The bacterial actin homologue MreB plays a key role in cell morphogenesis. In Bacillus subtilis MreB is essential under normal growth conditions and mreB mutants are defective in the control of cell diameter. However, the precise role of MreB is still unclear. Analysis of the lethal phenotypic consequences of mreB disruption revealed an unusual bulging phenotype that precedes cell death. A similar phenotype was seen in wild-type cells at very low Mg2+ concentrations. We found that inactivation of the major bi-functional penicillin-binding protein (PBP) PBP1 of B. subtilis restored the viability of an mreB null mutant as well as preventing bulging in both mutant and wild-type backgrounds. Bulging was associated with delocalization of PBP1. We show that the normal pattern of localization of PBP1 is dependent on MreB and that the proteins can physically interact using in vivo pull-down and bacterial two-hybrid approaches. Interactions between MreB and several other PBPs were also detected. Our results suggest that MreB filaments associate directly with the peptidoglycan biosynthetic machinery in B. subtilis as part of the mechanism that brings about controlled cell elongation.
Bacteria have a wide range of characteristic shapes, and these shapes are largely determined by the cell wall. The cell wall is also responsible for maintenance of the structural integrity of the cell by bearing its internal turgor pressure. The major load-bearing component of the cell wall is peptidoglycan (PG), which is composed of long glycan strands cross-linked by peptide cross bridges (Höltje, 1998; Vollmer et al., 2008). The precursors of PG are synthesized in the cytoplasm and transferred to the outside of the membrane by a recently discovered lipid II flippase (Inoue et al., 2008; Ruiz, 2008). They are incorporated into the existing PG meshwork, the murein sacculus, by a family of proteins known as penicillin-binding proteins (PBPs). The PBPs can be subdivided into three classes: class A high-molecular-weight bi-functional PBPs, which posse both transglycosylase and transpeptidase activities; class B high-molecular-weight PBPs, which have only transpeptidase activity; and a class of low-molecular-weight PBPs with carboxypeptidase or endopeptidase activities (Goffin and Ghuysen, 1998; Höltje, 1998; Vollmer and Bertsche, 2008). Expansion of the PG layer requires controlled cleavage of the existing structure, by a range of murein hydrolase activities, without compromising wall integrity.
Elimination of PBP1 suppresses the lethality of an mreB null mutation
It has been reported that the growth of mreB null mutants is restored in the presence of high concentrations of Mg2+ (Formstone and Errington, 2005). To gain insights into the molecular basis for the growth impairment and morphological defect caused by inactivation of MreB, we isolated extragenic transposon (TnYLB-1) insertions that restored viability of an mreB mutant plated under non-permissive conditions (without added Mg2+ or inducer). Strain YK400 (ΔmreB amyE::Pxyl-gfp-mreB) has a functional copy of mreB fused to gfp and controlled by the xylose inducible promoter, Pxyl. Strain YK400 was transformed with the transposon plasmid and transformants were plated on penicillin assay broth (PAB, Difco Antibiotic Medium 3) agar in the absence of Mg2+ or inducer. After screening a library that generated about 30 000 colonies in the presence of xylose, we identified nine mutants capable of growth in the absence of Mg2+ or inducer and which had stable suppressor mutations linked to a transposon insertion. Mapping and sequencing of five of the mutants showed that the transposon had inserted into the ponA gene. ponA encodes PBP1, which is a high-molecular-weight bi-functional PBP proposed to catalyse both the transglycosylation and transpeptidation of PG precursors (Murray et al., 1998; Pedersen et al., 1999). The other transposon insertions were in ptsI (three hits), and ccpA, and they are not discussed further here.
To confirm the ability of ponA disruption to rescue viability of the mreB mutant, a null mutation of ponA (ponA::spc) was transformed into the mreB mutant. As shown in Fig. 1A, an in-frame deletion mutant of mreB (3725, ΔmreB) grew on solid PAB medium supplemented with 10 mM MgSO4 (c), but not on unsupplemented medium (d). In contrast, the double mutant YK401 (ΔmreB ponA::spc) grew on solid PAB medium without added MgSO4 (Fig. 1A d). Growth of mreB mutant cells in liquid PAB after removal of extra Mg2+ resulted in cell lysis (Fig. 1B, open diamonds). However, the mreB ponA double mutant cells grew in liquid medium in the absence of Mg2+ (open circles), although more slowly than wild-type cells, confirming that lethality of the mreB mutation was largely suppressed by disruption of ponA.
