SGM Journals
Research Article

Characterization and subcellular localization of a bacterial flotillin homologue

Catriona Donovan, Marc Bramkamp

  • Institut für Biochemie, Universität zu Köln, Zülpicher Str. 47, D-50674 Köln, Germany
  • Correspondence
    Marc Bramkamp
    marc.bramkamp{at}uni-koeln.de

    • Microbiology 2009; 155(6):1786 · https://doi.org/10.1099/mic.0.025312-0

    View at publisher PubMed

    Abstract

    The process of endospore formation in Bacillus subtilis is complex, requiring the generation of two distinct cell types, a forespore and larger mother cell. The development of these cell types is controlled and regulated by cell type-specific gene expression, activated by a σ-factor cascade. Activation of these cell type-specific sigma factors is coupled with the completion of polar septation. Here, we describe a novel protein, YuaG, a eukaryotic reggie/flotillin homologue that is involved in the early stages of sporulation of the Gram-positive model organism B. subtilis. YuaG localizes in discrete foci in the membrane and is highly dynamic. Purification of detergent-resistant membranes revealed that YuaG is associated with negatively charged phospholipids, e.g. phosphatidylglycerol (PG) or cardiolipin (CL). However, localization of YuaG is not always dependent on PG/CL in vivo. A yuaG disruption strain shows a delay in the onset of sporulation along with reduced sporulation efficiency, where the spores develop to a certain stage and then appear to be trapped at this stage. Our results indicate that YuaG is involved in the early stage of spore development, probably playing a role in the signalling cascade at the onset of sporulation.

    Edited by: T. Abee

    INTRODUCTION

    Many proteins that were thought to exist exclusively in eukaryotes have, in fact, close homologues in prokaryotes. A milestone was certainly the discovery of a bacterial cytoskeleton (Jones et al., 2001). Today we know that, besides actin (MreB)- and tubulin (FtsZ)-like cytoskeletal structures, many proteins form polymer structures that serve as a scaffold for compartmentalization (Carballido-Lopez & Formstone, 2007; Graumann, 2007). In parallel to the new observations of subcellular organization of the bacterial cytoplasm, we also learned that the bacterial cell membrane is not the homogeneous liquid mosaic composed of lipids and proteins suggested by the original theory (Singer & Nicolson, 1972). In fact, it became apparent that membrane proteins are also localized to certain membrane areas. The most prominent region is certainly the curved membrane at the cell poles. Several proteins are known to localize exclusively to pole regions. Examples include the pole-determining protein DivIVA (Marston et al., 1998), bacterial chemoreceptors (Alley et al., 1992), transport proteins (Romantsov et al., 2008) and many more. Other membrane proteins, such as peptidoglycan-synthesizing proteins, localize in helical patterns along the cylinder of rod-shaped cells (Daniel & Errington, 2003). Similar to the non-homogeneous distribution of many membrane proteins, it has now been demonstrated conclusively that phospholipids and glycolipids also exhibit a laterally heterogeneous distribution within the membrane. The non-homogeneous distribution of phospholipids in bacterial membranes is best studied for negatively charged lipids, such as phosphatidylglycerol (PG) and cardiolipin (CL) (Kawai et al., 2004; Koppelman et al., 2001; Mileykovskaya & Dowhan, 2000; Nishibori et al., 2005). Various functions have been attributed to CL in bacterial cells. Negatively charged lipids such as PG and CL have been shown to play a role in the recruitment of membrane-integral proteins that contain a charged amphitropic helix (Matsumoto, 2001), in the activation of sensory proteins such as KdpD (Stallkamp et al., 1999) and in the localization and control of transporters involved in the osmostress response (Romantsov et al., 2008), and they are thought to be involved in the formation of non-bilayer structures that promote membrane invagination and fusion during division and sporulation events (Dowhan, 1997). Although the existence of heterogeneous lipid distribution in bacterial membranes is established, there is still controversy about lipid rafts in eukaryotic cells (Jacobson et al., 2007; Simons & Ikonen, 1997). Initial studies revealed that certain membrane compartments could not be solubilized with Triton X-100 at low temperatures. These membranes were termed detergent-resistant membranes (DRM) and they exhibit a buoyant density on sucrose gradients (Pralle et al., 2000). Interestingly, various proteins can be co-isolated with the DRM. Among these proteins, flotillins (or reggies) were found in the DRM and hence are termed ‘raft markers’ (Lang et al., 1998; Stuermer et al., 2001). Eukaryotic flotillins have conserved hydrophobic stretches and palmitoylation sites that target the protein to the DRM and anchor the proteins there (Morrow et al., 2002). Bacterial homologues of flotillins have been described (Langhorst et al., 2005; Zhang et al., 2005). Based on the primary sequence, the closest homologue of eukaryotic flotillins is found in the genus Bacillus. In the model Gram-positive bacterium Bacillus subtilis, the flotillin homologue, YuaG, is encoded in the yuaFGI operon. Expression of this operon is under control of the alternative sigma factor, σW (Huang et al., 1999). σW-controlled genes are upregulated upon envelope stress (Wiegert et al., 2001), usually mimicked by addition of alkali to the medium.

    A recent report about a flotillin homologue from Bacillus halodurans (Zhang et al., 2005) describes the isolation of the B. halodurans flotillin from DRM, suggesting that the bacterial flotillins may also serve as microdomain markers. A structure of YuaF has recently been published and similarities between YuaF and NfeD proteins were pointed out (Walker et al., 2008). NfeD proteins seem to coexist with SPFH domain-containing proteins. SPFH domains are found in all flotillins and hence a direct interaction between YuaF and YuaG was postulated (Walker et al., 2008).

    Here, we describe the subcellular distribution of the B. subtilis flotillin (YuaG) and its biological function. Our results show clearly that the σW-controlled yuaFGI operon is also transcriptionally upregulated during late stationary phase in a σW-independent manner. However, comparison of 10-N-nonyl-3,6-bis(dimethylamino)acridine orange (NAO)-stained CL/PG domains and YuaG localization revealed that YuaG is not found exclusively in PG/CL microdomains. A yuaG disruption mutant is impaired in sporulation, suggesting a role for the Bacillus flotillin in efficient cell-differentiation processes during spore formation.

    METHODS

    Media and growth conditions.

    Escherichia coli cells were grown at 37 °C in Luria–Bertani broth (LB) containing the appropriate antibiotics. Corynebacterium glutamicum cells were grown in brain–heart infusion medium (Difco) at 30 °C. B. subtilis strains were grown at 37 °C in CH medium (Partridge & Errington, 1993) or sporulation medium (Sterlini & Mandelstam, 1969), as indicated in the text. Antibiotics were used at the following concentrations when appropriate: spectinomycin, 50 μg ml−1; neomycin, 2.5 μg ml−1; erythromycin, 1 μg ml−1; chloramphenicol, 5 μg ml−1 (for B. subtilis); and carbenicillin, 100 μg ml−1 (for E. coli). Xylose was used at a concentration of 0.05–0.10 %, as indicated in the text.

    Bacterial strains, plasmids and oligonucleotides.

    Strains, plasmids and oligonucleotides used in this study are listed in Table 1⇓.

    Table 1.

    Strains, plasmids and oligonucleotides

    Strain construction.

