The FtsEX ABC transporter directs cellular differentiation in Bacillus subtilis

Authors

  • Sharon Garti-Levi,

    1. Department of Molecular Biology, Faculty of Medicine, POB 12272, The Hebrew University of Jerusalem, 9112 Jerusalem, Israel.
    Search for more papers by this author
  • Ronen Hazan,

    1. Department of Molecular Biology, Faculty of Medicine, POB 12272, The Hebrew University of Jerusalem, 9112 Jerusalem, Israel.
    2. Department of Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, MA 02114, USA.
    Search for more papers by this author
  • James Kain,

    1. Department of Molecular and Cellular Biology, Harvard University, 16 Divinity Avenue, Cambridge, MA 02138, USA.
    Search for more papers by this author
  • Masaya Fujita,

    1. Department of Biology and Biochemistry, University of Houston, Houston, TX 77204-5001, USA.
    Search for more papers by this author
  • Sigal Ben-Yehuda

    Corresponding author
    1. Department of Molecular Biology, Faculty of Medicine, POB 12272, The Hebrew University of Jerusalem, 9112 Jerusalem, Israel.
    Search for more papers by this author

*E-mail sigalb@ekmd.huji.ac.il; Tel. (+972) 2675 8600; Fax (+972) 2675 8311.

Summary

A fundamental challenge in developmental biology is to elucidate the regulatory events that trigger cellular differentiation. Sporulation in the Gram-positive bacterium Bacillus subtilis serves as a simple experimental model system to address this challenge. The hallmark of sporulation is the formation of an asymmetrically positioned septum that divides the cell into unequally sized progeny. Here we describe the role of an ABC transporter, comprising the FtsE and FtsX proteins, in the initiation of spore formation. We discovered that in the absence of this transporter, entry into sporulation is delayed and an atypical symmetric septum is formed instead of a polar one. We show that this phenotype can be suppressed by artificially activating the master regulator of sporulation, Spo0A, or by activating the histidine kinases that function upstream of Spo0A. Our data indicate that the FtsEX transporter is one of the top components in the hierarchy of factors required to initiate sporulation, and thus it is essential for establishing proper temporal activation of the process.

Introduction

In response to challenging environmental conditions, the Gram-positive bacterium Bacillus subtilis (B. subtilis) undergoes a complex differentiation process, termed sporulation, culminating in an adapted, highly resistant cell type, the spore. During the course of sporulation, the bacterial cell undergoes remarkable morphological changes. The most characteristic feature of this process is the formation of an asymmetrically positioned (polar) septum that divides the developing cell (sporangium) into unequally sized progeny, called the forespore (the small compartment) and the mother cell. Subsequently, the forespore is engulfed by the mother cell, which encases it, forming a cell within a cell. Following engulfment, a thick cell wall material, called the cortex, is synthesized in between the two membranes that separate the mother cell from the forespore. Concomitantly, a thick proteinaceous coat is assembled outside the cortex within the mother cell. Finally, after about 6–8 h of development, the fully ripened spore is liberated by lysis of the mother cell. The spore can convert back to an actively growing cell that divides symmetrically via a process called germination, which involves rehydration, cortex hydrolysis, and outgrowth (Stragier and Losick, 1996; Errington, 2003; Piggot and Hilbert, 2004). This morphological process serves as a powerful system for studying mechanisms that regulate cell fate and differentiation.

Bacillus subtilis begins to sporulate after integrating various environmental and physiological signals. A critical concentration of such signals is sensed by cell surface receptors that are able to initiate a sequential transfer of phosphate groups. The phosphorelay consists of three main histidine protein kinases, KinA, KinB and KinC, through which the phosphate is passed to the relay protein Spo0F, then to Spo0B, and finally to the key transcriptional regulator of sporulation, Spo0A (Jiang et al., 2000; Piggot and Losick, 2001; Piggot and Hilbert, 2004). The phosphorylated form of Spo0A (Spo0A∼P) directly controls the transcription of more than 120 genes by binding to their promoters at its recognition sequence. These target genes subsequently bring about more global changes in gene expression, indirectly altering the expression profile of over 500 genes and thus, triggering sporulation (Fawcett et al., 2000; Liu et al., 2003; Molle et al., 2003).

