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 Gram-positive pathogen Staphylococcus aureus divides by synthesizing the septum in three orthogonal planes over three consecutive division cycles. This process has to be tightly coordinated with chromosome segregation to avoid bisection of the nucleoid by the septum. Here we show that deletion of the nucleoid occlusion effector Noc in S. aureus results in the formation of Z-rings over the nucleoid, as well as in DNA breaks, indicating that Noc has an important role as an antiguillotine checkpoint that prevents septa from forming over the DNA. Furthermore, Noc deleted cells show multiple Z-rings which are no longer placed in perpendicular planes. We propose that the axis of chromosome segregation has a role in determining the placement of the division septum. This is achieved via the action of Noc which restricts the placement of the division septum to one of an infinite number of potential division planes that exist in S. aureus.


Introduction

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

Bacterial cell division requires coordination between cytokinesis and chromosome segregation to ensure placement of the division septum at the right place and time, avoiding bisection of the chromosome by the division apparatus.

Proper temporal and spatial control of the assembly of the division septum largely depends on the regulation of FtsZ localization, as polymerization of this tubulin homologue in a ring-like structure (Z-ring) constitutes the earliest known step in the process of cell division (Lutkenhaus, 2007; Adams and Errington, 2009). The Z-ring then serves as scaffold for the recruitment of the other components of the division apparatus, to form a highly complex protein machinery called the divisome, responsible for the synthesis of a new cell wall and for the constriction of the cell membrane (Errington et al., 2003; Adams and Errington, 2009). Many bacteria seem to use a dual mechanism to ensure that FtsZ polymerization occurs at the middle of the cell, between the two segregated chromosomes: the Min system and nucleoid occlusion (Barak and Wilkinson, 2007; Lutkenhaus, 2007; Bramkamp and van Baarle, 2009).

The Min system inhibits FtsZ polymerization, and thus inappropriate septation, at the cell poles, through the action of a complex formed by MinC and MinD (de Boer et al., 1989; Rothfield et al., 2005; Lutkenhaus, 2007). MinC is an inhibitor of division that is activated by the membrane-associated ATPase MinD (Rothfield et al., 2005; Lutkenhaus, 2007). In Escherichia coli, the inhibitory action of MinCD is retained at the vicinity of the cell poles as a result of a rapid pole-to-pole oscillation of this complex driven by the topological factor MinE (Raskin and De Boer, 1999), while in Bacillus subtilis it is achieved through the tethering action of the polar protein DivIVA (Cha and Stewart, 1997; Edwards and Errington, 1997).

Septation in mutants lacking the Min system is not totally unregulated, as these mutants form division septa randomly within DNA-free regions of the cell, but not on top of the nucleoid (Levin et al., 1998). This is due to the nucleoid occlusion effect, which results from a negative effect of the chromosome on FtsZ polymerization (Mulder and Woldringh, 1989; Woldringh et al., 1991; Wu and Errington, 2004; Bernhardt and de Boer, 2005). Nucleoid occlusion therefore prevents division through areas occupied by the nucleoid and protects the DNA against guillotining by the division septum. The mechanistic base for nucleoid occlusion was elucidated with the identification of two DNA-binding proteins, Noc in B. subtilis (Wu and Errington, 2004) and SlmA in E. coli (Bernhardt and de Boer, 2005), that act as inhibitors of FtsZ polymerization and inhibit Z-ring formation until DNA segregation has properly cleared chromosomal DNA from midcell. In the absence of a functional Min system, noc or slmA mutants form numerous abortive FtsZ structures that overlap the nucleoid and may lead to bisection of the chromosome (Wu and Errington, 2004; Bernhardt and de Boer, 2005). In fact, under different conditions that perturb DNA replication or segregation, these nucleoid occlusion effectors have a critical role in the antiguillotine checkpoint that prevents formation of the septum over the DNA (Wu and Errington, 2004; Bernhardt and de Boer, 2005). Although evolutionary unrelated, both Noc and SlmA display similar localization patterns, binding to the nucleoid with highest concentration at the pole-proximal regions (Wu and Errington, 2004; Bernhardt and de Boer, 2005). Recently, Wu et al. showed that this uneven distribution of Noc results from the binding of this protein to about 70 copies of a conserved 14 bp inverted repeat, which are scattered throughout the chromosome but essentially absent from a large region around the replication terminus (Wu et al., 2009). This strict localization of Noc activity, absent from the terminus region, suggests that this protein may not only be a spatial, but also a temporal regulator of cell division, inhibiting FtsZ polymerization during early stages of chromosome replication and segregation. This would allow the division machinery to start assembling at a defined moment, late in the DNA replication cycle, when the majority of Noc has been displaced from the cell centre (Wu et al., 2009).

Importantly, mutants devoid of both the Min system and nucleoid occlusion still show preferential localization of FtsZ rings in regions between the nucleoids, indicating that other(s), yet unknown factors may also contribute to the nucleoid occlusion effect (Wu and Errington, 2004; Bernhardt and de Boer, 2005).

Division site selection has been studied in detail in the two rod-shaped model organisms B. subtilis and E. coli, both of which divide by placing the division septum always at the same medial plane, equidistant from the two cell poles and perpendicular to the long axis of the cell, the only division plane that generates two identical daughter cells. The problem becomes more complex in cocci, which have an infinite number of theoretical division planes (all circles with maximum diameter within the cell) that can generate two identical daughter cells. Staphylococcus aureus, an important clinical pathogen, divides by switching the division plane in three consecutive perpendicular orientations in successive division cycles (Koyama et al., 1977; Tzagoloff and Novick, 1977) (Fig. S1), akin to the initial series of division planes in a fertilized egg. This implies a novel mechanism for the definition of the division plane and septum placement in bacteria, for which virtually no information is available. S. Foster and colleagues have recently proposed a model in which S. aureus uses epigenetic information contained in the cell wall, in the form of a large belt of peptidoglycan material that labels the division sites, to define the sequential orthogonal planes (Turner et al., 2010). As for the known systems that regulate Z-ring placement, S. aureus does not have a Min system (based on homology searches) but, although the existence of nucleoid occlusion has never been shown, a noc homologue is present in its genome, as well as in the genomes of other spherical bacteria that divide in either two or three planes. Importantly, most studies regarding choice of division planes in round cells used either Neisseria or E. coli round cells as models (Westling-Haggstrom et al., 1977; Begg and Donachie, 1998; Zaritsky et al., 1999; Pas et al., 2001; Ramirez-Arcos et al., 2001; Corbin et al., 2002). These organisms, however, have a Min system which is required for normal growth and division, and therefore are not the most appropriate models for the mode of division in three orthogonal planes used by S. aureus.

