MreB forms a cytoskeleton in many rod-shaped bacteria which is involved in cell shape determination and chromosome segregation. PCR-based and Southern analysis of various actinomycetes, supported by analysis of genome sequences, revealed mreB homologues only in genera that form an aerial mycelium and sporulate. We analysed MreB in one such organism, Streptomyces coelicolor. Ectopic overexpression of mreB impaired growth, and caused swellings and lysis of hyphae. A null mutant with apparently normal vegetative growth was generated. However, aerial hyphae of this mutant were swelling and lysing; spores doubled their volume and lost their characteristic resistance to stress conditions. Loss of cell wall consistency was observed in MreB-depleted spores by transmission electron microscopy. An MreB–EGFP fusion was constructed to localize MreB in the mycelium. No clearly localized signal was seen in vegetative mycelium. However, strong fluorescence was observed at the septa of sporulating aerial hyphae, then as bipolar foci in young spores, and finally in a ring- or shell-like pattern inside the spores. Immunogold electron microscopy using MreB-specific antibodies revealed that MreB is located immediately underneath the internal spore wall. Thus, MreB is not essential for vegetative growth of S. coelicolor, but exerts its function in the formation of environmentally stable spores, and appears to primarily influence the assembly of the spore cell wall.
Bacterial morphologies range from spherical and rod-shaped to curved, helical and filamentous. A major determinant of cell shape is the bacterial cell wall, which consists of glycan strands cross-linked by short peptides. Isolated peptidoglycan sacculi can retain the shape of the bacterial cell, and mutants defective in peptidoglycan synthesis typically show an altered morphology (Young, 2003; Cabeen and Jacobs-Wagner, 2005). However, not only the enzymes that are directly involved in the synthesis and assembly of peptidoglycan can affect the morphology. It is becoming clear that cytoskeletal elements exist in the bacterial cytoplasm (Löwe et al., 2004), and that they are determining the architecture of the cell wall and have strong impacts on cell shape. These include homologues of all three major types of eukaryotic cytoskeletons. Thus, MreB proteins are actin homologues that produce microfilament-like fibres and determine rod shape in many bacteria (van den Ent et al., 2001; Jones et al., 2001); crescentin (CreS) produces intermediate filament-like elements in Caulobacter crescentus and give rise to the curved shape of those cells (Ausmees et al., 2003); and FtsZ is a tubulin homologue that assembles into a cytokinetic ring at the site of cell division and directs cell division and formation of the septal peptidoglycan (Margolin, 2005).
MreB is a member of the HSP70-actin-sugar kinase (ASHKA) superfamily of proteins (Bork et al., 1992), and its crystal structure is strikingly similar to the structure of actin (van den Ent et al., 2001). Purified MreB of Thermotoga maritima polymerizes in vitro to form filaments with a spacing between the MreB monomers of 51 Å, which is reminiscent of the spacing between the subunits (55 Å) in actin filaments (van den Ent et al., 2001). MreB homologues have been studied primarily in Escherichia coli (mreB), Bacillus subtilis (mreB, mbl, mreBH) and C. crescentus (mreB), and been found with immunofluorescence microscopy and GFP tagging to form helical-like structures underneath the cell envelope (Jones et al., 2001; Shih et al., 2003; Figge et al., 2004; Soufo and Graumann, 2003). Deletion of mreB in E. coli resulted in loss of the rod shape and in the production of spherical cells (Doi et al., 1988). In B. subtilis, mreB is an essential gene, and depletion of MreB resulted in increased cell width, loss of cell shape and eventually lysis (Jones et al., 2001; Lee and Stewart, 2003; Formstone and Errington, 2005). Furthermore, the MreB-like Mbl protein is required for determination of the rod shape of B. subtilis, and forms helical cytoskeletal structures that are needed for incorporation of new peptidoglycan along the lateral wall during cell elongation (Daniel and Errington, 2003). Despite these observations that suggest that MreB proteins form actin-like fibres that define the bacterial cell shape, the specific function of MreB proteins remain still unclear. Recent studies link MreB to peptidoglycan synthesis (Daniel and Errington, 2003), correct chromosome segregation (Kruse et al., 2003; Soufo and Graumann, 2003; Gitai et al., 2005) and establishment of cell polarity (Gitai et al., 2004). Interestingly, in the rod- and coccobacillus-shaped Rhodobacter sphaeroides MreB failed to produce the typical helical structures observed in Bacillus but rather formed a ring at the mid-cell position of elongating cells, suggesting a role for MreB both in septation and in peptidoglycan synthesis in this organism (Slovak et al., 2005). Thus, it appears that MreB may have different functions in different organisms.
