Dimeric structure of the cell shape protein MreC and its functional implications

Authors


*E-mail fent@mrc-lmb.cam.ac.uk; Tel. (+44) 122 325 2969; Fax (+44) 122 321 3556.

Summary

The bacterial actin homologue MreB forms helical filaments in the cytoplasm of rod-shaped bacteria where it helps maintain the shape of the cell. MreB is co-transcribed with mreC that encodes a bitopic membrane protein with a major periplasmic domain. Like MreB, MreC is localized in a helical pattern and might be involved in the spatial organization of the peptidoglycan synthesis machinery. Here, we present the structure of the major, periplasmic part of MreC from Listeria monocytogenes at 2.5 Å resolution. MreC forms a dimer through an intimate contact along an N-terminal α-helix that connects the transmembrane region with two C-terminal β-domains. The translational relationship between the molecules enables, in principle, filament formation. One of the β-domains shows structural similarity to the chymotrypsin family of proteins and possesses a highly conserved Thr Ser dipeptide. Unexpectedly, mutagenesis studies show that the dipeptide is dispensable for maintaining cell shape and viability in both Escherichia coli and Bacillus subtilis. Bacterial two-hybrid experiments reveal that MreC interacts with high-molecular-weight penicillin-binding proteins (PBPs), rather than with low-molecular-weight endo- and carboxypeptidases, indicating that MreC might act as a scaffold to which the murein synthases are recruited in order to spatially organize the synthesis of new cell wall material. Deletion analyses indicate which domains of B. subtilis MreC are required for interaction with MreD as well as with the PBPs.

Introduction

The majority of bacteria have a rigid cell wall that maintains their cell shape and resists internal turgor pressure. A fine network of glycan strands linked together by short peptides forms a peptidoglycan (PG) layer (or murein sacculus) that is responsible for the rigidity of the bacterial cell wall (reviewed in Bhavsar and Brown, 2006). As isolated murein sacculi retain the shape of the cell from which they are derived, it is thought that the information for cell shape is implemented in the sacculus (Höltje, 1998). Indeed, many mutations that affect cell shape are located in genes involved in PG synthesis (Spratt, 1975; Spratt and Pardee, 1975; Tamaki et al., 1980; Henriques et al., 1998; Nelson and Young, 2000; Wei et al., 2003; Nilsen et al., 2004). The PG layer is assembled from precursors by a family of proteins known as penicillin-binding proteins (PBPs) that form and cross-link glycan strands through their glycosyl transferase and transpeptidase activities (Höltje, 1998; Popham and Young, 2003). Murein hydrolases create gaps in the PG layer, and in Gram-negative organisms it is thought that newly synthesized peptidoglycan strands are inserted into these gaps to enable cell elongation and septation to occur (Heidrich et al., 2002).

In order to maintain the shape of the cell, the PBPs as well as the murein hydrolases are likely to be temporally and spatially organized. It has been speculated that some of the proteins that affect cell morphology, such as MreB, MreC and MreD (encoded by the mreBCD operon), might position the peptidoglycan synthesis machinery, while others, such as RodA and FtsW, might be involved in transport of PG precursors (Tamaki et al., 1980; Wachi et al., 1987; Levin et al., 1992; Varley and Stewart, 1992; Henriques et al., 1998; Jones et al., 2001; Lee and Stewart, 2003; Figge et al., 2004; Gitai et al., 2004; Formstone and Errington, 2005; Kruse et al., 2005; Leaver and Errington, 2005).

Depletion of MreB, MreC or MreD showed that they are essential for viability and maintenance of the natural shape of the bacterial cell (Lee and Stewart, 2003; Soufo and Graumann, 2003; Formstone and Errington, 2005; Kruse et al., 2005). The mre (for murein region ‘e’) genes were initially identified in Escherichia coli mutants that were affected in their sensitivity to mecillinam, an antibiotic known to inhibit PBP2 (Wachi et al., 1987; 1989).