The fact that cocci have multiple potential division planes, but only a few that will not lead to bisection of the DNA, raises the possibility that chromosome segregation may have a crucial role in the selection of division planes in spherical cells devoid of Min. Here we show that initiation of chromosome segregation in S. aureus precedes assembly of the FtsZ ring. We also show that the nucleoid occlusion effector Noc has a vital role in determining the placement of the FtsZ ring in orthogonal planes that correspond to Noc-free regions, generated between the two segregated chromosomes, in each division cycle.

Results

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

Chromosome segregation initiates prior to septum assembly in actively dividing S. aureus cells

The volume of an S. aureus spherical cell is almost entirely filled by the nucleoid (see example cells in Fig. S4A). Therefore, if the chromosome is not segregated, there is no circular plane with a diameter equal to that of the cell, which can be used to generate two identical daughter cells, without bisecting the nucleoid. When chromosome segregation occurs, a nucleoid-free area ‘competent’ for Z-ring assembly is generated. To visualize chromosome position as well as FtsZ localization we labelled FtsZ and the chromosome origin of replication in S. aureus cells. For that purpose we constructed S. aureus strain BCBHV005 coexpressing a fluorescent derivative of FtsZ (FtsZ–CFP) and a fluorescent derivative of SpoOJ (SpoOJ–YFP), a member of the ParB family of DNA binding proteins, which colocalizes with the oriC region (Lewis and Errington, 1997; Pinho and Errington, 2004). In this strain, ftsZ–cfp and spo0J–yfp are expressed from their native loci, under the control of their native promoters. An extra, un-tagged, copy of each gene, controlled by Pspac, is expressed in the presence of 0.1 mM IPTG. The extra FtsZ copy is required because, as observed for other organisms (Levin et al., 1999), the fluorescent derivative of FtsZ is only partially functional (data not shown).

Visualization of BCBHV005 cells by fluorescence microscopy showed that the two origins of replication (marked by Spo0J) were already segregated at a time when FtsZ had not yet started to polymerize at the future division site (Figs 1A and S2), confirming that chromosome segregation initiates prior to Z-ring formation in staphylococcal cells. We can hypothesize that once chromosome segregation has proceeded to generate an area of low DNA concentration at mid cell, the FtsZ ring can only assemble between the nucleoids, in the only possible division plane which does not bisect the DNA and which is perpendicular to the division plane from the previous division cycle (a schematic representation of three S. aureus division cycles is shown in Fig. S1). Therefore, theoretically, the axis of chromosome segregation could be the only spatial cue required for specifying the division plane. We thus raised the possibility that the nucleoid occlusion effect may have an important role in determining the placement of the division plane in S. aureus spherical cells, which led us to study the role of Noc in this organism.

image

Figure 1. A. Chromosome segregation initiates prior to FtsZ assembly. Localization of the oriC marker Spo0J–YFP and of FtsZ–CFP in BCBHV005 cells. The overlay image (right) shows that the two new origins of replication in each daughter cell, segregate before FtsZ polymerization at the future division site, between the segregating nucleoids (note that the Z-ring shown corresponds to the division cycle being finished). B. Noc–YFP colocalizes with the nucleoid. Localization of Noc–YFP in BCBHV006 cells shows this protein colocalizing with the nucleoid, often as diffuse spot at places compatible with the origin of replication. The panels show, from left to right, phage-contrast images, Noc–YFP fluorescence, DNA (stained with Hoechst 33342 dye) and a merge of the two last channels. All images are false-coloured. Scale bars, 1 µm.

Download figure to PowerPoint

S. aureus Noc colocalizes with the origin proximal region of the chromosome

A good candidate for the role of nucleoid occlusion effector in S. aureus is the product of the gene identified in the NCBI database as SAOUHSC_03049 in the NCTC8325-4 genome. This gene encodes a 279-amino-acid protein, with a predicted molecular mass of 32 KDa, which shares 48.4% identity with the B. subtilis protein Noc.

To determine the subcellular localization of Noc, we visualized by fluorescence microscopy BCBHV006 cells, which express two copies of noc gene: a native copy controlled by the IPTG-induced Pspac promoter and a noc–yfp fusion, under the control of noc native promoter. Importantly, we have introduced a plasmid-encoded copy of lacI into BCBVH006 (resulting in strain BCBHV007) to allow stringent repression of the Pspac promoter. When BCBHV007 cells were grown in the absence of IPTG, therefore expressing Noc–YFP but not native Noc, they did not exhibit the characteristic Noc null mutant phenotype (data not shown), indicating that the Noc–YFP fusion is functional.

Noc–YFP colocalized with the nucleoid but, similarly to its B. subtilis counterpart (Wu and Errington, 2004), it seems absent from the mid cell region in elongated, bilobed chromosomes, most likely corresponding to the final segregating section of the chromosome, near the replication terminus region (Figs 1B and S3). Noc–YFP localization is reminiscent of Spo0J characteristic pattern of origin localization. However, Noc–YFP signal is more diffuse than that of Spo0J–YFP (comparison between Fig. 1A and B). In agreement with the findings in B. subtilis, Noc recognizes DNA sequences more distributed throughout the chromosome than Spo0J (Gruber and Errington, 2009; Wu et al., 2009).

Noc mutants fail to avoid bisection of the chromosome, which results in DNA breaks

To functionally characterize the S. aureus Noc homologue, we analysed the phenotypic effects of its absence. For that, a noc null mutant was constructed in the background of S. aureus strain NCTC8325-4 by removing the entire gene from the chromosome, leaving no resistance marker. Fluorescence microscopy analysis of the noc null mutant BCBHV001 cells showed that 18.5% (n = 649) of the cells show division defects due to the absence of Noc (Fig. 2). The most striking phenotype present, in 15.4% of the BCBHV001 cells, and not observed in WT cells (Fig. S4A), was the presence of septa over the DNA. We were also able to observe 2.3% of anucleated cells, most likely resulting from the degradation of DNA due to DNA breaks (see below) and a small percentage of cells (0.8%) with condensed DNA (data not shown). Moreover, the BCBHV001 Noc mutant cells have an increased average diameter, when compared with wild-type cells (Fig. 2D).

image

Figure 2. Effects of Noc deletion in S. aureus cells and rescue by noc ectopic expression. Images of BCBHV001 Δnoc cells stained with membrane dye Nile Red and DNA dye Hoechst 33342 (A) or with the cell wall dye Van-FL and Hoechst 33342 (B) showing cells with the septum closing over the DNA (labelled ‘1’) and anucleated cells (labelled ‘2’). Images are false-coloured. Scale bars, 1 µm. C. Quantitative analysis of the observed Noc deletion phenotypes in wild-type (NCTC8325-4) and Δnoc mutant (BCBHV001) strains, as well as in the Noc mutant complemented by ectopic expression of noc (induced with 0.1 mM IPTG) from the spa locus of the S. aureus chromosome (BCBHV010); n = total number of cells analysed. Phenotypes ‘1’ and ‘2’ as described above. D. Quantification of cells diameter in NCTC8325-4 and Δnoc BCBHV001 strains, showing that, in average, Noc mutant cells are bigger than wild-type cells.