The observation that the mre gene cluster is absent from coccoid species such as Streptococcus, Staphylococcus and Lactococcus suggested that a spherical shape may be correlated with the absence of an mre-dependent system, and that MreB proteins are particularly important in rod-shaped cells (Jones et al., 2001). However, in contrast to most other rod-shaped bacteria, the actinomycete Corynebacterium glutamicum lacks mreB-like genes in its genome. In accordance with this, it apparently establishes its rod shape in a different and MreB-independent way, through growth of the cell wall only at the cell poles (Daniel and Errington, 2003). Most other actinomycetes also lack MreBs, but the Streptomyces genus is a notable exception as they encode well conserved MreB homologues (see below). This raises the question of what the roles may be for the Streptomyces MreBs. Streptomycetes are Gram-positive filamentous actinobacteria known for their complex morphology and their ability to produce a wide range of secondary metabolites. Their morphological development resembles in many ways that of filamentous fungi. They produce a vegetative mycelium consisting of a network of branching, multinucleoidal, hyphal filaments, which grow in a highly polarized way by tip extension. This polarized growth differs from that of C. glutamicum by being independent of cell division and occurring at hyphal tips generated by branching or germ tube emergence, rather than at the poles created by cell division (Flärdh, 2003). Upon starvation colonies enter the reproductive phase by forming aerial hyphae that eventually produce chains of spores (Chater and Hopwood, 1993). This involves a specialized form of cell division that divides long multigenomic aerial hyphae into a large number of spore compartments, which then mature into thick-walled dormant spores (Chater and Losick, 1997; Flärdh, 2003). The genome of Streptomyces coelicolor contains two homologues of mreB, but the functions of them during growth and development are poorly understood. One of these genes, mreB, is located in a cluster with mreC and mreD, while the other one is located elsewhere. We have previously identified and studied the mreBCD homologues in S. coelicolor, and gene disruption experiments resulted in a viable mreC knockout mutant that showed growth retardation and reduced sporulation, while an mreB mutant could not be created by conventional gene replacement (Burger et al., 2000). Transcriptional analysis of the mreBCD gene cluster by promoter probing and S1 nuclease mapping suggested that the three genes are regulated as an operon from three promoters upstream of mreB (Burger et al., 2000). In this article, we provide evidence that MreB is not essential for vegetative growth in S. coelicolor but plays a vital role in formation of spores resistant to various forms of environmental stress.
MreB is found only in actinomycetes that convert aerial hyphae into spore chains
In S. coelicolor, S. avermitilis and S. scabies, the mreBCD operon lies between ndk, encoding a nucleoside diphosphate kinase, and pbp83 (= pbp2), encoding a penicillin-binding protein (PBP) (Burger et al., 1998). In order to test how widespread the mreB genes are among rare actinomycetes, genomic DNA was isolated from various actinomycetes and PCR was performed using primers PM13 and PM14 (see Table 4) that were designed at two highly conserved regions of mreB. The primers did not distinguish between the two homologues of mreB found in the streptomycete genomes, and should detect both genes, although the products did not differ in size. Following the example from B. subtilis, we designate the second homologue that is located at a position distant from the mre cluster, mbl. We obtained a band of the expected size in all tested Streptomyces strains, in Streptoverticillum mobaraense and in Streptosporangium roseum (Table 1). In contrast, we failed to detect such a band with actinomycetes that do not produce aerial hyphae (Miyadoh, 1997), such as Actinoplanes (A. friuliensis and Actinoplanes sp.), Micromonospora sp., C. glutamicum, Microbacterium testaceum, Nocardioides simplex, Gordonia sp. and Rhodococcus rhodochrous, suggesting the lack of an mreB homologue in these organisms. These findings were confirmed by Southern blot analyses using mreB of S. coelicolor as a probe (Table 1). MreB proteins therefore seem to be present only in strains that produce both aerial mycelium and spores (see Discussion).
Expression of MreB in S. coelicolor leads to growth impairment and lysis
Overexpression of E. coli MreB in E. coli leads to inhibition of cell division (filamentous phenotype), probably due to a reduction of FtsI activity, which is enhanced in mreB mutants (Wachi and Matsuhashi, 1989). To analyse the effects of overexpressed S. coelicolor MreB in S. coelicolor, the gene was cloned under the control of the thiostrepton-inducible tipA promoter in the E. coli-Streptomyces shuttle vector pGM190; the construct was designated pPM1. When S. coelicolor M145 transformed with pPM1 was plated on MS plates containing 10 µg ml−1 thiostrepton, the strain failed to grow (Fig. 1B), while the same strain harbouring the control plasmid pGM190 (without insert) showed normal growth (Fig. 1D); this indicated that overexpression of MreB on solid medium is lethal.
To study the effect of S. coelicolor MreB overexpression on the growth of substrate mycelium, spores of S. coelicolor carrying either pPM1 or the control plasmid were inoculated in S-medium and grown at 30°C. After 8–12 h of growth, MreB expression was induced by the addition of 10 µg ml−1 thiostrepton. Interestingly, it was possible to overexpress MreB in liquid-grown cultures, although these cultures grew significantly slower than those of the parental strain harbouring the empty vector pGM190 (Fig. 2). Swelling of the extremities of the hyphae and extensive lysis of the mycelium of pPM1 transformants were observed already 2 h after induction (Fig. 2C), while the pGM190 transformants remained unaffected. Overexpression of MreB in the pPM1 strain was confirmed by Western blot analysis with anti-MreB antibodies (data not shown).
