Role of penicillin-binding protein PBP 2B in assembly and functioning of the division machinery of Bacillus subtilis


Jeff Errington. E-mail; Tel. (+44) 1865 275561; Fax (+44) 1865 275556.


We have characterized the role of the penicillin-binding protein PBP 2B in cell division of Bacillus subtilis. We have shown that depletion of the protein results in an arrest in division, but that this arrest is slow, probably because the protein is relatively stable. PBP 2B-depleted filaments contained, at about their mid-points, structures resembling partially formed septa, into which most, if not all, of the division proteins had assembled. Although clearly deficient in wall material, membrane invagination seemed to continue, indicating that membrane and wall ingrowth can be uncoupled. At other potential division sites along the filaments, no visible ingrowths were observed, although FtsZ rings assembled at regular intervals. Thus, PBP 2B is apparently required for both the initiation of division and continued septal ingrowth. Immunofluorescence microscopy showed that the protein is recruited to the division site. The pattern of localization suggested that this recruitment occurs continually during septal ingrowth. During sporulation, PBP 2B was present transiently in the asymmetrical septum of sporulating cells, and its availability may play a role in the regulation of sporulation septation.


Cell division in bacteria involves the ordered assembly of about 10 fairly conserved proteins at the division site. Together, these proteins direct annular ingrowth or constriction of the cell envelope, including cell membrane and cell wall layers (Lutkenhaus and Addinall, 1997; Lutkenhaus, 1999). The most conserved and thoroughly studied component of the division machinery is a tubulin-like protein, FtsZ. This is a highly abundant protein that probably polymerizes to form a ring-like structure around the circumference of the cell, lying just under the cytoplasmic membrane (Bi and Lutkenhaus, 1991). During ingrowth of the septum, the ring constricts, being maintained close to the leading edge of the septum (Addinall et al., 1996; Levin and Losick, 1996). Another conserved soluble protein, FtsA, appears to assemble next, probably by interacting directly with FtsZ (Ma et al., 1997; Wang et al., 1997). The other proteins involved in septation all have at least one membrane span. In Escherichia coli, the sequence of assembly of the proteins after FtsZ can be summarized as follows (Ghigo et al., 1999): ZipA (which assembles independently of FtsA: Hale and De Boer, 1999; Liu et al., 1999); FtsQ (Chen et al., 1999); FtsL (Ghigo et al., 1999); FtsI (Weiss et al., 1999); FtsN (Addinall et al., 1997). The order of assembly of FtsK and FtsW proteins is not yet clear. In Bacillus subtilis, the hierarchy is less extensively characterized. This organism has no equivalent of the ZipA or FtsN proteins, but it has a probable FtsQ homologue, DivIB (Harry and Wake, 1997), and a second FtsL-like protein, DivIC (Levin and Losick, 1994; Katis et al., 1997). FtsZ again assembles early (Levin and Losick, 1996), but later steps in the pathway appear to be more concerted than in E. coli. Thus, assembly of both DivIB (FtsQ) and DivIC is dependent on FtsL (Daniel et al., 1998). The order of assembly of other B. subtilis division proteins remains to be elucidated.

PBP 2B is one of the few division proteins for which a biochemical function has been deduced. It belongs to a family of high-molecular-weight, penicillin-binding proteins, which catalyse the final stages of peptidoglycan synthesis (Ghuysen, 1991). PBP 2B appears to be specialized for the formation of septal peptidoglycan because, in E. coli at least, pbpB mutations prevent septation but not continued synthesis of the cylindrical part of the cell wall (Spratt, 1977; Pogliano et al., 1997). We have shown previously that the pbpB gene of B. subtilis, which lies in the equivalent chromosomal position to that of E. coli ftsI, is essential and that a small C-terminal deletion results in filamentous growth (Yanouri et al., 1993; Daniel et al., 1996). Here, we provide conclusive evidence that PBP 2B protein is required for cell division in B. subtilis. We have found that the protein is quite stable but, when it is depleted, most cells arrest with a partially formed septum-like structure containing most, if not all, of the division proteins. FtsZ rings still form at intermediate potential division sites but, unlike E. coli, at least two late-assembling division proteins, DivIB and DivIC, fail to assemble at sites other than mid-filament upon depletion of PBP 2B. Finally, we show that PBP 2B targets to division sites in both vegetative and sporulating cells and that this targeting requires all of the characterized division gene products.