    All plasmid cloning was carried out in E. coli DH5α. In the case of pSG1154-yuaG pSG-SNAP-yuaG, the whole yuaG gene was amplified by using primers yuaG-pSG1154_for and yuaG-pSG1154_rev or yuaG-SNAP_for and yuaG-SNAP_rev, respectively (Table 1⇑). Amplified DNA was purified, digested by using the indicated restriction enzymes and cloned into the pSG1154 or pSG-SNAP vectors. The recombinant plasmids were transformed into B. subtilis 168. Integration into the amyE locus by a double-crossover event and subsequent selection of spectinomycin-resistant strains yielded strains DB001 (YuaG–GFP+) and DB002 (YuaG–SNAP+). Similarly, pMUTIN4-yuaG′ construction involved amplification of a central region of the yuaG gene, using the primers yuaG-pMUTIN4_for and yuaG-pMUTIN4_rev (Table 1⇑) and cloning with HindIII and BamHI. Strain DB003 (yuaG : : pMUTIN4) was constructed by transformation of pMUTIN4-yuaG′ and subsequent selection for erythromycin resistance. The knockout of the yuaG allele was verified by PCR using primer pairs Pspac_for/yuaG-pMUTIN_rev and LacZ_rev/yuaG-pMUTIN_for. Isogenic strains harbouring a yuaG disruption and yuaG wild-type cells containing lacZ fusions to spoIIAB, spoIIQ and spoIVAB were constructed as follows. Chromosomal DNA from strains 700, BS5 : JH642 and 713 was transformed into yuaG : : pMUTIN4. Strains DB006, DB007 and DB008 were selected for chloramphenicol and erythromycin resistance and checked by PCR. A complementation strain, DB009, that expresses yuaG under the control of an inducible Pxyl promoter was constructed. To this end, the coding region of yuaG, including a Shine–Dalgarno sequence, was cloned into vector pJPR1, giving vector pJPR1-yuaG. Strain DB003 was transformed with plasmid pJPR1-yuaG and transformants were selected on chloramphenicol, giving strain DB009. Strains DB010 and DB011 were constructed by transformation of KP646 and KP647 chromosomal DNA into strain DB003. Transformants were selected on chloramphenicol. Strain DB012, expressing yuaG–SNAP under Pxyl control as the only allele of yuaG, was constructed by transformation of DB002 DNA into strain DB003 and subsequent selection on spectinomycin.

    A ‘knock-in’ mutant was constructed by amplification of the 3′ end of yuaG by using the primer pair yuaG-CFP_for and yuaG-pSG1154_rev. The PCR product was cloned with KpnI/HindIII into vector pSG1186 (Feucht & Lewis, 2001) to give plasmid pSG1186-yuaG. B. subtilis cells were transformed with pSG1186-yuaG and transformants were selected for chloramphenicol resistance. Integration of the plasmid was checked by PCR. The resulting strain (DB015) expresses a C-terminal YuaG–CFP fusion under control of the wild-type promoter. Strain DB016 (YuaG–CFP+ sigW : : neo) was constructed by transforming HB4247 (sigW : : neo) chromosomal DNA into strain DB015.

    NAO staining.

    Cell cultures were grown in CH medium with appropriate supplements. An aliquot of cells was stained with NAO to a final concentration of 100 nM and incubated at 37 °C with shaking for 20 min before mounting on an agarose-coated microscope slide.

    SNAP labelling.

    BG-430 (Covalys Biosciences AG) (excitation maximum, 429 nm; emission maximum, 484 nm) labelling solution was used to label cells of strain DB002. To this end, cells were pre-cultured in CH medium and 400 μl culture was incubated with 1 μl labelling solution at 37 °C with shaking for 30 min. Cells were washed twice in 400 μl CH medium and incubated at 37 °C with shaking for 15 min.

    Sporulation assays.

    Sporulation assays were carried out as described by Sterlini & Mandelstam (1969). Freshly grown cells were diluted serially in CH medium and incubated at 37 °C overnight. Cell culture with an OD600 of 1.0 was diluted to an OD600 of 0.1 in CH medium and grown to an OD600 of 0.8–1.0. Cells were sedimented at 7000 g at 37 °C for 4 min and resuspended in the same volume of warm sporulation medium. Samples were taken and flash-frozen at defined time intervals.

    Sporulation efficiency assay.

    Sporulation assays were carried out as described above. Cells were grown in sporulation medium for 9 h. Two aliquots of cells were diluted serially in sporulation medium. One aliquot was heated to 80 °C for 20 min and the other was heated to 37 °C for 20 min. Each serial dilution was plated onto nutrient agar plates in duplicate and grown at 37 °C overnight. The number of cells germinated from spores was compared with the total number of cells.

    Alkaline phosphatase assay.

    p-Nitrophenyl phosphate (333 μl of 1 mg ml−1) in Tris/HCl (pH 8) was added to 500 μl sample. The reaction was stopped by adding 333 μl 2 M NaOH when a yellow colour was visible or after 1 h. Samples were centrifuged in a tabletop centrifuge for 5 min. Absorption of the p-nitrophenolate anion (A420) was read against a blank. An enzyme unit is defined as the amount of enzyme releasing 1 nmol p-nitrophenol min−1 from the substrate at 30 °C (Grant, 1974).

    β-Galactosidase assay.

    β-Galactosidase assays were carried out essentially as described previously (Karow & Piggot, 1995). Briefly, 400 μl lysis solution (200 μg lysozyme ml−1, 100 μg DNase ml−1 and 1.25 % Triton X-100) was added to 200 μl sample and incubated for 10 min or until the cells were lysed. Two hundred microlitres of 4 mg ONPG ml−1 was added to samples. The reaction was stopped by adding 400 μl 1 M Na2CO3 when a yellow colour was seen or after 60 min. A420 was read against a blank.

    DRM fractionation.

    DRM fractions containing only integral proteins were separated as described previously (Salzer & Prohaska, 2001). Cells of strain DB001 were grown in CH medium containing 0.05 % xylose to an OD600 of 1.0, resuspended in ice-cold 1 % Triton X-100 in lysis buffer [10 mM Tris/HCl (pH 7.4), 1 μg RNase A ml−1, 1 μg DNase I ml−1, 0.5 mM PMSF, 5 mM MgCl2], disrupted by a French press, incubated for 20 min on ice (controls were incubated at 30 °C) and centrifuged at 15 000 g for 10 min at 4 °C. The pellet was resuspended in 800 μl ice-cold 1 % Triton X-100 in TBS (10 mM Tris/HCl, 150 mM NaCl, pH 7.5) and mixed with an equal volume of 80 % sucrose in 0.2 M Na2CO3. This suspension was transferred to ultracentrifuge tubes, overlaid with 4 ml 30 % sucrose in TBS plus 1 % Triton X-100 and then with 3 ml 10 % sucrose in TBS plus 1 % Triton X-100. The mixture was centrifuged in an SW41Ti rotor (Beckman) for 15 h at 4 °C (41 000 g). Fractions (1 ml) were collected from the top, TCA-precipitated and subjected to SDS-PAGE (12 % gel).

    Trichloroacetic acid (TCA) precipitation.

    One volume of 100 % TCA was added to 4 vols protein sample and incubated for 10 min at 4 °C. Samples were centrifuged at 10 000 g for 5 min at 4 °C. Supernatant was removed and the pellets were washed with 200 μl ice-cold acetone. Samples were centrifuged and washed again with acetone. The pellet was dried at 95 °C for 5–10 min and resuspended in buffer.

    Lipid extraction.

    Lipids were extracted from DRM fractions by using a modified Bligh–Dyer method (Bligh & Dyer, 1959) as described by Özcan et al. (2007). Cell-membrane fractions were incubated with chloroform/methanol (2 : 1, v/v) at 4 °C overnight and then filtered. Samples were centrifuged at 20 °C at 800 g for 10 min. The lower organic phase was isolated and washed with diethyl ether, a dextran gel slurry (Sephadex G-25 medium) and twice with chloroform. The extract was filtered and dried in a rotary evaporator, and samples were stored under nitrogen at −20 °C until they were analysed.

    TLC.

    For analysis of polar lipids, a chloroform/methanol/water (65 : 25 : 4, v/v) solvent system was used. This solvent system was added to the TLC tank at least 30 min before analysis. Lipid standards (Sigma-Aldrich) were diluted to a concentration of 5 mg ml−1. Silica gel 60 plates (Macherey & Nagel) were activated at 105 °C for 1 h. Sample (25 μl) and standards (5 μl) were spotted approximately 5 cm from the bottom of the plate. Plates were removed from the developing tank when the solvent front was about 2 cm from the top. The lipids were visualized by exposure to iodine vapours.

    Membrane preparation.

    Cells were grown in LB containing 0.05 % xylose to an OD600 of 0.8. Cells were centrifuged at 10 000 g for 10 min at 4 °C. The pellet was resuspended in 5 ml PBS (pH 7.4) containing EDTA-free protease inhibitor (Roche). Cells were disrupted by using a FastPrep (MP Biomedicals) and centrifuged at 10 000 g for 5 min at 4 °C. The supernatant was ultracentrifuged at 200 000 g for 30 min at 4 °C. The supernatant, representing the cytosolic fraction, was removed and stored at −20 °C until further analysis. The pellet was washed twice in 2 ml PBS buffer and ultracentrifuged at 200 000 g for 30 min at 4 °C. The pellet was finally resuspended in PBS (washed membrane fraction). The cytosolic and membrane fractions were subjected to SDS-PAGE (12 % gel) and immunoblotting.