Mutual suppression of mutations affecting MreB and the bifunctional PBPs
To test whether the effects described above were specific for ponA we examined the effects of mutations affecting the two other major vegetatively expressed genes encoding class A PBPs in B. subtilis, pbpD (PBP4) and pbpF (PBP2c) (Popham and Setlow, 1993; 1994). Various mutant combinations were constructed and first tested for viability on PAB plates supplemented with or without 10 mM Mg2+(Table 1 and Fig. 1A). A ponA pbpD double mutant, and a ponA, pbpD and pbpF triple mutant both showed a strong growth defect in the absence of Mg2+, and the triple mutant barely grew even in the presence of Mg2+ (Table 1 and Fig. 1A a and b). Under our conditions, the growth defect was more severe than previously reported (McPherson and Popham, 2003) (see Fig. S1 and Discussion). The severe growth defect was consistent with the bi-functional PBPs having partially redundant roles in the essential transglycosylation reaction of PG synthesis. Interestingly, suppression of the lethal mreB phenotype in the absence of Mg2+ was not observed for pbpD or pbpF mutations, nor by simultaneous disruption of pbpD and pbpF (Table 1); it occurred only when these mutations were combined with ponA (Table 1, Fig. 1A d and Fig. S1) (disruption of pbpA or pbpH, which encode class B PBPs, also did not suppress the lethality of mreB mutant; data not shown). These results suggested that PBP1 activity is involved in lethality of mreB mutations at low Mg2+.
Table 1. Viability of mreB mutant cells in the absence of bi-functional PBPs.
Viability on PAB plate
10 mM Mg2+
10 mM Mg2+
Strains were grown on PAB plates supplemented with or without 10 mM Mg2+ at 37°C.
Surprisingly, these experiments also revealed a strong mutual suppression effect, in that mutation of mreB strongly enhanced the growth of the ponA mutants and the derivatives in which other class A PBPs were knocked out. Note that the triple PBP mutant hardly grew at all on a Mg2+-supplemented plate when MreB was present, whereas the equivalent MreB- strain grew well even in the absence of Mg2+ (Fig. 1A b and d and Fig. S1). Therefore, the deleterious consequences of loss of the three major class A PBPs can be largely overcome by deletion of mreB.
Bulging and lysis of mreB mutants depends on PBP1 function
Previous experiments showed that mreB mutant cells undergo bulging and eventually lyse under normal Mg2+ conditions (Formstone and Errington, 2005). In addition, we recently reported that PBP1 activity is required for bulging of gpsB ezrA double mutants (Claessen et al., 2008). To test whether this was also the case for bulging of mreB mutants, various strains were examined by phase contrast microscopy of cultures grown in PAB medium supplemented with sucrose, which prevents lysis of the mreB mutant (Formstone and Errington, 2005). The mreB mutant cells maintained their rod shape reasonably well in the presence of 10 mM Mg2+ (Fig. 2B), whereas the cells became swollen after the removal of Mg2+, eventually undergoing a massive degree of bulging and lysis (Fig. 2C and D), as shown previously (Formstone and Errington, 2005). Ectopically expressed GFP-MreB (strain YK400, ΔmreB amyE::Pxyl-gfp-mreB) complemented the effects of the chromosomal mreB deletion preventing bulging and lysis (data not shown), confirming that the bulging phenotype is caused by inactivation of MreB. In contrast, in the absence of both PBP1 and MreB, no bulges were detected after the removal of Mg2+ (Fig. 2G and H), although many cells had an abnormal coiled morphology (Fig. 2H). As noted previously, cells of the ponA single mutant were slightly thinner and longer than wild-type cells (Murray et al., 1998; Claessen et al., 2008; compare Fig. 2A and E).
The morphology of mreB mutant cells was further examined in a synthetic minimal medium with various concentrations of Mg2+. In high-Mg2+ media (10 mM, Fig. 2M; or 5 mM, data not shown), the mreB mutant cells maintained their rod shape. However, at 1 mM the cells frequently showed polar bulges (Fig. 2N). Such bulges were not observed in cells of the wild-type strain in 1 mM (Fig. 2J) or 0.5 mM (see below) Mg2+ medium, but, interestingly, at lower concentrations of Mg2+ (0.1 mM, Fig. 2K; or 0.05 mM, Fig. 3C) polar bulges were observed, and the cells were not viable in 0.01 mM Mg2+ medium (data not shown). In contrast, no bulges were detected in ponA-disrupted cells cultivated under even lower-Mg2+ conditions (Fig. 2L and P). Thus, PBP1 is required for bulge formation in wild-type cells at low Mg2+ conditions, as well as in mreB-disrupted cells.
