SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

The earliest event in bacterial cell division is the formation of a Z ring, composed of the tubulin-like FtsZ protein, at the division site at midcell. This ring then recruits several other division proteins and together they drive the formation of a division septum between two replicated chromosomes. Here we show that, in addition to forming a cytokinetic ring, FtsZ localizes in a helical-like pattern in vegetatively growing cells of Bacillus subtilis. FtsZ moves rapidly within this helix-like structure. Examination of FtsZ localization in individual live cells undergoing a single cell cycle suggests a new assembly mechanism for Z ring formation that involves a cell cycle-mediated multistep remodelling of FtsZ polymers. Our observations suggest that initially FtsZ localizes in a helical pattern, with movement of FtsZ within this structure occurring along the entire length of the cell. Next, movement of FtsZ in a helical-like pattern is restricted to a central region of the cell. Finally the FtsZ ring forms precisely at midcell. We further show that another division protein, FtsA, shown to interact with FtsZ prior to Z ring formation in B. subtilis, also localizes to similar helical patterns in vegetatively growing cells.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

The earliest stage of cell division in bacteria is the polymerization of the tubulin-like FtsZ protein at the division site at midcell into a ring, called the Z ring. This ring forms on the inside of the cytoplasmic membrane and mediates division by recruiting several other division proteins which drive septum formation, resulting in the production of two newborn cells. Cell division is tightly controlled in time and space to ensure accurate partitioning of chromosomes into newborn cells. Likewise, the timing of assembly and the positioning of the Z ring must be precisely regulated. However, the mechanism for triggering Z ring assembly so that it occurs at the correct place and at the correct time is unclear.

FtsZ is an essential, highly conserved protein which, like tubulin, undergoes GTP-dependent polymerization (de Boer et al., 1992; RayChaudhuri and Park, 1992; Mukherjee et al., 1993). In the presence of GTP, FtsZ reversibly assembles into protofilaments in vitro that consist of a head-to-tail linear polymer of FtsZ (Mukherjee and Lutkenhaus, 1994; Erickson et al., 1996; Lowe and Amos, 1999; Oliva et al., 2004). Unlike tubulin, FtsZ protofilaments do not form microtubules in vitro but associate laterally to form bundles or sheets (see Margolin, 2005; Dajkovic and Lutkenhaus, 2006; Harry et al., 2006 and references therein). The arrangement of FtsZ protofilaments in the Z ring has not been established, as the in vivo polymer in the ring has never been observed directly by electron microscopy. The Z ring is highly dynamic, continually remodelling itself both before and during its constriction. FRAP (fluorescence recovery after photobleaching) experiments have estimated a Z ring turnover with a half-time of 8–9 s in Escherichia coli and Bacillus subtilis (Stricker et al., 2002; Anderson et al., 2004). This subunit exchange within the Z ring depends on GTP hydrolysis (Stricker et al., 2002; Anderson et al., 2004). Once formed, the Z ring persists for a considerable part of the cell cycle (Weart and Levin, 2003; Aarsman et al., 2005) before undergoing constriction. The mechanism of Z ring constriction and whether this ring provides the force for septum formation remain unsolved (Margolin, 2005; Shih and Rothfield, 2006).

FtsZ has been observed to form localization patterns other than the Z ring. However, the majority of helical-like localization patterns of FtsZ reported some time ago were observed under non-wild-type conditions, such as overexpression of an FtsZ–YFP fusion protein in E. coli (Ma et al., 1996), or in certain B. subtilis and E. coli mutants (Addinall and Lutkenhaus, 1996; Mileykovskaya et al., 1998; Jones et al., 2001; Stricker and Erickson, 2003). These assemblies were usually considered abnormal and were often non-functional, causing division defects. More recently, and probably as a result of improved fluorescence microscopy capabilities, FtsZ has been observed to form helical-like patterns in wild-type bacterial cells. Such helical-like patterns of FtsZ have been observed in wild-type exponentially growing E. coli cells and in Caulobacter cells (Thanedar and Margolin, 2004; Thanbichler and Shapiro, 2006). In the case of E. coli it has been proposed that upon disassembly of the Z ring, FtsZ is redistributed in a dynamic helical cytoskeleton. It was suggested that this dynamic helix created a reservoir for the rapid turnover of FtsZ in the ring (Thanedar and Margolin, 2004).

Similar helical patterns of FtsZ have been observed in wild-type B. subtilis and Streptomyces coelicolor cells undergoing the developmentally regulated division during sporulation (Ben-Yehuda and Losick, 2002; Grantcharova et al., 2005). In B. subtilis these helical assemblies are instrumental in the switch from medial to asymmetric division that occurs upon entry into sporulation in this organism (Ben-Yehuda and Losick, 2002). However, helical localization of FtsZ during vegetative growth in these Gram-positive organisms has not been reported. The question that remains is, do Gram-positive bacteria, such as B. subtilis, form non-ring FtsZ assemblies only during sporulation or do they also play a role in midcell Z ring formation in vegetatively growing cells?

Here we report previously unrecognized dynamic FtsZ helical assemblies in vegetatively growing wild-type B. subtilis cells. Our results suggest a new model for midcell Z ring assembly during vegetative growth in this organism that involves a cell cycle-regulated redistribution of this helical assembly of FtsZ in two distinct stages. We discuss the insights this new model provides to current models of Z ring regulation in bacterial cells.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

