INTRODUCTION

Top Abstract Introduction Materials & Methods Results & Discussion References

Bacillus subtilis is a gram-positive endospore-forming bacterium that when grown vegetatively divides by precise placement of a septum at midcell. However, during sporulation the generation of two cells with different development fates is preceded by an asymmetric division. Vegetative division involves FtsZ, FtsA, DivIB, DivIC, FtsL, and PBP2B, in addition to other proteins. It is believed that the asymmetric division in sporulation uses the same division proteins. FtsZ appears to play a pivotal role in selection of the division site prior to cell division in both Escherichia coli and B. subtilis (7, 8); at the onset of sporulation, assembly of an FtsZ ring shifts from midcell to the two potential polar sites, and this shift is dependent on the transcription factor Spo0A (25).

Three min genes (minC, minD, and minE) are known to exist in E. coli. Homologous MinC and MinD proteins have been identified in B. subtilis, but a MinE equivalent has not yet been found (23, 24, 32). The MinE protein appears to play a role in topological specificity in E. coli (29). Because of the similarity between minicell formation and asymmetric septation during sporulation, it was proposed that there were two MinE-like B. subtilis proteins, with different topological properties (29). The vegetative specific MinE protein would activate septation at midcell, whereas sporulation-specific MinE would allow division at the polar position and simultaneously allow the division inhibitor to block septation at midcell. This model also predicts that these different MinE proteins will have different topological properties. The vegetative MinE would counteract the action of the MinCD division inhibitor at midcell but not at the cell poles. In contrast, sporulation-specific MinE would counteract the division inhibitor at the polar sites and would block the midcell septation site. The search for MinE-like proteins in B. subtilis led to the discovery of the DivIVA protein, which has some characteristics expected of a MinE-like protein (9, 14). Although genetic evidence supports the hypothesis that in B. subtilis DivIVA has a function analogous to that of MinE in E. coli, there are important differences between these two proteins. Cha and Stewart (9) also proposed that DivIVA is involved in activation of polar septation sites during the sporulation process.

MinC and MinD proteins are inhibitors of septation at medial and polar sites in both E. coli and B. subtilis, and their absence leads to minicell formation. However, the role of MinCD in sporulation is unclear, as minCD mutants sporulate very efficiently, in contrast with mutations at another locus associated with the minicell phenotype, divIVA. Here we report that the lack of active MinCD during sporulation causes the creation of thin sporulation-like septa at the midcell position. This is a surprising discovery that has likely not been reported previously because this phenotype has only a minimal influence on the level of sporulation efficiency. We also report that SpoIIE, a sporulation-specific protein which allows activation of ^F in the forespore, localizes in these cells not only in a bipolar manner but also at the midcell position.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results & Discussion References

Bacterial strains, culture media, and genetic techniques. Bacterial strains used in this work are listed in Table 1. Isolation of genomic and plasmid DNA, methods for transformation of B. subtilis and E. coli strains, selection of antibiotic resistance markers, and other standard genetic techniques were carried out as described previously (10). Standard enzymatic reagents, including restriction enzymes, DNA ligase, and Taq DNA polymerase, were purchased from Boehringer Mannheim, New England Biolabs, or Perkin-Elmer, Inc. Amplification of DNA by PCR was carried out using standard methods (19). To construct plasmids pBGCDM3 and pBGCDdel, PCR products were amplified from PY184 and PY79 chromosomal DNA, respectively, using oligonucleotides MINCECO (5'-GTGGAATTCGTGAAGACCAAAAAGCAGC-3') and MINDBAM (5'-CGGTGGATCCAAGAACAAAGCAGGC-3'). These PCR products contained the nucleotide sequence coding for the entire minC and minD genes. The PCR fragment derived from PY184 was digested with EcoRI and BamHI and inserted into pGEM-3Zf(+)cat (35) digested with the same restriction enzymes, creating plasmid pBGCDM3. The PCR fragment derived from PY79 was inserted into plasmid pGEM-3Zf(+)cat as described above, creating plasmid pBGCD. This plasmid was cut with KpnI, and then a 5-kb backbone was purified by agarose gel electrophoresis and ligated, creating plasmid pBGCDdel. The presence of the divIVB1 mutation on plasmid pBGCDM3 was confirmed by DNA sequencing using the fmol DNA cycle sequencing system (Promega Corp.). Plasmids pBGCDM3 and pBGCDdel were used for transformation of PY79 cells by selection for chloramphenicol resistance, creating strains IB498 and IB515, respectively. The correct recombination of both plasmids in the minCD region of the chromosome was verified by Southern blot hybridization. Strain IB508 was prepared and its chromosomal structure was verified as described above for strain IB498. Plasmid pBGCDM3-kan, a derivative of plasmid pBGCDM3, was prepared by replacing the cat gene with the Km^r gene from plasmid pUK19 (a gift from W. Haldenwang). This plasmid was used for transformation of strain PY79, creating strain IB508. The chromosomal structure was also verified by Southern blot hybridization.