Polar bulging is associated with abnormal localization of PBP1
It has been shown that PBP1 dynamically relocates from the septum to the lateral wall during the cell cycle and suggested that incorrect localization of PBP1 was responsible for the bulging phenotype of ezrA gpsB mutants (Claessen et al., 2008). Therefore, we suspected that bulge formation in mreB mutant cells or in the wild-type cells in low-Mg2+ medium might be due to delocalization of PBP1. Cells expressing GFP–PBP1 were cultivated in minimal medium with various concentrations of Mg2+. In mreB+ cells grown in medium supplemented with 10 mM MgSO4, GFP–PBP1 (YK706, amyE::Pxyl-gfp-ponA) was detected at the septum as a band, or in dots along the cylindrical parts of the cell as described previously (Claessen et al., 2008; Fig. 3A). At 60 min after reduction to 0.05 mM Mg2+, GFP–PBP1 remained predominantly in the cylindrical parts of the cell (Fig. 3B), although abnormal polar localizations were also detected (arrows). Further cultivation of mreB+ cells at this lower level of Mg2+ resulted in polar bulges and the discrete localizations of GFP–PBP1 along the lateral cell wall had virtually disappeared, being replaced by prominent fluorescence around the bulging poles (Fig. 3C). In mreB mutant cells, typical localization patterns of GFP–PBP1 were also observed in the presence of 10 mM Mg2+ (Fig. 3D). However, 60 min after reduction of the Mg2+ concentration, the cells already showed severe bulging and GFP–PBP1 fluorescence was predominantly at division sites and the bulging poles (Fig. 3E). By 90 min the cells were starting to lyse and GFP–PBP1 had largely condensed into single foci (Fig. 3F). Use of the fluorescent penicillin analogue Bocillin FL (Zhao et al., 1999) suggested that PBP1 is properly folded and inserted into the membrane in mreB mutant cells under low-Mg2+ conditions (Fig. S2). We conclude that PBP1-dependent polar bulging is associated with abnormal localization of PBP1 at the cell poles.
Lateral wall localization of PBP1 depends on MreB
Recent finding that simultaneous disruption of gpsB and ezrA results in abnormal polar localization of GFP–PBP1 suggests that GpsB and EzrA have some role in the localization dynamics of PBP1 during the cell cycle (Claessen et al., 2008). However, in that work, the localization of GFP–PBP1 in a gpsB and ezrA double mutant was only tested in medium with high Mg2+ (Claessen et al., 2008). Therefore, we examined the localization of PBP1 in gpsB and ezrA double mutant cells cultivated with lower concentrations of Mg2+. Cells of the mutant cultivated in minimal medium showed bulging poles by 90 min after reduction of the Mg2+ concentration (from 10 to 0.5 mM) (Fig. 3H and arrows), and accumulation of GFP–PBP1 fluorescence at the bulging poles was observed as shown previously (Claessen et al., 2008; Fig. 3H b). However, the discrete localizations of GFP–PBP1 along the lateral cell wall were still present in the cells (Fig. 3H). In contrast, in the mreB mutant cells, GFP–PBP1 fluorescence in the lateral wall was almost undetectable under the same conditions (Fig. 3I), suggesting that MreB is important for recruitment of PBP1 to the cylindrical parts of the cell.