MreB forms cytoskeletal filaments that follow a helical path just underneath the cell membrane (Jones et al., 2001; Kruse et al., 2003; Defeu Soufo and Graumann, 2004; Figge et al., 2004; Gitai et al., 2004; Kruse and Gerdes, 2005). The structure of MreB resembles that of actin (van den Ent et al., 2001) and the polymer is reminiscent of the one strand of filamentous actin (Amos et al., 2004; Löwe et al., 2004), indicating that an actin-like cytoskeleton in bacteria is required for cell shape determination. Bacillus subtilis has two additional MreB-like genes, Mbl and MreBH, which are now thought to colocalize and therefore to form a single helical cable system (Jones et al., 2001; Carballido-Lopez and Errington, 2003; Carballido-Lopez et al., 2006; Defeu Soufo and Graumann, 2004). All three isologues affect morphogenesis of the cylindrical part of the cell wall (Jones et al., 2001; Soufo and Graumann, 2003), suggesting that they all participate, directly or indirectly, in organization of the wall synthetic system. Using fluorescently labelled vancomycin (that binds the d-ala-d-ala ends of nascent peptidoglycan subunits before they are cross-linked to peptides), it has been shown that new peptidoglycan in B. subtilis is incorporated into the pre-existing cell wall in a helical pattern that is reminiscent of and influenced by the MreB/Mbl/MreBH cables (Daniel and Errington, 2003; Tiyanont et al., 2006). PBP2, involved in longitudinal growth of the peptidoglycan in Caulobacter crescentus, exhibits a helical pattern similar to that of MreB (Figge et al., 2004), further supporting a link between the Mre proteins and the PG-synthesizing machinery. Apart from directing cell shape, the MreB cytoskeleton might also be involved in other functions, such as chromosome partitioning (Soufo and Graumann, 2003; Gitai et al., 2005; Kruse et al., 2006).

The widely conserved genes mreC and mreD are usually arranged immediately downstream of mreB (Wachi et al., 1989; Levin et al., 1992; Varley and Stewart, 1992) in the same operon (Formstone and Errington, 2005) and are associated with the membrane and the division site (Lee et al., 2003; Leaver and Errington, 2005). In contrast to the polytopic membrane protein MreD, MreC is predicted to have a single transmembrane spanning helix with the major C-terminal domain located in the periplasm. Indeed, fractionation studies have shown that MreC is associated with the inner membrane (Kruse et al., 2003). Topology experiments using a PhoA fusion to MreC have indicated that the major C-terminal part is located in the periplasm, whereas the N-terminal helix anchors the protein in the membrane (Lee et al., 2003).

Depletion of MreC or MreD results in the formation of spherical cells that eventually lyse, possibly because of a deficiency in peptidoglycan synthesis in the cylindrical part of the cell (Lee and Stewart, 2003; Leaver and Errington, 2005). MreC in C. crescentus interacts with several PBPs (Divakaruni et al., 2005) and colocalizes with PBP2 in Rhodobacter sphaeroides (Slovak et al., 2006). PBP2 forms helical structures that depend on MreB (Figge et al., 2004). Also, MreC has been found to localize in a helical fashion (Dye et al., 2005; Leaver and Errington, 2005), suggesting that it might form a bridge between the intracellular cytoskeleton and the periplasmic cell wall synthesis machinery. An indication that MreC does interact with the cytoskeleton has come from bacterial two-hybrid screens in E. coli, where MreC might interact with MreB as well as MreD (Kruse et al., 2005). In addition, the helical organization of the MreB cytoskeleton in E. coli is dependent on the presence of MreC, again suggesting a link between MreB and MreC (Kruse et al., 2005). Recently, it has been shown that in C. crescentus MreB and MreC might localize independently of each other, whereas the localization of PBP2 depends on both of them (Dye et al., 2005).

Here, we present the crystal structure of the major, periplasmic part of MreC from Listeria monocytogenes at 2.5 Å resolution. MreC forms a dimer through an intimate contact along the N-terminal α-helix that connects the transmembrane region with the C-terminal β-domains. The second β-domain has structural similarity to the chymotrypsin family of proteins and possesses highly conserved Thr and Ser residues (TS dipeptide). To investigate whether these residues are important for MreC function, they were mutated in E. coli and B. subtilis. Despite their high degree of conservation these residues are not essential for viability and maintenance of cell shape. Bacterial two-hybrid studies revealed that MreC interacts with the high-molecular-weight murein synthases rather than with the low-molecular-weight peptidases and deletion analyses indicate the domains of B. subtilis MreC that are required for interaction with MreD and several of the PBPs.