Download figure to PowerPoint

The phenotype observed in noc null mutant BCBHV001 cells could be rescued by ectopic expression of the noc gene, under the control of Pspac, at the distant spa locus in the chromosome, in strain BCBHV010, which results in a 10-fold reduction in the percentage of cells with bisected nucleoid (Fig. 2C).

While both in B. subtilis and in E. coli Noc and SlmA have a role in preventing septum formation over the nucleoid only under conditions of DNA replication/segregation perturbation (Wu and Errington, 2004; Bernhardt and de Boer, 2005), our data indicate that in S. aureus the presence of Noc is critical to avoid bisection of the chromosome under normal growth conditions. A possible consequence of the formation of the septum over the nucleoid is the occurrence of DNA breaks with the consequent loss of genomic integrity and/or cell death. Using a TUNEL (Terminal deoxynucleotidyl transferase mediated X-dUTP nick end labelling) assay, which allows the visualization of DNA breaks, we showed that deletion of Noc eventually results in the formation of DNA breaks, which could be detected in 16% of the BCBHV001 noc mutant cells (Fig. 3A). In contrast, wild-type NCTC8325-4 cells were not labelled in the TUNEL assay (data not shown), while 100% of NCTC8325-4 cells pre-treated with the DNA cleavage protein DNaseI were labelled (Fig. 3B).

image

Figure 3. Noc deletion results in DNA breaks. DNA breaks detected by catalytic incorporation of dUTPs-fluorescein (TUNEL assay) in Δnoc BCBHV001 cells (A) and NCTC8325-4 wild-type cells treated with DNaseI (positive control, B). Left panels: phase-contrast images. Middle panels: fluorescence of dUTPs-fluorescein incorporated in DNA breaks. Right panels: Merge of the two last fields with dUTPs-fluorescein fluorescence false coloured in green. Scale bars, 1 µm.

Download figure to PowerPoint

FtsZ polymerizes in multiple ring/arc structures in the absence of Noc

The results shown above indicate that the presence of Noc is important to prevent Z-ring assembly over the DNA. To further evaluate the role of Noc in the localization of FtsZ, we examined the subcellular localization of this protein using a strain expressing an FtsZ–CFP fusion protein. For that purpose, we ectopically expressed ftsZ–cfp under the control of the Pspac promoter, from the spa region of the chromosome, leaving the native ftsZ locus intact. In the parental strain BCBHV011, the FtsZ–CFP fusion protein assembles, as expected, as a ring at the future division site (Fig. 4A). However, in 137 out of 906 (15%) of the BCBHV012 Noc mutant cells, FtsZ polymerized in multiple ring/arc structures, usually with weak fluorescence, disposed at different angles in the spherical S. aureus cells (Fig. 4B). These aberrant FtsZ structures polymerized over the nucleoid (Fig. 4C).

image

Figure 4. In the absence of Noc, FtsZ forms multiple ring/arc structures on top of the nucleoid. A. FtsZ–CFP localized in mid-cell rings (seen as rings or lines across the cell) in BCBHV011 cells that produce Noc. B. FtsZ–CFP localized in multiple rings/arcs (white arrowheads) in BCBHV012 Δnoc cells. In both cases ftsZ–cfp was expressed from the spa locus of S. aureus chromosome by induction with 0.1 mM IPTG. C. Localization FtsZ–CFP in BCBHV012 Δnoc cells with labelled DNA (using Hoechst 33342) showing FtsZ–CFP ring/arc structures polymerized over the nucleoid (red arrowheads). D. Western Blot analysis of FtsZ cellular levels in noc deletion mutants and parental strains. 1: NCTC8325-4; 2: BCBHV001; 3: BCBHV011 grown without IPTG; 4: BCBHV011 grown in the presence of 0.1 mM IPTG; 5: BCBHV012 grown without IPTG; 6: BCBHV012 grown with 0.1 mM IPTG. Comparison of lanes 1 and 2, and 4 and 6 shows that deletion of noc does not result in significant alteration of FtsZ levels. E. FtsZ–CFP and EzrA–mCherry localization in BCBAJ029 cells showing that EzrA colocalizes with FtsZ during S. aureus cell division. F. Localization of EzrA–CFP at mid-cell in BCBHV015 cells that produce Noc. G. Localization of EzrA–CFP in the Noc null mutant BCBHV016 showing cells with multiple non-orthogonal FtsZ rings/arcs (arrowheads). All images are false-coloured. Scale bars, 1 µm.

Download figure to PowerPoint

It should be noted that in 119 (87%) of the BCBHV012 cells with accumulation of FtsZ in multiple sites (n = 137), the Z-ring structures formed are not perpendicular to each other (forming angles less than 80°). These results suggest an important role of Noc in defining the orthogonal planes in S. aureus, as one could argue that if a second factor would have a major role in promoting the assembly of FtsZ rings in orthogonal planes, then, in the absence of Noc, Z-rings would be assembled over the nucleoid, but they would still be mostly perpendicular to each other.

Importantly, Western blotting showed that the intracellular levels of FtsZ were not significantly altered as a consequence of Noc deletion (Fig. 4D), indicating that the formation of multiple FtsZ rings does not result from an excess of FtsZ, but most likely from a lack of the spatial confinement that is usually imposed by the presence of Noc on FtsZ polymerization. The FtsZ multiple ring/arc structures observed in Noc null mutants usually have weaker fluorescence than FtsZ ring in wild-type cells, presumably because similar amounts of FtsZ protein are distributed throughout a larger number of structures.

As the FtsZ multiple ring/arc structures were difficult to visualize by immunofluorescence (data not shown) we used EzrA as a surrogate marker for FtsZ localization, to confirm the presence of these structures in the absence of Noc, using an independent strain. EzrA is an early component of the division apparatus and a regulator of FtsZ polymerization (Levin et al., 1999), which colocalizes with FtsZ during S. aureus cell division (Fig. 4E). Localization of an EzrA–CFP fusion in noc deleted BCBHV016 strain showed the formation of structures containing two or more rings/arcs in 12% (n = 480) of the cells (Fig. 4G), similar to those observed for FtsZ (Fig. 4B).