Among the genes controlled by Spo0A∼P are those responsible for switching from medial to polar division at the onset of sporulation. Developing cells that have passed the point of asymmetric division are committed to complete spore formation, whereas cells that have begun to sporulate but have not yet formed a polar septum are capable of resuming vegetative growth (Parker et al., 1996; Ben-Yehuda and Losick, 2002; Dworkin and Losick, 2005). The switch in division plan is brought about by a sporulation-specific increase in transcription of ftsZ, which encodes the key bacterial division protein, and by activation of the spoIIE gene encoding a FtsZ-associated protein. During vegetative growth, FtsZ forms a medial Z ring that is converted into a medial septum. However, during sporulation the medial Z ring is replaced by a helical filament of FtsZ that grows along the long axis of the cell, redeploying FtsZ to the poles. Finally, only one Z ring is converted into a polar septum, consequently allowing sporulation to progress (Levin and Losick, 1996; Khvorova et al., 1998; Ben-Yehuda and Losick, 2002). Bacteria carrying Spo0A mutations do not form polar Z rings or polar septa; instead they form medial Z rings and divide symmetrically when suspended in sporulation medium (Dunn et al., 1976; Levin and Losick, 1996; Ben-Yehuda and Losick, 2002).

The observation that cells become committed to sporulation only once the polar septum has formed implies that sporulation initiation must be determined before this irreversible stage. Indeed, it has been largely demonstrated that the phosphorelay components provide multiple targets for controlling a gradual initiation of sporulation (Jiang et al., 2000; Piggot and Losick, 2001; Piggot and Hilbert, 2004). Moreover, during exponential growth, an efficient induction of sporulation can be achieved by artificial activation of the phosphorelay kinases, but not by artificially activating their ultimate target, Spo0A. Thus, an important characteristic of the phosphorelay is that it causes the level and activity of Spo0A to increase gradually (Fujita and Losick, 2005). Accordingly, the genes controlled by Spo0A largely consist of two categories: those activated or repressed by a low dose of Spo0A∼P (low-threshold genes) and those that require a higher dose to be turned on or off (high-threshold genes). Genes directly involved in sporulation generally fall into the high-threshold category (Fujita et al., 2005). Based on these data, the current model is that incrementally increasing Spo0A activity is critical for triggering sporulation by allowing low-threshold genes to be switched on or off before high-threshold genes (Fujita and Losick, 2005).

To better understand the molecular mechanisms that trigger cellular differentiation, we searched for novel components that function upstream of Spo0A. We identified a mutant that forms a medial septum instead of a polar one under sporulation conditions, implying that the mutant is defective in initiating sporulation. This abnormal symmetrical division creates small daughter cells that are approximately half the size of a normal sporangium. Interestingly, at later time points, these small cells undergo polar division and eventually form functional mature spores. The mutation conferring this unique phenotype is located in a gene termed ftsE, which is part of an operon encoding FtsE and FtsX, which together comprise an ATP-binding cassette transporter (ABC transporter) (de Leeuw et al., 1999). We provide evidence that this transporter acts upstream of the phosphorelay cascade, and that it is required for proper spatial and temporal activation of sporulation.

Results

FtsEX is required for the switch from medial to polar division at the onset of sporulation

To better understand the molecular mechanisms governing the onset of sporulation, we screened for genes that affect polar division. We performed a visual microscopic screen, searching for mutants impaired in polar septation upon a shift to sporulation conditions. We examined a collection of strains bearing knockout alleles in early sporulation genes, as classified by genome-wide analysis (Fawcett et al., 2000; Molle et al., 2003 and see Experimental procedures). Using this approach, we identified a mutant that forms a medial, instead of a polar septum, at the onset of sporulation (compare Fig. 1A with Fig. 1B, t1.5). The mutant allele conferring this unusual phenotype is a knockout of a gene called ftsE, in which the DNA coding region for amino acids 9–228 was replaced by the erm resistance gene. ftsE is the first gene in an operon encoding the FtsE and FtsX proteins (henceforth, called FtsEX). A complementation assay revealed that the expression of both the ftsE and ftsX genes is required to rescue the mutant phenotype (Fig. S1), presumably as the effect of the mutation in ftsE is polar on ftsX (for additional details, see Table S1). Sequence analysis of FtsEX reveals that together the two proteins form an ABC transporter. ftsEX was initially identified in E. coli, where mutations within this operon caused the formation of long aseptate filaments (fts, for filamentation temperature-sensitive phenotype) (Ricard and Hirota, 1973; Taschner et al., 1988), yet the function of this operon in B. subtilis has not been investigated. Recent experiments suggest that the E. coli FtsEX localizes to the division ring and that FtsE physically interacts with FtsZ; however, the precise role of FtsEX in the E. coli division process remains elusive (Schmidt et al., 2004; Corbin et al., 2007). Examination of the B. subtilis FtsEX mutant cells during vegetative growth revealed that the cells exhibit a shape defect and frequently appear curved (compare Fig. 1A with Fig. 1B, t0). In addition, measuring the width of the mutant cells revealed that, on average, they are slightly wider than the wild-type cells (1.2 ± 0.04 μm and 1.0 ± 0.02 μm, respectively). However, despite these morphological defects, the mutation does not appear to have any significant effect on growth rate under various conditions examined (data not shown).