Results and discussion

Crystal structure of MreC

The crystal structure of the periplasmic part of MreC from the Gram-positive bacterium L. monocytogenes reveals two β-strand domains arranged as wings of a butterfly around a central, N-terminal helix (Fig. 1A). Following the N-terminal helix, the first wing encompasses two β-sheets in an orthogonal arrangement, each consisting of three antiparallel β-strands. The second wing is inserted in front of the C-terminal β-strand of the first wing and contains six antiparallel β-strands, wrapped around each other as in the canonical β-barrel fold. The crystal structure shows MreC in close dimeric contact through a long N-terminal helix (Fig. 1B) as well as through its globular domains (Fig. 4). The helices are held together via hydrophobic interactions between leucine residues, some of them participating in a hepta repeat typical of leucine zippers (Figs 1B and 3). The β-domains of the second molecule are simply translated along the longitudinal axis of the coiled coil domain. This has the implication that the surface of the dimer interface in the β-domain belonging to one molecule is solvent exposed in the other molecule. In principle this could enable filament formation, resulting in a step-like chain of alternating first and second β-domains, with the pairs of helices protruding from every second molecule. The helices point in a similar direction that favours anchoring of both N-termini in the cytoplasmic membrane. A major part of the dimeric interface is dependent on the extended helices, which have a buried surface of 1892 Å2 at the interface, whereas the globular domains have a buried surface 1304 Å2. It needs to be established whether the latter is enough to exert filament formation or whether additional interactions may be needed to support this kind of arrangement.

Figure 1.

A. Ribbon plot of the crystal structure of the major, periplasmic part of MreC (aa 74–283) from L. monocytogenes at 2.5 Å resolution. The secondary structure elements are labelled according to their appearance in the primary sequence.
B. MreC dimerizes through its extended N-terminal helix (in white). The second molecule (depicted in cyan) is a simple translation of the first molecule (blue).
Figures were prepared using molscript and raster3d (Kraulis, 1991; Merritt and Bacon, 1997).

Figure 4.

Conservation plot of MreC dimer. Based on an alignment of over 30 different mreC genes the degree of conservation of each residue was calculated and plotted onto the surface of the structure of the MreC dimer. A mini-ribbon plot is added to facilitate orientation of the molecule (with colour code as in Fig. 1A).
A. MreC dimer from the front, arrow pointing into the tunnel towards the conserved Ser residue. Highly conserved residues are in red, least conserved residues in blue.
B. MreC dimer 180° rotated along the y-axis. To facilitate discrimination between the two monomers, one of them is depicted in blue. The N-terminal helix of this monomer, shown in ribbon representation, blocks access to the conserved diad (arrow) upon dimerization.

Figure 3.

Sequence alignment of MreC from L. monocytogenes, B. subtilis, E. coli, Streptococcus mutans and C. crescentus with the secondary structure elements of L. monocytogenes MreC depicted on top (cylinders for α-helices and arrows for β-strands). The blue coloured residues are conserved among the five sequences shown. The red arrows point towards two absolutely conserved residues among 30 MreC sequences. Numbers are relative to the L. monocytogenes MreC sequence. The figure was prepared using alscript (Barton, 1993).

A similarity search among the currently known structures in the protein databank shows that the first β-domain of MreC has structural similarity with a small nucleolar RNP similar to Gar-1 (Rashid et al., 2006). The second β-domain of MreC is similar to the β-barrel fold characteristic of the family of Trypsin-like serine proteases, of which two of the most similar structures to MreC are shown in Fig. 2 (α-lytic protease and the serine protease SPL). Encouraged by this finding, we investigated whether MreC would have any of the residues that make up the catalytic triad in this family of proteases. The triad of α-lytic protease (Fuhrmann et al., 2004) consists of Asp102, His57 and Ser195, where Ser195 (depicted in yellow in Fig. 2) is the primary nucleophile. Sequence analysis of over 30 bacterial genomes revealed that MreC contains highly conserved Ser and Thr residues (TS dipeptide) (Fig. 3). These residues are located in the second β-domain, in a loop connecting strand 9 and 10 (shown in red in Fig. 2). Superposition of the second β-domain of MreC on the two proteases shows that the conserved Ser of MreC overlaps with a Ser in both proteases (shown in red in Fig. 2). However, the Ser residues in the proteases are not part of the active site, making it unlikely that the conserved dipeptide in MreC is involved in a Trypsin-like protease activity.

Figure 2.