    Immunoblotting.

    Samples were subjected to SDS-PAGE through a 12 % gel and blotted onto a PVF membrane. The blot was incubated with anti-GFP (1 : 2000) at 4 °C for at least 1 h. The blot was then washed with sodium phosphate buffer and incubated with the secondary antibody, anti-rabbit conjugated with alkaline phosphatase (1 : 10 000) at 4 °C for at least 1 h. The blot was again washed with sodium phosphate buffer and developed with NBT/BCIP. FtsH antibodies (a gift from T. Wiegert, Department of Genetics, University of Bayreuth, Germany) were used at a 1 : 10 000 dilution and antibodies against subunit C of ATP synthase (a gift from G. Deckers-Hebestreit, Department of Microbiology, University of Osnabrück, Germany) were used at a 1 : 2000 dilution. ATP synthase antibodies were raised against the E. coli enzyme, but we observed cross-reactivity with the homologous Bacillus protein and hence used it in this study to detect ATP synthase in Bacillus membranes.

    Fluorescence microscopy.

    Cells for microscopy were grown in CH medium with appropriate supplements and grown to the exponential-growth phase. An aliquot of cells was labelled, if necessary, with dyes as described above. For phase-contrast and fluorescence microscopy, 1–3 μl of a culture sample was placed on a microscope slide coated with a thin 1 % agarose layer and covered by a coverslip. For membrane or DNA staining, a 10 μl culture sample was mixed with 1 μl Nile red (12.5 μg ml−1; Molecular Probes) or 2 μl DAPI solution (1 μg ml−1, 50 % glycerol; Sigma), respectively. Images were taken on a Zeiss AxioImager M1 equipped with a Zeiss AxioCam HRm camera. GFP fluorescence was monitored by using filter set 38 HE eGFP, BG-430 fluorescence and CFP were monitored by using filter 47 HE CFP, red fluorescence (NAO, Nile red) was monitored by using filter 43 HE Cy3 and DAPI fluorescence was examined with filter set 49. An EC Plan-Neofluar ×100/1.3 Oil Ph3 objective was used. Digital images were acquired and analysed with AxioVision 4.6 software (Carl Zeiss). Final image preparation was done by using Adobe Photoshop 6.0 (Adobe Systems Inc.).

    RESULTS

    Subcellular localization of YuaG

    Many bacterial cells harbour a gene that encodes a flotillin-like protein. In B. subtilis, the yuaG gene encodes a protein that is highly similar (34.8 % identity and 55.6 % similarity to human flotillin-1) to eukaryotic flotillins. The yuaG gene is encoded in the yuaFGI operon and driven by a promoter in front of yuaF. This promoter is induced strongly by σW (Huang et al., 1999; Wiegert et al., 2001). Topology predictions of YuaG suggest that the protein has an N-terminal transmembrane helix. The N terminus is predicted to face the extracellular space, whilst the C terminus is predicted to be cytoplasmic.

    Eukaryotic flotillins are usually localized in a punctate pattern within the membrane (Lang et al., 1998). To study the subcellular distribution of the B. subtilis YuaG protein, we constructed a translational YuaG–GFP fusion. To this end, the coding sequence of yuaG was cloned into the pSG1154 vector (Lewis & Marston, 1999). The vector integrates into the chromosomal amyE locus and places the fusion gene under control of the inducible Pxyl promoter to give strain DB001. To visualize the distribution of YuaG–GFP, cells were grown in CH medium with various inducer concentrations (0.05–0.1 %). In vegetatively growing cells, YuaG–GFP localized in a punctate pattern within the cell membrane (Fig. 1b⇓). The frequency with which the YuaG–GFP foci formed was dependent on inducer concentration. Overexpression of YuaG had no obvious phenotype. To study colocalization of YuaG and, for example, NAO-stained membrane domains, we wanted to have a more versatile fluorescent tag fused to YuaG and therefore constructed a YuaG–SNAP fusion protein. SNAP-tag labelling offers the advantage of labelling the fusion protein with a variety of different fluorophores that are coupled to O6-alkylguanine substrates. Therefore, we removed the coding sequence for GFP from plasmid pSG1729 (Lewis & Marston, 1999) and replaced it with the coding sequence for the SNAP protein (Covalys Biosciences AG) (Keppler et al., 2004). The yuaG gene was cloned into pSG-SNAP and the resulting plasmid was transformed into B. subtilis. Transformants (DB002) carried the yuaG–SNAP fusion gene (note that this is an N-terminal fusion) at the ectopic amyE locus under the control of the Pxyl promoter. Similar to the YuaG–GFP fusion, the YuaG–SNAP fusion protein was distributed in discrete foci along the cell membrane (Fig. 1a⇓).

    Figure image not available in archive
    Fig. 1.

    Subcellular localization of YuaG. (a) Localization of YuaG–SNAP stained with BG-430 substrate in exponentially growing B. subtilis (strain DB002). Shown are images of YuaG–SNAP foci, membrane and a merged image side by side. Bar, 2 μm. (b) Localization of YuaG–GFP (strain DB001). (c, d) YuaG–CFP localization (strain DB015) in exponentially growing cells (c) and in stationary cells (d). Cells for (a) and (b) were grown in CH medium and induced with 0.05 % xylose; cells of strain DB015 (c, d) express YuaG–CFP as the only allele from the wild-type promoter and were grown in CH medium. (e) Membrane localization of YuaG–GFP in B. subtilis was analysed by cell fractionation into cell membrane and cytoplasm. Samples were separated by SDS-PAGE and subsequently analysed by immunoblotting using α-GFP. Lane A contains the washed membrane fraction and lane B contains the cytoplasm. *Expected migration site of YuaG–GFP; **potential oligomers. (f) Growth curve (□) of strain DB003 (yuaG : : pMUTIN4). β-Galactosidase activity (•) shows gene transcription of the yuaG promoter. Background activities of stationary-phase cells were subtracted.

    Expression of YuaG–GFP and YuaG–SNAP was driven by xylose induction and hence does not reflect the native transcription mediated by the yuaF promoter. To circumvent this problem, we constructed a ‘knock-in’ strain that expresses a YuaG–CFP fusion (strain DB015). Examination of YuaG–CFP foci revealed that only very few foci were visible during early exponential growth; however, an increasing number of foci were observed when cells entered the stationary phase (Fig. 1c⇑). Quantification of the YuaG–CFP foci revealed that the total number of foci increased steadily with the onset of the stationary-phase transition (Fig. 1d⇑; Table 2⇓).

    Table 2.

    Formation of YuaG–CFP foci in strain DB015

    Values are given as a percentage of cells (for each OD600, ≥150 cells have been counted).

    To address the question of how the YuaG foci were distributed in the 3D space of the cells, a series of Z stacks were made. After deconvolution of the obtained Z stacks (iterative algorithm; Zeiss AxioVision), a 3D reconstruction of the YuaG–GFP fluorescence was made. The resulting 3D image revealed that YuaG is localized in spots around the cytoplasmic membrane (Fig. 2b, c⇓). The spots are always individual, but it seems that they follow a spiralling track within the membrane. Obviously, the spots are not just distributed randomly.

    Figure image not available in archive
    Fig. 2.

    YuaG foci are dynamic. (a) Time-lapse analysis of YuaG–GFP foci. Cells of strain DB001 (YuaG–GFP+) were grown in CH medium with 0.05 % xylose. Exponentially growing cells were mounted on agarose-coated slides supplemented with CH medium. YuaG–GFP foci were analysed within the same cell for 8 min. Shown is a phase-contrast image (left side) of the cell. Time points where YuaG–GFP foci were imaged are indicated. Bar, 2 μm. (b) A Z-stack image series though three cells of strain DB001 (YuaG–GFP+). Bars, 2 μm. A 3D reconstruction of the Z stack and a cartoon of the cell outline can be seen in (c).

    To verify the membrane insertion of YuaG–GFP, cytoplasmic and membrane fractions of YuaG–GFP-expressing cells were prepared and subsequently analysed by immunoblotting. Using anti-GFP antibodies, the YuaG–GFP fusion protein was detected exclusively in the membrane fraction (Fig. 1e⇑). A prominent band around 66 kDa was observed in the membrane fraction, which fits with the expected size of the YuaG–GFP fusion protein. Interestingly, a higher-molecular-mass band (around 170 kDa) was also stained in the immunoblot. This might hint at a homo-oligomer of YuaG in the membrane. The cytoplasmic fraction did not show any signal in the immunoblot, confirming that YuaG is exclusively a membrane-integral protein.