To analyse whether MreB overexpression affects spore germination, spores of S. coelicolor M145 transformed with pPM1 were inoculated in liquid culture containing thiostrepton (Fig. 2D–F). In this case the spores were able to germinate, but the elongation of germ tubes was inhibited (Fig. 2F) and the hyphae failed to elongate properly.
A mutant deficient in mreB was constructed by PCR targeting
Previous attempts to inactivate mreB by gene disruption using the temperature-sensitive Streptomyces vector pGM9 were unsuccessful (Burger et al., 2000). However, the use of REDIRECT technology (see Experimental procedures) allowed the isolation of an insertion mutant (mreB-IM) in which mreB was replaced with the apramycin resistance cassette aacC4 (1384 bp). The replacement of mreB with the apramycin cassette was verified with PCR and Southern blot analysis (Fig. 3A and B), and the loss of expression of MreB was confirmed with Western blot analysis using anti-MreB antiserum (Fig. 3C). The mreB-IM mutation is expected to have polar effects to mreC and mreD, as these appear to be transcribed as an operon (Burger et al., 2000). In order to avoid such polar effects, an mreB in-frame deletion mutant (mreB-IFD) was also constructed removing the apramycin cassette from the chromosome of the mreB-IM mutant (see Experimental procedures). The loss of the resistance marker and the exact location of the deletion were confirmed by PCR (data not shown) and Southern blot analysis (Fig. 3B).
The mreB deletion mutant produces swollen aerial hyphae and swollen spores
Unexpectedly, in the case of the mreB-IM mutant, colony morphology and vegetative growth of the substrate mycelium was normal on plates and in liquid-grown cultures (data not shown). However, differentiation was significantly affected: many aerial hyphae had a swollen appearance, were irregular in shape (Fig. 4B) and showed a large degree of lysis. Spore dimensions were also abnormal. Measurement of spore sizes (Table 2) using the analysis® program (SIS, Soft Imaging System GmbH) showed that the mreB-IM mutant spores had an average width of 1.31 µm while the spores of the parental strain S. coelicolor M145 had an average width of only 0.75 µm. Mutant spores were also longer (Table 2), as the average length of mreB-IM mutant spores was 1.65 µm while the average length of parental spores measured 1.09 µm. In addition, both light (Fig. 4C) and scanning electron microscopy (data not shown) frequently showed premature germination of spores already in the spore chains. The phenotype of the in-frame mreB deletion mutant (mreB-IFD) strongly resembled that of the mreB-IM replacement mutant, again with swollen, kinky and lysing aerial hyphae (data not shown). Spores were also significantly larger and wider than the parental strain showing average values of width (1.16 µm) and length (1.64 µm) comparable to the insertional mutant (Table 2).
Table 2. Average values of spore length and width of spores in chains of M145, MreB-depleted strains, complemented mutants and mre-IM complemented with the mreB–egfp fusion gene.
Average values for 60 spores measured by phase contrast microscopy.
To confirm that the absence of an intact mreB was responsible for the observed phenotype, a single copy of the mreB gene under its own promoter (construct pPM6; see Experimental procedures) was inserted at the φC31 attachment site in the chromosome of the mreB-IFD mutant, generating the strain mreB-IFDc. Insertion of mreB complemented the phenotype of the mreB mutant. Aerial hyphae of the strain mreB-IFDc did not swell and lyse (data not shown). Length (1.23 µm) and width (0.84 µm) of the spores of mreB-IFDc were almost completely restored to parental values (Table 2), suggesting that indeed the altered spore morphology was due to deletion of mreB.
Integrity of the envelope is severely compromised in mreB mutant spores
To analyse the defects of the mreB mutant of S. coelicolor in more detail, thin sections of sporulating aerial mycelia of the mutant and its parental strain M145 were analysed by high-resolution transmission electron microscopy (TEM). Typical examples of dividing aerial hyphae, prespore chains and mature spores are presented in Fig. 5. In line with the observations with light microscopy, the mreB mutant showed extensive lysis of the aerial hyphae, producing bloated structures (data not shown). Especially remarkable was the altered appearance of the nucleoids, invariably surrounded by electron-lucent (white) material (Fig. 5B). Spores had highly heteromorphous shapes, and collapsed, dented and broken spores evidenced loss of spore wall consistency (Fig. 5B), when compared with the parental M145 (Fig. 5A). Apparently, both nucleoid organization and integrity of the spore envelope were severely compromised in the mreB mutant.
The spores of the mreB mutants are sensitive to heat and detergent
Streptomyces spores are dormant cells which are relatively resistant to desiccation, sonic vibration, enzymatic digestion and exposure to moderately high temperature (McBride and Ensign, 1987). To test whether the mutant spores were as heat resistant as those of the parental strain, spore suspensions (105 spores ml−1) were incubated at 60°C for different intervals and then plated on SFM and further cultivated for 2 days at 30°C. After 5 min of heat treatment, 84% of the spores of the parental M145 survived, while of the mreB mutant only 7.5% (mreB-IFD) had survived (Fig. 6). Extended heat treatment (60 min) was required to decrease the survival rate of M145 spores to 30%, and during those conditions only 0.01% of the mreB-IFD mutant could survive (Fig. 6).