Figure 1.

FtsEX is required for the efficient switch from medial to polar division.
A and B. Wild-type (PY79) (A) and ΔftsE (RL2422) (B) cells were induced to sporulate and sporulation progression was analysed. At the indicated time points, samples were taken from the sporulating cultures, stained with a fluorescent membrane dye (FM4-64), and visualized by fluorescence microscopy (t0 = vegetative cells; t1.5; t2). White arrows denote polar septa, whereas yellow arrows indicate the presence of medial septa. Spore formation and morphology were examined by phase contrast microscopy (t24). The dashed red ovals highlight the spore borders, whereas the middle red line indicates the length of the spore as measured in (D). Scale bar corresponds to 1 μm.
C. Wild-type (PY79) and ΔftsE (RL2422) cells were induced to sporulate. At the indicated time points, samples were taken, stained with FM4-64, and visualized by florescence microscopy. The percentage of polar septa formation was assayed by evaluating the proportion of cells exhibiting polar septum and engulfing membrane morphologies. At least 1000 cells in each sample were scored.
D. Wild-type (PY79) and ΔftsE (RL2422) spores (t24) were visualized by phase contrast microscopy and the distribution of spore length was measured as depicted in (A). Length was determined using MetaMorph 6.2r4 software (Molecular Devices). At least 600 spores were analysed for each strain.

Further investigation of FtsEX mutant cells during sporulation revealed that inappropriate formation of the medial septum results in small daughter cells that are approximately half the size of a normal sporangium (Fig. 1B, t1.5). The de novo synthesis of medial septa in the mutant cells was further confirmed by the observation that FtsZ–GFP was almost entirely localized to the mid-cell position when the cells were suspended in sporulation medium (Fig. S2). However, at later time points, these small cells do undergo polar division and eventually form mature spores (Fig. 1A–C). These spores are, on average, smaller than normal and exhibit a broader size distribution (Fig. 1D). Nevertheless, a heat resistance colony-forming unit assay revealed that there is no significant difference between the viability of wild-type and FtsEX mutant spores (see Experimental procedures and data not shown). Indeed, when germination was monitored by time-course and time-lapse microscopy, we found that the mutant spores exit dormancy and germinate in a manner indistinguishable from wild-type spores (Fig. S3). Thus, in the absence of FtsEX, a medial septum is initially formed during sporulation, leading to the production of small sporangia. Nonetheless, later, a polar septum is formed by the FtsEX mutant sporangia, creating small mother cell and forespore compartments, which are finally capable of generating mature and functional spores. Based on these observations, we concluded that FtsEX ensures accurate temporal polar division at the onset of sporulation.

FtsEX is essential for efficient entry into sporulation

The results presented so far indicate that FtsEX regulates the switch from medial to polar division at the onset of sporulation. Accordingly, it is possible that FtsEX is part of the division machinery and interacts directly with the division components to determine the position of the septum. Thus, FtsEX may act either by inhibiting medial septation or by activating polar division. Alternatively, FtsEX may be required to initiate sporulation and thus, indirectly regulates the timing of polar septum formation. The absence of FtsEX causes an overall delay in the initiation of sporulation, allowing cytokinesis to take place at mid-cell by a default mechanism.

To distinguish between these two scenarios, we compared the expression of early sporulation genes in wild-type and mutant cells. If the first possibility is correct, we would expect the expression patterns of both strains to be similar, such that early sporulation genes are expressed concomitantly with inappropriate medial septation in the mutant cells. On the other hand, if the second possibility is correct, we would expect the expression of sporulation genes to be delayed in the mutant cells, parallel to the observed delay in the formation of a polar septum. Initially, we investigated whether expression of the early sporulation gene spoIIE (Piggot, 1973; Barak et al., 1996) is affected by the FtsEX mutation. To this end, monitoring the β-galactosidase activity from wild-type and FtsEX mutant cells harbouring the spoIIE–lacZ reporter revealed a significant delay in the expression of the reporter in the mutant cells (Fig. 2A). Next, polar septa formation was examined simultaneously with signal from SpoIIE–GFP fusion by fluorescence microscopy. We observed that at early sporulation time points, the production of both SpoIIE–GFP and polar septa formation is largely delayed in the mutant relative to the wild-type (Fig. 2B). However, at later time points, fluorescence from SpoIIE–GFP was increased in the mutant cells, correlating with the tardy formation of polar septa. Similar results were obtained when the sporulation proteins RacA and SpoIIQ (Londono-Vallejo et al., 1997; Ben-Yehuda et al., 2003) were fused to GFP and their production was examined by fluorescence microscopy (data not shown). Thus, we concluded that FtsEX is required for efficient entry into sporulation. In its absence, sporulation is delayed, and consequently, an additional medial division takes place before the initiation of sporulation. Consistent with this idea, when lacZ was placed under the control of the ftsEX promoter, it was highly expressed at the onset of sporulation and its expression decreased at later developmental stages (Fig. S4), suggesting that FtsEX activity is important primarily during early sporulation stages.