Depletion of PBP 2B blocks cell division

To clarify the function of PBP 2B protein, we made use of strain 804, in which pbpB expression is regulated by the IPTG-inducible Pspac promoter. This strain exhibits a strong dependence on IPTG when grown on solid media. The strain was grown in liquid PAB medium containing IPTG, then washed with fresh medium lacking IPTG and divided into two portions, only one of which was supplemented with IPTG. In parallel, a culture of wild-type strain SG38 was studied. The optical density (OD) of each culture was followed, and samples of cells were examined using phase-contrast light microscopy. No obvious phenotypic effect was seen for the first 40–60 min after the removal of the inducer. However, from 60 min onwards, an impairment of cell division became evident (Fig. 1A). Cell length measurements confirmed the increase in cell length compared with the control (+IPTG) culture (Fig. 1B). Inhibition of cell division continued until stationary phase, when the cells began to lyse, as measured by both culture OD and microscopy (not shown). Cell length remained more or less constant in the IPTG-supplemented culture of strain 804, whereas it fell gradually in the wild-type culture as the cells approached stationary phase (see below). This experiment confirmed our expectation that PBP 2B is required for cell division during vegetative growth of B. subtilis.

Figure 1.

. Depletion of PBP 2B causes a block to cell division resulting in cell filamentation. A. Phase-contrast images of representative fields of cells of strain 804 taken 0, 30, 60 and 90 min after removal of IPTG. Scale bar = 5 μm. B. The mean cell length of strain 804 after removal of IPTG (triangles) compared with a parallel culture grown in the presence of IPTG (squares) and a wild-type culture (circles). C. Immunodetection of PBP 2B from total protein extracts of strain 804 grown in the presence of IPTG and a wild-type strain 168. For each strain, a protein sample was taken at early exponential phase (1), mid-exponential phase (2) and early stationary growth phase (3). Protein quantities were equalized on the basis of the OD of the culture from which they were taken. D. Depletion of PBP 2B in strain 804 after the removal of IPTG. Lanes are labelled according to the time (min) after the removal of IPTG from the culture (−IPTG). + IPTG represents results with a parallel culture, from which IPTG was not removed. Lanes were loaded with equivalent amounts of protein on the basis of the culture OD.

We found previously that repression of the ftsL division gene (which lies adjacent to pbpB) gives rise to a much more rapid block in cell division (Daniel et al., 1998). The relatively lengthy delay between pbpB repression and the arrest of division suggested either that PBP 2B is quite a stable protein, or that it is required only in small quantities, and that almost enough protein is made from the Pspac promoter even when repressed. To examine the effect of repression on PBP 2B protein levels, antibodies were raised against affinity-tagged, full-length PBP 2B. In control Western blot experiments, the serum reacted strongly (at a dilution of 1:10 000) with a protein of the mobility expected for PBP 2B, and cross-reacting bands were only detected in highly overexposed images (not shown). The band detected was shifted to a faster migrating position in extracts of strain 168::pL92, in which PBP 2B is slightly truncated, although still functional (Yanouri et al., 1993; not shown). A strong band of PBP 2B was detected in extracts of vegetative cells of B. subtilis, and the levels of protein increased about twofold on entry into stationary phase (Fig. 1C), perhaps related to the relative increase in cell division that occurs in this phase of growth. Interestingly, PBP 2B was present at slightly lower levels in strain 804 (grown in the presence of IPTG) than in the wild type (strain 168) (Fig. 1C), probably explaining the slight difference in cell length between the wild type and 804 (+ IPTG) in 1Fig. 1B. Immunoblots of samples of strain 804 taken after the removal of IPTG showed that the PBP 2B protein was depleted slowly, approximately in parallel with cell growth (Fig. 1D; the doubling time of the culture was ≈ 20 min in this experiment). In other experiments (not shown), equal volumes of culture were analysed after removal of IPTG, and then almost no change in PBP 2B amount was detected. These results suggest that the inhibition of cell division involves dilution of pre-existing PBP 2B protein and that the protein is quite stable.