PBP1 localization is independent of the MreB paralogues, Mbl and MreBH
Bacillus subtilis has three MreB isoforms, MreB, Mbl and MreBH. They have been demonstrated to colocalize in a single helical structure and to have important roles for cell shape determination (see Introduction). Therefore, we thought that Mbl and MreBH might also have a role in the regulation of PBP1 localization. To test this possibility, we examined cell morphology and localization of PBP1 in mbl and mreBH mutants in minimal media. In the presence of 0.5 mM Mg2+, mreB mutant cells showed polar bulges (Fig. 4B), but such bulges were not observed in the mbl and mreBH mutant cells (Fig. 4C and D). Under these conditions, discrete localizations of GFP–PBP1 on the lateral cell wall were clearly visible in mbl or mreBH mutants (Fig. 4E and F), but not in the mreB mutant cells (Fig. 3I). However, further reduction of the available Mg2+ (below 0.1 mM) resulted in a bulging phenotype and loss of lateral localizations of GFP–PBP1 in mbl or mreBH mutant cells (data not shown). Therefore, Mbl or MreH might be required for normal cell morphology and localization of PBP1 at very low Mg2+ concentrations (< 0.1 mM), while MreB is required at higher Mg2+ concentrations (< 1 mM: Fig. 2N).
PBP1 associates with MreB in a complex
The above results showed that disruption of mreB under low Mg2+ causes delocalization of PBP1, bulging and finally cell death. Previous work suggested that PBP1 associated with the MreB helix of the lateral wall indirectly via an interaction with MreC (Claessen et al., 2008). To test for the existence of complexes containing both MreB and PBP1, we made an mreB-histidine tag construct, under the control of the xylose-inducible promoter, which was integrated into the amyE locus of ΔmreB strain (3738, ΔmreB amyE::Pxyl-mreB-his10). The growth rate and cell shape of 3738 were indistinguishable from that of the wild type in the presence of inducer, indicating that the fusion is functional (data not shown). Using this background, the only copy of ponA was modified such that it was under the control of the isopropyl β-D-thiogalactoside (IPTG)-inducible Pspac promoter (YK781, ΔmreB amyE::Pxyl-mreB-his10 Pspac-ponA). Cells expressing the MreB-His fusion were treated with the fluorescent penicillin analogue Bocillin FL (Zhao et al., 1999), and formaldehyde to cross-link protein, then disrupted by sonication. MreB complexes were purified using Ni resin under denaturing conditions, then heated to disrupt cross-linked proteins and separated by SDS-PAGE. As a negative control, cells expressing His-tagged DnaC [a component of the replisome that interacts with DnaI (Imai et al., 2000; Bruand et al., 2001; Ishikawa et al., 2006)] were analysed in parallel. As reported previously, DnaI was detected as a component of DnaC complex by liquid chromatography-tandem mass spectrometry analysis (Fig. 5A, lane 2). However, no DnaI was detected in equivalent samples from the MreB-His complexes (data not shown). In contrast, TufA, which has been detected as non-specific background in FtsA and DnaC complexes purified by similar methods (Ishikawa et al., 2006), was readily detectable in both complexes (Fig. 5A). Therefore, DnaI, at least, is specific to the DnaC complex. As an another control, the presence of MreC and SpoIIIE, which is DNA translocase and has N-terminal transmembrane segments that are required for proper septal localization (Wu and Errington, 1997; Sharp and Pogliano, 1999; 2002), in MreB complexes was analysed by Western blotting (Fig. 5B). MreC was readily detected in both the whole-cell extracts and in purified MreB complexes (Fig. 5B a), showing that MreB and MreC are physically associated in a complex. In contrast, SpoIIIE was not detected, or only present in trace amounts (Fig. 5B b), indicating that SpoIIIE is not closely associated with MreB.
The scan for Bocillin fluorescence in the whole-cell extracts of B. subtilis strains revealed proteins corresponding to the molecular masses of PBP1, PBP2a, PBP2b, PBP2H, PBP2c, PBP4 and PBP5 (Fig. 5C, arrows). The most highly labelled and highest-molecular-weight band was not detected in cells expressing MreB-His cultivated in the absence of IPTG (i.e. with synthesis of PBP1 repressed) (Fig. 5C, lane 1), showing that this band corresponds to PBP1. When an MreB complex was purified in the presence of IPTG (i.e. in the presence of PBP1) using YK781, PBP1 was readily detected (Fig. 5C, lane 6). In contrast, PBP1 was not detected in the DnaC complexes (Fig. 5C, lane 7). Thus, we concluded that PBP1 is specific to the MreB complexes. In addition, several extra bands were also detected in the MreB complexes (Fig. 5C, arrow heads). Three of these were detected in complexes isolated from cultures with IPTG (Fig. 5C, open arrow heads), but not without IPTG (Fig. 5C, lane 5), suggesting that they are either degradation products of PBP1, or as yet unidentified PBPs that only associate with MreB when PBP1 is present. Three other fluorescently labelled bands corresponding approximately to the molecular masses of PBP2A, PBP2c and PBP4 were also detected in the MreB complexes, independent of the presence or absence of PBP1 (Fig. 5C, solid arrow heads) [note that the mobility of the bands in the right part (purified protein complexes) of B are slightly different from the left part (whole-cell extract), probably because of the greatly reduced protein concentrations]. A band possibly corresponding to PBP4 was also detected in the DnaC complex, although the signal was much weaker than that in the MreB complex (Fig. 5C, solid arrow head and asterisk).