These results constitute strong evidence of the negative regulatory role of Noc in preventing the polymerization of FtsZ in random planes, by confining the formation of the Z-ring to the only free space that does not bisect the nucleoid, the future division site.

Perturbation of DNA replication/condensation results in the assembly of non-orthogonal Z-rings

If proper chromosomal segregation is determinant for the correct assembly of FtsZ rings in orthogonal planes, in a Noc-dependent manner, then inhibition of chromosome segregation/replication should have an effect on the correct placement of the Z-rings. To test this hypothesis we incubated BCBHV013 cells, which express FtsZ–CFP, in the presence of the replication elongation inhibitor HPUra. As predicted, we were able to observe multiple non-orthogonal Z-rings in 32% (n = 613) of HPUra treated cells (Fig. 5A). Furthermore, one could expect that condensation of the nucleoid would result in the availability of extra space within the cell, which could allow for the formation of Z-rings in places other than at midcell. Addition of chloramphenicol to wild-type NCTC8325-4 S. aureus cells results in the condensation of the nucleoid (Fig. S4), an effect previously described for E. coli cells (Zimmerman, 2006). When chloramphenicol was added to BCBHV013 cells, we could observe FtsZ rings around condensed chromosomes, which do not present a bilobed morphology (Fig. 5B, cells labelled ‘2’), possibly because the distance between the condensed nucleoid and the membrane was too large for Noc to have an inhibitory effect on FtsZ polymerization. Even more surprisingly, some cells showed FtsZ rings with smaller diameters than that of the cell, which were not located at mid cell (Fig. 5B, cells labelled ‘1’). Importantly, after 30 min treatment with chloramphenicol, the growth rate of BCBHV013 culture was not altered (data not shown) and examination of BCBHV006 strain showed that Noc–YFP was still expressed and colocalized with the condensed nucleoids (Fig. S4C).

image

Figure 5. FtsZ forms abnormal, non-orthogonal structures when DNA replication/condensation is impaired. A. FtsZ–CFP localization in BCBHV013 cells that produce Noc, grown for 1h in the presence of the DNA replication inhibitor HPUra. Top panels: phase-contrast images. Bottom panels: FtsZ–CFP fluorescence. Arrowheads point to examples of cells with multiple non-orthogonal FtsZ structures. B. FtsZ–CFP localization in BCBHV013 cells that produce Noc presenting a condensed nucleoid as a result of a 30 min treatment with chloramphenicol. FtsZ forms abnormal structures in cells with condensed DNA. Cells labelled ‘1’ show FtsZ rings of small diameters which are not located at mid cell; cells labelled ‘2’ show FtsZ rings forming in the DNA free area around condensed, non-bilobed nucleoids presumably because the large distance between the nucleoid and the membrane impairs the action of Noc. DNA was visualized using Hoechst 33342. ftsZ–CFP expression was induced by the addition of 0.1 mM IPTG. Images are false-coloured. Scale bars, 1 µm.; WT: Schematic representation of a wild-type cell with Z-ring (green) at mid cell, between the chromosomes (blue). 1 and 2: Schematic representation of cells labelled ‘1’ and ‘2’ described above.

Download figure to PowerPoint

These results described above support the idea that the position occupied by the nucleoid is determinant for the positioning of the Z-ring in S. aureus.

Discussion

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

Similarly to what occurs in S. aureus, initiation of chromosome segregation in B. subtilis precedes FtsZ assembly. However, establishment of the axis of chromosome segregation does not constitute, on its own, sufficient information for definition of the division site in this rod-shaped bacteria. Accordingly, B. subtilis cells are able to form septa not only at mid-cell, but also near the poles, resulting in the formation of anucleated minicells, a phenotype prevented by the action of the Min system (Levin et al., 1992; Levin et al., 1998). In fact, the Min system seems sufficient to guarantee FtsZ assembly only at mid cell, even in the absence of nucleoid occlusion. For this reason the function of Noc in B. subtilis is apparently redundant under normal growth conditions. Only under conditions that perturb DNA replication/segregation or in the absence of the Min system was it possible to observe a strong phenotype of bisection of the nucleoid by the septum, in B. subtilis noc mutants (Wu and Errington, 2004). As discussed below, the situation is very different in S. aureus, despite the likelihood that the mechanism by which Noc inhibits FtsZ polymerization in S. aureus is similar to the mechanism in B. subtilis.

The localization of the nucleoid occlusion effector in S. aureus overlaps the nucleoid, with highest concentration at places compatible with the localization of the origins of replication. We therefore infer that, like in rod shaped bacteria, the progression of chromosome segregation in S. aureus leaves space at midcell where Noc inhibition of FtsZ polymerization is released, allowing the divisome to assemble at that position, even before replication/segregation is completed. However, contrary to rod-shaped bacteria, deletion of noc has dramatic consequences for the spherical staphylococcal cells, even under normal growth conditions, in the absence of any perturbations in DNA replication/segregation. We have observed Z-rings associated with membrane constriction assembled over the nucleoid in 15% of the BCBHV001 noc mutant cells (Fig. 2) and multiple FtsZ rings/arcs, also polymerized on top of the DNA, in 15% of the cells (Fig. 4B and C). The fact that the septum assembles over the nucleoid in BCBHV001noc mutant cells does not necessarily imply that DNA breaks will occur, as cells have means of clearing trapped DNA, mainly through the action of the DNA translocases SpoIIIE in B. subtilis (Wu and Errington, 1994) and FtsK in E. coli (Aussel et al., 2002). In fact, Bernhardt and de Boer (2005) have seen apparent movement of the nucleoid through the central septal pore in DnaA-SlmA- cells. We have now shown that lack of Noc in S. aureus does indeed lead to actual breakage of the chromosomal DNA in 16% of the noc null cells (Fig. 3), implying that Noc acts as an important antiguillotine checkpoint during normal growth of S. aureus.