FtsZ forms a helical structure during vegetative growth

Previous immunofluorescence microscopy (IFM) studies examining FtsZ localization in vivo in vegetatively growing B. subtilis cells have often used the severe fixative formaldehyde. This provides clear images of a single band at midcell representing the Z ring (Levin and Losick, 1996; Harry et al., 1999; Regamey et al., 2000). However, we noticed that, when compared with IFM experiments localizing other proteins, there is a higher than normal level of background staining in FtsZ-immunostained cells that is a direct result of specific staining of FtsZ (our unpublished observations). A more careful examination of this non-ring localization of FtsZ suggested a helical-like localization pattern. We modified our IFM technique to optimize resolution of this localization. We anticipated that methanol might allow better preservation of these FtsZ polymers, if they existed. The wild-type strain, B. subtilis 168 (SU5), was grown in Penassay broth (PAB) at 34°C, which yields a doubling time of 30 min, and exponentially growing cells were fixed with methanol and subjected to IFM using anti-FtsZ antibodies (Fig. 1). The majority of cells (64%) contained Z rings at the cell centre, as expected (Fig. 1B). In addition, however, we noticed that all cells contained a different pattern of FtsZ staining throughout the cell (Fig. 1B and C). This pattern consisted of a series of dots preferentially on the sides of the rod-shaped cell that were sometimes seen to be connected by a fluorescent line, suggesting an organized FtsZ structure. To obtain a higher resolution of this non-ring localization pattern of FtsZ, we performed three-dimensional (3D) deconvolution. This involves acquiring a series of two-dimensional images of the fluorescently stained cells in different focal planes, and then deconvolving the image to eliminate out-of-focus fluorescence. The result is a higher resolution reconstruction of the immunostained pattern in 3D. Examples of raw and deconvolved images are shown in Fig. 2A–C. The deconvolved images show fluorescent lines along the cell in a helical-like pattern. Although the pattern of FtsZ localization was not always clearly defined, in the majority of cells (77%) FtsZ appeared as a double helical structure, with lines of fluorescence overlapping diagonally across the cell (Fig. 2Aii and Bii). The helical-like nature of this FtsZ localization is most readily observed by examining each successive focal plane of the deconvolved images. This revealed the curvature of the fluorescent track of FtsZ in the cell. This is illustrated in Fig. 2D which shows lines of fluorescence in different orientations in the different focal planes, consistent with the presence of a helical structure.

image

Figure 1. Immunofluorescence analysis of FtsZ localization. Vegetative wild-type B. subtilis cells (SU5) were grown to mid-exponential phase in Penassay broth at 34°C. (A) Phase contrast image and (B) fluorescence image of the same cells shown in (A). Cells outlined by the white box are enlarged in (C). Scale bar, 5 μm for (A) and (B); 2 μm for (C).

Download figure to PowerPoint

image

Figure 2. Projected two-dimensional images of FtsZ structures in cells obtained by 3D deconvolution. Wild-type B. subtilis cells were grown to mid-exponential phase in Penassay broth at 34°C, then immunostained for FtsZ using anti-FtsZ antibodies. A–C. Image (i) shows the original fluorescence image and image (ii) shows the corresponding deconvolved image as a maximum intensity projection. A cartoon for each cell is shown depicting the interpreted location of FtsZ within the cell. (A) Cell with an FtsZ helix, but no Z ring. (B) Cell with a Z ring and FtsZ helix. (C) Cell with a constricting Z ring and an FtsZ helix. D. Different focal planes of a deconvolved cell showing the top (i); middle (ii); and bottom (iii). Scale bar, 1 μm.

Download figure to PowerPoint

We also performed confocal microscopy on the FtsZ-immunostained cells to see if we could obtain an even higher resolution of the helical-like FtsZ structure. The reconstructed images obtained using confocal optics show a helical-like structure very similar to the reconstructed images obtained with wide-field fluorescence microscopy (not shown). Interestingly we noticed that with both the confocal and the deconvolved wild-field images the fluorescence of the helical-like pattern of FtsZ was less intense in cells that had a Z ring compared with those without a Z ring. We compared the helical immunostaining intensity in each type of cell, that is, those containing a Z ring and those without (n = 40 cells; 20 for each cell type) as described in Experimental procedures. The average total fluorescence intensity of the helical-like FtsZ structure was 42 ± 1.6 [± standard error of the mean (SEM)] units in cells without a Z ring, compared with only 31 ± 1.7 units for cells containing a Z ring. This represents a 25% decrease in fluorescence intensity of the helix in cells that also contain a Z ring. This difference is highly significant (P = 0.000; see Experimental procedures). Fluorescence intensity decreased to a similar value towards the end of the image captured in both types of cells (Z ring and no Z ring) as the plane of focus extended well below fluorescence detected from the cells (see right-hand side of curves in Fig. 3A). The Z ring itself had an average area under the curve of 153 ± 9.3 units, which is three- to fivefold that of the helix. These results show that when a Z ring is present in vegetative cells, there is a significant concomitant decrease in the amount of FtsZ in the helical assembly.

image

Figure 3. A. Quantification of FtsZ-immunostaining intensity of the FtsZ helical structures in wild-type (SU5) B. subtilis cells that contained a Z ring and in those that did not. Cells grown in Penassay broth at 34°C were collected at the mid-exponential phase of growth and then immunostained using anti-FtsZ antibodies. Confocal microscopy images were acquired at 0.12 μm intervals in the axial plane. The images on the graph indicate the approximate positions of the successive axial planes relative to the cell circumference. The fluorescence intensity of the helical structure at each of the 22 axial planes was calculated for cells either containing a Z ring (grey) or not containing a Z ring (black) (20 cells of each). The intensity of the helix at each axial plane was averaged and is plotted in this graph. *The lateral axial distance on the x-axis refers to the sequential steps from the top of the cell to the bottom of the cell in the z-plane. B. Cell length distribution of vegetatively growing wild-type (SU5) B. subtilis cells that contain both a Z ring and an FtsZ helical structure (grey; n = 216), and those that contain only a helical structure (black; n = 124). Cells were grown to mid-exponential phase in Penassay broth at 34°C and immunostained for FtsZ using anti-FtsZ antibodies.

Download figure to PowerPoint

The helical assembly of FtsZ forms prior to the Z ring

It has previously been shown that Z rings form prior to septum formation, subsequently constricting as the cell envelope layers invaginate, and then disassembling once division is complete to produce newborn cells. The discovery here that all cells contain a helical pattern of FtsZ localization, but not all cells contain a Z ring, predicts that the shortest cells would have a helical-like pattern with no Z ring and that the longest cells would have both a Z ring and a Z helical structure. To confirm that this is the case using our modified IFM technique, we measured cell lengths of wild-type B. subtilis vegetative cells (SU5) that had been grown in PAB at 34°C and stained for FtsZ by IFM. Figure 3B shows the range of lengths for cells containing either a helical-like structure, or both a helix and a ring. Cells that contained only a helical pattern comprised 36% of the population, and had a mean cell length of 3.9 ± 0.05 μm (SEM), with 87% being shorter than 4.5 μm. In contrast, cells that did contain a Z ring comprised 64% of the population and were longer on average: 5.1 ± 0.07 μm, with the majority (69%) being longer than 4.5 μm.