The PBP1 has a molecular weight of about 99 kDa comprising a 37-amino-acid cytoplasmic N-terminus, a single 23-amino-acid transmembrane region, and an 854-amino-acid extracellular catalytic domain. Strong interactions between PBP1 with MreC and EzrA were detected previously, requiring only the transmembrane domain of PBP1 (van den Ent et al., 2006; Claessen et al., 2008). Interaction between PBP1 and GpsB was also detected previously, requiring both of the cytoplasmic tail and the transmembrane domain of PBP1 (Claessen et al., 2008; Fig. S3). As a test of whether the putative interaction between MreB and PBP1 was direct, we performed a bacterial two-hybrid experiment in E. coli host cells (which should not contain proteins sufficiently homologous or abundant to bridge the B. subtilis proteins under test). As shown in Fig. 6, MreB showed self-interaction, indicating that the fusions were functional, and a reproducible strong interaction was also detected between MreB and PBP1, although the interaction was only detectable when full-length PBP1 was coexpressed with MreB (Fig. S3). In addition, we tested whether MreB also interacted directly with other PBPs. MreB gave a positive two-hybrid signal for PBPs belonging to both class A and class B high-molecular-weight PBPs (PBP2a, PBP2b, PBP2c, PBP2d, PBP3, PBP4, PbpH and PbpI) (Fig. 6), providing support for the idea that MreB and several PBPs interact directly in the complexes detected in the pull-down experiment.
Mutual suppression of MreB and transglycosylase lesions
Bacillus subtilis has four bi-functional PBPs, PBP1, PBP2c, PBP2d and PBP4. It was previously reported that triple or even quadruple mutants, although compromised for growth, are still viable (Popham and Setlow, 1996; McPherson and Popham, 2003). We found that elimination of the three vegetative proteins resulted in an almost complete growth arrest, even in the presence of high Mg2+. We noticed that this mutant was difficult to maintain, and readily picks up suppressor mutations that improve its growth. Most likely, suppression or differences in strain background account for the differences between our results and those of McPherson and Popham (2003). Surprisingly, our data also identify mreB as a source of suppression. Indeed, disruption of mreB substantially restored the growth of the triple mutant. Therefore, mutations in mreB and ponA have a mutually suppressive action. The deregulated localization of PBP1 can now explain the suppression of mreB mutants by ponA. However, the reverse effect – restoration of growth of a multiple transglycosylase mutant by inactivation of mreB– is more difficult to explain. One possibility would be that mreB controls the export or activity of an autolytic enzyme that is deleterious when cells have reduced PG synthetic (transglycosylase) activity.