PBP 2B-depleted cells show an unusual septal morphology

When samples from a PBP 2B-depleted culture taken soon (60 min) after the removal of IPTG were viewed by electron microscopy, it appeared that septation became arrested at a late stage. Septal structures were clearly deficient, with most elongated segments of cells bearing no visible signs of constriction. Although filaments lying completely in the plane of the section were rare, of 12 such filaments seen, one had no visible invagination, and the other 11 all had only a single invagination, which was positioned at about the mid-point of the filament (Fig. 2A). Interestingly, the septal ingrowths that were seen had an unusual ultrastructure, particularly in being thinner than normal, and the insertion of wall material was clearly deficient. Several such structures from cells taken at two different time points are shown at higher magnification in 2Fig. 2B and C, and examples of the typical progression of septation in the wild type are shown for comparison in 2Fig. 2D. In the PBP 2B-depleted cells, an inward growth of at least some cell wall material was associated with all the structures observed. In some of the earlier samples (B1–3), the space between the wall ingrowths appeared to be typically cytoplasmic, with granular and fibrous material, presumably representing ribosomes and nucleic acids respectively. In others, thin electron-dense lines, probably membrane, appeared to span the two wall ingrowths. In the later samples (Fig. 2C), this membrane-like material was more prevalent and often formed whorls extending out laterally into the cytoplasm (e.g. C3), or vesicles (C4). The latter structures seemed to be particularly prevalent in sections above or below the medial axis of the cell, such as the one shown. If the electron-dense lines do represent membrane, as seems likely, it appears that membrane invagination can run ahead of wall growth in this mutant. Depletion of PBP 2B thus appears to uncouple membrane invagination from wall synthesis during septation.

Figure 2.

. Electron micrographs illustrating typical cell morphologies after the depletion of PBP 2B. A. A typical cell filament of strain 804 grown in the absence of IPTG for 90 min. Scale bar = 1 μm. B and C. Enlarged images of aborted septa contained within filaments taken 90 min (B) and 120 min (C) after removal of IPTG. D. A compilation of images of septa at different stages of completion in the wild type.

Effect of PBP 2B depletion on the formation of the division complex

If the unusual structures formed at mid-cell in PBP 2B-depleted filaments corresponded to slowly forming or stalled septa, they might comprise a fully assembled division complex. To investigate this possibility, immunofluorescence microscopy (IFM) and specific antisera were used to examine the distribution of various division proteins. The conditions used for growth were slightly different in these experiments (cells were grown in Luria–Bertani (LB) medium; see Experimental procedures) to facilitate the IFM. The first protein expected to assemble at the division site was the tubulin-like protein FtsZ. As shown in 3Fig. 3A–C), FtsZ assembled into bands positioned at regular intervals along most of the filaments. Presumably, these bands corresponded to both partially constricted septa at mid-cell (arrow in Fig. 3B) and potential sites for subsequent divisions. Quantification of the number of localizations of FtsZ per unit of cell length showed that there was little or no change in the frequency of FtsZ band formation in the presence or absence of IPTG (Table 1). Thus, the formation of FtsZ bands appears to be independent of PBP 2B.

Figure 3.

. Localization of FtsZ (A–C) DivIB (D and E) and DivIC (F and G) 90 min after depletion of PBP 2B by the removal of IPTG (strain 804). A–G. Phase-contrast images. The corresponding FITC-immunofluorescence images are labelled A′ to G′. Arrows point to the positions of septa (when visible) at about mid-filament in these cells. Scale bar = 5 μm.

Table 1. . Effects of depletion of PBP 2B on localization of other cell division proteins. a. Time zero represents the time at which a portion of each culture (grown with IPTG present) was washed, divided into two portions and resuspended in medium with (+) or without (−) IPTG.b. For the +IPTG samples, these values are normalized to the time 0 sample, which is set at 100%. For the −IPTG samples, the values are normalized to the +IPTG sample of the same time point. Number of cells counted in parentheses.Thumbnail image of

When cells from the same culture were examined for other division proteins, quite different results were obtained. [Note that slightly different conditions were used for optimal detection of FtsZ compared with the other proteins (see Experimental procedures), and this alters the appearance of the cells by phase-contrast microscopy. However, control experiments showed that this did not affect the relative number of fluorescent bands detected.] As shown in 3Fig. 3D–G, PBP 2B-depleted filaments tended to show only single localizations of DivIB or DivIC proteins, approximately at the mid-point of each filament, where the ‘stalled’ septal structures were observed by electron microscopy. Frequently, these positions corresponded to places where invaginations of the filaments were visible in the phase-contrast images (arrows in Fig. 3D–F). In both cases, quantification of the localizations of DivIB and DivIC showed a substantial reduction in frequency per unit of cell length (Table 1), showing that these ‘late-assembling’ division proteins require PBP 2B for localization. On the other hand, the general detection of mid-cell bands for these proteins and FtsZ (Fig. 3) supports the notion that the structures detected by electron microscopy represent partially formed, possibly stalled septal structures.

Analysis of total cellular protein in the PBP 2B-depleted cells showed that the amounts of the three division proteins examined were similar to those of cells in which septation was not blocked (not shown). Thus, the failure to complete septation or to form bands at all potential division sites was not caused by an indirect effect on protein synthesis or stability.