It is known that Streptomyces spores are resistant to treatment with SDS and this detergent has also been reported to activate spore germination (Grund and Ensign, 1982). The effect of SDS on mre-depleted S. coelicolor spores was therefore tested. SDS (final concentration 5%) or water (as a control) was added to spore suspensions of M145 and of the mreB-IFD mutant and incubated at room temperature for 1 h. Spores were then diluted, plated on MS agar and incubated at 30°C for 2 days. Viable count of wild-type spores did not decrease after incubation either in water or in SDS. However, the viable count of the mreB mutant was reduced to 0.1% after SDS treatment. As the spores of the mreB mutant strains are, unlike the wild type, sensitive to some physiological stresses, the response of these spores to mechanical stress was also analysed. Spore suspensions were treated with ultrasound as described in Experimental procedures. The wild type and the mutants were equally resistant to this treatment, and even after five cycles of ultrasonication the survival rate was more than 50% in both cases (data not shown).
MreB is a membrane-associated protein that localizes at the septa of aerial hyphae and at the spore wall
Recent results showed that MreB in B. subtilis and E. coli forms helical structures located on the inner surface of the cytoplasmic membrane (Jones et al., 2001; Kruse et al., 2003; Shih et al., 2003; Formstone and Errington, 2005) and that it is only weakly associated with the membrane (Slovak et al., 2005), while MreC and MreD are membrane proteins (Kruse et al., 2005). In order to localize MreB in S. coelicolor M145, the entire mreB gene and its three promoters were positioned upstream of the egfp gene, to generate a fusion between the two genes, encoding an MreB–EGFP fusion protein under the control of the natural mreB promoters (see Experimental procedures). The fusion gene was cloned in pSET152, allowing integration at the φC31 attachment site of the S. coelicolor chromosome generating the strain SCPM6. The same plasmid (pPM4) was also integrated into the chromosome of the mreB-IM mutant to analyse whether the fusion protein is functional and able to complement the MreB deficiency. The presence of a single copy of mreB–egfp complemented the defect causing the pre-germination of spores (data not shown) and reduced the swollen spore size to almost parental values (Table 2), suggesting that the MreB–EGFP fusion protein was active.
During vegetative growth of SCPM6, only diffuse fluorescence was observed in the hyphae although expression of MreB–EGFP was confirmed by Western blot analysis with both anti-MreB and anti-GFP antibodies (data not shown). In contrast, study of surface-grown cultures revealed very clear fluorescence signals that were specifically localized (Fig. 7) at the septa of sporogenic aerial hyphae. As shown in Fig. 7B, bands of fluorescence coincided with the constrictions caused by sporulation septation. Simultaneous labelling of septa with fluorescent conjugates of wheat germ agglutinin confirmed that MreB–EGFP overlapped with the septa (data not shown). However, it is not clear whether MreB–EGFP also localized to the basal septa sometimes seen at the bottom of sporogenic hyphal cells (Kwak et al., 2001). In prespore chains, foci were localized at the cell poles generated by sporulation septation (white arrows in Fig. 7D). Successively, in more mature spores, the MreB–EGFP signal completely surrounded the spores, giving rise to ring-like appearance of the fluorescence (Fig. 7F). However, in fully mature spores most of the fluorescence from the MreB–EGFP fusion had disappeared (data not shown).
Closer inspection of the localization of MreB was performed by immunogold electron microscopy of spore preparations from S. coelicolor M145 (Fig. 8). As a negative control we used the mreB insertion mutant (mreB-IM). While in the mutant only background gold labelling was obtained, similar to when anti-MreB antibody was omitted (not shown), in the parental strain clear and specific labelling, particularly close to the inside of the spore walls was detected.
The cellular localization of MreB was further examined by cell fractionation and Western analysis using anti-MreB antibodies. For this, the mycelium and spores grown on MS agar were disrupted by French press treatment and the cytoplasmic, the membrane-associated and the membrane fractions were individually isolated. MreB was detected both in the soluble fraction and mostly in the membrane-associated fraction, while we failed to detect MreB in the membrane fraction (data not shown). Cellular localization of the MreB–EGFP fusion protein was also conducted. Indeed, the protein was localized in the same fractions as the native MreB, namely in the soluble fraction and in the membrane associated fractions (data not shown), indicating that the fusion protein was localized equivalent to the parent protein.