MreB-dependent localization of PBP1 into the cylindrical part of the cell
MreB proteins form helical cables just under the cytoplasmic membrane (Jones et al., 2001; Kruse et al., 2003; Shih et al., 2003; Defeu Soufo and Graumann, 2004; Figge et al., 2004; Formstone and Errington, 2005; Carballido-López et al., 2006). Recent studies using a GFP–PBP1 fusion have demonstrated a dynamic redistribution of the protein from division sites to the cell cylinder and into a helix-like configuration (Claessen et al., 2008). In addition, it has been shown that septal localization of PBP1 is dependent on the cell division machinery (Scheffers and Errington, 2004). However, GFP–PBP1 was still observed all over the cylindrical parts of the cell after depletion of the key division protein FtsZ (data not shown), suggesting that cell division proteins are not required for the lateral distribution of PBP1. We have now found that lateral wall-associated fluorescence is virtually absent in mreB mutant cells, but present at the septum, as well as at cell poles. A bacterial two-hybrid experiment has identified GpsB and EzrA, which are thought to have partially differentiated roles in the localization cycle of PBP1, as interacting partners for PBP1 (Claessen et al., 2008). Simultaneous disruption of gpsB and ezrA has a synthetic phenotype with defects in both cell division and elongation, and an unusual bulging of the cell poles, apparently due to aberrant polar localization of PBP1 (Claessen et al., 2008). The phenotypic similarities between mreB single mutants and gpsB ezrA double mutants suggest that they have overlapping functions in regulation of the PBP1 localization cycle. Nevertheless, discrete localizations of GFP–PBP1 along the lateral cell wall were still present in gpsB and ezrA double mutant cells. Furthermore, the lateral wall localization of GFP–PBP1 was also observed in mbl and mreBH mutant cells, at least under certain experimental conditions, but not in mreB mutant cells. The dependence of PBP1 localization to the lateral wall by MreB probably reflects a direct interaction, as judged by our in vivo pull-down and bacterial two-hybrid experiments. Therefore, MreB appears to be the key player in recruitment of PBP1 to the lateral wall. GpsB, EzrA and MreC may provide secondary or alternative sites of association. The previous results of Claessen et al. (2008) suggest that the contribution of GpsB may be especially important during the transition from elongation to division, or during high salt stress.
The Mg2+ effect on PBP1 localization
We showed in this study that high concentrations of Mg2+ could compensate for delocalization of PBP1 in mreB mutant. Although the detailed mechanism remains unclear, it has been suggested that the effect of Mg2+ might be mediated by changing the structure of the cell wall (Formstone and Errington, 2005; Leaver and Errington, 2005). On the other hand, examination of strains with combinations of mutations showed that simultaneous disruption of mreB and mbl, or mreB and mreBH, is lethal, even with the presence of high concentrations of Mg2+, and the double mutant of mbl and mreBH is only viable at high Mg2+ (Defeu Soufo and Graumann, 2006; A. Formstone and J. Errington, unpublished). These results suggest that the three MreB paralogues have partially overlapping functions. Perhaps Mbl and/or MreBH support the correct localization of PBP1 in the absence of MreB, under high Mg2+ conditions. Unfortunately, we could not test for interactions between PBP1 and Mbl or MreBH, because the two-hybrid constructs appear not to be functional. In addition, the dependence of growth on Mg2+ in the multiple mutants of mreB paralogues requires the presence of at least one MreB paralogue. Mg2+ might affect the activity or expression levels of MreB proteins. Indeed, wild-type cells also showed a bulging phenotype under low-Mg2+ conditions, similar to mreB mutant cells.
A complex web of interactions among factors in the cell elongation machinery
Use of fluorescent derivatives of vancomycin and ramoplanin has revealed helical incorporation of new PG in the lateral cell wall (Daniel and Errington, 2003; Tiyanont et al., 2006). In addition, previous reports have shown that MreB provides positional information for PG-synthesizing enzymes in the lateral walls of other rod-shaped bacteria (Figge et al., 2004; Divakaruni et al., 2007; Mohammadi et al., 2007). Based on these observations, it is currently believed that MreB is involved in spatial organization of the PG synthetic machinery along the cylindrical cell wall. MreB is also required for proper localization of other membrane-associated cell morphogenesis proteins, such as MreC and MreD (Divakaruni et al., 2005; Dye et al., 2005; Kruse et al., 2005; Leaver and Errington, 2005; van den Ent et al., 2006). MreC of B. subtilis is required for growth of the cylindrical cell wall and is tightly associated with MreB proteins (Leaver and Errington, 2005). The detection of direct interactions with the high-molecular-weight PBPs suggested that MreC could provide the bridge between the intracellular cytoskeleton and the extracellular cell wall synthetic machinery in B. subtilis (van den Ent et al., 2006). Indeed, a direct interaction between MreC and MreB has been detected by bacterial two-hybrid experiments (data not shown), and such an interaction has been also shown in E. coli (Kruse et al., 2005). On the other hand, our new results suggest that several high-molecular-weight PBPs, including PBP1, PBP2a, PBP2c and PBP4, at least, associate directly with MreB. PBP1, PBP2a, PBP2c and PBP4 are predicted to have cytoplasmic N-terminus tails of 37, 20, 10 and 6 amino acids, respectively, which might participate in direct interactions with MreB. However, at the present time, we cannot exclude the possibility that MreC and/or MreD act as secondary or alternative bridging molecules for interactions between MreB and the PBPs. Further work is needed to confirm these proposed interactions biochemically, using purified components.