Figure 2.

FtsEX is essential for efficient entry into sporulation.
A. β-Galactosidase activity assay from sporulating wild-type (SB355) and ΔftsE (SB356) cells harbouring the spoIIE–lacZ reporter. Samples were taken at the indicated time points (hours), and β-galactosidase activity was calculated in Miller units.
B. Wild-type (SB201) and ΔftsE (SB343) strains producing the SpoIIE–GFP fusion were induced to sporulate and analysed by fluorescence microscopy at the indicated time points. Shown are overlays of FM4-64 (red) and SpoIIE–GFP (green) fluorescence images. Scale bar corresponds to 1 μm.

Spo0A activation is mediated by FtsEX

Similar to the FtsEX mutant phenotype, mutation in the key transcription factor, Spo0A, results in inappropriate medial divisions upon induction of sporulation (Dunn et al., 1976; Levin and Losick, 1996). However, unlike FtsEX mutant cells that eventually proceed through sporulation, Spo0A mutant cells arrest immediately after the formation of the medial septum. These observations led us to hypothesize that FtsEX might act upstream of Spo0A and is required for its efficient activation. Therefore, the delayed activation of Spo0A causes the improper medial septation exhibited by the FtsEX mutant cells.

To investigate whether FtsEX acts upstream of Spo0A, we tested whether a constitutively active form of Spo0A can suppress the ftsE mutation. To this end, an IPTG-inducible allele of spo0A, Phyper-spank-spo0A-sad67, was introduced into a ftsE null strain. This spo0A allele produces a protein that does not require phosphorylation to be active, and it is locked in a constitutively active state due to a short internal truncation (Ireton et al., 1993; Fujita and Losick, 2005). We examined whether the production of Spo0A-Sad67 is sufficient to bypass the septation defect characteristic of the FtsEX mutant under sporulation conditions. The spo0A-sad67 allele was induced when the cells were suspended in sporulation medium, and the formation of the polar septa was monitored by fluorescence microscopy. As shown in Fig. 3A (upper panel), inducing Spo0A-Sad67 synthesis in FtsEX mutant cells blocked medial septa formation and instead, polar septa were formed (t2 = 26%). In contrast, only a small fraction of the mutant cells (t2 = 4%) harbouring the inducible construct produced polar septa in the absence of the inducer at the same sporulation time point (Fig. 3A, data not shown). These results indicate that production of Spo0A-Sad67 suppresses the FtsEX phenotype. This supports the aforementioned hypothesis that FtsEX activity is required for efficient activation of Spo0A at the onset of sporulation; in its absence, sporulation is impeded due to delayed Spo0A activation.

Figure 3.

FtsEX functions upstream of the sporulation phosphorelay.
A. Strains SG52 (spo0AΩPhyper-spank-sad67-spc, ftsE::erm; upper panels) and SG46 (kinAΩPhyper-spank-kinA-spc, ftsE::erm; lower panels) were induced to sporulate with and without the addition of 20 μm IPTG. Two hours after the shift to sporulation conditions, cells were stained with FM4-64 and visualized by fluorescence microscopy. Scale bar corresponds to 1 μm.
B. Strains SG46 (kinAΩPhyper-spank-kinA-spc, ftsE::erm), SG48 (kinBΩPhyper-spank-kinB-spc, ftsE::erm) and SG50 (kinCΩPhyper-spank-kinC-spc, ftsE::erm) were induced to sporulate with and without the addition of 20 μM IPTG. Two hours after the shift to sporulation conditions, cells were stained with FM4-64 and visualized by fluorescence microscopy as in (A). The percentage of polar septa formation was assayed by evaluating the proportion of cells exhibiting polar septum and engulfing membrane morphologies. At least 800 cells were scored for each treatment. SD was calculated using values generated from several independent cultures.
C. Strains RL2411 (rapA::erm) and SG72 (rapA::erm, ftsE::cat) were induced to sporulate. At the indicated time points, samples of cells were taken, stained with FM4-64 and visualized by fluorescence microscopy. The percentage of polar septa formation was assayed by evaluating the proportion of cells exhibiting polar septum and engulfing membrane morphologies. At least 800 cells in each sample were scored. SD was calculated using values generated from several independent cultures.