Immunolocalization of PBP 2B at the division site and its dependence on other division proteins

IFM was used to determine the localization of PBP 2B in vegetatively growing cells of the wild-type strain 168 (Fig. 4A and Table 2[link]). We were not able to eliminate a weak peripheral signal that was seen in all cells, even when PBP 2B was depleted (not shown). We assume that this was caused by cross-reaction with one or more membrane-associated proteins, candidates being other PBPs, particularly PBP 1, which was also shown recently to be targeted to the division site (Pedersen et al., 1999). However, by far the strongest signal obtained, which we assume represents specific staining of PBP 2B, was in the form of fluorescent spots and bands, which were detected at about the mid-points of most of the cells (e.g. the cell labelled a in Fig. 4A). In some cells with a PBP 2B band, cell constriction did not appear to have started (as judged by scrutiny of the equivalent phase-contrast image) but, in the majority of such cells, a distinct constriction was visible, suggesting that division was complete, or nearly so (e.g. the cell labelled b in Fig. 4A). As septation progressed, the form of PBP 2B staining changed to that of a spot (e.g. the cell labelled c; Table 2) as reported previously for DivIB and DivIC proteins (Harry and Wake, 1997; Katis et al., 1997). Finally, a small spot of PBP 2B protein was frequently visible at mature cell poles after the completion of division. In summary, these observations showed that PBP 2B is at least partly targeted to the division site and that recruitment to these sites is a relatively late event in the cell cycle.

Table 2. . Localization of PBP 2B during cell cycle progression.Thumbnail image of
Figure 4.

. Localization of PBP 2B in conditional cell division mutants, grown in the non-permissive state for 40 min. A. Wild type (strain 168). B. ftsZ (strain ts1). C. divIB (strain MB76). D. divIC (strain SU349). E. ftsL (804; grown in the presence of IPTG but in the absence of xylose). For descriptions of the features defined by the labels a, b and c, see text. Scale bar = 5 μm.

To determine the dependence of PBP 2B localization on the other cell division proteins, similar localization experiments were performed with strains bearing conditional alleles of various division genes (Fig. 4B–E). Three of the mutant strains used were temperature sensitive (bearing mutations in the ftsZ, divIB and divIC genes). Control experiments (not shown) demonstrated that temperatures of up to 48°C had no discernible effect on the localization of PBP 2B in the wild type; also, that localization was normal in the mutant strains at the permissive temperatures (30°C). However, as shown in 4Fig. 4B–D, PBP 2B bands were not detected in filamentous cells of any of these mutants grown at the non-permissive temperature (48°C) (80–100 filaments counted in each case). To test for localization in the absence of FtsL protein, a repressible allele of ftsL was used (strain 804). Again, no PBP 2B localizations were detected in the cell filaments (Fig. 4E). Thus, targeting of PBP 2B to the sites of cell division requires the function of all the cell division proteins we have tested.

To test whether PBP 2B was stable in the absence of the other division proteins, samples from cultures prepared as in Fig. 4 were examined by Western blotting. The levels of PBP 2B detected were similar irrespective of the growth temperature or the protein function eliminated (not shown).

Role of PBP 2B in the formation of the asymmetric septum during sporulation

In contrast to previous experiments with division genes, such as ftsZ (Beall and Lutkenhaus, 1991), divIC (Levin and Losick, 1994) and ftsL (Daniel et al., 1998), depletion of PBP 2B had little effect on sporulation when IPTG was removed from the culture at the time of induction of sporulation. This was presumably because of the stability of the protein. Removal of the inducer at earlier time points, before sporulation was induced, resulted in a reduction in sporulation, as judged by the expression of alkaline phosphatase (a marker for stage II of sporulation; Errington and Illing, 1992; data not shown). Unfortunately, this prolonged period of time needed to deplete the protein made it difficult to demonstrate conclusively a specific requirement for PBP 2B in the asymmetric cell division of sporulation.

As an alternative approach to this problem, immunolocalization was used to determine whether PBP 2B was targeted to the asymmetric septum. Cells of wild-type strain 168 were stained for PBP 2B at various time points during sporulation. Progression through the early stages of sporulation can be followed conveniently by DNA staining (Setlow et al., 1991; Hauser and Errington, 1995). First, the DNA becomes spread out to form an ‘axial filament’ (Stage I). The septum then forms (stage II), closing around one end of the extended DNA. Soon afterwards, the remainder of the chromosome is packed into the small prespore compartment. Asymmetrically positioned bands of PBP 2B staining were indeed detected in wild-type cells during the early stages of sporulation. They began to be detected at about the time when the sporulation septum would be expected to form (60–65 min; Partridge and Errington, 1993) (Fig. 5A). In general, these bands were fainter than the central bands of vegetative cells (e.g. the cell marked v). They were often visible in cells at stage I of sporulation (cells marked i) but, by stage II (e.g. cell marked ii), they had usually disappeared (quantified in Table 3). This indicates that the protein is present at the asymmetric division site only transiently, probably explaining why they were relatively difficult to detect. Interestingly, we did not detect any cells with PBP 2B bands at both poles of the cell, in contrast to the common bipolar pattern of FtsZ staining (Levin and Losick, 1996).