In this study, an mreB deletion mutant was obtained in the filamentous Gram-positive bacterium S. coelicolor, and it was demonstrated that MreB is dispensable for vegetative growth, but essential for spore formation and assembly of the spore cell wall. Previous attempts to inactivate mreB in S. coelicolor by gene disruption using the temperature-sensitive Streptomyces vector pGM9 (Muth et al., 1989) were unsuccessful (Burger et al., 2000). However, using a PCR-targeting method (REDIRECT method; Gust et al., 2003) we were able to generate deletion mutants in mreB. The discrepancy may be explained by a higher frequency of allelic exchange in the REDIRECT method, due to the much larger regions flanking the mreB locus that were available for recombination. As several mreB mutants were isolated in independent experiments with frequencies that are typical of gene inactivation experiments using the REDIRECT procedure, it appears unlikely that mreB is essential in Streptomyces and that lethality of the mreB mutations was masked by (a) suppressor mutation(s), as has been described for other bacteria (Kruse et al., 2005). This is further supported by the observation that mreB is absent from many non-sporulating actinomycetes.
Deletion of mreB in S. coelicolor gave no obvious phenotype during vegetative growth in liquid culture or on solid plates. However, transcriptional analyses by S1 nuclease mapping (Burger et al., 2000) and by microarray and reverse transcription polymerase chain reaction (RT-PCR) studies (E.E. Noens and G.P. van Wezel, unpublished) revealed that S. coelicolor mreB is transcribed during vegetative growth, and expression was confirmed by Western blot analysis of MreB protein in liquid-grown cultures (data not shown). Thus, the possible role of MreB in young mycelium is unclear and needs further analysis, and it can not be excluded that there could be some degree of redundancy between MreB and the second MreB homologue Mbl. However, based on the lack of a clear phenotype of the mreB mutant during vegetative growth, and the lack of MreB–EGFP signals at hyphal tips, it seems unlikely that MreB would have a direct role in the elongation of the hyphal cell wall, which occurs primarily at the tips. Sequence analysis of genomic DNA of different actinomycetes showed that species forming only substrate mycelium, such as Rhodococcus sp. and Nocardia sp. (http://Rhodococcus.ca; http://nocardia.nih.go.jp), do not have an mreBCD cluster, supporting the idea that the Mre proteins may not be required for vegetative growth in actinomycetes.
While inactivation of mreB did not compromise vegetative growth of S. coelicolor, ectopic expression of mreB was highly toxic. After germination of spores the elongation of germ tubes was inhibited and hyphae lysed. Obviously, the presence of (too much) MreB at the wrong spatial compartment severely interferes with normal growth, perhaps by recruiting PBPs and preventing them from their function in cell wall synthesis.
Mutational analysis clearly demonstrated that MreB in S. coelicolor is required for correct sporulation, and appears to affect spore wall formation. The aerial hyphae and spores of the mreB deletion mutant were swollen (Fig. 4; Table 2), spores were sensitive to heat and treatment with SDS (Fig. 6), and irregularities in the spore cell walls were observed using TEM (Fig. 5B). This is consistent with the observation made in B. subtilis that deletion of mreB causes increased cell width and cell lysis (Formstone and Errington, 2005). Formstone and Errington (2005) proposed that loss of MreB lead to a change in resistance of the cell wall to osmotic or mechanical stress, due to the absence or incorrect assembly of cell wall components (Formstone and Errington, 2005). In analogy, the depletion of MreB in S. coelicolor appears to impair assembly of the cell wall, but only during sporulation. The spore wall composition of streptomycetes is not well studied. However, it is different in thickness from vegetative hyphal walls and has often two layers, although qualitative differences in peptidoglycan components have not been reported (Glauert and Hopwood, 1961; Ensign, 1978). The importance of the spore wall may be indicated by the fact that spores acquire resistance to different physiological and mechanical stresses, which the substrate mycelium does not have. Consistently, the mreB mutants failed to mount resistance against two types of stress.
The involvement of MreB in spore formation is consistent with the PCR and Southern blot analysis of different actinomycetes. Using primers corresponding to a highly conserved region of mreB, genomic DNA of sporulating actinomycetes gave rise to an amplified fragment of the expected size, while genomic DNA of non-sporulating actinomycetes did not (Table 1). Surprisingly two sporulating genera, Actinoplanes and Micromonospora, did not give an amplified product (Table 1). These two genera share an important characteristic: they sporulate, but they do not form aerial mycelium. The genus Actinoplanes produces motile spores enveloped in structures called sporangia (Waksman, 1961). The genus Micromonospora grows as straight or curved branching substrate hyphae without cross-walls (Waksman, 1961). Like in Actinoplanes, aerial mycelium is not formed and multiplication occurs by means of fragments of mycelium and special spores formed singularly. The genus Actinoplanes is a member of the family of the Actinoplanaceae together with the genus Streptosporangium which in contrast produces an aerial mycelium which resembles that of most species of Streptomyces (Waksman, 1961). Unlike Actinoplanes, Streptosporangium DNA was shown to encode an mreB homologue (Table 1). Apparently, the differentiation process in the two genera is highly different, suggesting that the mre genes are needed only in Streptosporangium sporulation that indeed resembles that of Streptomyces. From these observations, MreB seems not to be required in sporulation processes of actinomycetes in general, but only in a sporulation process which involves septation of aerial mycelium.