The FtsEX transporter acts upstream of the sporulation phosphorelay

Having established that FtsEX acts upstream of Spo0A, we were next concerned with determining how FtsEX exerts its effect. As outlined in the introduction, Spo0A is activated by phosphorylation and obtains its phosphate via a phosphorelay comprising histidine kinases (mainly KinA, B and C) and the phosphotransfer proteins Spo0F and Spo0B. Therefore, we next investigated whether FtsEX acts upstream of the histidine kinases, by testing whether induced histidine kinase activity suppresses the FtsEX mutation. Initially, we artificially triggered the phosphorelay using an IPTG-inducible form of KinA, the major sporulation kinase (Fujita and Losick, 2005) and assayed polar septa formation under sporulation conditions. As shown in Fig. 3A and B, induced synthesis of KinA in the mutant cells resulted in efficient formation of polar septa without any additional medial divisions. Similar results were observed using inducible forms of KinB or KinC in the FtsEX mutant background (Fig. 3B). In addition, we investigated the effect of a mutation in RapA on the FtsEX mutant phenotype. RapA inhibits sporulation by specifically dephosphorylating the phosphotransfer protein Spo0F, which obtains its phosphate from all three kinases (Perego and Hoch, 1996; Perego et al., 1996), and therefore RapA mutation is phenotypically similar to activating the phosphorelay. In accordance with our previous results, introducing a mutation in RapA into the FtsEX mutant background largely suppresses the mutant division defect (Fig. 3C). Taken together, our findings indicate that FtsEX acts upstream of the sporulation phosphorelay and functions by activating the histidine kinases, which subsequently transfer phosphate to the phosphotransferases and finally to Spo0A. FtsEX deficiency delays induction of the phosphorelay and, as a consequence, postpones Spo0A activation.

FtsEX is membrane associated

It has been demonstrated that the E. coli FtsEX forms an ABC transporter in which FtsX is an integral membrane protein, whereas FtsE is a cytoplasmic unit containing the ATPase cassette (de Leeuw et al., 1999). Interestingly, the E. coli FtsEX was found to localize preferentially to the division septum (Schmidt et al., 2004). To investigate the localization pattern of the FtsEX transporter in B. subtilis, we fused the endogenous ftsX to gfp as a sole chromosomal copy and the fusion protein was monitored by fluorescence microscopy. GFP fluorescence was observed mainly at the peripheral membrane, with a few protein molecules apparently localizing into a helical filament (Fig. 4A), resembling other bacterial membrane-associated proteins that were reported to form helical structures (for example, Jones et al., 2001; Ben-Yehuda and Losick, 2002; Espeli et al., 2003; Shiomi et al., 2006). Further investigation revealed that this GFP fusion protein is non-functional because the strain exhibits a defective sporulation phenotype similar to that of the FtsEX mutant (data not shown). To observe FtsX–GFP in the context of a wild-type phenotype, we constructed a merodiploid strain harbouring unmodified ftsX and the ftsX–gfp fusion, both under the control of the natural ftsEX promoter. Importantly, growth and sporulation of this merodiploid strain were indistinguishable from that of wild-type (data not shown). In these merodiploid cells, the FtsX–GFP fusion was again localized to the cell membrane; however, unlike the localization of FtsX–GFP as a single copy, the helical filaments were less visible (Fig. 4B).

Figure 4.

Localization pattern of FtsEX. Strains (A) SB374 (ftsX–gfp–kan), (B) SB377 (ftsX–gfp–kan; amyE::ftsEX-spc), (C) SB380 (ftsE–gfp–kan ftsE) and (D) SB704 (ftsE–gfp–kan ftsE, chrΩPspac-ftsZ-ble) were grown in CH medium and GFP fluorescence was visualized at 0.5 OD600 by fluorescence microscopy. Similar localization patterns were obtained when the strains were suspended in sporulation medium (data not shown). For (D) cells were grown O/N in the presence of 0.2 mM IPTG and diluted to a medium that does not contain IPTG. Scale bar corresponds to 1 μm.