Our data are consistent with Z rings forming at a specific stage in the cell cycle and disappearing once division is complete. However, it is not clear using vegetatively growing cells whether the FtsZ helical pattern of localization relies upon prior Z ring formation or whether it can form de novo. To resolve this, we analysed FtsZ localization during spore outgrowth using our modified IFM procedure. Unlike vegetative growth, spore outgrowth is uncomplicated by previous division events and allows the opportunity to determine if the helical structures of FtsZ are able to form prior to, and independently of Z rings. Spores of wild-type B. subtilis cells (SU5) were grown out in GMD at 34°C and FtsZ localization was visualized using IFM at various times after germination. As with vegetative cells, helical-like patterns of FtsZ were observed in all cells during spore outgrowth (data not shown). At the earliest time point, 180 min after incubation in GMD medium, the average cell length was 3.2 ± 0.07 μm (n = 133), increasing to 3.6 ± 0.10 μm (n = 121) and 3.7 ± 0.09 μm (n = 183) at 210 and 270 min respectively. At 180 min, the vast majority of cells (82%) displayed a helical-like localization pattern of FtsZ but did not contain a Z ring. Only 18% of cells contained both a helix and a ring at 180 min. The proportion of cells containing Z rings increased to 25% at 210 min and then to 43% at 270 min. As with vegetatively growing cells, at all time points outgrown cells without a Z ring were significantly shorter, on average, than those containing a Z ring (not shown). These data demonstrate that the FtsZ helical structure is present prior to division. Furthermore, these results strongly suggest that the helical structure forms prior to and independently of the Z ring.

The helical structure of FtsZ in live cells

To observe the FtsZ helical localization pattern in live cells, we monitored the localization of an FtsZ–YFP fusion protein. Outgrown spores were initially used in these experiments as the FtsZ helix was more intense in fluorescence in these cells than in vegetative cells. SU434 contains, in addition to the wild-type copy of the ftsZ gene, a xylose-inducible ftsZ–yfp fusion inserted into the chromosome at the amyE locus. Quantitative immunoblotting indicated that with the xylose induction conditions used here, approximately 30–40% of FtsZ in these cells was FtsZ–YFP, and total cellular FtsZ levels (FtsZ plus FtsZ–YFP) were only 1.3-fold higher than in wild-type cells (data not shown). Most importantly, Z ring assembly and cell division in this strain are completely normal (Migocki et al., 2002; 2004). SU434 (FtsZ–YFP) spores were grown out at 34°C in GMD containing 0.2% xylose and harvested after 150–210 min of incubation. The difference in harvest time compared with the wild-type strain simply reflects the different rates of germination and spore outgrowth routinely observed for each spore preparation. As with IFM, in addition to Z rings we observed significant fluorescence in other regions of the cell at all time points. This signal consisted of a regular pattern of dots, most prominent along the sides of the cell that were often linked by fainter lines of fluorescence, consistent with a helical-like structure (Fig. 4). As with IFM, at an early time point following incubation in GMD (150 min), the vast majority of SU434 (FtsZ–YFP) cells contained only the helical-like structure of FtsZ (81%). The remaining 19% of cells contained a Z ring, and of these cells, 60% also showed a helical pattern of FtsZ. The frequency of cells containing Z rings increased to 58% and 69% at 180 min and 210 min respectively. We also noticed that generally the helical-like FtsZ structure was far less intense in fluorescence in cells containing a Z ring, than in cells not containing a Z ring. This is consistent with the IFM quantification of FtsZ in the helix in these two types of cells, although the difference in intensity of the helix is far more obvious in the live cells containing FtsZ–YFP.

image

Figure 4. Visualization of FtsZ–YFP in cells outgrown from spores. Spores of B. subtilis SU434, which contains a xylose-inducible copy of ftsZ–yfp, were germinated and grown out at 34°C in GMD medium containing 0.2% xylose (w/v). Cells were then collected at various times of outgrowth, applied to agarose pads and viewed by fluorescence microscopy. Cartoons of various cells in (B), (C) and (D) are shown below these panels depicting the interpreted localization of FtsZ in the cell. Incubation in GMD medium was for 150 min (Ai and Aii); 180 min (B); 210 min (Ci and Cii); and 240 min (D). Z rings are shown in the bottom cells in (Cii) and (D). The constricting Z ring is marked with an arrowhead in (D) and is accompanied by helices of FtsZ at the approximate one-fourth and three-fourths positions (the future division sites). The asterisk in (Aii) marks an autofluorescent spore coat. Scale bar, 1 μm.

Download figure to PowerPoint

The FtsZ helical structure is highly dynamic

Although the helical-like FtsZ structures observed in live cells were similar to those observed in cells processed for IFM, the FtsZ–YFP signal appeared not to fluoresce along the whole length of the cell. Rather the helical pattern was mainly observed in a region (or regions) of the cell. It is highly likely that this is due to the significantly lower intensity of the YFP signal compared with the strong immunostained FtsZ signal, such that where no helical pattern is observed in the live cells, the YFP signal was too weak to detect. This difference in intensity between FtsZ immunostaining and the FtsZ–YFP signal has also been observed in sporulating B. subtilis cells (Ben-Yehuda and Losick, 2002) and is probably due to the significantly lower (approximately fourfold) cellular level of FtsZ–YFP in cells containing this fusion compared with the cellular level of wild-type FtsZ protein in the wild-type strain. However, to rule out the possibility that the presence of the FtsZ–YFP fusion protein itself caused a different pattern of FtsZ localization in these cells compared with wild-type cells, we immunostained vegetatively growing SU434 cells that had been induced to produce the FtsZ–YFP fusion. The immunostained helical pattern of FtsZ localization in these cells was identical to that of wild-type cells (SU5; data not shown), demonstrating that the FtsZ–YFP fusion does not alter the helical distribution of FtsZ within the cell, nor does it appear to change its structure.