In B. subtilis and E. coli, which both are rod-shaped bacteria, MreB forms helical filaments lying beneath the cell surface (Jones et al., 2001; Shih et al., 2003). In C. crescentus, MreB undergoes two distinct patterns of localizations: during the cell elongation phase it appears like spirals that traverse along the longitudinal axis of the cell, while during cell division it forms an FtsZ-dependent transverse band at the mid-cell (Figge et al., 2004). Therefore, it was speculated that MreB has a role in co-ordinating the switch between longitudinal and septal peptidoglycan synthesis. In S. coelicolor, we failed to detect the typical helical-like structures described for E. coli, B. subtilis and C. crescentus. The signal generated from the MreB–EGFP fusion during vegetative growth resulted in diffused fluorescence in the hyphae, indicating that MreB monomers are randomly distributed in the vegetative mycelium without a specific localization. In contrast, during differentiation the MreB–EGFP localized at the septa of aerial hyphae and successively, when spores were formed, in the spores (Fig. 7) probably underneath the cytoplasmic membrane, as indicated by immunogold electron microscopy (Fig. 8). Considering that MreB is organized in helical structures in rod-shaped bacteria, it is possible that in S. coelicolor MreB forms dense spirals just underneath the cytoplasmic membrane, but no such patterns could be discerned in our analyses. As it is possible that fusions of GFP to the C-terminus of MreB can interfere with its function, we can not exclude that our C-terminal MreB–EGFP fusion did not fully reproduce the localization patterns of MreB itself. However, we could demonstrate that MreB–EGFP localized to the same subcellular fractions of S. coelicolor as the MreB protein did. Furthermore, MreB–EGFP was able to complement the mutant phenotype of the mreB-IM mutant by restoring the spore size to almost parental values (Table 2) as well as by preventing pre-germination of spore chains.
In S. coelicolor, septation occurs during both vegetative growth and sporulation, but MreB localization was specific for sporulation septa. It is clear that MreB is not required for the formation of sporulation septa per se, as mreB mutants were still able to septate (Fig. 5B) and produced viable spores. On the other hand, it should be tested whether MreB localization could be FtsZ dependent, as already seen in C. crescentus (Figge et al., 2004). It will be therefore interesting to test in future experiments MreB localization in different S. coelicolor ftsZ mutants; for example, an ftsZ mutant impaired in Z-ring formation in sporogenic aerial hyphae, but not in vegetative mycelium (Grantcharova et al., 2003). In R. sphaeroides, MreB localizes predominantly at the mid-cell position (Slovak et al., 2005). The authors speculate that this mid-cell localization for MreB might reflect the fact that this region is the main site of peptidoglycan synthesis rather than that MreB plays a direct role in septation. In an analogous manner, MreB in S. coelicolor which was shown to localize at the sporulation septa, could be needed in subsequent steps for spore formation, rather than playing a crucial role in the sporulation-specific cell division. Thus, MreB in S. coelicolor has a novel localization pattern as well as a somewhat unique mode of action that has not been observed in the rod-shaped bacteria and may be typical of actinomycetes that form an aerial mycelium and spores.
Based on the localization of MreB at the spore wall and its important role in maintaining the integrity of the spore wall, we anticipate that MreB is involved in thickening of the spore wall and may be recruiting PBPs and other peptidoglycan-related proteins during the sporulation process. The mre genes in S. coelicolor and S. avermitilis are physically linked with pbp83 and sfr (Burger et al., 2000). Pbp83 (Pbp2; Bentley et al., 2002), encodes a PBP with high similarity to E. coli PBP2, which is involved in cell elongation in E. coli (Vinella et al., 1993), and with SpoVD, a PBP important for sporulation of B. subtilis (Daniel et al., 1994). The product of sfr shows similarity to RodA, a protein involved in peptidoglycan synthesis during cell elongation in E. coli, and SpoVE, a protein required for the synthesis of the spore cortex peptidoglycan in B. subtilis (Henriques et al., 1998). Inactivation of pbp83 by gene disruption lead to an impairment in differentiation (A. Burger and W. Wohlleben, unpublished), a phenotype that resembles that of the mreB mutant that is reported in this article. A protein interaction model of the MreBCD, RodA and PBPs in B. subtilis has been proposed by Errington (2003). According to this model a multi-PBP complex interacts with RodA, MreC and MreD and this assembly is spatially controlled by the MreB cables in the cytosol. These proposed interactions between the Mre proteins were partially confirmed in E. coli (Kruse et al., 2005). We observed that in S. coelicolor, deletion of mreB, mreC, mreD individually, or of the whole mreBCD cluster, leads to a similar phenotype (K. Schirner, P. Mazza, G. Muth, and W. Wohlleben, unpublished) suggesting that also in this organism the three proteins may participate in the same process. A complex that includes PBPs, a RodA homologue, and the Mre proteins, and which primarily is acting during spore wall assembly, is therefore imaginable in S. coelicolor. This leads to the following hypothesis on the mode of action of MreB in S. coelicolor: it is expressed at a basal level during vegetative growth and it is localized rather dispersely in the cytosol. When sporulation septation occurs, MreB condenses at the sporulation septa, possibly in an FtsZ-dependent manner. An increased expression of mreB before sporulation that was detected by S1-mapping experiments (Burger et al., 2000) may allow the cells to accumulate a sufficient number of MreB units to form a ring- or a shell-like structure underneath the spore wall. As MreB most likely is a membrane-associated protein, this structure may be anchored to the cell membrane through the two membrane proteins MreC and MreD and used to recruit and localize proteins responsible for spore wall formation, perhaps including Pbp83 and the product of sfr. Further studies to localize Pbp83 and Sfr in mre mutants and interaction studies between the Mre proteins and the products of pbp83 and sfr are underway to understand the unique role of MreB in S. coelicolor A3(2). It will also be important to clarify the role of Mbl in this organism, and to determine whether there could be some functional redundancy between this protein and MreB.