To investigate the localization of the FtsE cytoplasmic unit, we constructed a strain carrying both unmodified ftsE and ftsE–gfp fusion, and the localization of the fusion protein was followed by fluorescence microscopy. (The duplication of the ftsE gene was necessary to avoid a polar effect on the downstream ftsX gene by the fusion.) As shown in Fig. 4C, the FtsE–GFP signal was somewhat faint; however, the protein was clearly localized into the membrane highlighting the cell borders. Thus, the FtsE cytoplasmic unit seems to follow the localization pattern of the larger FtsX membrane subunit. Importantly, this localization pattern was independent of FtsZ, as FtsE–GFP was evident at the cell membrane when FtsZ was depleted (Fig. 4D). In summary, our localization studies revealed that the FtsEX transporter is located at the cell membrane and that, unlike its E. coli counterpart, its localization is neither restricted to the division septum nor dependent on FtsZ.

The FtsEX mutant phenotype cannot be complemented extracellularly

Given that FtsEX forms an ABC transporter located at the membrane and is required for efficient entry into sporulation, we speculated that this transporter stimulates the sporulation phosphorelay either by importing or secreting an extracellular signal. We reasoned that if FtsEX forms an exporter, then the mutant FtsEX phenotype should be complemented ‘extracellularly’ by co-culturing with wild-type cells. However, if FtsEX forms an importer, then the defective sporulation phenotype should be maintained regardless of the presence of wild-type cells.

To investigate these possibilities, it was necessary to differentiate wild-type from mutant cells by fluorescence microscopy; thus we generated a wild-type strain that constitutively expresses a gfp construct (Fig. 5A). Next, equal proportions of wild-type and FtsEX mutant cells were inoculated, grown to mid-log phase, and induced to sporulate. The formation of polar septa by non-fluorescent (FtsEX mutant cells) and fluorescent (wild-type) cells was scored at early times (t2) after exposure to sporulation conditions. Interestingly, we found that polar septa formation by wild-type cells was notably delayed in the co-culture relative to the timing of polar septa formation by wild-type cells grown alone, possibly due to the presence of numerous non-sporulating mutant cells that do not produce sporulation signals and thus reduce their overall concentration. Nonetheless, the efficiency of polar septa formation by FtsEX mutant cells in the co-culture was similar to that of mutant cells grown alone; in both cases sporulation was equally impaired (Fig. 5B). Moreover, increasing the ratio of wild-type versus mutant cells in the co-culture (even up to 4:1, respectively) did not improve the sporulation efficiency by the mutant cells (data not shown). These results are consistent with the scheme whereby FtsEX acts as an importer, as its absence is not complemented extracellularly by co-culturing mutant with wild-type cells. However, we can not rule out the possibility that FtsEX exports a sporulation factor that is not secreted into the medium, but instead is tethered to the membrane, where it acts to facilitate kinase phosphorylation. In addition, an ABC transporter that is phylogenetically related to FtsEX was found to release lipoproteins from the plasma membrane and not to act as a transporter (Yakushi et al., 2000; Bouige et al., 2002). Therefore, it is feasible that FtsEX is not an importer, but instead it could function by changing lipoprotein organization on the surface of the cell in such a way that it stimulates kinase activity at the onset of sporulation.

Figure 5.

The FtsEX mutant phenotype cannot be complemented extracellularly. Equal numbers of wild-type SG10 cells (amyE::Pspac–gfp–spc) constitutively producing GFP and FtsEX mutant cells of strain RL2422 (ftsE::erm) were co-inoculated into CH medium, grown to mid-log phase and then induced to sporulate. As a control, each strain was grown alone.
A. Shown is an overlay of FM4-64 (red) and GFP (green) fluorescence images of the co-cultures at t0 of sporulation. Scale bar corresponds to 1 μm.
B. Two hours after sporulation induction, samples were taken, stained with FM4-64 and visualized by fluorescence microscopy. The percentage of polar septa formation was assayed by evaluating the proportion of cells exhibiting polar septum and engulfing membrane morphologies. (a) SG10 grown alone, (b) SG10 grown in co-culture, (c) RL2422 grown alone, and (d) RL2422 grown in co-culture. At least 1000 cells were scored from each culture. SD was calculated using values generated from several independent cultures.

Discussion

Sporulation in B. subtilis is a relatively well-characterized process, yet little is known about the identity of the specific signals required for its initiation, or about the cellular components channelling them. Here we provide evidence suggesting that the ABC transporter FtsEX imports a signal required for activating the sporulation phosphorelay. Our data indicate that FtsEX acts at the top of a hierarchical cascade of events responsible for initiating sporulation, and contributes to proper temporal activation of the process. Based on our study, we propose the following model for FtsEX function at the onset of sporulation (Fig. 6). During vegetative growth, only low concentrations of an unidentified sporulation signal delivered by the FtsEX transporter exist. When sporulation conditions are induced, the level of this signal is elevated, stimulating its import via FtsEX. The import of this signal into the cells triggers the activation of the major kinases required to induce the phosphorelay and finally to phosphorylate Spo0A. Accumulation of phosphorylated Spo0A enables sporulation to proceed and consequently polar division is carried out. In the absence of the FtsEX transporter, activation of the kinases is delayed, Spo0A is not phosphorylated sufficiently to initiate sporulation and inappropriately, a default medial septum is formed. However, eventually, inputs from parallel pathways are channelled into the phosphorelay to produce threshold concentrations of phosphorylated Spo0A needed to successfully initiate sporulation.