    We wanted to test whether the putative N-terminal transmembrane domain of YuaG is necessary and/or sufficient for localization. Therefore, experiments were performed in which the transmembrane helix alone (aa 1–60) and the soluble part (aa 60–509) of the YuaG protein were expressed independently as GFP fusions. Neither YuaG truncation was able to localize to the membrane correctly and GFP fluorescence was only visible in the cytoplasm (data not shown).

    YuaG foci are dynamic

    We were interested to see whether the observed YuaG foci would be static or dynamic within the bacterial cell. Therefore, we used time-lapse analysis to visualize YuaG foci over a defined time period. The YuaG–GFP-expressing strain was grown to mid-exponential phase in CH medium supplemented with 0.05 % xylose. An aliquot of these cells was mounted on CH agar-coated slides and GFP fluorescence was monitored in a time-lapse assay. Results are shown for a series of images that were taken every 120 s (Fig. 2a⇑). It becomes immediately obvious that the YuaG–GFP foci are reorganized dynamically within the cell. Identical results have been obtained with the YuaG–CFP strain (DB015) that expresses YuaG–CFP under control of the wild-type promoter (data not shown).

    YuaG is co-purified with DRM

    Reports of a YuaG homologue from B. halodurans (Zhang et al., 2005) and of eukaryotic flotillins (Pralle et al., 2000) describe that flotillins reside in Triton X-100-insoluble membranes. However, DRM extraction from bacterial membranes is not an established technique compared with DRM purification from eukaryotic membranes. Therefore, we analysed the effect of Triton X-100 treatment under different conditions. We purified membrane fractions from YuaG–GFP-expressing B. subtilis (strain DB001; see Methods), extracted DRM with 1 % Triton X-100 at 4 °C and separated fractions by sucrose density-gradient centrifugation. The classical DRM extraction is performed at 4 °C, whereas higher temperatures result in complete solubilization of proteins. Therefore, we performed control experiments where the detergent treatment was performed at 30 °C and controlled for the effect of Triton X-100 in an experiment without detergent. Different fractions of the sucrose gradient were subjected to SDS-PAGE analysis and immunoblotting. We identified a band corresponding to YuaG–GFP in one of the top fractions of the sucrose gradient (Fig. 3a⇓), showing that YuaG is a buoyant protein that is co-purified with DRM. YuaG–GFP was also found in the higher-density fraction, but never in the bottom fractions. It might be possible that overexpression of YuaG–GFP via the Pxyl promoter resulted in an excess of YuaG and, hence, not all YuaG was inserted into the DRM. However, a similar distribution has been described for eukaryotic flotillin purified from DRM of phagosomes (Dermine et al., 2001). Interestingly, free GFP was detected only in the bottom fraction (band around 27 kDa, not shown), indicating that the band in the buoyant fraction contains exclusively proteins that were within DRM. In control experiments, we analysed the distribution of FtsH and ATP synthase (F0 part). Some of the FtsH was also found in the buoyant fractions, whereas the majority of ATP synthase subunit C was recovered in the bottom fraction (Fig. 3a⇓). In identical experiments without detergent, YuaG and FtsH were found in the high-density fractions, probably indicating higher-molecular-mass structures. Furthermore, we controlled for the effect of temperature in these experiments. When we performed the Triton X-100 treatment at 30 °C, the FtsH protein was shifted towards the bottom fractions (Fig. 3a⇓). YuaG was still found in floating fractions; however, the blot signals became very weak, indicating that the majority of the protein might have been completely solubilized. Immunoblot analysis of the supernatant obtained after detergent treatment at 30 °C revealed that YuaG was indeed almost completely solubilized under these conditions (data not shown).

    Figure image not available in archive
    Fig. 3.

    YuaG resides in DRM that are composed of negatively charged lipids. DRM of B. subtilis strain DB001 (YuaG–GFP+) were prepared and were fractionated by sucrose-gradient centrifugation. Nine fractions were collected and TCA-precipitated. The precipitated proteins were resuspended and subjected to SDS-PAGE analysis and immunoblotting. (a) Shown are immunoblots developed with different antibodies. YuaG–GFP was detected by using an α-GFP antibody; FtsH and ATP synthase were detected by using α-FtsH and α-ATPase antibodies, respectively. Extraction with 1 % Triton X-100 (TX-100) was done at 4 and 30 °C as indicated. A control without detergent at 4 °C is shown. FtsH and subunit C of the ATP synthase were used as control proteins. Note that YuaG–GFP is present in the buoyant fractions, but not in the high-density fraction when detergent was added at 4 °C, whereas in contrast, a significant amount of the ATP synthase and to some extent FtsH is always found in the high-density fractions. (b) Phospholipid content of the different sucrose-gradient fractions was analysed by TLC. Controls show migration of phospholipids PI, CL, PG and PE. Numbers underneath the TLC plate correspond to the fractions of the sucrose gradient. Fractions 3–4 and 5–6 were pooled and analysed together. TLC results show that PG is present in the low-buoyant fraction (arrowhead). The higher-density fractions are enriched in PE. Distribution of negatively charged phospholipids in bacterial membranes was detected by NAO fluorescence. (c) C. glutamicum possesses only negatively charged phospholipids. Fluorescence of NAO binding to negatively charged phospholipids (PL) was detected in the green channel, whereas binding of NAO to CL emitted red fluorescence. A merge is shown in the right panel. (d) Membranes of B. subtilis cells show a punctate distribution of negative PL and CL. Most of these membrane domains colocalize (merge). Bars, 1 μm.

    In order to analyse which phospholipids are present in the floating fractions, a TLC analysis was performed. Therefore, total lipids were extracted from different fractions of the sucrose gradient (see Methods). After separation of the lipids on TLC plates and subsequent iodine staining, it became apparent that the floating fractions on the sucrose gradient revealed a phospholipid spot that migrated identically to PG, whilst the less buoyant fractions were enriched in PE (Fig. 3b⇑). The amount of PE seems to decrease with decreasing sucrose-gradient density, whereas inversely, the concentration of PG increases.

    YuaG- and CL-rich microdomains have different localization patterns

    In eukaryotic cells, detergent resistance in membranes is due to the incorporation of cholesterol and sphingolipids. However, the B. subtilis membrane is devoid of these lipids. Previous reports demonstrated the existence of CL microdomains in B. subtilis membranes (Kawai et al., 2004, 2006; Matsumoto et al., 2006). Distribution of the CL domains in B. subtilis membranes was described as polar- or septal-localized; however, CL patches also formed foci (Kawai et al., 2004). Hence, we wanted to examine whether the YuaG foci that we observed (compare Fig. 1a–d⇑) would colocalize with the CL patches in B. subtilis membranes. Therefore, we first used the well-established technique of NAO labelling to visualize the CL domains in B. subtilis (Kawai et al., 2004; Matsumoto et al., 2006; Mileykovskaya & Dowhan, 2000). As described above, NAO binds electrostatically to negatively charged lipids such as PG and CL in B. subtilis membranes. Binding to PG is thought to occur in a molar ratio of 1 : 1 and leads to emission of a green fluorescence. In contrast, binding of NAO to CL results in a binding ratio of 2 : 1, giving a red fluorescence (Petit et al., 1992). To visualize CL domains, we grew B. subtilis wild-type cells in CH medium at 37 °C to mid-exponential phase and added NAO to a final concentration of 100 nM (see Methods). Subsequently, cells were analysed microscopically. We readily observed green- and red-fluorescent patches in the cell membrane, which colocalized (Fig. 3d⇑; Pearson–Manders coefficient, 0.92436 and 0.9846, respectively). This confirms earlier studies showing that B. subtilis membranes contain patches of negatively charged lipids (Kawai et al., 2004). As a control for the NAO stain, we used C. glutamicum cells, because C. glutamicum contains only negatively charged phospholipids and has a high CL content (Özcan et al., 2007). Exponentially growing C. glutamicum cells were stained with NAO. Examination of the green and red fluorescence revealed that the plasma membrane of C. glutamicum was labelled uniformly with NAO (Fig. 3c⇑). Thus, in contrast to B. subtilis, C. glutamicum possesses no PG/CL microdomains. This control suggests that the inhomogeneous, patch-like distribution of PG/CL domains in B. subtilis is not a staining artefact. However, we realized that NAO-stained cells also exhibited a fluorescence that was detectable with the cyan filter set (Zeiss 47 HE, 436 nm/480 nm). Hence, colocalization experiments with the YuaG–CFP or the YuaG–SNAP strain were hampered. Therefore, we decided to analyse the distribution of YuaG–CFP during sporulation. It was described previously that almost all CL is localized to the prespore membrane in sporulating bacilli (Kawai et al., 2006). We found that sporulating cells that were stained with NAO exhibited a strong fluorescence at the prespore membrane (Fig. 4⇓), confirming earlier results. In contrast, YuaG–CFP foci (the same was found for YuaG–GFP, data not shown) were still moving dynamically along the cell membrane (Fig. 4⇓, right panel). Localization of YuaG–CFP in sporulating cells suggests that YuaG is restricted to the mother-cell membrane.

