g/l yeast extract, 30 g/l glucose, 225 mg/l adenine) was used to culture cells until midlog phase at 24°C before the temperature shift. LatA (Enzo Life Sciences; BML-T119) was used at a final concentration of 150 µM to perturb the actin dynamics in spheroplasts. Brefeldin A (Fisher Scientific; 15526276) was used at a final concentration of 75 μM to slow plasma membrane invagination. Calcofluor White Stain for cell wall staining was purchased from Sigma.

Preparation of Schizosaccharomyces pombe spheroplasts for live-cell imaging (spheroplasting)

The cps1-191 cells used in this study were first cultured in YEA medium at 24°C to midlog phase (OD595 = 0.2–0.5), and then were shifted to 36°C for 6 h 15 min (nonpermissive conditions). Twenty milliliters of culture were spun down at 3000 rpm for 1 min and washed once with an equal volume of E-buffer (50 mM sodium citrate, 100 mM sodium phosphate, pH 6.0). After the cells were spun down and resuspended in 5 ml of E-buffer containing 1.2 M sorbitol, the cell suspension was incubated with 30 mg of lysing enzymes from Trichoderma harzianum or Glucanex (Sigma, L1412; an enzymatic mixture with at least glucanase, cellulase, protease, and chitinase activities) at 36°C with shaking at 80 rpm for 90 min. This was followed by continuous incubation with 40 µl of LongLife Zymolyase (G-Biosciences, 1.5 U/µl; an enzymatic mixture with at least β-glucanase, protease, and mannanase activities) at 36°C with shaking at 80 rpm for 60 min. After enzymatic digestion, the cell suspensions were spun down at 450 × g for 2 min and washed once with 5 ml of E-buffer containing 0.6 M sorbitol. After being spun at 450 × g for 2 min, the spheroplasts were recovered in 10 ml EMMA (Edinburgh minimal medium with all amino acid and nucleotide supplements) containing 0.8 M sorbitol and 0.5% (vol/vol) of 1 M 2-deoxyglucose (Sigma, D6134) for 30 min at 36°C before microscopy imaging.

Sample preparation for light microscopy

One to two milliliters of spheroplast suspensions in EMMA containing 0.8 M sorbitol and 0.5% (vol/vol) of 1 M 2-deoxyglucose (Sigma, D6134) were concentrated to 20–100 µl by centrifugation at 450 × g for 2 min. About 10 µl of concentrated spheroplasts were loaded onto an Ibidi µ-Slide 8-well glass-bottomed dish (Cat. No. 80827), and covered with mineral oil (Sigma, M5310) to prevent evaporation during the imaging process.

To image cells in Figure 1 and Supplemental Figure 4, the cps1-191 cells and rlc1Δ cells after shifting to nonpermissive conditions were treated with buffers used to prepare spheroplasts, but with the lysing and lytic enzymes omitted to preserve the cell wall integrity. After the buffer washing, the cps1-191 cells in Figure 1 were recovered in EMMA with full supplements containing 0.8 M sorbitol and 0.5% 2-DG. For the rlc1Δ cells in the Supplemental Figure 4, after the buffer washing, the cells were recovered in EMMA with full supplements containing 0.8 M sorbitol but not 2-DG to allow septation.

Sample preparation for electron microscopy

Two hundred fifty milliliters of cells with OD595 0.2 were collected for spheroplasting. Spheroplasts were prepared by the spheroplasting method described above. Spheroplasts were spun down from EMMA with 0.8 M sorbitol and resuspended in phosphate-buffered saline (PBS) with 2.5% glutaraldehyde and 1.2 M sorbitol. Fixation solution was prepared by adding 2% glutaraldehyde and dissolving 1.2 M sorbitol in PBS. After 2 h incubation at room temperature, spheroplasts were spun down in round-bottomed tubes. The following procedures were done at 4°C and gently (vortex mixer was avoided). Spheroplasts were resuspended in fixation solution and stood on ice for 2 h. The spheroplasts were separated into 2 tubes: washed and unwashed samples. Unwashed samples were stored at 4°C. The washed samples were washed three times with 1 ml PBS containing 1.2 M sorbitol. Last, the spheroplasts were resuspended in 1 ml PBS containing 1.2 M sorbitol and stored at 4°C before electron microscopy.

For electron microscopy, glutaraldehyde-fixed cells were placed on a slide glass whose surface was pretreated with 0.1% poly-l-lysine. They were washed with 0.1 M phosphate buffer (pH 7.2), postfixed with 1% osmium tetroxide at 4°C for 1 h, dehydrated with graded series of ethanol, and critical point–dried with a Leica EM CPD030 apparatus (Leica Microsystems, Vienna). The specimens were coated with osmium tetroxide by an osmium coater (Vacuum Device, Japan) and observed with S-3400N and SU8020 scanning electron microscopes (Hitachi High Technologies, Tokyo) at 10.0 kV and 1.0 kV, respectively (Namiki ).

Light microscopy

An Andor Revolution XD spinning disk confocal microscope was used to image the spheroplasts and cells at 36°C. The microscope was equipped with a Nikon ECLIPSE Ti inverted microscope, a Nikon Plan Apo Lambda 100 ×/1.45N.A. oil immersion objective lens, a spinning-disk system (CSU-X1; Yokogawa), and an Andor iXon Ultra EMCCD camera 897 or 888. Andor IQ3 software was used to acquire images at a pixel size of 80 or 69 nm/pixel, depending on the camera model. Laser lines at wavelengths of 405, 488, or 561 nm were used for the excitation of fluorophores. Most images were acquired with Z-step sizes of 0.5 µm as listed here: Figure 2A (12 µm/25 Z-sections); Figure 2B (10 µm/21 Z -sections); Figure 3A (10 µm/21 Z-sections); Figure 3B (15 µm/31 Z-sections); Figure 3C (15 µm/31 Z-sections); Figure 4A (15 µm/31 Z-sections); Figure 4B (12 µm/25 Z-sections); Figure 4C (13 µm/27 Z-sections); Figure 4D (10 µm/21 Z-sections); Supplemental Figure 1 (10 µm/21 Z-sections); Supplemental Figure 3A (15 µm/31 Z-sections); Supplemental Figure 3B (13 µm/27 Z-sections); Supplemental Figure 4 (13 µm/27 Z-sections).

Image analysis

Images were processed using Fiji. The time-lapse montages are maximum-intensity projections of Z-stacks of specified time points. All images analyzed were prepared in this study, except images for quantification of the rate of ring sliding in wild-type spheroplasts, in which the data were based on time-lapse images acquired in a previous study (Lim ).

Click here for additional data file.

Click here for additional data file.