Staphylococcus aureus divides in three orthogonal planes over three successive division cycles leading to the formation of cubic packs of eight cells (Koyama et al., 1977; Tzagoloff and Novick, 1977). The mechanism by which selection of orthogonal division planes occurs is an intriguing question. We have shown that in actively dividing S. aureus cells, the replicating chromosomes in each of the two daughter cells are seen elongated, with two segregated origins of replication, in cells that have a division septum but in which FtsZ assembly in the next division plane is not yet observed (Fig. 1A). Therefore, the axis for chromosome segregation in actively dividing cells is established well before Z-ring placement. This is illustrated in more detail in Fig. 6A, a schematic representation of a dividing S. aureus cell. The left panel represents a cell that has recently formed a septum at the equatorial plane. In theory, at this stage conformational entropy alone could determine the direction of chromosome segregation along an axis parallel to the equator: immediately after septum formation, each of the daughter cells is highly asymmetrical, having one short axis (perpendicular to the equator) and one long axis (parallel to the equator). It has been suggested that under strong confinement conditions, replicated circular chromosomes could partition spontaneously as a result of conformational entropy and excluded-volume interactions of the replicating nucleoid, and this segregation would always occur along the longer axis of a cell (Jun and Mulder, 2006; Jun and Wright, 2010). This is represented in the middle panel of Fig. 6A, where the nucleoids start to segregate in both daughter cells along one of the longer axis of the cell, releasing the mid cell from the Noc action which inhibits FtsZ polymerization. At this stage, only one division plane becomes available for FtsZ-ring assembly without bisection of the nucleoid (plane shown in green in Fig. 6A, right panel), therefore explaining how round cells would divide in two (but not in three) orthogonal planes. However, the system in S. aureus has to be considerably more complex, as theoretically there is an infinite number of possible axis for chromosome segregation (all longer axis in daughter cells) and therefore an infinite number of possible division planes (all the meridians in Fig. 6A). Although all these planes are perpendicular to the previous division plane, only one is chosen by the cell, a choice that cannot be explained by the sole action of entropy.

image

Figure 6. Proposed model for cell division in three orthogonal planes. A. Schematic representation of a cell that has just formed the septum (red) at the equatorial plane of division (left panel). Each daughter cell is asymmetric and has a longer and a shorter axis (black arrows). All potential division planes (meridians) perpendicular to the equatorial plane would result in bisection of the nucleoid (blue). With the progression of the cell cycle, chromosomes segregate along the longer axis of each daughter cell (middle panel). The localization of the chromosome terminus region (free of Noc inhibition of FtsZ polymerization), defines the only division plane available for Z-ring formation, which is shown in green and is perpendicular to the previous division plane in red (right panel). B. Schematic representation of a cell where the orientation of the current division plane (n, green), the previous division plane (n-1, red) and the division plane from two divisions ago (n-2, red) are represented. The chromosome origins of replication (yellow spheres) always segregate towards the intersection points (small red circles) of the last two planes of division, n-1 and n-2. This automatically defines the next division plane n, as the only plane that does not bisect the nucleoid and which is perpendicular to the two previous division planes. It is possible that an unknown origin-binding factor accumulates at the cross-junction between division planes n-1 and n-2. Notice that green and red lines are not intended to represent complete circles of material in the cell surface, but merely indicate the orientation of the division planes. As each cell divides, it generates daughter cells in which one hemisphere is constituted of new material. However, a T-junction between the two previous division planes remains present in each daughter cell.

Download figure to PowerPoint

If chromosome segregation determines the division plane, as we suggest, then in order to understand the basis for the definition of the three orthogonal division planes characteristic of S. aureus, we first need to understand what determines the choice of one particular axis for chromosome segregation. We have observed that after complete segregation of the origins of replication, they localize parallel to the previous division plane (Fig. 1A) near the points which we can infer (based on the geometry of the division planes) correspond to cross junctions of the two previous division planes. This is illustrated in Fig. 6B, which represents one cell that has previously divided along the two perpendicular planes shown in red (n-1 and n-2) and, in the next division, is going to divide along the meridian plane shown in green (labelled n). It has been previously reported that S. aureus cells have ‘scars’ of the two previous division cycles, with the major autolysin Atl, for example, localizing along two perfectly orthogonal rings around the cell (Yamada et al., 1996). If an oriC binding protein would localize at the scars from the two previous division planes, then its concentration would be highest at the two poles where these planes cross each other (red circles in Fig. 6B), where it could anchor the origins of replication. One plausible candidate for this function would be the DivIVA protein. In B. subtilis, the polar anchored DivIVA protein is involved in anchoring the origin of replication during sporulation (Ben-Yehuda et al., 2003; Wu and Errington, 2003). However, we have previously deleted DivIVA in S. aureus and, under the tested conditions, we were unable to detect any phenotype in cell division or chromosome segregation in these mutants (Pinho and Errington, 2004). Recently, S. Foster and colleagues have suggested that S. aureus uses epigenetic information, contained in a large belt of peptidoglycan with a ‘piecrust’ texture, to divide in orthogonal planes (Turner et al., 2010). It is possible that this information is used to define the axis of chromosome segregation. An alternative explanation is the existence of a yet unknown and novel mechanism responsible for defining the axis of chromosome segregation in S. aureus.

We propose that the following conditions provide sufficient information to establish division in three orthogonal planes over successive division cycles, as illustrated in Fig. 6B: (i) the origins of replication segregate towards the point where the two previous division planes (n-1 and n-2) cross each other (as we could infer), thereby establishing the axis of chromosome segregation; (ii) the division plane is placed (as we have shown) in the Noc-free region generated upon chromosome segregation, and is therefore perpendicular to the axis of chromosome segregation.

Interestingly, the proposed model does not require asymmetry in daughter cells in order for the axis of chromosome segregation to be established. Therefore, although cells emerging from stationary phase may divide in a randomly chosen plane, it is also possible that the information regarding the localization of the two previous planes is maintained in these cells, and is used to determine the first division plane.

If the model we propose is correct, then one would predict that interfering with chromosome segregation would have an effect in Z-ring placement. In fact, addition of the replication elongation inhibitor HPUra to wild-type S. aureus cells expressing FtsZ–CFP resulted in the formation of multiple, non-orthogonal Z-rings (Fig. 5A). Moreover, our model predicts that condensation of the nucleoid would result in free space in the cell, devoid of Noc-mediated inhibition of FtsZ polymerization. In theory, this space could be used for incorrect assembly of Z-rings, which is what we have observed in cells treated with chloramphenicol that have condensed nucleoids (Fig. 5B).

If nucleoid occlusion is so important in determining division planes in S. aureus cells, why are not Z-rings formed in all cells? Several lines of evidence point to the likely existence of additional nucleoid occlusion effectors in B. subtilis and E. coli, besides Noc and SlmA. The original nucleoid occlusion model proposed by Woldringh suggested that active transcription/translation around the nucleoid exerts an inhibiting effect on cell division (Mulder and Woldringh, 1989; Woldringh et al., 1991). In the first report of B. subtilis Noc, Wu and Errington suggest the existence of a Noc-independent pathway for nucleoid occlusion based on the bias of septum formation toward internucleoid regions in mutants lacking both Noc and the Min system (Wu and Errington, 2004). Similar observations were made in E. coli mutants lacking SlmA and the Min system (Bernhardt and de Boer, 2005). More recently, Harry and colleagues proposed an additional mechanism for Z-ring positioning in which the mid cell becomes increasingly ‘potentiated’ for the formation of a Z-ring upon completion of initiation of replication (Moriya et al., 2010), while Rudner and co-workers showed evidence for Noc-independent nucleoid occlusion in the prevention of inappropriate cell division during replication fork arrest (Bernard et al., 2010). Therefore, it is not unlikely that S. aureus also uses other system(s) to ensure nucleoid occlusion, particularly because nucleoid occlusion has such a relevant role in cell division in this bacterium.