The question remains then, as to why the FtsZ–YFP signal would be stronger in certain regions of the cell than others in live cells. In fact, on more careful examination of the helical FtsZ patterns obtained with IFM, we observed a brighter stained region of the helical pattern in a significant number of cells (see Fig. 2Bii and Di–iii). This raised the question of whether, as in E. coli (Thanedar and Margolin, 2004), the helix was a dynamic structure in which FtsZ distribution within it changed with time. Time-lapse microscopy was performed using the FtsZ–YFP fusion in live outgrown SU434 cells. In these live-cell experiments, a heated stage was used and was required for significant growth (extension) of cells over the time-lapse period. SU434 (FtsZ–YFP) spores were incubated for 150 min in GMD containing 0.2% xylose at 34°C and cells were mounted onto agarose pads for visualization at 34°C. Figure 5 shows time-lapse images of an outgrown cell with the corresponding movie shown in Supplementary material (Fig. S1A). Intervals of 90 s were used over a 61 min period. These images show that FtsZ within the helical assemblies is in constant motion, with the distribution of fluorescence changing in every frame. This mobility was also observed in every frame when a time interval of only 15 s was used, indicating the high degree of mobility of FtsZ within this structure (data not shown). The movement of FtsZ within the helical assembly appeared random. It has been reported that, in addition to the mobility of FtsZ within the helical-like cytoskeleton in E. coli, there is a longer range pole-to-pole oscillation of FtsZ–GFP helical structures in live cells of this organism (Thanedar and Margolin, 2004). We did not observe such oscillatory fluorescence in B. subtilis cells.

image

Figure 5. Time-lapse microscopy of FtsZ–YFP localization in outgrown spores. Spores of B. subtilis strain SU434, which contain a xylose-inducible copy of ftsZ–yfp, were germinated and grown out at 34°C in GMD containing 0.2% xylose. Cells were then collected following 150 min of outgrowth, applied to agarose pads and viewed by fluorescence microscopy using a stage heated to 34°C. Images were taken every 90 s over a 61 min period to show the dynamic nature of the initial FtsZ helix. Numbers represent time in minutes. Scale bar, 1 μm. The movie can be viewed in Supplementary material (Fig. S1A).

Download figure to PowerPoint

Cell cycle analysis of FtsZ localization in live cells

The discovery that FtsZ can form a dynamic helical-like structure de novo in B. subtilis and is present prior to the assembly of a Z ring at the division site raised the possibility that the helical structure is a precursor to the Z ring and that Z ring assembly may arise through the redistribution or remodelling of FtsZ within the helical-like structure. The concomitant decrease in FtsZ levels in the helix when a Z ring is also present in the cell (observed when visualizing FtsZ using both IFM and FtsZ–YFP in live cells) supports this idea. To test this possibility, time-lapse microscopy was performed using outgrown spores of SU434 (FtsZ–YFP) to follow FtsZ localization in a single cell cycle in individual outgrown cells. Cells were collected after 150 min incubation in GMD at 34°C, placed on an agarose pad and images were taken every 90 s over a 150–180 min period. We used longer time intervals in this case to prevent bleaching of the sample over this longer time period. Figure 6 shows time-lapse images of outgrown cells and the corresponding movie for the cell on the left is shown in Fig. S1B (Supplementary material). Fig. S2 in Supplementary material shows the region of the cell occupied by FtsZ in the time-lapse sequence shown in Fig. 6. Image 0 in Fig. 6 shows that initially in the outgrown cell, FtsZ forms a helical pattern of localization. In the movie (Fig. S1B), initially FtsZ appears to be able to move within this helical structure along the entire length of the cell. To reduce significant loss of fluorescence intensity due to bleaching over time, in all time lapse experiments we used shorter exposures which required binning to maintain sensitivity, while sacrificing some resolution. For this reason the longer helical-like patterns of FtsZ obtained in all time-lapse sequences are less clear in these images compared the static images shown in Fig. 4. The helical pattern of FtsZ, while remaining dynamic, was then predominantly restricted to a smaller, central region of the cell for a significant amount of time and consisted of about two to three helical turns (Fig. 6, images 16, 22 and 28). Finally, this shorter helical structure was displaced with a Z ring at midcell (Fig. 6, image 49). All cells (total of 19) monitored in this way displayed significant cell extension during the time-course, showed similar FtsZ localization behaviour and exhibited these three basic stages of FtsZ localization. The time taken for Z ring formation to occur ranged from 21 to 158 min, as time lapse was initiated with cells of different lengths (1.8–4.3 μm). In the 19 cells that were monitored by time-lapse microscopy, the average cell length at which the FtsZ helix became spatially restricted to the central region of the cell was 3.9 ± 0.1 μm. This is significantly different to the cell length at which Z rings formed at midcell (4.6 ± 0.1 μm) and suggests that the restriction of FtsZ to this short helical structure is a cell cycle-regulated event. Importantly, in the time-lapse experiments Z rings formed at a similar cell length to that when a Z ring forms in the IFM experiments (4.5 ± 0.2 μm) and in live cells producing FtsZ–YFP (5.0 ± 0.1 μm).

image

Figure 6. Time-lapse microscopy of FtsZ–YFP localization in outgrown spores. Spores of B. subtilis SU434, which contain a xylose-inducible copy of ftsZ–yfp, were germinated and grown out at 34°C in GMD containing 0.2% xylose. Cells were collected for imaging following 150 min of incubation, applied to agarose pads and viewed by fluorescence microscopy using a stage heated to 34°C. Images were acquired at 90 s intervals over a 150 min period. The cell shows the redistribution of FtsZ within the helix (images 0, 5, 16, 22, 28) into a ring (image 49). This ring then constricts (images 91, 106; arrows) and is accompanied by FtsZ emanating in loops from the Z ring (images 106, 115; open arrow). Two new Z rings are then observed to assemble in the daughter cells (images 135, 150; arrowheads). The autofluorescent spore coat is marked with an asterisk in images 91–150. Numbers represent time in minutes. Scale bar, 1 μm. The movie can be viewed in Supplementary material (Fig. S1B).