    Figure image not available in archive
    Fig. 4.

    Localization of YuaG–CFP in sporangia. Cells of strain DB015 (YuaG–CFP+) were induced to sporulate. Localization of YuaG–CFP foci is depicted in cyan and membranes are shown in red (Nile red). Shown are characteristic cells 2 h (top panel), 4 h (middle panel) and 9 h (lower panel) after the onset of sporulation. Note that, in the lower panel, a phase-contrast image is shown in order to illustrate the refractive mature endospore. Localization of CL/PG in sporangia of B. subtilis wild-type cells visualized by NAO staining is shown in the right panel (green). Cartoons give the outline of the cells.

    Quantification of yuaG expression in vivo

    A yuaG disruption mutant was constructed by insertion of the pMUTIN4 plasmid (Vagner et al., 1998) into the yuaG gene, giving strain DB003. The correct insertion of the pMUTIN4 plasmid was checked by PCR using specific primer pairs (see Methods). With the pMUTIN4 integration, we had a tool in hand that allowed assaying the transcription of YuaG during growth of strain DB003. Integration of pMUTIN4 places the lacZ gene under the control of the native yuaG promoter, which is in front of yuaF (Wiegert et al., 2001). Growing cells were collected at different time points and β-galactosidase activity was measured. The results are shown in Fig. 1(f)⇑. Transcription of the yuaFGI operon starts during stationary phase (Fig. 1f⇑). During later stages of sporulation, we observed weak (up to 6 Miller units) transcription of the yuaFGI operon (data not shown). This notion is important because we used the yuaG disruption mutant (DB003) to assay different sporulation-specific promoter activities, again with lacZ activity (Fig. 5b⇓, ▴). This was necessary because we have not yet succeeded in constructing a clean yuaG-null mutant. The reasons for this are not yet understood. We also analysed expression of YuaG in vivo by counting YuaG–CFP foci during growth of strain DB015. Similar to the results obtained with lacZ reporter-gene expression, we found that the number of YuaG–CFP foci (in strain DB015) increased dramatically during stationary phase (Table 2⇑). The number of YuaG foci also increased upon alkaline stress (data not shown), suggesting increased expression under these conditions, as described previously (Wiegert et al., 2001). We also analysed the expression of YuaG–CFP in a strain background lacking sigW (strain DB016) and found that YuaG–CFP foci were present in the stationary phase, comparable to wild-type cells (data not shown).

    Figure image not available in archive
    Fig. 5.

    YuaG is involved in sporulation. (a) A yuaG disruption mutant (strain DB003) was assayed for alkaline phosphatase production during sporulation. Alkaline phosphatase production of a wild-type control (▪) and the yuaG mutant strain (□) is shown. The x-axis shows the time after resuspension in sporulation medium; the y-axis shows enzymic activity in nmol ml−1. (b–d) Sporulation block was tested with isogenic strains harbouring a yuaG disruption (yuaG : : pMUTIN4, □) and yuaG wild-type cells (▪) containing lacZ fusions to spoIIAB (b, strain DB008), spoIIQ (c, strain DB006) and spoIVAB (d, strain DB007). Strains were induced to sporulate and samples were taken at intervals to analyse β-galactosidase activity. As a negative control, β-galactosidase activity for yuaG : : pMUTIN4 (strain DB003) is indicated by ▴ in (b). All experiments were performed at least three times independently, with similar results.

    YuaG is involved in sporulation

    As eukaryotic flotillins are often involved in developmental processes, we wanted to analyse yuaG mutant cells during sporulation. We reasoned that proteins that are expressed during stationary-phase transition could be involved in the onset of sporulation. We assayed the production of alkaline phosphatase (Sterlini & Mandelstam, 1969) in wild-type B. subtilis 168 and in the yuaG mutant strain DB003. Reproducibly, the production of alkaline phosphatase was delayed in strains lacking a functional YuaG protein (Fig. 5a⇑).

    As the onset of sporulation is controlled by the master regulator Spo0A, we set out to investigate the phosphorylation state of Spo0A. Transcription of spoIIAB is driven directly by high levels of Spo0A∼P. In order to measure indirectly the level of Spo0A∼P, we utilized a spoIIAB–lacZ fusion originating from strain 700 (Partridge et al., 1991). To this end, DNA of strain 700 was transformed into DB003 to give strain DB008 (spoIIAB-lacZ ΔyuaG). Strains 700 and DB008 were induced to sporulate. LacZ production was monitored constantly. Results are shown in Fig. 5(b)⇑. The onset of spoIIAB transcription was similar in strains 700 and DB008 (ΔyuaG). However, the peak of β-galactosidase activity was reached much later in the absence of YuaG.

    Next, we combined a spoIIQ–lacZ fusion originating from strain BS5 : JH642 (Londono-Vallejo et al., 1997) with the yuaG mutant strain. Chromosomal DNA from strain BS5 : JH642 was transformed into strain DB003, giving strain DB006 (ΔyuaG spoIIQ-lacZ). The onset of β-galactosidase activity was delayed in strain DB006 (Fig. 5c⇑), indicating that progression of sporulation is delayed slightly in a strain lacking yuaG.

    As a third, independent experiment to assay the involvement of YuaG in sporulation, we performed promoter–lacZ assays with a spoIVA–lacZ construct. Therefore, chromosomal DNA of strain 713 was transformed into the yuaG mutant strain (DB003). Strains 713 and DB007 (ΔyuaG spoIVA-lacZ) were induced to sporulate and β-galactosidase activity was monitored. SpoIVA is transcribed from a σE-dependent promoter and thus depends on a closed spore septum (Stevens et al., 1992). In the absence of YuaG, the onset of spoIVA transcription is delayed compared with the control strain 713 (Fig. 5d⇑). This result also supports the notion that sporulation is delayed and that the asymmetrical septum is not formed as efficiently in the absence of YuaG compared with wild-type cells.

    Control experiments were designed to test whether expression of YuaG in trans would recover the ΔyuaG phenotype. To this end, strain DB009, which expresses yuaG under Pxyl control as the only allele, was constructed. As an effective test for correct onset of sporulation, we counted the number of asymmetrical septa formed at different time points after the onset of sporulation. The results are summarized in Fig. 6(a)⇓. Wild-type cells start to produce asymmetrical septa at around 90 min. In contrast, cells of strain DB003 (ΔyuaG) had almost no visible septa at this stage. Insertion of pMUTIN4 into the yuaG gene could cause polar effects on the downstream gene yuaI that might result in this phenotype. The pMUTIN4 insertion places yuaI under the control of an IPTG-inducible Pspac promoter (Vagner et al., 1998), hence addition of IPTG might alter the profile. The onset of sporulation in strain DB003 with IPTG present was shifted slightly (Fig. 6a⇓), which might indicate that YuaI could have some function with YuaG; however, the main phenotype observed in strain DB003 is due to loss of YuaG. When YuaG or YuaG–SNAP was expressed in trans upon addition of 0.05 % xylose (higher concentrations of inducer did not increase the complementation rate) in strains DB009 and DB012, respectively, the sporulation phenotype was recovered to a significant degree (Fig. 6a⇓).

    Figure image not available in archive
    Fig. 6.