The fact that lack of a conserved protein, Noc, results in such different phenotypes in cocci and rods, highlights the importance of studying cell division mechanisms in different model organisms, particularly those with different shapes and modes of division from the more traditional model organisms E. coli and B. subtilis.

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 and growth conditions

All strains used in this study are listed in Table 1. Sequences of primers used are listed in Table S1. S. aureus strains were grown in tryptic soy agar (TSA, Difco) at 37°C or in tryptic soy broth (TSB, Difco) at 37°C with aeration. The medium was supplemented, when necessary, with erythromycin (10 µg ml−1, Sigma), chloramphenicol (10 µg ml−1, Sigma), kanamycin and neomycin (50 µg ml−1 each, Sigma), 5-bromo-4-chloro-3-indolyl β-d-galactopyranoside (X-Gal, 100 µg ml−1, BDH Prolabo) or isopropyl β-d-thiogalactopyranoside (IPTG, at various concentrations as indicated below, BDH Prolabo).

Table 1.  Plasmids and bacterial strains.
Plasmids and strainsRelevant characteristicsSource or reference
  1. lacImc– cells expressing multicopies of the lacI gene (encoded by pMGPII).

Plasmids  
 pMADE. coliS. aureus shuttle vector with a thermosensitive origin of replication for Gram-positive bacteria; Ampr, EryrArnaud et al. (2004)
 pBCB13pMAD derivative with up- and downsteam regions of spa and Pspac-lacI;Ampr, EryrPereira et al. (2010)
 pMUTINCFPB. subtilis integrative vector for C-termini CFP fusions; Ampr, EryrKaltwasser et al. (2002)
 pMUTINYFPB. subtilis integrative vector for C-termini YFP fusions; Ampr, EryrKaltwasser et al. (2002)
 pMUTINCFPKanB. subtilis integrative vector for C-termini CFP fusions; Ampr, KanrThis study
 pMUTINYFPKanB. subtilis integrative vector for C-termini YFP fusions; Ampr, KanrAtilano et al. (2010)
 pMUTINCFPΔermBpMUTINCFP with ermB gene deletedThis study
 pDG792Plasmid containing a kanamycin resistance geneGuerout-Fleury et al. (1995)
 pMGPIIPlasmid encoding lacI gene; CmrPinho et al. (2001)
 pBCB4-ChES. aureus integrative vector for N- and C-termini mCherry fusions; Ampr, EryrAtilano et al. (2010)
 pBCBHV001pMAD containing up- and downsteam regions of noc; Ampr, EryrThis study
 pBCBHV002pBCB13 containing Pspac-noc-lacI; Ampr, EryrThis study
 pBCBHV003pBCB13 containing Pspac-ftsZ–CFP-lacI; Ampr, EryrThis study
 pBCBHV004pMUTINYFP containing spo0Jyfp; Ampr, EryrThis study
 pBCBHV005pMUTINYFPKan containing noc–yfp; Ampr, KanrThis study
 pBCBHV006pMUTINCFPKan containing ftsZ–cfp; Ampr, KanrThis study
 pBCBAJ001pMAD containing the 3′ end and the downstream region of ezrA; Ampr, EryrThis study
 pBCBAJ002pBCBAJ001 containing ezrA–mCherry; Ampr, EryrThis study
E. coli  
 DH5αrecA endA1 gyrA96 thi-1 hsdR17 supE44 relA1_80_lacZ_M15Lab strain
S. aureus  
 RN4220MSSA strain. Mutagenized strain derivated from NCTC8325-4 that accepts foreign DNAR. Novick
 NCTC8325-4MSSA strainR. Novick
 COLMRSA strainGill et al. (2005)
 BCBHV001NCTC8325-4 ΔnocThis study
 BCBHV002NCTC8325-4 lacImc; CmrThis study
 BCBHV003NCTC8325-4 ΔNoc lacImc; CmrThis study
 BCBHV004RN4220 ftsZ::pBCBHV006, expressing ftsZcfp fusion; KanrThis study
 BCBHV005RN4220 ftsZ::pBCBHV006 spo0J:: pBCBHV004, expressing ftsZcfp and spo0Jyfp fusions; Kanr, EryrThis study
 BCBHV006NCTC8325-4 noc::pBCBHV005, expressing nocyfp fusion; KanrThis study
 BCBHV007NCTC8325-4 noc::pBCBHV005 lacImc, expressing nocyfp fusion; Kanr, CmrThis study
 BCBHV008NCTC8325-4 Δspa::Pspac-lacI lacImc; CmrThis study
 BCBHV009NCTC8325-4 ΔnocΔspa::Pspac-lacI lacImc; CmrThis study
 BCBHV010NCTC8325-4 ΔnocΔspa::Pspa-noc-lacI lacImc; CmrThis study
 BCBHV011NCTC8325-4 Δspa::Pspac-ftsZ–cfp-lacI lacImc; CmrThis study
 BCBHV012NCTC8325-4 ΔnocΔspa::Pspac-ftsZ–cfp-lacI lacImc; CmrThis study
 BCBHV013NCTC8325-4 Δspa::Pspac-ftsZ–cfp-lacIThis study
 BCBHV014NCTC8325-4 Δspa::Pspac-ftsZ–cfp-lacI lacImcnoc:: pBCBHV005, expressing ftsZcfp and nocyfp fusions; Kanr; CmrThis study
 RNpEzrACFPRN4220 ezrA::pEzrA–CFP, expressing ezrAcfp fusion; EryrPereira et al. (2010)
 BCBHV015NCTC8325-4 ezrA::pEzrA–CFP, expressing ezrAcfp fusion; EryrThis study
 BCBHV016NCTC8325-4 Δnoc ezrA::pEzrA–CFP, expressing ezrAcfp fusion; EryrThis study
 BCBAJ025NCTC8325-4 ΔezrA::ezrA–mCherryThis study
 BCBAJ029NCTC8325-4 ΔezrA::ezrA–mCherryΔspa::Pspac-ftsZ–cfp-lacIThis study