Download figure to PowerPoint

To monitor what happens after Z ring assembly at midcell, time-lapse analysis was continued on some of the 19 cells (see Fig. 6, left-hand cell in images 91–150; Figs S1B and S2). Z rings were initially observed to constrict, indicating formation of the division septum (Fig. 6, image 91). When Z ring constriction was almost complete, very little FtsZ–YFP fluorescence was observed at the central division site (Fig. 6, image 106). FtsZ then appeared to emanate from this region in both directions as a series of dots, resembling the dynamic helical structure observed in the previous cycle (Fig. 6, image 115; Fig. S1B), although it was less clear at this time point due to bleaching of the signal. FtsZ movement then became restricted within the cell again so that the dominant signal localized to the central region of each cell (arrow in Fig. 6, image 115). Subsequently Z rings were seen to form at these sites (Fig. 6, left-hand cell in images 135, 150; Fig. S1B). Sometimes the formation of Z rings at the future division sites was more rapid so that localization of FtsZ to the shorter helices occurred at the one-fourth and three-fourths positions of the undivided cell almost immediately upon constriction of the Z ring at midcell.

Similar FtsZ–YFP visualization and time-lapse microscopy experiments were performed using vegetatively growing cells to ensure that our observations were not just a phenomenon specific to outgrowing spores. Initially SU434 (FtsZ–YFP) vegetative cells were grown in PAB; however, the helix-to-ring redistribution of FtsZ was hard to capture in cells grown in such a rich medium as the cell cycle was progressing rapidly (doubling time of 30 min), with very little time between Z ring constriction and Z ring formation in the two successive division cycles. The PAB itself was also slightly autofluorescent which made it hard to acquire reasonable images over the required period. To overcome both these complications we switched to using a medium that allowed slower growth [Spizizen minimal medium (SMM); doubling time ∼45 min]. SU434 cells grown in the presence of 0.2% xylose (w/v) were collected at the mid-exponential phase of growth to visualize FtsZ localization using FtsZ–YFP. Under these conditions, 70% of cells contained a Z ring. Of these, 21% had a clear non-Z ring pattern that was similar to that observed in outgrown spores and consistent with a helical structure. The remaining 30% of cells displayed the same helical pattern of FtsZ but had no Z ring (data not shown). For time-lapse microscopy, a total of 46 exponentially growing cells were imaged every 90 s over a 150 min period. This time frame allowed observation of up to three rounds of cell division. Figure 7B shows a time-course of a typical vegetative cell that contained a Z ring at the beginning of the time-course (arrow in Fig. 7B, image 0). This is shown as a time-lapse movie in Supplementary material (Fig.  S3A). As was observed with outgrown cells, during Z ring constriction, FtsZ appeared to emanate from this ring in a manner that is consistent with a helical pattern (Fig. 7, right-hand cell in images 67 and 70). This is more clearly seen in the movie in Fig. S3A. A shorter, dynamic helical structure then appeared at the future division at the one-fourth and three-fourths position sites (Fig. 7, images 76 and 81), and subsequently a Z ring formed at these sites (Fig. 7, asterisks in image 103). As the FtsZ–YFP signal was much harder to detect in vegetatively growing cells, it was not possible to firmly conclude that the non-ring localization pattern of FtsZ was helical in vegetative cells. However, in all 46 cells we observed a dot pattern that was consistent with a dynamic helix and in all cells this pattern became restricted to a smaller, central region of the cell prior to Z ring formation at the division site.

image

Figure 7. Time-lapse images of FtsZ–YFP localization in SU434 vegetatively growing cells. SU434 cells, which contain a xylose-inducible copy of ftsZ–yfp, were grown at 34°C in SMM supplemented with 0.2% (w/v) xylose, collected at exponential phase, applied to agarose pads and viewed by fluorescence microscopy using a stage heated to 34°C. Images were acquired at 90 s intervals over a 103 min period. A. Phase contrast image at time 0. B. Fluorescence time-lapse montage. The right-hand cell shows a midcell Z ring, marked by an arrow (image 0), which then constricts (images 58, 67, 70), and is followed by release of FtsZ from the ring (image 67; arrowheads) and formation of two new Z rings (marked with an asterisk in image 103) in the daughter cells via a helical intermediate. Prior to Z ring formation FtsZ localized to a helix-like pattern at the midcell position of the newborn cells (images 67, 70, 76, 81; arrowhead). Numbers represent time in minutes, scale bar, 1 μm. The movie can be viewed in Supplementary material (Fig. S3A).

Download figure to PowerPoint

To obtain a bright, clear image of FtsZ localization that was confined to a helical pattern in the central region of vegetative-growing cells we performed time-lapse microscopy with cells containing these structures at the beginning of the time-course. An example is shown in the Supplementary material, Fig. S3B (left-hand cell). As with outgrown cells, FtsZ movement within this helical pattern was mainly confined to the central region of the cell for a significant amount of time (see Fig. S2) prior to the formation of a Z ring at midcell.

The cell division protein, FtsA, also forms a predivisional helical structure

It has been known for some time that another cytosolic bacterial division protein, FtsA, interacts directly with FtsZ (Ma et al., 1997; Wang et al., 1997; Din et al., 1998; Ma and Margolin, 1999; Yan et al., 2000; Haney et al., 2001; Jensen et al., 2005). Previous research has shown that, in B. subtilis, FtsA is complexed with FtsZ prior to Z ring formation (Jensen et al., 2005). Furthermore, FtsA is required for the efficient formation of Z rings in B. subtilis (Jensen et al., 2005). We therefore examined whether, like FtsZ, FtsA is able to form a helical structure in growing B. subtilis cells. To examine this possibility, B. subtilis wild-type cells (SU5) were grown in PAB at 34°C and IFM was performed using anti-FtsA antibodies. The majority of cells (65%) contained a sharp band of FtsA at the division site, a similar proportion to those containing a Z ring (64%; Fig. 8B and C). However, in addition to this, a pattern reminiscent of the helical structures of FtsZ observed in cells processed for IFM, was observed in all cells stained for FtsA. This fluorescent pattern was not as intense as that observed for FtsZ, and this may be due to the different cellular concentrations of the proteins: one molecule of FtsA for every five molecules of FtsZ in B. subtilis (Feucht et al., 2001). Three-dimensional deconvolution was performed to obtain a higher resolution of the FtsA staining pattern (Fig. 8A–C). The reconstructed image of FtsA is strikingly similar to that of FtsZ. Clear connections were often observed between the fluorescent dots present at the cell edges, and many cells contained a helical-like localization pattern (see, for example, Fig. 8Aii). Control IFM experiments using an ftsA-deletion strain (Jensen et al., 2005) showed essentially no FtsA staining, confirming that these helical-like structures are indeed composed of FtsA, and are not due to some non-specific staining of another protein (data not shown).