    Loss of YuaG leads to ineffective production of asymmetrical septa. (a) As a measure of effective onset of sporulation, the formation of asymmetrical septa (percentage of total cells) was quantified for different strains. Wild-type (•), strain DB003 (ΔyuaG) (□), strain DB003 plus 1 mM IPTG (▵), strain DB012 (YuaG–SNAP+ ΔyuaG) plus 1 mM IPTG (⧫) and strain DB009 (YuaG+ ΔyuaG) supplemented with xylose (0.05 %) and 1 mM IPTG (▴). (b) Strains KP646 and DB010 (ΔyuaG) express GFP under the control of the σF-dependent PspoIIQ promoter and hence GFP expression is restricted to the prespore compartment [cf. (d)]. The intensity of the GFP fluorescence was quantified at different time points after the onset of sporulation. Filled bars represent results obtained for a yuaG wild-type strain (KP646) and empty bars represent results for a strain lacking yuaG (DB010). (c) Strains KP647 and DB011 (ΔyuaG) express GFP under the control of the σE-dependent PspoIID promoter. GFP expression is restricted to the mother-cell compartment [cf. (c)]. The intensity of the GFP fluorescence was quantified at different time points after the onset of sporulation (n=20–40 sporangia). Filled bars represent results obtained for a yuaG wild-type strain (KP647) and empty bars represent results for a strain lacking yuaG (DB011). (d) Shown is a typical field of cells expressing GFP (green) from the σF-dependent PspoIIQ promoter either in the wild-type background (left) or in the yuaG mutant background (right). (e) Shown is a typical field of cells expressing GFP (green) from the σE-dependent PspoIID promoter either in the wild-type background (left) or in the yuaG mutant background (right). Membranes are depicted in red. Arrows point to malformed prespore compartments that seem to have a defect in membrane fusion. Note that disporic cells can be seen in (d) and (e) if YuaG is absent.

    In order to gain more data that can be quantified, we used strains that produce GFP under the control of spoIIQ- and spoIID-specific promoters. PspoIIQ is activated by the prespore-specific sigma factor σF and PspoIID by the mother cell-specific sigma factor σE. Strains KP646 and KP647 express GFP from an ectopic locus from PspoIIQ and PspoIID, respectively (Sharp & Pogliano, 2002). PspoIIQ-driven GFP expression is restricted to the prespore compartment (Fig. 6d⇑), whereas PspoIID-controlled GFP expression is only in the mother cell (Sharp & Pogliano, 2002) (Fig. 6e⇑). We combined the yuaG disruption mutant with the KP646 and KP647 backgrounds, giving strains DB10 and DB11, respectively. After induction of sporulation, we quantified the amount of GFP fluorescence, counted the number of cells expressing GFP and recorded the timing of the onset of GFP expression. The results are summarized in Fig. 6(b–e)⇑. Loss of YuaG leads to a decrease of σE- and σF-driven GFP expression, indicating that the closure of the asymmetrical septum is less efficient in cells that lack YuaG (Fig. 6b, c⇑). Strikingly, we also observed a significant increase in disporic cells [cells in which prespore compartments are formed at both cell poles; see Fig. 6⇑(d, e)]. Usually, disporic cells are only seen in σE mutants (Illing & Errington, 1991; Lewis et al., 1994). This observation shows that loss of YuaG leads to late and inefficient activation of σE. Quantification of the production of heat-resistant spores revealed that the yuaG mutant strain has 66 % reduced sporulation efficiency compared with that of B. subtilis 168. It should be noted that 85 % (114 colonies tested) of the colonies that grew out from spores derived from the yuaG mutant strain lost the pMUTIN4 plasmid (percentage values of viable spores were therefore only based on the spores that maintained the pMUTIN integration into the yuaG locus). Because of the Campbell-type integration mechanism of the plasmid, the disruption can be reverted with the simultaneous loss of plasmid integration. Therefore, all colonies that grew from spores derived from the yuaG mutant strain were plated a second time on erythromycin plates. The high frequency with which we observed loss of the pMUTIN integration in the yuaG mutant strain during the sporulation assays supports the idea that YuaG plays an important role during sporulation of B. subtilis.

    DISCUSSION

    Here, we describe the characterization of a bacterial flotillin homologue from the Gram-positive model bacterium B. subtilis. Despite a number of publications describing the putative functions of flotillins in eukaryotes, their function even in well-understood systems remains largely obscure. A role of flotillin in the secretion of Wnt and Hedgehog, ensuring effective spreading of these signalling molecules, has been described (Katanaev et al., 2008). Flotillins are involved in amyloid precursor protein clustering (Schneider et al., 2008) and T-cell activation (Simons & Toomre, 2000). Recently, it has been shown that flotillins regulate cytoskeletal remodelling during neuronal differentiation (Langhorst et al., 2008). Hence, directly or indirectly, eukaryotic flotillins are involved in developmental processes. Thus, it seems not surprising that the closest homologues of eukaryotic flotillins are found in Bacillus species such as B. subtilis. Bacilli are endospore-forming bacteria. Sporulation is a simple developmental programme of cell differentiation in which a cell differentiates into a spore compartment and a mother cell. This process has similarities to endocytotic processes and requires a complex signalling network (reviewed by Errington, 2003). Sporulation is a severe commitment and is consequently regulated by a complex network. Different sporulation signals (e.g. nutrient shortage) are sensed by signal histidine kinases that feed phosphoryl goups into a phosphorelay. The final acceptor is the master regulator Spo0A. Spo0A∼P acts as a transcriptional activator of a large regulon. Activation of the Spo0A regulon prepares the cell for spore formation. Here, we show that the B. subtilis flotillin (YuaG) plays a role in Spo0A∼P formation. Although a yuaG knockout strain is only slightly impaired in the production of viable spores (34 % compared with wild-type cells), we found that the onset of sporulation is delayed. Formation of asymmetrical septa was delayed in the yuaG mutant strain, and promoter–lacZ fusions confirmed that transcription of the spoIIAB locus was initiated later in the yuaG mutant than in the wild-type. This suggests that the phosphorylation cascade leading to phosphorylation of the master regulator Spo0A is less efficient in the absence of YuaG. When a threshold level of Spo0A∼P is reached, Spo0A∼P initiates transcription directly of the Spo0A regulon (of which spoIIAB is a part) (Molle et al., 2003). Hence, a delay in spoIIAB transcription is a direct measure of the cellular level of Spo0A∼P. Thus, YuaG might have a role in a signalling cascade early on in sporulation. At this stage, we can only speculate whether YuaG is involved directly in the phosphorelay that leads to Spo0A∼P formation or whether YuaG complexes are modulating the membrane and/or membrane proteins, leading to efficient sensing of environmental signals leading to sporulation. It is known that eukaryotic flotillins can be phosphorylated by EGF-induced tyrosine phosphorylation (Neumann-Giesen et al., 2007). Similar to flotillins, the corresponding Src kinases were also shown to reside in DRM (de Diesbach et al., 2008); thus, concentrating kinases and flotillins in rafts might provide a platform where signal transduction is enhanced. However, so far we have not obtained evidence that YuaG is phosphorylated in B. subtilis. In eukaryotic cells, flotillins have been described as so-called ‘raft markers’ because the protein was identified in a DRM fraction isolated from murine lung tissue and it was found in the buoyant fraction of a sucrose density gradient (hence the name flotillin) (Lang et al., 1998). We show that B. subtilis YuaG can also be isolated from DRM and that it is associated with negatively charged phospholipids. Although some of the FtsH protease was detected in the buoyant fractions, the majority of FtsH was found in the high-density fractions, whereas the ATP synthase was less prominent in low-density fractions. Thus, FtsH and ATP synthase behaved differently from YuaG in the DRM-extraction experiments. Lipid analysis of the sucrose gradients obtained after Triton X-100 treatment revealed that the floating fractions were enriched in negatively charged PG, whereas the less buoyant fractions were enriched in the neutral phospholipid PE. Interestingly, localization studies of FtsH showed that the protein is enriched at the constricting septa (Wehrl et al., 2000), where PG/CL concentrations are high (Kawai et al., 2004). Probably, the localization of FtsH to sites with negatively charged lipids results in the isolation of a fraction of FtsH in the low-density fractions. ATP synthase is localized within the entire cytoplasmic membrane and is therefore unlikely to be concentrated in lipid microdomains (Johnson et al., 2004).