Bacterial strains and media
The E. coli and S. coelicolor A3(2) strains used are listed in Table 3. S. coelicolor strains were cultivated on MS agar plates or in S-medium (Kieser et al., 2000). Cultivation of strains and procedure for DNA manipulation were performed as previously described for E. coli (Sambrook et al., 1989) and S. coelicolor (Kieser et al., 2000).
Table 3. Strains of E. coli and S. coelicolor used in this work.
Plasmid pPM1 was obtained by amplifying the mreB gene from the S. coelicolor M145 genome using the primers PM1 and PM2 (Table 4), digesting the PCR product with NdeI and HindIII and cloning the fragment under the control of the thiostrepton-inducible promoter PtipA into the Streptomyces multicopy expression plasmid pGM190 (G. Muth, unpubl. data). To construct a C-terminal fusion with the EGFP protein, the mreB gene was amplified using the primers PM5 and PM6 (Table 4), cut with NdeI and BglII and cloned in front of the egfp gene in the plasmid pTST101 (J. Altenbuchner, pers. comm.) cut with NdeI and BamHI. To transform S. coelicolor M145 the resulting mreB–egfp fusion was cut out from pTST101 and cloned in the integrative Streptomyces vector pSET152 (Bierman et al., 1992) generating pPM4. Plasmid pPM6 was constructed by amplifying the mreB gene with its promoter region from S. coelicolor M145 with the primers PM11 and PM12 (Table 4) and cloning the PCR product as BglII–EcoRI fragment in pSET152 (Bierman et al., 1992).
Protein expression, purification and antibody production
mreB from S. coelicolor M145 was amplified by PCR with the primers PM7 and PM2 that include BamHI and HindIII site, respectively, and cloned in pRSETB (Invitrogen) so that it forms an in-frame fusion with the His-tag (pPM5). His-tagged MreB was expressed in E. coli BL21 cells containing pPM5 and purified with Ni-Nta spin columns (Qiagen) under denaturing condition as described in the Ni-Nta manual (Qiagen). Rabbit polyclonal antibodies were raised against the purified protein (Eurogentec).
Overexpression of MreB
Protoplasts of S. coelicolor M145 were transformed with the plasmids pPM1 or pGM190 as described previously (Kieser et al., 2000). Spores of transformants were inoculated in S-medium containing 50 µg ml−1 kanamycin and incubated with shaking at 30°C. To overexpress MreB, thiostrepton was included at a concentration of 10 µg ml−1.
Overexpression of MreB was confirmed by Western blot analysis with anti-MreB serum. Mycelium was collected after 18 h of growth with or without thiostrepton and broken with French press. Crude extracts were mixed with an equal volume of 2× sample buffer (125 mM Tris-HCl pH 6.8; 4% SDS; 20% glycerol; 2.0 mM EDTA; 0.02% bromophenol blue; 3% dithiothreitol), boiled for 5 min and loaded on SDS-polyacrylamide gels (12.5%). Proteins were capillary transferred to a pure nitrocellulose blotting membrane (Pall Corporation) and MreB was detected using a 1:2500 dilution of polyclonal anti-MreB serum.
Streptomyces coelicolor M145 and the strain SCPM6 carrying the mreB–egfp fusion were grown for 2–3 days in 150 ml of S-medium with or without antibiotics at 30°C and harvested by centrifugation. Mycelium was resuspended in 6 ml of 25 mM Tris-HCl (pH 7.5), 100 mM NaCl and 1 mM protease inhibitor (Complete EDTA-free tablets, Roche), and the cells were broken using the French press. The cell extract was centrifuged at 90 000 g for 30 min at 4°C and the supernatant, containing the cytoplasmic fraction, was saved and stored at −20°C. Pellet was resuspended in 4 ml of 25 mM Tris-HCl (pH 7.5), 1 M NaCl and 20% glycerol, stirred for 2 h at 4°C and centrifuged at 90 000 g for 30 min at 4°C. The supernatant containing membrane-associated proteins was saved and stored at −20°C. To obtain the membrane fraction, pellet was resuspended in 1.5 ml of 25 mM Tris-HCl (pH 7.5), 1 M NaCl, 20% glycerol and 2% Triton X-100, stirred overnight at 4°C and again centrifuged for 30 min at 90 000 g. The samples obtained from cell fractionation were loaded on SDS-gels and Western blot was performed as described above.