Three top scores of a structural similarity search with MreC among all known structures in the PDB (using dali; Holm and Sander, 1993). The first β-domain of MreC (depicted in blue) has structural resemblance to a small nucleolar RNP similar to Gar-1 [PDP entry 2ey4-c, Z-score 7.2, root mean square deviation (Rmsd) of superimposed 64 CA atoms of 2.2 Å]. The second β-domain of MreC (depicted in white) has structural similarity to α-lytic protease (PDB entry 1qq4, Z-score 7.3, Rmsd 3.2 Å over 85 residues) and the serine protease SPL from Staphylococcus aureus (PDB entry 2as9, Z-score 6.7, Rmsd 3.1 Å over 68 residues). Colour codes: blue domains show structural similarity to each other as do the white domains, structural elements depicted in cyan are distinct. Conserved Ser 226 in MreC is depicted in red and superimposes on the red Ser residues in the proteases, whose active sites are depicted in yellow. The figure was prepared using pymol (DeLano, 2002).

The degree of conservation of all residues was calculated based on an alignment of over 30 different mreC genes and was mapped onto the surface of the structure of the MreC dimer (Fig. 4). The conserved Ser226 is located at the bottom of a channel accessible from the front of the molecule (see arrow in Fig. 4A). At the back of the molecule, access to the cleft containing these conserved residues (arrow in Fig. 4B) is hampered by the N-terminal helices, involved in dimer formation. The N-terminal helix of the second molecule interacts with that of the first molecule and hence the cleft with the conserved Ser and Thr residues in the second molecule remains accessible upon dimer formation.

The conservation plot also shows that the conserved stretch of amino acids (in red) at either end of the N-terminal helix in one molecule makes contact with the similar region in the other molecule, supporting that dimer formation as observed in the crystal structure is functionally relevant (Fig. 4B). Size-exclusion chromatography suggests that the periplasmic portion of MreC from E. coli, B. subtilis and L. monocytogenes all form a dimer in solution at a concentration of 3 mg ml−1 (data not shown). As discussed below, two-hybrid studies also show that B. subtilis MreC self-interacts, supporting dimer formation in a more physiological context.

Mutational analysis of MreC in vivo

To investigate whether the conserved diad Thr225 and Ser226 is required for MreC function, the corresponding residues in B. subtilis and E. coli MreC were changed to Ala Ala, and the phenotypes of the mutant alleles were determined. Because mreC is essential, a B. subtilis strain (Pxyl-mreC) expressing an ectopic copy of native mreC under the control of an inducible promoter was constructed. Subsequently, the original copy was either deleted (ΔmreC, strain 3461) or mutated (strain 3473), and cells were grown in the absence of xylose to reveal the phenotype of the mutations. In contrast to the complete deletion mutant of mreC, changing the conserved Ser and Thr residues into alanines did neither affect cell shape nor affect cell viability (Fig. 5A). To confirm that the mutation was not lethal, the mreCTS230AA allele was introduced in a wild-type background. The resulting strain (3474) did not show any shape defects and grew as wild type.

Figure 5.

A. Strains 3427 (neo mreB+), 3473 [amyE::(spc Pxyl-mreC) neo mreB+mreCTS230AA] and 3461 [amyE::(spc Pxyl-mreC) neo mreB+ΔmreC] were grown overnight at 37°C on nutrient agar with and without 0.5% xylose. Cells of strain 168 (wild-type B. subtilis) are shown for comparison.
B. Images of strains from (A) grown at 37°C in CH medium with and without 0.5% xylose.
C. Western blot quantifying the total number of MreC molecules per cell in B. subtilis. Total cell lysate diluted twofold (loaded approximately 6.7 × 107 and 3.35 × 107 cells per lane (lanes 1 and 2) and pure MreC protein (periplasmic domain, from B. subtilis) with a concentration of 105 to 0.5 ng (lanes 4–12). Lane 3 (M) shows markers of 33 and 48 kDa.

Surprised by the finding that these highly conserved residues are clearly not essential for MreC function in B. subtilis, the comparable mutations were introduced in mreC of E. coli, which lacks the extra MreB-like genes of B. subtilis. To this end, the double mutant (T222A,S223A) was introduced in MreC fused to RFP (cherry) and tested for its ability to convert the spherical shape of ΔmreC cells back to rods. Briefly, plasmids pFB211[Plac::mreC-rfp], pFB218[Plac::mreC(T222A,S223A)-rfp], pFB223[Plac::mreC(1–104)-rfp] and a vector control (pMLB1113) were transformed into the ΔmreC strain FB15/pFB112/pFB128[ΔmreC/sdiA/cI857 PλR::mreD] and cells were grown at 37°C in LB in the presence of IPTG (a range from 0 to 250 μM) to induce expression of the mreC fusions. Transformants carrying pFB211 or pFB218 underwent shape conversion from sphere to rod in an IPTG-dependent manner (10 μM, or higher), while those carrying pFB223 or pMLB1113 remained spherical (even at 250 μM inducer). RFP fused to the wild-type as well as the double mutant protein correctly localized peripherally (Kruse et al., 2005). Thus, whereas MreC(1–104) was insufficient for function, the MreC(T222A,S223A)–RFP fusion appeared to be fully functional.