    The punctate localization of YuaG and the fact that eukaryotic homologues colocalize with specialized membrane microdomains tempted us to examine whether YuaG would colocalize with CL-rich domains that were described for B. subtilis (Kawai et al., 2004). Visualization of CL domains in bacterial membranes has been achieved by using the lipophilic dye NAO (Kawai et al., 2004; Mileykovskaya & Dowhan, 2000). Similar to previous reports, we observed that binding of NAO to bacterial membranes results in emission of a green and a red fluorescence. This effect has been attributed to binding of NAO to either PG (green fluorescence) or CL (red fluorescence) (Gallet et al., 1995; Mileykovskaya et al., 2001; Petit et al., 1992). We could confirm earlier results that CL forms microdomains in B. subtilis (Kawai et al., 2004). In contrast, NAO-stained microdomains have not been observed in the control strain C. glutamicum. We found, in B. subtilis, that the foci of green and red fluorescence colocalize, suggesting that the patches stained by NAO contain negatively charged phospholipids, but probably not exclusively CL. These foci were also found along the lateral axis of the cells. Thus, the localization of YuaG foci and CL domains has a similar pattern in vegetative cells. During sporulation, most of the CL and PG is localized to the prespore membrane; however, YuaG foci remained scattered all over the mother-cell membrane. Therefore, we conclude that YuaG is at least not always localized to CL/PG microdomains. In order to test whether the bacterial flotillin needs the N-terminal transmembrane domain to cluster into punctate foci, we constructed truncation mutants of YuaG. Experiments with truncated YuaG versions have shown that the transmembrane domain and the soluble domains of YuaG are necessary for correct subcellular localization.

    Previous reports described expression of the yuaFGI operon by the alternative sigma factor σW (Huang et al., 1999). Induction of σW-dependent gene expression has been attributed to envelope stress, such as alkaline shock. Here, we showed that transcription of the yuaFGI operon is upregulated in B. subtilis cells during the stationary-phase transition without external envelope stress. YuaG foci were observed in a strain lacking sigW; hence, it can be concluded that the yuaFGI operon has a σW-independent activation during the stationary phase.

    Acknowledgments

    We thank Drs Kit Pogliano (Division of Biological Sciences Faculty, University of California at San Diego, CA, USA), Ling Juan Wu, Jeff Errington (both at the Institute for Cell and Molecular Biosciences, University of Newcastle, UK) and John Helmann (Department of Microbiology, Cornell University, NY, USA) for providing strains. We acknowledge Drs Gabriele Deckers-Hebestreit (Department of Microbiology, University of Osnabrück, Germany) and Thomas Wiegert (Department of Genetics, University of Bayreuth, Germany) for the kind gift of antibodies. Anja Wittmann, Inga Wadenpohl and Frank Bürmann at the Institute of Biochemistry, University of Cologne, Germany, are acknowledged for help with strain construction. Dr Reinhard Krämer (Institute of Biochemistry, University of Cologne, Germany) is acknowledged for continuous support. This work was supported financially by the Deutsche Forschungsgemeinschaft (SFB 635, project C6).