image

Figure 8. Visualization of FtsA in vegetatively growing cells. Wild-type (SU5) B. subtilis cells were grown to mid-exponential phase in PAB at 34°C and immunostained using anti-FtsA antibodies. A and B. Image (i) shows the original fluorescence image and image (ii) shows the corresponding deconvolved image as a maximum intensity projection. (A) Cell with an FtsA helix, but no ring of FtsA. (B) Cell with a ring of FtsA and an FtsA helix. C. Different focal planes of a deconvolved cell showing the top (i); middle (ii); and bottom (iii). Scale bar, 1 μm.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Here we show that in addition to forming a Z ring at the division site at midcell, FtsZ also forms dynamic helical-like assemblies in vegetatively growing wild-type cells and outgrowing cells of B. subtilis. This helical-like localization pattern occurs prior to and independently of Z ring formation. Time-lapse microscopy of individual outgrown cells during a single cell cycle revealed three sequential FtsZ localization patterns. This is illustrated in a model in Fig. 9. In newborn cells (or early outgrown cells) FtsZ moves randomly within a helical pattern that extends the length of the cell (Fig. 9, upper cell). FtsZ is then redistributed so that the majority of FtsZ molecules, while still highly mobile, are restricted to a smaller, central region of the cell (Fig. 9, middle cell). Finally the shorter helix is replaced by a Z ring which forms precisely at midcell (Fig. 9, lower cell). Upon contraction of the Z ring, FtsZ appears to emanate from the ring to form a dynamic helical structure once again. Emanation of FtsZ from the ring in a helical pattern has been observed in exponentially growing E. coli cells (Thanedar and Margolin, 2004). Our data suggest that formation of the Z ring at midcell arises from a cell cycle-mediated succession of different FtsZ cytoskeletal structures.

image

Figure 9. Model for the pathway of FtsZ polymerization during the cell cycle, leading to establishment of the Z ring at the division site at midcell. A detailed description of the model is given in Discussion. Solid lines indicate the dominant location and pattern of FtsZ at the different stages of the cell cycle, with dotted lines denoting a lower concentration of FtsZ, representing a permanent cytoskeletal coil.

Download figure to PowerPoint

Although these three different localization patterns of FtsZ dominate distinct stages of the cell cycle leading up to cytokinesis, we always observed a fainter helical pattern of FtsZ localization along the length of the cell. We infer from this, as suggested for E. coli (Thanedar and Margolin, 2004), that the helical structure of FtsZ throughout the cell is a permanent cytoskeletal feature of growing B. subtilis cells, even when the Z ring is present. The helical pattern of dotted lines in the cells shown in Fig. 9 denotes this permanent helical cytoskeletal FtsZ when present at a lower concentration.

How does the FtsZ helical structure become a ring? Although we cannot rule out the possibility that the Z ring forms completely separately from the helical structures, we favour the idea that one type of FtsZ polymer is directly derived from the other, so that FtsZ molecules (or at least a fraction of them) already in the longer helix are redistributed into the shorter helix and, subsequently, the ring. This suggestion is consistent with the observed decrease in intensity of each preceding FtsZ assembly as the subsequent assembly appears, as well as the specific timing of the appearance of a shorter FtsZ helical structure, and the subsequent Z ring in the cell cycle. As shown in Fig. 9, it is conceivable that the shorter helical structure and the Z ring itself are part of the underlying helical cytoskeleton of FtsZ, which is somewhat analogous to the proposal that the MinE ring occurs within a MinDE cytoskeleton in E. coli (Shih et al., 2003). In this case the Z ring would be part of a helix (Ben-Yehuda and Losick, 2002; Thanedar and Margolin, 2004). FtsZ-binding proteins known to regulate Z ring formation could play a role in regulating FtsZ distribution in these different cytoskeletal structures by altering the association states of FtsZ. For example, if the ring is a helix then remodelling of the short helix into the ring may involve an increase in lateral associations of FtsZ protofilaments, perhaps by ZapA, to decrease the pitch in the helix. Further investigation of a temperature-sensitive ftsZ mutant strain of B. subtilis (ts1) that appears unable to undergo the final remodelling step from the short helical structure into the Z ring could provide valuable clues regarding this mechanism in wild-type cells (Michie et al., 2006).

Our new model for Z ring assembly is different to an earlier model proposing that the Z ring arises from polymerization of FtsZ from a specific nucleation site on the membrane (Lutkenhaus, 1993; Addinall and Lutkenhaus, 1996). Our proposal that FtsZ ring formation in B. subtilis arises from previously formed, helical-like polymers of FtsZ is consistent with new models of Z ring formation proposed for E. coli (Thanedar and Margolin, 2004; Pichoff and Lutkenhaus, 2005). It has recently been shown that, in the Gram-negative aquatic bacterium, Caulobacter crescentus, FtsZ polymers undergo a cell cycle-regulated re-organization (Thanbichler and Shapiro, 2006), and this may become a common theme for bacteria.

The helical FtsZ structures, as with the Z ring, presumably form just under the surface of the cytoplasmic membrane. Current evidence from E. coli suggests that FtsA tethers FtsZ to the membrane via a membrane targeting sequence and that this is required for Z ring formation (Pichoff and Lutkenhaus, 2005). FtsA interacts directly with FtsZ, and in B. subtilis this interaction occurs prior to formation of the Z ring (Ma et al., 1997; Wang et al., 1997; Din et al., 1998; Ma and Margolin, 1999; Yan et al., 2000; Haney et al., 2001; Jensen et al., 2005). Consistent with this idea, we have now shown that FtsA also forms helical structures of essentially identical appearance to those of FtsZ in exponentially growing B. subtilis cells. In B. subtilis Z ring formation is drastically decreased in the absence of FtsA (Jensen et al., 2005) suggesting a principle role for FtsA in this organism in tethering both the Z rings and helices to the inside of the cytoplasmic membrane.