Figure 5.

. PBP 2B localization during sporulation. Samples were taken 90 min after the induction of sporulation. Left. DNA distribution in the cells. Centre. Overlays of phase-contrast and FITC-immunofluorescence images. Right. FITC-immunofluorescence only. A. Wild type (strain 168). B. spo0H17(strain 17.4). C. spoIIB (strain 131.5). D. spoIIG49 (strain 49.4). Letters in the centre identify cells at various characteristic stages in sporulation: v, vegetative; I, stage I; ii, stage II. Scale bar = 2.5 μm.

Table 3. . Localization of PBP 2B during sporulation. a. As judged by DNA morphology from DAPI fluorescence images. 0, no PBP 2B localization; C, central band; A, asymmetrically positioned band; PP, band at prespore pole; OP, band at pole opposite prespore; BP, bands at both prespore poles.b. Two cells had a central PBP 2B band and two condensed prespores.Thumbnail image of

To clarify the regulation of PBP 2B assembly at the asymmetric septum, we examined protein localization in several sporulation mutants. spo0H mutants form FtsZ bands at one or both cell poles, but they do not go on to make polar septa (Levin and Losick, 1996). In contrast to this result, we only detected PBP 2B at central positions (Table 3), usually where there was an obvious division constriction (Fig. 5B). The failure of PBP 2B to be recruited to the polar FtsZ rings could explain why these mutant cells fail to make polar septa.

It was possible that the later stages of prespore development, during which it becomes engulfed within the cytoplasm of the mother cell, might be somehow hastening the removal of PBP 2B from the sporulation septum. We therefore examined mutants in which this development is blocked. In cells of a spoIIB mutant, asymmetrically positioned bands of PBP 2B were brighter and more readily detected than in the wild type (Fig. 5C), although again, they were prominent only in cells at stage I (cells marked i) and absent from cells at stage II (e.g. the cell marked ii) (Table 3). Analogous results were obtained with a spoIIG mutant, in which septa form sequentially at the two polar positions (Lewis et al., 1994) (Fig. 5D). Thus, polar PBP 2B bands were readily detected in stage I cells (marked i) and, by the time that both polar septa were formed, staining had generally disappeared (Table 3). More interestingly, in cells with one completed prespore (as judged by their completely segregated prespore chromosome, e.g. the cell labelled ii), which should be about to form a second septum at the distal pole, a band of PBP 2B was frequently detected at that pole (Table 3). In summary, these results showed that PBP 2B is recruited to the prespore septum, but it seems that the protein is only present (or at least visible) transiently at about the time that the septum is being formed. PBP 2B does not appear to be recruited to FtsZ rings that do not go on to form a septum, such as those at the prespore-distal pole of wild-type cells or at either of the poles of spo0H mutant cells.


Targeting of PBP 2B to division sites and its dependence on various division genes

Antibodies raised against His-tagged PBP 2B were used to examine the subcellular localization of PBP 2B in B. subtilis. In vegetatively growing cells, the protein was found to target in the form of bands at mid-cell, and it was also faintly visible at mature cell poles. A weak signal probably representing protein not targeted to these sites was present around the cell periphery, consistent with the expected transmembrane topology of the protein. Determination of the precise time at which PBP 2B is targeted was difficult because of the relatively weak fluorescence signal obtained. However, the strongest signals for the protein tended to be detected in cells in which septation appeared to be complete or nearly so (with a visible constriction in the phase-contrast image; Fig. 4A). This contrast with results obtained for FtsZ, in which the signal tends to reduce as division proceeds and the septal annulus closes (Levin and Losick, 1996). However, it is similar to the patterns described previously for the membrane-associated DivIB (Harry and Wake, 1997) and DivIC (Katis et al., 1997) proteins of B. subtilis. Most probably, PBP 2B tends to accumulate in increasing amounts at the septum as it invaginates. Presumably, the protein remains spread out in a disc, rather than constricting with the leading edge of the septum. Eventually, however, as the cell pole matures, the PBP 2B is released, giving a diminished signal at old cell poles.