Construction of the Noc null mutant

The Noc null mutant was constructed in the background of S. aureus NCTC8325-4 strain using the thermosensitive plasmid pMAD (Arnaud et al., 2004). PCR fragments containing the upstream and downstream flanking regions of the noc gene were amplified from chromosomal DNA of NCTC8325-4, using primer pairs NocP1/NocP2 and NocP3 /NocP4. These two PCR products were joined in a second PCR reaction, using primers NocP1/NocP4. The resulting fragment was restricted with BamHI and EcoRI and cloned into pMAD, giving rise to pBCBHV001. The insert was sequenced and the plasmid was electroporated as previously described (Veiga and Pinho, 2009) into the transformable strain RN4220 at 30°C (using erythromycin selection) and subsequently transduced to NCTC8325-4 using phage 80α (Oshida and Tomasz, 1992). The deletion of noc was completed after a two-step homologous recombination process. In the first step, recombinants in which pBCBHV001 was integrated into the chromosome were selected at the non-permissive temperature of 43°C. In the second step, recombinants in which the integrated plasmid (and consequently the lacZ and erm genes) had been excised were selected at the permissive temperature (30°C) in the absence of antibiotic selection. The absence of noc gene was confirmed by PCR screening of the obtained erythromycin sensitive, white colonies, using the primers NocP5/NocP6 and the strain was named BCBHV001.

Complementation of Noc mutation

To complement the noc null mutant strain BCBHV001, a full copy of noc gene was placed in the spa locus of S. aureus chromosome, under the control of the IPTG-inducible Pspac promoter, using pBCB13 plasmid (Pereira et al., 2010). For that purpose, an 865 bp DNA sequence, containing the ribosomal binding site (RBS) and the noc gene, was amplified by PCR using the primers NocP7 and NocP8. After restriction with XmaI and XhoI, the insert was cloned into pBCB13, downstream of the Pspac promoter, and sequenced. The resulting pBCBHV002 plasmid was electroporated into RN4220 at 30°C (using erythromycin selection) and subsequently transduced to a Noc mutant strain containing an extra, plasmid-encoded, copy of the lacI gene (BCBHV003). To exchange the spa gene for the Pspac-noc DNA fragment and obtain strain BCBHV010, the transductants were incubated at the non-permissive temperature of 43°C, in the presence of erythromycin (to select for plasmid integration) and then incubated at the permissive temperature of 30°C, in the absence of antibiotic selection, to select for white colonies in which the vector had been excised. As a control, the pBCB13 vector was also transduced from RN4220 into BCBHV002 (NCTC8325-4 strain containing an extra, plasmid-encoded, copy of the lacI gene) and BCBHV003 strains, which went through the two-steps recombination process, resulting in strains BCBHV008 and BCBHV009 in which the spa gene was deleted.

Construction of fluorescent derivatives of S. aureus proteins

For the construction of YFP fusions to the C-terminal of Spo0J and Noc we used, respectively, the pMUTINYFP vector (Kaltwasser et al., 2002) and its derivative pMUTINYFPKan in which the erythromycin-resistance marker was replaced by a kanamycin-resistance marker (Atilano et al., 2010).

DNA fragments of spo0J and noc genes without the stop codon and connected to a 5-codon linker (Table S1, highlighted in bold) were amplified by PCR using, respectively, primer pairs Spo0JP1 /Spo0JP2 and NocP9/NocP10.

The spo0J insert was restricted with HindIII and EagI and cloned into pMUTINYFP, generating pBCBHV004. The insert was sequenced and the plasmid was electroporated into RN4220 (selection with 10 µg ml−1 erythromycin and 0.1 mM IPTG) and transduced to BCBHV004 strain (see below) using phage 80α (selection with 50 µg ml−1 kanamycin, 50 µg ml−1 Neomycin, 10 µg ml−1 erythromycin and 0.1 mM IPTG) to obtain BCBHV005 strain. In this strain, the expression of spo0J and ftsZ fusions from their native loci, is regulated, respectively, by spo0J and ftsZ native promoters while an extra un-fused copy of each gene is expressed under control of Pspac by the addition of 0.1 mM IPTG.

The noc insert was restricted with KpnI, cloned into pMUTINYFPKan and sequenced. The resulting vector, pBCBHV005, was electroporated into RN4220 (selected with 50 µg ml−1 of kanamycin, 50 µg ml−1 of Neomycin and 0.1 mM IPTG) and transduced into NCTC8325-4. The integration of pBCBHV005 in the noc locus originates S. aureus strain BCBHV006, in which the noc–yfp fusion is controlled by its native promoter and an extra copy of noc (controlled by Pspac) is expressed by addition of 0.1 mM IPTG to the culture medium.

To fully repress the expression of the Noc un-tagged protein in BCBHV006 strain and test the functionality of Noc–YFP, we introduced plasmid pMGPII (encoding lacI gene) into strain BCBHV006 to originate strain BCBHV007 that was grown in the absence of IPTG.

For co-visualization of FtsZ–CFP and Noc–YFP, pBCBHV005 was also transduced into the background of strain BCBHV011 to originate BCBHV014. This strain, expresses at the same time, an FtsZ–CFP fusion from the ectopic spa locus and a Noc–YFP protein from noc native locus.

To construct a S. aureus strain expressing an FtsZ–CFP C-terminal fusion from the ftsZ native locus, a PCR fragment encompassing the ftsZ gene (without the stop codon) and a 5-codon linker (Table S1, highlighted in bold) was amplified using primers FtsZP1 and FtsZP2, digested with KpnI, cloned into pMUTINCFPKan plasmid (see below) and sequenced. The resulting plasmid, pBCBHV006, was electroporated into RN4220 competent cells (selected with 50 µg ml−1 kanamycin, 50 µg ml−1 Neomycin and 0.1 mM IPTG) giving rise to BCBHV004. This strain expresses both an ftsZ–cfp fusion from the ftsZ native promoter and ftsZ from the Pspac promoter by addition of 0.1 mM IPTG to the culture medium.

The pMUTINCFPKan vector is a derivative of pMUTINCFP (Kaltwasser et al., 2002) in which the erythromycin-resistance marker was replaced by a kanamycin-resistance marker from pDG792 vector (Guerout-Fleury et al., 1995). For this replacement, primers KanP1 and KanP2 were used to amplify the entire pMUTINCFP, except the ermB cassette, which was then auto-ligated to generate pMUTINCFPΔermB. The kanamycin marker was excised from pDG792 by restriction with NcoI and BglII and cloned into pMUTINCFPΔermB, generating the pMUTINCFPKan plasmid.