To conclude, the major challenge ahead will be to understand how the complex web of interactions between the proteins in the cell wall synthetic machinery is organized in such a way as to generate controlled cell elongation under different environmental conditions.
Bacterial strains, plasmids
Bacillus subtilis strains used in this study are listed in Table 2 and plasmids are listed in Table S1.
Bacillus strains were grown at 37°C in LB medium, Difco antibiotic medium 3 (PAB) or minimum [modified Spizizen's (Spizizen, 1958)] medium supplemented, where required, with xylose, IPTG and/or MgSO4. When necessary to ameliorate the lytic phenotype of mreB mutants, 0.3 M sucrose was added to PAB or minimal medium (Figs 2–4). Bacillus transformants were selected on PAB agar supplemented with 50 μg ml−1 spectinomycin, 5 μg ml−1 chloramphenicol, 5 μg ml−1 kanamycin, 0.2 μg ml−1 phleomycin, 1 μg ml−1 erythromycin and/or 25 μg ml−1 lincomycin. 0.25% xylose or 10 mM MgSO4 was also supplemented as necessary. DNA manipulation and E. coli transformation were performed by standard methods (Sambrook et al., 1989).
Depletion of Mg2+
To remove Mg2+, cells were grown in PAB or minimal medium supplemented with 10 mM MgSO4 to mid-exponential phase and then collected by centrifugation. The cells were then washed once in PAB without added MgSO4, or in minimal medium containing the usual low levels of MgSO4, and then diluted back to an OD600 of 0.02–0.05 into fresh low-Mg2+ media.
Transposon mutagenesis was performed essentially as described previously (Claessen et al., 2008). To screen for mutants that restored viability of a strain YK400 (ΔmreB amyE::Pxyl-gfp-mreB) in the absence of inducer (xylose), cells were transformed with the transposon plasmid pMarB and transformants were selected on PAB agar containing 0.5% xylose at 37°C (Le Breton et al., 2006). Several transformants were picked and individually grown for 8 h in PAB medium supplemented with 0.5% xylose at 37°C. The cells from each culture were then plated and incubated overnight at 50°C on plates containing kanamycin and 0.5% xylose, erythromycin and 0.5% xylose. We then selected the plate that gave the highest ratio of kanamycin-resistant colonies versus erythromycin-resistant colonies. The selected culture was used to generate a library of about 30 000 colonies, which was plated on PAB plates containing kanamycin at 50°C. Mutants that restored the viability in YK400 strain in the absence of inducer were streaked on PAB plates. Genomic DNA of the mutants was isolated and backcrossed into strain 4281 (ΔmreB cat) to confirm that the suppression was not due to a second site mutation. Mutants that had stable suppressor mutations linked to a transposon insertion were subjected to inverse PCR amplification and sequencing of the transposon insertion site as described previously (Le Breton et al., 2006).
Construction of a GFP–PBP1 fusion pSG1729 plasmid (Lewis and Marston, 1999) allows a translational fusion of gfp under the control of Pxyl and have been used to insert into amyE locus on B. subtilis chromosome. The coding sequence of ponA gene was amplified by PCR from the wild-type strain 168 genomic DNA using the primers pSG-ponAF (which contains a XhoI site in bold, 5′-CTCCTCGAGATGTCAGATCAATTTAACAG-3′) and pSG-ponAR (which contains a EcoRI site, 5′-GAAGAATTCTTAATTTGTTTTTTCAATGG-3′), then cloned between the XhoI and EcoRI sites of plasmid pSG1729, creating pSG1729-ponA. The resulting plasmid was used to transform 168, with selection for spectinomycin resistance, to generate YK706.
Construction of IPTG-inducible ponA mutants
The first 300 bp of the ponA gene containing Shine–Dalgarno sequence was amplified by PCR from the wild-type strain 168 genomic DNA using the primers pM4-ponAF (which contains a HindIII site, 5′-CAAGAAGCTTTTTTCGCCATCATCTGGTGC-3′) and pM4-ponAR (which contains a BamHI site 5′-GGAGGATCCAGACGTAGGTCCGTTTTTCG-3′), then cloned between the HindIII and BamHI sites of plasmid pMutin4 (Vagner et al., 1998), creating pMutin4-ponA. The resulting plasmid was used to transform 168, with selection for erythromycin resistance, to generate YK684 in which the full-length ponA gene is expressed from the IPTG-inducible promoter Pspac.