Figure 6.

A model for the role of FtsEX in initiating sporulation. The figure depicting the potential role of FtsEX in triggering sporulation. The different components are listed in the legend. See the text for a further explanation.

The observation that in the absence of FtsEX, a medial division takes place instead of a polar one, suggests that medial septum formation is the default mechanism. Notably, however, this inappropriate medial division is not completely normal, as the cells do not elongate as much as they do during vegetative growth. Thus, the medial septum divides the cell into two small daughter cells, similar to the phenotype observed in the absence of Spo0A (Dunn et al., 1976; Levin and Losick, 1996). Based on these observations, we suggest that distinct molecular mechanisms are responsible for inhibiting cell elongation and for forming a medial septum. In the absence of Spo0A or FtsEX, the signals responsible for inhibiting cell elongation are active, whereas the signals for inhibiting medial divisions are absent or defective.

Our discovery that in the absence of FtsEX, sporulation is not aborted but postponed, raises the possibility that multiple transporters exist, which import various sporulation signals. A delicate balance between sporulation signals and transporter levels most likely determines the timing of sporulation initiation. In support of this idea, mutation in the ybdA gene encoding an ATPase subunit of a putative ABC transporter has been shown to reduce Spo0A activation (Isezaki et al., 2001), and the ABC transporter composed of YheH and YheI proteins has been shown to bind KinA and to affect Spo0A activation (Fukushima et al., 2006). Furthermore, Phr peptides, reported to induce sporulation by inhibiting the activity of phosphatases that dephosphorylate Spo0F (Lazazzera, 2001; Perego and Brannigan, 2001; Piggot and Hilbert, 2004), are transported into the cell via oligopeptide permeases that belong to the family of ABC transporters. Taken together, these studies corroborate that several ABC transporters are involved in activating sporulation (and accordingly, probably an assortment of signals), ultimately via induction of phosphorelay components.

ABC transporters are present in both prokaryotes and eukaryotes and constitute one of the largest superfamilies of proteins. ABC transporters are involved in import or export of a wide variety of substances, and thus could affect various cellular processes (Saurin et al., 1999; Bouige et al., 2002; Davidson and Maloney, 2007). Our co-culture experiments support the idea that FtsEX acts as an importer. Accordingly, sequence comparisons of several hundred ABC transporters classified FtsEX as a putative importer (Saurin et al., 1999). We therefore, prefer the possibility that FtsEX acts as an importer of an extracellular factor; however, we cannot exclude the possibilities that FtsEX acts either as an exporter of a membrane-tethered factor or as a lipoprotein release factor, as mentioned above (see Results). The variety of signals delivered by ABC transporters makes it technically challenging to discover the nature of the substance imported specifically by the FtsEX transporter. Multiple studies indicate that B. subtilis produces, secretes and imports several cell–cell signalling molecules that serve to monitor cell density and are required for entry into sporulation (Grossman and Losick, 1988; Perego and Hoch, 1996; Lazazzera, 2001; Perego and Brannigan, 2001). In preliminary experiments, we employed Fourier Transform InfraRed (FTIR) spectroscopy to determine whether there are differences in the chemical composition of supernatant fluids derived from sporulating wild-type versus mutant cells. FTIR spectroscopy is a useful tool for characterizing compounds, as in general, each compound within a mixture exhibits a distinct absorption spectrum that serves as a ‘fingerprint’, ultimately enabling its identification. Preliminary results indicate that conditioned medium of wild-type cells and that of FtsEX mutant cells exhibit detectably different spectra. In future studies, we intend to identify the distinctive compounds and test their effect on sporulation.