The observation that a small amount of PBP 2B protein sometimes remains at the mature cell poles was interesting, because such a retention is important for functioning of the division site selection protein, DivIVA, and DivIVA targeting to the pole is a late event depending on all division proteins including PBP 2B (Marston et al., 1998; H. B. Thomaides and J. Errington, unpublished). Thus, PBP 2B could be the immediate target for DivIVA protein.

It was difficult to demonstrate clearly the targeting of PBP 2B to the asymmetric septum of sporulating cells. The relatively low proportion of wild-type cells with a subpolar band of PBP 2B suggests that the protein is only present transiently. Nevertheless, the use of sporulation mutants provided strong evidence for the recruitment of PBP 2B to this kind of septum, particularly the spoIIB mutant. We do not understand why the PBP 2B bands were brighter in the spoIIB mutants, but it is probably unrelated to the block in later development of the prespore, as the bands still disappeared in cells with a completely segregated prespore chromosome (Fig. 5C, cell ii). Also, the PBP 2B bands were not brighter than the wild type in the spoIIG mutant (Fig. 5D), in which mother cell development is completely blocked (Piggot and Coote, 1976; Illing and Errington, 1991a). Whatever the explanation for the differences in band intensity, it seems that PBP 2B is recruited to the asymmetric septum, but that it remains there only briefly.

The finding that the PBP 2B signal was generally detected at only one of the polar potential division sites was unexpected. FtsZ and the sporulation-specific protein SpoIIE, at least, are recruited to sites at each of the cell poles (Arigoni et al., 1995; Levin and Losick, 1996) (Fig. 5A). Normally, only one of these division assemblies matures to form a septum, although it was shown recently that the second apparatus sometimes initiates but then aborts ingrowth (Pogliano et al., 1999). In spo0H mutants, moreover, FtsZ rings form at each of the cell poles, but neither goes on to form a septum (Levin and Losick, 1996). Thus, control over septation seems to be exerted at a point after assembly of an FtsZ ring in these cells. In wild-type, spoIIB and spoIIG mutant cells we invariably detected polar PBP 2B bands only at one end of the cell. Furthermore, in spo0H mutants, strong bands of PBP 2B were only seen at central positions, so this protein does not seem to be recruited to sites at which septation does not occur. These observations suggest that PBP 2B recruitment might play a role in the regulation of asymmetric septation during sporulation, although it should be remembered that, in vegetative cells, it is difficult to detect PBP 2B associated with the division site until division is under way. Nevertheless, the recent detection of abortive, partially formed septal structures at the prespore-distal pole of the cell (Pogliano et al., 1999) and our findings of analogous structures in vegetative cells depleted for PBP 2B (see below) suggest that recruitment of PBP 2B may be an important step in the regulation of sporulation septation.

Separate effects of PBP 2B depletion on assembly of the division machinery and septum ingrowth

To investigate the phenotypic consequences of PBP 2B deficiency, we used a strain in which transcription of pbpB was placed under the control of the Pspac promoter. The effects of repression of this promoter were slow to appear, in contrast to the very rapid arrest in septation in similar experiments with the adjacent ftsL gene (Daniel et al., 1998). Western analysis showed that the main reason for slow development of the PBP 2B phenotype is that the protein is quite stable.

Eventually, however, a clear arrest in cell division became apparent (Fig. 1). High-resolution (electron microscopical) examination of the filamentous cells (Fig. 2) revealed another difference from ftsL-depleted cells (Daniel et al., 1998) (and from other division mutants examined previously; Callister and Wake, 1981), in that an ingrowth of the cell envelope, presumably corresponding to an incomplete septum, was usually found at about mid-filament. No other wall ingrowths were found in the long filaments, so it seems that septation is finally arrested at the level of the initiation of septal ingrowth. As the mid-point of the filament corresponds to the oldest potential division site in the cell, we must presume that these structures were generated when PBP 2B was just limiting. If so, it appears that PBP 2B is required not only for the initiation of septation, but also for the duration of septal ingrowth.

IFM showed that the ‘early assembling’ division protein FtsZ still formed bands at appropriate regular intervals along the length of PBP 2B-depleted filaments, indicating that FtsZ ring assembly does not require PBP 2B. However, the later assembling proteins, DivIB and DivIC, generally formed bands only at mid-filament. The dependence of these ‘late’ proteins on PBP 2B was unexpected because, in E. coli, PBP 2B is not required for the assembly of any other division protein, except possibly FtsN (Addinall et al., 1997).