Purification of protein complexes and Bocillin FL labelling
An overnight liquid culture of B. subtilis cells expressing histidine-tagged protein in LB medium containing 0.25% xylose, 0.5 μg ml−1 erythromycin and 0.5 mM IPTG, as necessary, at 30°C was inoculated into 400 ml of the same medium. When the cells, grown at 37°C, reached an OD600 of 0.5–0.6, the culture was treated with 0.05 μg ml−1 Bocillin FL (Invitrogen) for 3 min, and then formaldehyde (1% final concentration) for 10 min. The culture was incubated for an additional 10 min after addition of 150 mM glycine. Purification of protein complex with His-tagged protein was performed as described (Ishikawa et al., 2006). To identify proteins in the complex, eluates were separated by SDS-PAGE (NuPAGE system with 4–12% gradient Bis-Tris Midi Gel and MOPS SDS Running Buffer, Invitrogen) after heating (60 min at 90°C). Separated proteins were visualized by Colloidal Coomassie staining of NuPAGE Bis-Tris Gels (Invitrogen). Prior to staining, Bocillin FL-labelled PBPs were visualized with a Typhoon Imager (Amersham Biosciences); the images were acquired and analysed with ImageQuant TL (Amersham Biosciences). Separated proteins were also analysed by Western blotting as described (Ishikawa et al., 2006). MreC and SpoIIIE proteins were detected with anti-MreC (F. van den Ent and M. Leaver, unpublished) and anti-SpoIIIE antisera (Wu and Errington, 1994) (both at 1/1000 dilution). In-gel enzymatic digestion, peptide analysis and protein identification by liquid chromatography-tandem mass spectrometry were carried out as described by Kuwana et al. (2002).
Bacterial two-hybrid analysis
To detect interactions of MreB with various PBPs, the coding sequence of the mreB gene was amplified by PCR from the wild-type 168 genomic DNA using the primers B2H-mreBF (which contains a SalI site, 5′-GTCGTCGACTATGTTTGGAATTGGTGCT-3′) and B2H-mreBR (which contains a EcoRI site, 5′-GAAGAATTCTTATCTAGTTTTCCCTTTG-3′), then cloned between the SalI and EcoRI sites of plasmid pKT25 and pUT18C. PBP fusions were derived from plasmid pUT18C (van den Ent et al., 2006).
Bacterial two-hybrid analysis was developed by Karimova et al. (1998), and adapted as described by Daniel et al. (2006). A 10 μl aliquot from each transformation reaction was spotted onto minimal medium plates (Daniel et al., 2006) containing 100 μg ml−1 ampicillin, 50 μg ml−1 kanamycin and 0.008% Xgal. Pictures were taken after 48 h of growth at 30°C. Under these conditions, control transformations with the empty vectors gave colonies that remained white for up to 72 h of incubation.
For fluorescence microscopy, cells from overnight culture or from a culture grown to mid-exponential stage from a fresh overnight plate were diluted into fresh minimal [Spizizen's (Spizizen (1958)] medium with 10 mM or 50 μM MgSO4, or casein hydrolysate medium (Sterlini and Mandelstam, 1969) containing 0.5% xylose and grown to mid-exponential phase at 37°C. For live cell imaging, cells were mounted on microscope slides covered with a thin film of 1.2% agarose in water, essentially as described previously (Glaser et al., 1997). Images were acquired with a Sony Cool-Snap HQ cooled CCD camera (Roper Scientific) attached to a Zeiss Axiovert 200 M microscope. The images were acquired and analysed with metamorph version 6 software.
We thank D. Claessen for helpful discussions and members of the Errington, Daniel and Hamoen lab for advice. We thank K. Schirner for the gift of strains 4261, 4262 and 4281, A. Formstone for the gift of strain 3738, R. Emmins for the gift of plasmids for bacterial two-hybrid experiments, M. Leaver for the gift of MreC antibody and L.J. Wu for the gift of SpoIIIE antibody. We thank S. Ishikawa and M. Kuwano for technical assistance. We thank K. Gerdes for critical reading of the manuscript. This work was supported by a grant from the UK Biotechnology and Biological Sciences Research Council.