The existence of the FtsEX transporter during vegetative growth implies that the signal transduced by the transporter is required at low levels during vegetative growth, or that the transporter is kept ‘on alert’, ready to capture and transmit the sporulation signal at any time. Notably, FtsEX homologues exist in many Gram-positive and Gram-negative bacteria that do not undergo sporulation, suggesting that homologues must mediate divergent activities, a premise supported by several lines of evidence. In E. coli, FtsEX was shown to biochemically interact with FtsZ (Corbin et al., 2007), to localize to the septal ring, and to recruit several cell division proteins (Schmidt et al., 2004). In contrast, the B. subtilis counterpart localizes to all membranes and not exclusively to the division site. Moreover, depletion of E. coli FtsE results in inhibition of cell division and growth and causes the formation of filamentous cells (Taschner et al., 1988), a phenomenon not seen in B. subtilis FtsEX mutant cells grown vegetatively. Furthermore, the division defect observed in E. coli FtsE mutant cells can be suppressed by high salt concentrations (de Leeuw et al., 1999; Reddy, 2007), whereas the vegetative growth and sporulation phenotype of B. subtilis FtsEX mutant cells cannot be suppressed by increasing the osmotic strength (data not shown). Taken together, these differences support the view that the two homologous transporters mediate diverse activities in these distinct bacterial species. In E. coli, FtsEX may deliver a factor required to stabilize the division ring or a ring structural component, whereas the B. subtilis homologue may transport an inducer of sporulation. Whether or not these homologues transduce the same factor that is utilized differently in each bacterial species or deliver dissimilar factors remains to be resolved.

Experimental procedures

Strains and plasmids

Bacillus subtilis strains were derivatives of the wild-type strain PY79 (Youngman et al., 1984) and were constructed using standard methods (Harwood and Cutting, 1990). Strains are listed in Table S1. Plasmid constructions are described in the Supplementary material.

General methods

Growth and sporulation were carried out at 32°C. Cells were grown in hydrolyzed casein (CH) growth medium. The cultures were inoculated at 0.05 OD600 from an overnight culture. At mid logarithmic phase, sporulation was induced by suspending cells in Sterlini and Mandelstam medium (Sterlini and Mandelstam, 1969; Harwood and Cutting, 1990). Sporulation progression was determined by fluorescence microscopy. Sporulation efficiency was assessed by incubating cells in DSM at 37°C for 24 h, and then evaluating the number of spores by heat-resistant-colony-forming units. Heat resistance was assayed by incubating the spores 20 min at 80°C or 20 min at 90°C. Germination was performed at 37°C in LB medium and was monitored by time-course and time-lapse microscopy. For induction of IPTG-inducible promoters, IPTG was added at a final concentration of 20 μM. Additional general methods were carried out as described previously (Bejerano-Sagie et al., 2006).

Fluorescence microscopy

Fluorescence microscopy was carried out as described previously (Bejerano-Sagie et al., 2006). Briefly, samples (0.5 ml) of culture were removed, gently centrifuged and re-suspended in 10 μl of PBS ×1 (Phosphate-Buffered Saline) supplemented with the fluorescent membrane stain FM4-64 (Molecular Probes, Eugene, OR) at 1 μg ml−1. We have verified that this procedure of sample preparation has no effect on cell morphology either for wild-type or mutant cells. For observing FtsE–GFP, a chamber slide containing 1% agarose was used (Bejerano-Sagie et al., 2006). Cells were visualized and photographed using an Axioplan2 microscope (Zeiss) equipped with a CoolSnap HQ camera (Photometrics, Roper Scientific) or an Axioobserver Z1 microscope (Zeiss) equipped with a CoolSnap HQII camera (Photometrics, Roper Scientific). System control and image processing were performed using MetaMorph software (Molecular Devices).

Screening for mutants impaired in polar septation

In an attempt to identify novel genes that are required for polar division during sporulation, we performed a visual microscopic screen (using the fluorescent membrane dye FM4-64), searching for mutants impaired in polar septation upon a shift to sporulation conditions. We examined a collection of 80 strains bearing knockout alleles in Spo0A-controlled genes, as classified by genome-wide analyses (Fawcett et al., 2000; Molle et al., 2003; S.B.Y. and R.L., unpubl. data). This knockout collection was generated in PY79 by using a long-flanking-homology PCR strategy (Wach, 1996).

Acknowledgements

We thank R. Losick (Harvard University, USA) in whose laboratory the preliminary experiments for this study were performed and V. Molle (IBCP, France) for constructing the original ΔftsE strain. We thank D. Rudner (Harvard University, USA), Alex Rouvinski (Hebrew University, IL), and members of the Ben-Yehuda laboratory for valuable comments on the manuscript. This work was supported by the Focal Initiatives in Science and Technology, the Israel Science Foundation (FIRST, ISF) (Grant No. 203/06) to S.B.-Y., by the Bruno-Goldberg Endowment Foundation to S.B.-Y., and by the Robert A. Welch Foundation (E-1627) to M.F.

Ancillary