Electron microscopy. Sporulation was induced by nutrient exhaustion in DSM (Difco sporulation medium) broth as described previously (10). At various times after cessation of growth, samples were harvested by centrifugation, washed in a buffer containing 0.1 M sodium cacodylate (pH 7.2), and fixed for 1 h at 4°C in the same buffer containing 2% glutaraldehyde and 1% OsO₄ (17). Following fixation, cells were pelleted by centrifugation, washed in 0.1 M cacodylate buffer, and dehydrated by a series of washes and incubations (10 min each at ambient temperature) in graded concentrations of ethanol (30, 50, 70, 85, 95, and 100% [twice in 100% ethanol]). This was followed by washing (twice for 30 min each time) in acetone. Dehydrated cells were embedded in Spurr medium (31) and polymerized for 8 h at 70°C. Samples were sectioned (60- to 120-nm thickness) on a NOVA3 ultramicrotome (LKB) and floated onto copper grids. Sections were stained on uranyl acetate drops for 20 min and on lead citrate drops for 5 min. Stained thin sections were examined and photographed on a JEOL-100 CX or Tesla Brno (T541) electron microscope.

Scoring criteria. Thickness of septa was scored from digitalized electron micrographs scanned with a Hewlett-Packard 4C scanner and using PowerPoint 7.0 software (Microsoft) or by direct measurement of thickness from photographs taken at higher magnification. For quantification of morphological classes by electron microscopy, at least 100 complete longitudinal sections were scored from random fields for each sample.

Fluorescence microscopy and light microscopy. Cells were prepared for fluorescence microscopy as described previously (4). B. subtilis IB509, containing a copy of plasmid pBsIIE-mutGFP integrated at the chromosomal spoIIE locus, was cultured in 50 ml of DSM broth at 37°C. At selected times during growth and sporulation, 400-µl culture samples were transferred to 1.5-ml Eppendorf tubes and collected by centrifugation, and the pellet was resuspended in preservation buffer (40 mM NaN₃, 50% [wt/vol] sucrose, 0.5 M Tris [pH 7.5], 0.77 M NaCl). Samples were examined or photographed after a few hours. Epifluorescence micrographs were recorded on Kodak transparency film with an Olympus B201 microscope using a filter set (420- to 480-nm excitation, DM500 dichroic reflector, and 515-nm emission filters). Film transparencies were digitalized, and the epifluorescence micrographs were prepared for printing by using CorelDraw 6.0 software.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials & Methods Results & Discussion References

Sporulation of minD mutants. We prepared divIVB1 mutant (IB498) and minD deletion (IB515) strains as described in Materials and Methods. The divIVB1 strain contains two mutations in the last gene of the divIVB operon, which is minD. The first mutation results in substitution of lysine for methionine at amino acid residue 87, and the second results in an isoleucine-to-threonine change at amino acid 147 of MinD (32). The minD deletion is at a KpnI site which leaves only the first 23 amino acid residues at the N terminus of MinD.

Ultrastructure of minD mutant cells. Sporulating cultures of the divIVB1 mutant (IB498), the minD deletion strain (IB515), and the wild-type strain (PY79) were harvested at various times during sporulation and examined by electron microscopy to test whether the small decrease in sporulation efficiency in divIVB1 cells was caused by release of MinCD repression at all possible sites of septation (middle and both polar sites). We used samples harvested at t₄ (4 h after the beginning of sporulation) for quantitative determinations of different types and localization of septa. As expected, in the wild-type strain we observed two kinds of septa: asymmetrically positioned thin sporulation septa and thick vegetative septa at the midcell site (Fig. 1A and C). In strains IB498 and IB515, we observed cells with unusually thin septa at the midcell position, which we call sporulation-like septa (Fig. 1A to C). We also observed these thin septa at the midcell position in different stages of forespore development. Figures 1A and C show cells with a thin midcell septum at a very early time in sporulation. Some midcell thin septa also were observed in later stages of forespore development (Fig. 1B). In a few cross sections, cells with multiple septa were observed. Some of the thin septa are not positioned at the precise midpoint of the cell (Fig. 1A), possibly a consequence of the nature of the specific minD mutation carried by the cells. Heterogeneity of cell length was observed previously in E. coli minD mutants, and it is believed that this is caused by the role of MinD protein in nucleoid partition, which is coupled with septation (12). Dassain and Bouché (11), in contrast, proposed that polar division in minD mutants causes an abrupt and temporary decrease in the amount of FtsZ available for nonpolar division, which in turn has effects such as the cell length heterogeneity noted above.