The helical structures reported here are similar to those in sporulating cells of the same organism, with respect to both their appearance and their apparent ability to act as intermediates between successive Z rings (Ben-Yehuda and Losick, 2002). However, the spiral-like FtsZ filaments in sporulating cells appear to undergo directional growth towards the poles from the cell centre, compared with the more random movement of FtsZ within the helical structures observed in exponentially growing cells (Ben-Yehuda and Losick, 2002). Interestingly, the assembly of FtsZ into a ladder of regularly spaced Z rings during sporulation in S. coelicolor also involves the formation and remodelling of helical FtsZ intermediates (Grantcharova et al., 2005). Formation of these FtsZ intermediates in both sporulating organisms has been proposed to coincide with an increase in FtsZ levels (Ben-Yehuda and Losick, 2002; Grantcharova et al., 2005). In growing cells of E. coli and B. subtilis FtsZ levels do not change significantly during the cell cycle (Rueda et al., 2003; Weart and Levin, 2003) suggesting that, unlike the situation in sporulating cells, the remodelling of the Z helix during vegetative growth is not regulated by a change in FtsZ level.

Z ring formation at the correct site is negatively regulated by the Min and Noc systems, which inhibit ring formation in regions of the cell other than midcell (Wu and Errington, 2004; Bernhardt and de Boer, 2005; Goehring and Beckwith, 2005; Margolin, 2005; Rothfield et al., 2005). It has been proposed that these systems act by preventing Z ring formation at sites other than midcell. Interestingly, the dynamic helical-like assemblies of FtsZ in E. coli showed a slower bulk oscillatory movement remarkably similar to, and possibly dependent on, the known pole-to-pole oscillation of the Min proteins in this organism (Thanedar and Margolin, 2004). It has been proposed that this MinCD inhibitor oscillation, which also follows a helical path, causes the Z helix to disassemble at each cell pole in turn, allowing the Z ring to form at midcell (Shih et al., 2003; Margolin, 2005). In B. subtilis the Min proteins, MinC, MinD and DivIVA, do not oscillate. Rather, DivIVA acts to pilot the MinCD complex to the poles (Edwards and Errington, 1997; Marston et al., 1998; Marston and Errington, 1999). Consistent with this idea, we did not observe any oscillation of the FtsZ helix in B. subtilis. We also did not observe a significant decrease in intensity of the FtsZ helical structure at the cell poles, where the concentration of Min proteins is the highest. This raises the possibility that the Min system in B. subtilis does not prevent Z ring assembly at the poles by preventing or reducing the assembly of helical FtsZ polymers at the cell poles. Rather, the Min proteins may prevent redistribution of FtsZ within this structure specifically at the poles, so that a Z ring cannot form there (equivalent to stages shown in Fig. 9B and C). Further work is needed to clarify the situation.

The Noc and SlmA proteins that are involved in nucleoid occlusion of Z ring formation in B. subtilis and E. coli, respectively, have been proposed to block Z ring assembly over the nucleoid by promoting FtsZ polymer disassembly (Wu and Errington, 2004; Bernhardt and de Boer, 2005). However, as FtsZ can apparently form helical assemblies over nucleoids in both these organisms, these proteins may not destabilize FtsZ polymers per se, but may prevent Z ring assembly over the nucleoid by specifically preventing the redistribution of FtsZ into the more centrally located short helix (or the ring) that forms closer to the time of division. The formation of this shorter FtsZ helix may actually coincide with the relief of nucleoid occlusion, which appears to be accompanied by a lower concentration of the Noc and SlmA proteins in this region, prior to bulk nucleoid segregation (Wu and Errington, 2004; Bernhardt and de Boer, 2005).

The movement of FtsZ within a helical-like structure in B. subtilis is similar to that observed previously for E. coli (Thanedar and Margolin, 2004). Presumably, the later assembling division proteins are recruited once the Z ring forms (Aarsman et al., 2005). There are a growing number of proteins that form helices in bacterial cells (Lewis, 2004; Shih and Rothfield, 2006). It will be interesting to see how many of them follow the same path, what proteins in these structures are interacting directly, and whether particular proteins orchestrate this increasingly popular localization pattern.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Bacterial strains, growth conditions, spore preparation, germination and outgrowth

Bacillus subtilis strains used in this study include the wild-type strain, SU5 (168 trpC2; E. Nester) and SU434 [thyA thyB trpC2 amyE::(Pxyl-ftsZ–yfp spc); Migocki et al., 2002; 2004]. SU434 is a derivative of 168 thyA thyB trpC2, known as SB566 (Adler et al., 1967). It contains a xylose-inducible ftsZ–yfp fusion inserted into the chromosome at the amyE locus in addition to the wild-type copy of ftsZ (Feucht and Lewis, 2001). Vegetative B. subtilis SU5 and SU434 cells were grown at 34°C in either antibiotic medium 3 (PAB) supplemented with thymine (20 μg ml−1), and with xylose (0.2% w/v) where required, or SMM (0.5% w/v glucose, 0.02% w/v MgSO4, 1× mineral salts A) supplemented with thymine (20 μg ml−1), l-tryptophan (50 μg ml−1), casamino acids (0.05% w/v) and 1× trace metals (Harwood and Cutting, 1990), and with xylose (0.2% w/v) where required, and harvested at mid-exponential phase (A600∼ 0.4) for fluorescence microscopy. Spores of the wild-type (SU5) and SU434 strains were prepared as described previously (Migocki et al., 2004). Spore germination and outgrowth was performed at 34°C in GMD medium using 2 × 108 spores ml−1 (Harry et al., 1999; Regamey et al., 2000) supplemented with thymine (20 μg ml−1) and xylose (0.2% w/v). Cells were harvested after 150–270 min of incubation in GMD for fluorescence microscopy.