The structures formed at mid-cell in PBP 2B-depleted filaments may contain a complete division assembly; they certainly include FtsZ, DivIB and DivIC. As discussed above, it seems likely that they also contain residual PBP 2B protein, although not sufficient for the septum to go rapidly to completion. Taken together with our observations on the development of PBP 2B bands (see above), we suggest that a small amount of PBP 2B protein, possibly synthesizing a small amount of wall material, is needed for the initiation of septation and for the initial co-assembly of the other transmembrane division proteins. Septal progression then requires continual addition of PBP 2B molecules near the leading edge of the annulus. These presumably help to fill in the membrane invagination with wall material. The PBP 2B molecules recruited may tend to remain localized at their site of insertion, which would explain why the mass of PBP 2B at the division site appears to increase during septation, as evidenced by the brighter PBP 2B bands being associated with cells that have clear constrictions. It would also explain why the PBP 2B accumulations retain the form of a band late in division, in contrast to FtsZ, which constricts to a point along with septal ingrowth (Levin and Losick, 1996). So far, the functions of the smaller transmembrane division proteins, DivIB, DivIC and FtsL, have not been clear. The above considerations suggest that one or more of these proteins may act to recruit PBP 2B molecules to the leading edge of the septum as it progresses. Consistent with these ideas, in the depletion experiments, once PBP 2B has been used up, further ingrowth of the septum slowed or stopped.

A final unexpected feature of the PBP 2B depletion phenotype was the ready detection of cells in which wall synthesis was incomplete, possibly arrested (Fig. 2B and C) but membrane invagination appeared to have gone to completion. This suggests that wall and membrane ingrowth, although normally tightly co-ordinated, are separable processes, at least after the full set of division proteins have been assembled, and wall ingrowth has been initiated. We are currently investigating the possible roles of the smaller transmembrane division proteins in this co-ordination.

Experimental procedures

Bacterial strains and plasmids

The B. subtilis strains used are listed in Table 4. Other bacterial strains and plasmids are dealt with in the appropriate sections below.

Table 4. . Bacterial strains and plasmids (Illing and Errington (1991b); Errington and Mandelstam (1986); Katis et al. (1997)).Thumbnail image of

General methods

E. coli stains were grown at 37°C in 2 × YT (Sambrook et al., 1989), supplemented with ampicillin (100 μg ml−1) and/or kanamycin (25 μg ml−1), as necessary. B. subtilis strains were grown in Difco antibiotic medium 3 (PAB) or LB medium (without glucose; Morrison, 1979) supplemented with xylose (0.5%) and IPTG (0.5 mM), as necessary.

B. subtilis strains were transformed using the method of Anagnostopoulos and Spizizen (1961) as modified by Jenkinson (1983). To examine PBP 2B localization during sporulation, strains were grown in CH medium and induced to sporulate by resuspension in sporulation medium (SM), as has been described previously (Sterlini and Mandelstam, 1969; Nicholson and Setlow, 1990; Partridge and Errington, 1993).

All figures were prepared using Adobe Photoshop.

Depletion of PBP 2B

Strain 804 was grown overnight at 30°C in PAB with xylose (0.5%) and IPTG (0.5 mM). This starting culture was then diluted 1:10 in the same medium and incubated for a further 1 h at 37°C. After this time, the culture was diluted to an OD600 of 0.05 and allowed to grow to an OD600 of 0.2. The culture was centrifuged, and the cells were washed with medium lacking IPTG, then resuspended to the original volume in PAB containing xylose (0.5%). This culture was divided into two portions, one of which was supplemented with IPTG as a control.

Changes in cell morphology were determined using phase-contrast microscopy of ethanol-fixed samples (Hauser and Errington, 1995). Samples taken during the depletion time course were centrifuged and resuspended in an equal volume of 70% ethanol and stored at 4°C overnight. These samples were then resuspended in an equal volume of H2O and mounted on polylysine-treated slides. Images of the cells were acquired using a cooled CCD camera (System 3000; Digital Pixel Advanced Imaging Systems). Analysis of cell length was carried out with Object-Image software (Vischer et al., 1994).

For IFM of FtsZ, DivIB and DivIC in PBP 2B-depleted cells, strain 804 was grown at 25°C overnight in L broth (without glucose; Morrison, 1979) containing xylose (0.5%) and IPTG (0.5 mM), then diluted to an OD600 of 0.02 and incubated at 37°C until an OD600 of ≈ 0.2 was reached. A 10 ml sample of culture was washed, and the cells were resuspended in 40 ml of L broth with xylose and either with or without IPTG, and incubation was continued at 37°C.