    References

    1. Alley, M. R., Maddock, J. R. & Shapiro, L. (1992). Polar localization of a bacterial chemoreceptor. Genes Dev 6, 825–836.
    2. Bligh, E. G. & Dyer, W. J. (1959). A rapid method of total lipid extraction and purification. Can J Biochem Physiol 37, 911–917.
    3. Cao, M., Bernat, B. A., Wang, Z., Armstrong, R. N. & Helmann, J. D. (2001). FosB, a cysteine-dependent fosfomycin resistance protein under the control of σW, an extracytoplasmic-function sigma factor in Bacillus subtilis. J Bacteriol 183, 2380–2383.
    4. Carballido-Lopez, R. & Formstone, A. (2007). Shape determination in Bacillus subtilis. Curr Opin Microbiol 10, 611–616.
    5. Daniel, R. A. & Errington, J. (2003). Control of cell morphogenesis in bacteria: two distinct ways to make a rod-shaped cell. Cell 113, 767–776.
    6. de Diesbach, P., Medts, T., Carpentier, S., D'Auria, L., Van Der Smissen, P., Platek, A., Mettlen, M., Caplanusi, A., van den Hove, M. F. & other authors (2008). Differential subcellular membrane recruitment of Src may specify its downstream signalling. Exp Cell Res 314, 1465–1479.
    7. Dermine, J. F., Duclos, S., Garin, J., St-Louis, F., Rea, S., Parton, R. G. & Desjardins, M. (2001). Flotillin-1-enriched lipid raft domains accumulate on maturing phagosomes. J Biol Chem 276, 18507–18512.
    8. Dowhan, W. (1997). Molecular basis for membrane phospholipid diversity: why are there so many lipids? Annu Rev Biochem 66, 199–232.
    9. Errington, J. (2003). Regulation of endospore formation in Bacillus subtilis. Nat Rev Microbiol 1, 117–126.
    10. Feucht, A. & Lewis, P. J. (2001). Improved plasmid vectors for the production of multiple fluorescent protein fusions in Bacillus subtilis. Gene 264, 289–297.
    11. Gallet, P. F., Maftah, A., Petit, J. M., Denis-Gay, M. & Julien, R. (1995). Direct cardiolipin assay in yeast using the red fluorescence emission of 10-N-nonyl acridine orange. Eur J Biochem 228, 113–119.
    12. Grant, W. D. (1974). Sporulation in Bacillus subtilis 168. Control of synthesis of alkaline phosphatase. J Gen Microbiol 82, 363–369.
    13. Graumann, P. L. (2007). Cytoskeletal elements in bacteria. Annu Rev Microbiol 61, 589–618.
    14. Huang, X., Gaballa, A., Cao, M. & Helmann, J. D. (1999). Identification of target promoters for the Bacillus subtilis extracytoplasmic function sigma factor, σW. Mol Microbiol 31, 361–371.
    15. Illing, N. & Errington, J. (1991). Genetic regulation of morphogenesis in Bacillus subtilis: roles of σE and σF in prespore engulfment. J Bacteriol 173, 3159–3169.
    16. Jacobson, K., Mouritsen, O. G. & Anderson, R. G. (2007). Lipid rafts: at a crossroad between cell biology and physics. Nat Cell Biol 9, 7–14.
    17. Johnson, A. S., van Horck, S. & Lewis, P. J. (2004). Dynamic localization of membrane proteins in Bacillus subtilis. Microbiology 150, 2815–2824.
    18. Jones, L. J., Carballido-Lopez, R. & Errington, J. (2001). Control of cell shape in bacteria: helical, actin-like filaments in Bacillus subtilis. Cell 104, 913–922.
    19. Karow, M. L. & Piggot, P. J. (1995). Construction of gusA transcriptional fusion vectors for Bacillus subtilis and their utilization for studies of spore formation. Gene 163, 69–74.
    20. Katanaev, V. L., Solis, G. P., Hausmann, G., Buestorf, S., Katanayeva, N., Schrock, Y., Stuermer, C. A. & Basler, K. (2008). Reggie-1/flotillin-2 promotes secretion of the long-range signalling forms of Wingless and Hedgehog in Drosophila. EMBO J 27, 509–521.
    21. Kawai, F., Shoda, M., Harashima, R., Sadaie, Y., Hara, H. & Matsumoto, K. (2004). Cardiolipin domains in Bacillus subtilis Marburg membranes. J Bacteriol 186, 1475–1483.
    22. Kawai, F., Hara, H., Takamatsu, H., Watabe, K. & Matsumoto, K. (2006). Cardiolipin enrichment in spore membranes and its involvement in germination of Bacillus subtilis Marburg. Genes Genet Syst 81, 69–76.
    23. Keppler, A., Kindermann, M., Gendreizig, S., Pick, H., Vogel, H. & Johnsson, K. (2004). Labeling of fusion proteins of O6-alkylguanine-DNA alkyltransferase with small molecules in vivo and in vitro. Methods 32, 437–444.
    24. Koppelman, C. M., Den Blaauwen, T., Duursma, M. C., Heeren, R. M. & Nanninga, N. (2001). Escherichia coli minicell membranes are enriched in cardiolipin. J Bacteriol 183, 6144–6147.
    25. Lang, D. M., Lommel, S., Jung, M., Ankerhold, R., Petrausch, B., Laessing, U., Wiechers, M. F., Plattner, H. & Stuermer, C. A. (1998). Identification of reggie-1 and reggie-2 as plasmamembrane-associated proteins which cocluster with activated GPI-anchored cell adhesion molecules in non-caveolar micropatches in neurons. J Neurobiol 37, 502–523.
    26. Langhorst, M. F., Reuter, A. & Stuermer, C. A. (2005). Scaffolding microdomains and beyond: the function of reggie/flotillin proteins. Cell Mol Life Sci 62, 2228–2240.
    27. Langhorst, M. F., Jaeger, F. A., Mueller, S., Sven Hartmann, L., Luxenhofer, G. & Stuermer, C. A. (2008). Reggies/flotillins regulate cytoskeletal remodeling during neuronal differentiation via CAP/ponsin and Rho GTPases. Eur J Cell Biol 87, 921–931.
    28. Lewis, P. J. & Marston, A. L. (1999). GFP vectors for controlled expression and dual labeling of protein fusions in Bacillus subtilis. Gene 227, 101–110.
    29. Lewis, P. J., Partridge, S. R. & Errington, J. (1994). Sigma factors, asymmetry, and the determination of cell fate in Bacillus subtilis. Proc Natl Acad Sci U S A 91, 3849–3853.
    30. Londono-Vallejo, J. A., Frehel, C. & Stragier, P. (1997). SpoIIQ, a forespore-expressed gene required for engulfment in Bacillus subtilis. Mol Microbiol 24, 29–39.
    31. Marston, A. L., Thomaides, H. B., Edwards, D. H., Sharpe, M. E. & Errington, J. (1998). Polar localization of the MinD protein of Bacillus subtilis and its role in selection of the mid-cell division site. Genes Dev 12, 3419–3430.
    32. Matsumoto, K. (2001). Dispensable nature of phosphatidylglycerol in Escherichia coli: dual roles of anionic phospholipids. Mol Microbiol 39, 1427–1433.
    33. Matsumoto, K., Kusaka, J., Nishibori, A. & Hara, H. (2006). Lipid domains in bacterial membranes. Mol Microbiol 61, 1110–1117.
    34. Mileykovskaya, E. & Dowhan, W. (2000). Visualization of phospholipid domains in Escherichia coli by using the cardiolipin-specific fluorescent dye 10-N-nonyl acridine orange. J Bacteriol 182, 1172–1175.
    35. Mileykovskaya, E., Dowhan, W., Birke, R. L., Zheng, D., Lutterodt, L. & Haines, T. H. (2001). Cardiolipin binds nonyl acridine orange by aggregating the dye at exposed hydrophobic domains on bilayer surfaces. FEBS Lett 507, 187–190.
    36. Molle, V., Fujita, M., Jensen, S. T., Eichenberger, P., Gonzalez-Pastor, J. E., Liu, J. S. & Losick, R. (2003). The Spo0A regulon of Bacillus subtilis. Mol Microbiol 50, 1683–1701.
    37. Morrow, I. C., Rea, S., Martin, S., Prior, I. A., Prohaska, R., Hancock, J. F., James, D. E. & Parton, R. G. (2002). Flotillin-1/reggie-2 traffics to surface raft domains via a novel Golgi-independent pathway. Identification of a novel membrane targeting domain and a role for palmitoylation. J Biol Chem 277, 48834–48841.
    38. Neumann-Giesen, C., Fernow, I., Amaddii, M. & Tikkanen, R. (2007). Role of EGF-induced tyrosine phosphorylation of reggie-1/flotillin-2 in cell spreading and signaling to the actin cytoskeleton. J Cell Sci 120, 395–406.
    39. Nishibori, A., Kusaka, J., Hara, H., Umeda, M. & Matsumoto, K. (2005). Phosphatidylethanolamine domains and localization of phospholipid synthases in Bacillus subtilis membranes. J Bacteriol 187, 2163–2174.
    40. Özcan, N., Ejsing, C. S., Shevchenko, A., Lipski, A., Morbach, S. & Kramer, R. (2007). Osmolality, temperature, and membrane lipid composition modulate the activity of betaine transporter BetP in Corynebacterium glutamicum. J Bacteriol 189, 7485–7496.
    41. Partridge, S. R. & Errington, J. (1993). The importance of morphological events and intercellular interactions in the regulation of prespore-specific gene expression during sporulation in Bacillus subtilis. Mol Microbiol 8, 945–955.
    42. Partridge, S. R., Foulger, D. & Errington, J. (1991). The role of σF in prespore-specific transcription in Bacillus subtilis. Mol Microbiol 5, 757–767.
    43. Petit, J. M., Maftah, A., Ratinaud, M. H. & Julien, R. (1992). 10N-Nonyl acridine orange interacts with cardiolipin and allows the quantification of this phospholipid in isolated mitochondria. Eur J Biochem 209, 267–273.
    44. Pralle, A., Keller, P., Florin, E. L., Simons, K. & Horber, J. K. (2000). Sphingolipid-cholesterol rafts diffuse as small entities in the plasma membrane of mammalian cells. J Cell Biol 148, 997–1008.
    45. Romantsov, T., Stalker, L., Culham, D. E. & Wood, J. M. (2008). Cardiolipin controls the osmotic stress response and the subcellular location of transporter prop in Escherichia coli. J Biol Chem 283, 12314–12323.
    46. Salzer, U. & Prohaska, R. (2001). Stomatin, flotillin-1, and flotillin-2 are major integral proteins of erythrocyte lipid rafts. Blood 97, 1141–1143.
    47. Schneider, A., Rajendran, L., Honsho, M., Gralle, M., Donnert, G., Wouters, F., Hell, S. W. & Simons, M. (2008). Flotillin-dependent clustering of the amyloid precursor protein regulates its endocytosis and amyloidogenic processing in neurons. J Neurosci 28, 2874–2882.
    48. Sharp, M. D. & Pogliano, K. (2002). Role of cell-specific SpoIIIE assembly in polarity of DNA transfer. Science 295, 137–139.
    49. Simons, K. & Ikonen, E. (1997). Functional rafts in cell membranes. Nature 387, 569–572.
    50. Simons, K. & Toomre, D. (2000). Lipid rafts and signal transduction. Nat Rev Mol Cell Biol 1, 31–39.
    51. Singer, S. J. & Nicolson, G. L. (1972). The fluid mosaic model of the structure of cell membranes. Science 175, 720–731.
    52. Stallkamp, I., Dowhan, W., Altendorf, K. & Jung, K. (1999). Negatively charged phospholipids influence the activity of the sensor kinase KdpD of Escherichia coli. Arch Microbiol 172, 295–302.
    53. Sterlini, J. M. & Mandelstam, J. (1969). Commitment to sporulation in Bacillus subtilis and its relationship to development of actinomycin resistance. Biochem J 113, 29–37.
    54. Stevens, C. M., Daniel, R., Illing, N. & Errington, J. (1992). Characterization of a sporulation gene, spoIVA, involved in spore coat morphogenesis in Bacillus subtilis. J Bacteriol 174, 586–594.
    55. Stuermer, C. A., Lang, D. M., Kirsch, F., Wiechers, M., Deininger, S. O. & Plattner, H. (2001). Glycosylphosphatidyl inositol-anchored proteins and fyn kinase assemble in noncaveolar plasma membrane microdomains defined by reggie-1 and -2. Mol Biol Cell 12, 3031–3045.
    56. Vagner, V., Dervyn, E. & Ehrlich, S. D. (1998). A vector for systematic gene inactivation in Bacillus subtilis. Microbiology 144, 3097–3104.
    57. Walker, C. A., Hinderhofer, M., Witte, D. J., Boos, W. & Moller, H. M. (2008). Solution structure of the soluble domain of the NfeD protein YuaF from Bacillus subtilis. J Biomol NMR 42, 69–76.
    58. Wehrl, W., Niederweis, M. & Schumann, W. (2000). The FtsH protein accumulates at the septum of Bacillus subtilis during cell division and sporulation. J Bacteriol 182, 3870–3873.
    59. Wiegert, T., Homuth, G., Versteeg, S. & Schumann, W. (2001). Alkaline shock induces the Bacillus subtilis σW regulon. Mol Microbiol 41, 59–71.
    60. Zhang, H. M., Li, Z., Tsudome, M., Ito, S., Takami, H. & Horikoshi, K. (2005). An alkali-inducible flotillin-like protein from Bacillus halodurans C-125. Protein J 24, 125–131.