Fluorescence microscopy

Immunofluorescence microscopy was performed as described previously (Harry et al., 1999) with the following fixation modification. Cells (500 μl) were fixed in 10 ml of ice-cold (−20°C) methanol for 1 h prior to lysozyme treatment. Affinity-purified rabbit-α-FtsZ antibodies were used at a 1:10 000 dilution to detect FtsZ, and affinity-purified rabbit-α-FtsA antibodies were used at a 1:50 dilution to detect FtsA. Goat-α-rabbit IgG conjugated to Alexa 488 (Molecular Probes) at a dilution of 1:5000 or 1:10 000 was used for detection. For all fluorescence microscopy experiments, wild-field fluorescence microscopy was used except to obtain the quantitative data presented in Fig. 3A. For wide-field microscopy cells were viewed with either an Olympus BX60 or Zeiss Axioplan 2 fluorescence microscope. Olympus BX60 was equipped with a 100× phase-contrast objective and images were captured and analysed using a Hamamatsu Orca-ER camera and V++ Image Processing and Analysis Software (Digital Optics). Typical exposure times were 200–500 ms. For visualization of Alexa 488 a U-MWIB (Olympus) filter set was used. The Zeiss Axioplan 2 fluorescence microscope was equipped with a 100× phase-contrast objective and an AxioCam MRm cooled CCD camera controlled through AxioVision software (version 4.4 and 4.5; Carl Zeiss). Filter Set 09 (Zeiss) was used to visualize Alexa 488 fluorescence.

Deconvolution (constrained iterative algorithm) of immunostained cells was performed and analysed using a Zeiss Axioplan 2 fluorescence microscope with AxioVision 4.4. Images were taken in 0.2 μm steps for 15 sections in the axial plane, and deconvolved using a simulated point spread function based on microscope settings. Z-stack correction was disabled, as the Alexa 488 dye is very photostable with minimal bleaching occurring during image acquisition under these conditions (Peters et al., 2006).

All live cell microscopy was performed by placing cells on 2% (w/v) agarose pads (prepared with identical media to that which the cells were grown in) within a 65 μl Gene Frame (AB Genes). Time-lapse studies were performed using a Zeiss Axioplan 2 fluorescence microscope with a Zeiss heated stage equipped with an objective heater, heatable universal mounting frame and an achromatic condenser, set at 34°C. Binning of 2 × 2 and a gain factor of 3 were used to obtain all live cell images, with a 200 ms exposure time. YFP fluorescence was visualized with filter set 41029 (Chroma Technology). Image analysis, processing and preparation for publication were performed using AxioVision version 4.5.

Confocal microscopy was performed on immunostained cells to quantify fluorescence of the FtsZ helical localization (see Fig. 3A). Images were acquired on a Leica TCS SP2 confocal microscope and quantification was performed using LASAF 1.3 (Leica Application Suite Advanced Fluorescence). Z-stacks were taken at a zoom factor of 4 (voxel size 73 nm × 73 nm) at 0.12 μm intervals in the axial plane throughout the field of cells. To quantify the fluorescence intensity of the FtsZ helical pattern in cells with and without a Z ring (n = 40; 20 cells of each type), we first calculated the average intensity of this fluorescence in each of the 40 cells along the entire longitudinal axis of the cell in a total of 22 focal planes. To exclude fluorescence from the Z ring when it was present, the average intensity of the helical-like structure at each focal plane was determined using only the region between the Z ring and the cell pole. We then calculated the average fluorescence intensity at each focal plane for each cell type (i.e. those with Z rings and those without), shown in Fig. 3A. The average total fluorescence for each cell type was determined by calculating the area under the average fluorescence intensity curves for each cell type (using GraphPad Prism 4). The quantification data were statistically analysed using a Repeated-Measures General Linear Model anova (systat Version 9, SPSS Inc. 1998) with the between-subjects factor cell type which had two levels (Z ring and no Z ring), and the within-subjects factor cell section (22 axial planes).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

We are grateful to Arne Müller (Carl Zeiss) for technical assistance with 3D deconvolution and image analysis, Fraser Torpy for performing statistical analyses, and Shigeki Moriya for the FtsZ antibody. We thank Peter Lewis and Gerry Wake for critical reading of the manuscript. This work was supported by an Australian Research Council Discovery Project Grant (DP0450770) to E.J.H.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Fig. S1. Movies of FtsZ-YFP localization patterns in outgrown spores. Spores of B. subtilis strain SU434, which contain a xylose-inducible copy of ftsZ-yfp, were germinated and grown out at 34?C in GMD containing 0.2% xylose. Cells were then collected following 150 min of outgrowth, applied to agarose pads and viewed by fluorescence microscopy using a stage heated to 34?C. (A) Images were taken every 90 sec over a 61-min period to show the highly dynamic nature of the initial FtsZ helix (B) Images were taken every 90 sec over a 150 min period to show a longer period of time. FtsZ can be seen to localize first as a helix, which then becomes restricted to the central region of the cell, then rearranges into a Z ring, constricts and two new Z rings are formed at the future division sites via an FtsZ helical intermediate.

Fig. S2. FtsZ-YFP localization during two cell cycles in an outgrown spore. Spores of B. subtilis strain SU434, containing a xylose-inducible ftsZ-yfp fusion, were grown out at 34?C in GMD containing 0.2% xylose (w/v). The graph shows the regions of the cell occupied by FtsZ over the time period, starting from the time the Z helix is visible in the outgrown cell, to the formation of Z rings at the future division sites in the two newborn cells. This graph was plotted from the timelapse images shown in Fig. 6 and Fig. S1B. The region of the cell occupied by FtsZ is expressed as the length of the cell occupied by FtsZ-YFP staining as a percentage of the total cell length. The y-axis shows the progression of the timelapse analysis in minutes. Gaps near the end of the time lapse indicate a frame in which significant fluorescence in the cell was not observed.

Fig. S3. Movies of FtsZ-YFP localization in B. subtilis vegetative cells. SU434 cells, containing a xylose-inducible ftsZyfp fusion, were grown at 34?C in SMM supplemented with 0.2% xylose (w/v), collected at exponential phase, applied to agarose pads and viewed by fluorescence microscopy using a viewing stage heated to 34?C. (A) Images were taken every 90 sec over a 103 min period. FtsZ can be seen in a ring which then constricts and forms two new Z rings via a helical intermediate. (B) Images were taken every 90 sec over a 133 min period. In the left-hand cell FtsZ can been seen in an extremely dynamic helical form that is restricted to the central region of the cell, and is then redistributed into a midcell Z ring.

FilenameFormatSizeDescription
MMI_5673_sm_FigS1A.zip2517KSupporting info item
MMI_5673_sm_FigS1B.zip22469KSupporting info item
MMI_5673_sm_FigS3A.zip2656KSupporting info item
MMI_5673_sm_FigS3B.zip12587KSupporting info item
MMI_5673_sm_FigS2.tif29KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.