View larger version (87K): [in this window] [in a new window]   View larger version (93K): [in this window] [in a new window]   FIG. 1.   Electron micrographs of representative examples of septum morphologies in strains IB498 (divIVB1) (A to C), IB507 (spoIIE::Tn917HU7 divIVB1) (D), and IB506 (spoIIE71 divIVB1) (E). The insets to panels A and C show examples of wild-type (wt) vegetative and sporulation septa, respectively. Samples were harvested at 2 (A and C), 4 (D and E), and 7 h after cessation of logarithmic growth (B). Each scale bar represents 0.3 µm.

View larger version (32K): [in this window] [in a new window]   FIG. 2.   Quantitation of morphological classes in sporulating cells of strains PY79, IB498, IB506, and IB507. Cells were cultured in DSM broth at 37°C and harvested for examination by electron microscopy 4 or 7 h after cessation of logarithmic growth. Morphological classes were used as indicated at the right: 1, no septa; 2, thin septum at polar position; 3, thick septum at polar position; 4, thin septum at midcell position; 5, thick septum at midcell position; 6, cell with forespore at stage III or later.

Localization of SpoIIE-GFP fusion in the minD strain. A few sporulation-specific proteins are known to be associated specifically with the sporulation septum: SpoIIE (2, 4), SpoIIGB (20), SpoIIGA (15), and SpoIIIE (34). The SpoIIE protein has an important role as a serine phosphatase that releases ^F activity specifically in the forespore compartment (13). We investigated the localization of SpoIIE protein in minD cells by using a SpoIIE-green fluorescent protein (GFP) fusion in strain IB509. The main pattern of GFP localization at t₂ was bipolar (Table 4), as observed previously in a MinCD⁺ strain (2, 4). In addition to polar localization of SpoIIE, we expected to find cells in which the signal localized at the midcell position. Indeed, in some cells we observed a signal that represents the localization of SpoIIE at the unusual midcell position (Fig. 3), although the number of such cells was lower (9% of total cells) (Table 4) than we expected (30% of total cells). The unexpectedly low frequency of cells with SpoIIE localized to the midcell may well have been caused by the fast disappearance of the protein from this position, as earlier work has shown that SpoIIE rapidly disappears from the sporangium, first from the mother cell pole and then from the forespore pole (28). Another possibility for the low level of cells with SpoIIE at the midcell position could be lower affinity of the protein for this site than for polar sites at onset of sporulation. However, localization of SpoIIE phosphatase at the midcell position might inhibit sporulation and thus cause a delay or arrest of further forespore development.

View larger version (34K): [in this window] [in a new window]   FIG. 3.   Fluorescence microscopy of cells with SpoIIE-GFP-associated fluorescence at sites of septation in strain IB509 from cultures harvested 2 h after cessation of logarithmic growth, photographed as described in Materials and Methods. Cells with SpoIIE-GFP localization are marked. Arrows indicate the position of SpoIIE-GFP signals. Each scale bar represents 2 µm.