Electron microscopy

Samples (10 ml) from the PBP 2B-depleted cultures were fixed and sectioned as has been described previously (Illing and Errington, 1991a). The sections produced were observed in a Zeiss 912 Omega electron microscope, and images were collected by CCD camera (Fig. 1B,C) or direct plate (Fig. 1A) photography.

Overexpression and purification of PBP 2B

Plasmid pQE PBP 2B was made by cloning a 2172 bp polymerase chain reaction (PCR) fragment, generated using the two oligonucleotides 9600 (5′-AACATACAGGGATCCTTCAAATGCCAAAAA-3′) and 4030 (5′-CGACGGCTTTGGTA CCAATCAGGATTTTTAAAC-3′) into the pQE30 expression vector (Qiagen). The resultant plasmid, pQE-PBP 2B, was used for the expression of 6 His-tagged PBP 2B in E. coli strain NM554 (Raleigh et al., 1988) containing plasmid pRep4 (Qiagen). Overnight cultures were diluted 1:100 and grown to an OD600 of 0.5, at which point the culture was induced with IPTG (final concentration 1 mM). Three hours after induction, the cells were harvested by centrifugation and lysed in 6 M urea, 50 mM sodium phosphate, pH 7. The lysate was mildly sonicated to break the released chromosomal DNA and then clarified by centrifugation at 20 000 r.p.m. The resulting solute was subjected to affinity purification on Chelating Sepharose Fast Flow (Pharmacia) loaded with Ni, and eluted with 50 mM EDTA in 6 M urea, 50 mM sodium phosphate, pH 7.0. The protein was precipitated with acetone and redissolved in phosphate-buffered saline (PBS). Rabbit polyclonal antiserum was raised against the purified PBP 2B using standard methods (Harlow and Lane, 1988).


Culture samples (1 ml) were centrifuged and the pellets were frozen. Later, the cell pellets were thawed and resuspended in 200 μl of sample loading buffer and lysed by sonication. Equal amounts of protein (adjusted according to the OD of the original culture) were separated using 6% SDS–PAGE. The proteins were then transferred to polyvinylidene difluoride (PVDF) membranes (Amersham, Hybond-P) by electroblotting and probed with a dilution of the primary antiserum in PBS, 0.1% Triton-20 and 5% dried milk powder. The dilutions were 1:10 000 for anti-PBP 2B, 1:5000 for DivIB and DivIC and 1:10 000 for FtsZ. The secondary antibody was a 1:8000 dilution of anti-rabbit horseradish peroxidase conjugate (Sigma).

Immunolocalization experiments

For immunolocalization of PBP 2B, cell samples were taken from exponentially growing cells in PAB or from sporulating cells in SM. For temperature-sensitive strains, overnight cultures were diluted 1:20 and allowed to reach an OD600 of 0.2 at 30°C, at which point the culture was split, and one aliquot was moved to the non-permissive temperature, 48°C, for 40 min. Samples were then taken from both cultures and processed in parallel.

Cells were fixed and permeabilized essentially as described by Harry and Wake (1997) and Katis and Wake (1999). After mounting on multiwell slides, the cells were blocked for 1 h in PBS containing 0.05% Tween-20 (PBST) and 2% BSA, and then left overnight at 4°C with a 1:1000 dilution of anti-PBP 2B antiserum in PBST containing 2% BSA. The wells were washed at least 20 times with 20 μl of PBST and blocked with PBST containing 2% BSA, 2% whole goat serum for 1 h. The secondary antiserum, a 1:1000 dilution of anti-rabbit fluorescein isothiocyanate (FITC) conjugate (Sigma), in PBST, 2% BSA, was then applied, and the slides were incubated in a humid chamber. After 1.5 h at room temperature, the wells were washed with 20 μl of PBST as above. Finally, the cell samples were mounted in Vector Shield (Vector Laboratories) with 4′,6-diamino-2-phenylindole (DAPI) and allowed to equilibrate for 30 min. Slides were either stored at −20°C or visualized immediately using epifluorescence and phase-contrast microscopy. Images were acquired as described above.

For IFM of FtsZ, DivIB and DivIC, affinity-purified rabbit anti-DivIC, anti-DivIB and anti-FtsZ antibodies were used as has been described previously (Daniel et al., 1998), except that the fixed cells were treated with polylysine before and after lysozyme treatment for DivIB and DivIC detection, or with polylysine only before lysozyme treatment for FtsZ detection.


This work was supported by grants from the Biotechnology and Biological Sciences Research Council and the BIOTECH programme of the European Community (J.E.) and from the Australian Research Council (E.J.H.). We are grateful to Dr L. J. Wu for helpful comments on the manuscript.