To construct a strain in which the FtsZ–CFP fusion is ectopically expressed, under the control of the Pspac promoter, from the spa locus of S. aureus chromosome, the pBCB13 vector was used. For that purpose, a ftsZ–cfp DNA fragment was amplified from pBCBHV006 plasmid, using primers FtsZP3 and FtsZP4. The fragment was restricted with XmaI and XhoI, cloned into previously digested pBCB13, downstream of the Pspac promoter, and sequenced. The resulting pBCBHV003 plasmid was electroporated into RN4220 at 30°C (selected with erythromycin) and subsequently transduced to NCTC8325-4, BCBAJ025 (see below; selection at 30°C with erythromycin) and BCBHV002, BCBHV003 (selection at 30°C with erythromycin and chloramphenicol). The replacement of the spa gene for the Pspac-ftsZ–cfp DNA fragment was completed after a two-step homologous recombination process as described above and confirmed by PCR. The resulting strains expressing ftsZ–cfp from the spa locus were named BCBHV013, BCBAJ029, BCBHV011and BCBHV012 respectively.

To construct S. aureus strains expressing an EzrA–CFP fusion protein, the genomic region containing ezrA-cfp was transduced from RNpEzrA–CFP (Pereira et al., 2010) into NCTC8325-4 and BCBHV001 (using erythromycin selection). In the resulting strains, named, respectively, BCBHV015 and BCBHV016, EzrA–CFP fusion is expressed under the control of ezrA native promoter from its native chromosomal locus.

To allow the co-visualization of EzrA and FtsZ in the same cells, a functional ezrA–mCherry fusion, was expressed from its native locus. For that purpose, two PCR fragments containing the truncated 3′ end of ezrA gene (1099 bp) without its stop codon and the 1107 bp sequence downstream of ezrA gene, were amplified from S. aureus COL genome using primer pairs EzrAP4/EzrAP5 and EzrAP1/EzrAP2 respectively. The two fragments were joined by overlap PCR using primers EzrAP4 and EzrAP2, which resulted in the introduction of a 5-codon linker at the 3′ end of ezrA gene. The resulting fragment was digested and cloned into the EcoRI and BamHI restriction sites of pMAD vector, generating plasmid pBCBAJ001. The mCherry coding sequence (711 bp) was amplified from pBCB4-ChE (Pereira et al., 2010) using primers mCherryP3 and mCherryP4. The PCR product was cloned into pBCBAJ001 using enzymes NheI and XhoI, generating plasmid pBCBAJ002. This plasmid was introduced into RN4220 cells by electroporation and transduced to NCTC8325-4 using phage 80α. Integration of pBCBAJ002 into the ezrA locus was selected for by growing cells at 43°C with erythromycin. After a second recombination event leading to plasmid excision, cells in which ezrA was replaced by an ezrA–mCherry fusion, were identified by PCR and the resulting strain was named BCBAJ025.

TUNEL assay for the detection of DNA breaks

Staphylococcus aureus cells were prepared for TUNEL labelling of DNA breaks using a fixation and permeabilization protocol adapted from a previously described method for immunofluorescence (Pinho and Errington, 2003). NCTC8325-4 and BCBHV001 cultures were grown to mid-exponential phase (OD600 0.5), fixed with Histochoice (Amresco) washed three times with PBS and resuspended in GTE buffer (50 mM glucose, 20 mM Tris-HCl pH = 7.5, 10 mM EDTA). The fixed cells were then gently lysed with 2.5–5 µg ml−1 lysostaphin for 40 s on a poly-l-lysine-treated slide, washed with PBS, air dried and rehydrated with PBS (Phosphate Buffer Saline). An aliquot of fixed and permeabilized NCTC8325-4 cells was treated at room temperature, for 20 min, with 100 U ml−1 DNaseI, to obtain cells with DNA breaks, used as positive control.

The DNA breaks were detected by catalytic incorporation of dUTPs fluorescein conjugates at the 3′-hydroxyl ends of the fragmented DNA using an In Situ Cell Death Detection Kit (Roche). Fixed cells were incubated for 1h at room temperature in the dark, with a TUNEL reaction mixture containing the terminal deoxynucleotidyl transferase (TdT) diluted 1/1000 in the dUTPs mixture labelling solution. After washing 8 times with PBS, Vectashield mounting medium (Vector Laboratories) was added and the cells were visualized by fluorescence microscopy.

Fluorescence microscopy

Strains were grown overnight, in TSB at 37°C, with appropriate antibiotic selection. Cells were then diluted 1/1000 into fresh TSB, supplemented when necessary with 0.1 mM IPTG. When required, cells were stained with the cell wall dye Van-FL (1 µg ml−1), the DNA dye Hoechst 33342 (1 µg ml−1) and the membrane dyes Nile Red (5 µg ml−1) or FM5-65 (1µg ml−1).

To inhibit DNA replication, BCBHV013 cells were grown with 0.1 mM of IPTG until OD600nm 0.3. At that time, HPUra (40 µg ml−1) was added to the culture, which was then incubated for 1 h, prior to FtsZ–CFP visualization.

Chloramphenicol was used to condense the DNA. For that purpose, NCTC8325-4, BCBHV006 and BCBHV013 cells were grown until OD = 0.4 (BCBHV013 cells were grown in the presence of 0.1 mM IPTG). At that point, chloramphenicol (2µg ml−1) was added to the culture, which was then incubated for 30 min. BCBHV013 cells were stained with DNA Hoechst 33342. NCTC8325-4 and BCBHV006 cells were stained with Hoechst 33342 and, respectively, with the membrane dyes Nile Red and FM5-65.

Cells were mounted on a thin film of 1% agarose in PBS and observed by fluorescence microscopy using a Zeiss Axio observer.Z1 microscope. Image acquisition was performed using a Photometrics CoolSNAP HQ2 camera (Roper Scientific) and Metamorph 7.5 software (Molecular Devices).

The measurement of the angles established between the multiple FtsZ–CFP ring/arc structures in BCBHV012 strain was performed using the ImageJ angle tool. ImageJ software was also used to quantify the cell diameter of approximately one thousand NCTC8325-4 and BCBHV001 cells.

Acknowledgements

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

We thank Elisabeth Harry for providing the FtsZ antibody, Simon Foster for the pDG792 plasmid and George Wright for HPUra. We are grateful to Suckjoon Jun, Adriano Henriques, Sérgio Filipe, James Yates and Pedro Matos for critical reading of the manuscript and helpful discussions. We also thank S. Filipe and J. Yates for making Figs 6 and S1 respectively. This work was supported by grant POCI/BIA-BCM/66449/2006 from Fundação para a Ciência e Tecnologia, Portugal, awarded to M.G. Pinho. H. Veiga was supported by FCT fellowship SFRH/BD/38732/2007 and A. Jorge was supported by FCT fellowship SFRH/BD/28480/2006.

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
FilenameFormatSizeDescription
MMI_7651_sm_FigureS1-S4_TableS1.pdf320KSupporting 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.