Ultrastructure of minD and spoIIE double-mutant cells. It was shown previously that SpoIIE is essential for normal formation of the asymmetric sporulation septum and that a spoIIE null mutant produced cells containing thick asymmetric septa similar to those formed at the midcell site during normal vegetative growth (5, 16, 18, 27); in addition, the formation of these thick asymmetric septa is usually delayed (5, 16). Two strains with missense mutations, spoIIE64 and spoIIE71, that block sporulation at stage II have also been identified, but they have a strikingly different phenotype, characterized by the presence of only thin asymmetric septa, frequently at both polar positions (5, 16, 27). These two mutations are in the region of the spoIIE domain important for encoding the protein's phosphatase activity, but they have no effect on the morphology of sporulation septa (1, 5, 16). We have investigated septation defects in double-mutant strains, IB506 and IB507, each with the divIVB1 mutation and with one of the two different kinds of spoIIE mutations, i.e., the null mutation spoIIE::Tn917HU7 and the missense mutation spoIIE71, respectively. Sporulating cells of strain IB506 displayed characteristic phenotypic features of both minD and spoIIE null mutations. We observed only thick asymmetric septa, many cells having thick septa at the midcell position (about 16% of total cells examined) (Fig. 2C). Figure 1D shows a cross section of an IB506 cell, harvested 4 h after the beginning of sporulation, with minicell and vegetative septum beginning to be formed. Approximately 2% of the cells examined in cross sections contained midcell septa that by their thickness (14 to 26 nm) were classified as thin septa. An interesting feature of this double mutant was that we observed no significant delay of asymmetric septation (Fig. 2C) as was described previously for the spoIIE null mutant (5, 16). This observation suggests that the SpoIIE protein can be a part of the control mechanism that activates septation at the polar sites, probably through action of the MinCD complex. When SpoIIE was expressed earlier in development from an isopropyl--D-thiogalactopyranoside-inducible P_spac promoter, cells produced many septation abnormalities and also an increase in the number of septa at different positions (6). However, it is not likely that SpoIIE is able to directly recognize the MinCD complex at the poles because, as has been shown recently, localization of SpoIIE is FtsZ dependent (26) and thus FtsZ and probably other division proteins are directed to possible sites of septation earlier than SpoIIE itself. In contrast with this observation, Khvorova et al. (21) demonstrated that SpoIIE is involved in the Spo0A-dependent switch in the location of FtsZ rings.

Model of MinCD function during asymmetric septation. The main contribution of this work is the discovery that the MinCD complex blocks the midcell site of septation during sporulation. There are two main lines of evidence that support this finding. First, in the absence of functional minD gene products, about 30% of cells create thin sporulation-like septa at the midcell position, although the polar septation event and subsequent stages of spore morphogenesis proceed normally. Our findings confirm the model proposed recently by Rothfield and Zhao (29), which predicts that the midcell position is released from inhibition in the absence of MinC and MinD products during sporulation. Second, the SpoIIE protein, normally localized in a bipolar manner in wild-type sporulating cells, can also be recruited to the midcell septal position in minD cells. Previous work has shown that SpoIIE localization is FtsZ dependent and that SpoIIE may be recruited to the potential division sites directly by FtsZ or may interact with other FtsZ-associated division proteins (26). Localization of SpoIIE protein to the midcell position in minD cells clearly indicates that this position can be recognized not only by division proteins involved in binary fission but also by SpoIIE and probably other sporulation-specific proteins as well. It was shown previously that SpoIIE expressed during vegetative growth can be localized to the midcell division site (26) where unusually thin septa are formed (6). Levin and Losick (25) have shown that Spo0A protein switches the localization of FtsZ from medial to a bipolar pattern. Additionally, these authors showed that a mutation in minD does not restore bipolar localization of FtsZ in sporulating cells of a spo0A mutant. They observed bands of FtsZ positioned at poles as well as at medial locations (but only one band per cell length, although theoretically all three sites are available) and argued that Spo0A is not acting through the MinCD proteins. In contrast, our results indicate that bipolar localization of SpoIIE is influenced by MinCD, and FtsZ should be localized earlier than SpoIIE in a similar manner. The mechanism that serves to sequester FtsZ to a single division site, in the absence of active Spo0A and MinCD, is likely to be regulated by the level of FtsZ. This suggestion is supported by the finding that in E. coli, increased levels of FtsZ can override the inhibition imposed by the min system at the cell poles (33). One explanation for the increase in concentration of FtsZ at the onset of sporulation is that Spo0A, as a transcription activator, positively regulates its transcription. This regulation also can be indirectly mediated by AbrB and/or Spo0H proteins.

View larger version (20K): [in this window] [in a new window]   FIG. 4.   Possible models demonstrating how the MinCD complex controls the switch from midcell division to bipolar initiation of septation during the onset of sporulation. Triangles represent possible sites of septation; open triangles represent sites not blocked by MinCD, and filled triangles represent sites blocked by active MinCD complex (CD). 0A, Spo0A protein. (A) The midcell-stabilizing model, whereby Spo0A controls factor x, which specifically stabilizes the MinCD complex at midcell position. (B) The bipolar activation model. Here Spo0A controls factor y, which specifically releases MinCD inhibition at both polar positions. For details, see Results and Discussion.