Creation of an mreB null allele
To create an mreB mutant, the PCR-targeting procedure described by Gust et al. (2003) was used. The ΩaacC4 cassette (which confers apramycin resistance) from the plasmid pIJ773 was amplified using the oligonucleotides PM9 and PM10 (Table 4) designed so that the 3′-ends can anneal to the resistance cassette gene, whereas the 5′-ends are homologous to the sequences flanking S. coelicolor mreB. The PCR products was transferred into E. coli BW 25113/pIJ790 carrying the cosmid C88 and the λ RED system induced as described (Gust et al., 2003). The correct gene replacement on the cosmid C88 isolated from an apramycin-resistant clone was confirmed with PCR using a set of oligonucleotides priming outside the region of recombination. This cosmid was designated pPM11. pPM11 was introduced into E. coli ET12567 containing the plasmid pUZ8002 and then transferred to S. coelicolor M145 by conjugation as previously described (Kieser et al., 2000). Exconjugants were selected for kanamycin and apramycin resistance and then replicated on plates with apramycin and with or without kanamycin. DNA of kanamycin-sensitive and apramycin-resistant clones were isolated and tested with PCR and Southern blot to confirm the absence of the chromosomal mreB.
In order to remove the disruption cassette, the cosmid pPM11 was introduced in the E. coli strain DH5α/pCP20 that expresses the FLP recombinase (Gust et al., 2003). The resulting cosmid carrying the in-frame deletion was named pPM12 and used to transform the S. coelicolor mutant strain containing the marked deletion (Kieser et al., 2000). Transformants were first selected for insertion of the cosmid by single cross-over (kanamycin resistant) and then screened for the double-cross-over event (kanamycin sensitive and apramycin sensitive). The loss of the disruption cassette was confirmed with PCR and Southern blot analysis.
Heat and detergent treatment
To test whether the mutant spores were as heat resistant as the ones of the parental strain, spore suspensions of different strains (105 spores ml−1) were incubated at 60°C for different time. After incubation spore suspensions were diluted 1:10 or 1:100 and 100 µl of the dilution was plated on MS agar. Spore suspension of the mreB mutant incubated longer than 20 min at 60°C were plated without being diluted. Plates were further cultivated for 2 days at 30°C. Survival rate was calculated as number of colonies grown on plates after treatment at 60°C divided by the number of colonies grown on plates without treatment at 60°C in percentage.
To test the effect of the detergent SDS on mreB-depleted spores, 50 µl of spore suspensions (109 ml−1) of M145 and of the mreB mutant (mreB-IFD) was incubated in 5% SDS, 1% SDS or water for 1 h at room temperature. Appropriate spore dilutions were plated on MS agar and incubated at 30°C for 2 days. Titre of colonies surviving SDS treatment were determined in relation to those incubated in water.
To test spore resistance to ultrasonication, spore suspension (105 spores ml−1) was treated with ultrasounds (SONOPULS GM2070 digital ultrasonic homogenizers, 20 kHz; BANDELIN electronic GmbH) five cycles of 10 s and after each cycles 100 µl of spore suspension was collected, diluted, plated on MS agar and further incubated at 30°C for 2 days.
Phase contrast microscopy. For light microscopy, sterile coverslips were inserted at a 45° angle into MS agar and spores were inoculated in the acute angle along the glass surface. Coverslips were removed after 3- to 4-day incubation at 30°C and mounted in PBS containing 50% glycerol on poly l-lysine-coated slides. Alternatively liquid culture was spotted directly on microscope slides covered with 1% agarose and covered with coverslips. Samples were observed with an Olympus System Microscope BX60 and pictures taken with an Olympus F-view II camera.
Electron microscopy. TEM for the analysis of ultrathin cross-sections of hyphae and spores was performed as described previously (van Wezel et al., 2000).
For immunoelectron microscopy spores were fixed in 2% paraformaldehyde with 0.2% glutaraldehyde in PHEM buffer for 2 h at room temperature. After washing in PBS the spores were pelleted and embedded in 12% gelatine. The pellet was cut into 1 mm3 cubes, cryoprotected in 2.3 M sucrose and snapfrozen in liquid nitrogen. Ultrathin cryosections were labelled with rabbit anti-MreB (1:2000) followed by 15 nm protein A-gold particles. The labelled sections were embedded and contrasted in methylcellulose with uranyl acetate. All samples were viewed with a Philips EM 410 electron microscope (Eindhoven, the Netherlands).
Thanks to Eriko Takano for fruitful discussion and many helpful comments on the manuscript. This work was supported by grants from the DFG (Forschergruppe 449 ‘Bakterielle Zellhülle’) to W.W. and from the Netherlands Royal Society for Arts and Sciences (KNAW) to G.P.v.W.