Based on the above findings, it is very unlikely that these conserved residues are part of an active site. The high degree of conservation of Thr225 might be explained by its structural role. Thr225 is located at the end of β-strand S9 where it interacts with a moderately conserved Gly235. Similarly, Gly227 flanking Ser226 interacts with a conserved Pro233 in the loop connecting S9 with S10. However, the reason for the presence of the highly conserved Ser226 is not so obvious. It might be involved in interactions with a binding partner or has some other function that is redundant and therefore not detectable in the in vivo assays described above.

Interaction partners for MreC

Both the crystal structure and size exclusion chromatography prove MreC to be a dimer. To test for dimerization in vivo we turned to the bacterial two-hybrid system of Karimova et al. (1998). Full-length copies of MreC (MreC1-290) did indeed interact strongly, consistent with dimer formation in vivo (Fig. 6). A C-terminal deletion mutant lacking the two β-domains (MreC1–113) was still able to interact with full-length MreC, whereas further truncations of the N-terminal periplasmic helix H1 (MreC1–77, 1–41, 1–25) eliminated a detectable interaction. This would be consistent with the crystal structure, in which dimerization occurs via intertwining of the extended N-terminal helices of the monomers.

Figure 6.

Bacterial two-hybrid data on the interactions between deletion mutants of MreC and PBPs as well as MreD. Strains of E. coli BTH101 expressing truncations of MreC from plasmid pUT18c and either of the PBPs from pKT25 were spotted onto minimal media supplemented with X-gal and incubated at 30°C until a blue colour developed. The results are from a single experiment performed on the same day with strains spotted on one plate.

Full-length, dimeric MreC also interacted strongly with integral membrane protein MreD (Fig. 6). However, this interaction was lost in all of the truncated constructs, suggesting that a more extensive segment of MreC, including at least part of the β-domain region, is needed for MreC–MreD interaction.

Previously, it has been shown that the phenotype of cells doubly mutated for pbpH and pbpA is similar to that of an MreC depletion strain (Wei et al., 2003; Leaver and Errington, 2005). This suggested that MreC and the two PBPs act in the same pathway of cylindrical cell wall synthesis. Moreover, affinity purification using C. crescentus extracts showed that MreC interacts with several PBPs (Divakaruni et al., 2005). We therefore decided to test whether B. subtilis MreC interacted directly with various PBPs. Full-length MreC gave a positive two-hybrid signal for those PBPs that belong to class A and class B high-molecular-weight PBPs (PBP2b, PBP1, PBP2a, PBP3, PBP4, PBP2c, PBP2d, PbpH and PBP4b), but not for the low-molecular-weight peptidases (PBP4*, PBP5*, PbpX and PBP5; Fig. 6, Table 1). The low-molecular-weight carboxypeptidase, PBP4a, showed weak interaction with MreC (Fig. 6, Table 1). This supports a model in which MreC modulates cylindrical cell wall synthesis via the murein synthases, rather than the hydrolases. As quiet a number of the high-molecular-weight PBPs interact with MreC it is plausible that MreC might have a PBP recognition sequence that establishes the interaction. Most of the interactions between MreC and the PBPs required the β-domains of MreC (PBP1, PBP2a, PBP3 PBP4, PBP2c and PBP4b; Fig. 6). However, the N-terminal domain of MreC, including the periplasmic and transmembrane helices, seemed to be sufficient for interaction with PBP2b, PBP2d and perhaps PbpH (Fig. 6). As reported previously, the helical localization of MreC might or might not be dependent on the MreB cytoskeleton (Figge et al., 2004; Divakaruni et al., 2005; Dye et al., 2005; Kruse et al., 2005; Leaver and Errington, 2005). The sequence of the cytoplasmic tail of MreC is quite variable and is very short in some organisms (only six aa residues in B. subtilis), so it seems unlikely that this makes a strong interaction with MreB. Possibly, the third protein encoded by the mre operon, MreD, could function as an intermediate, establishing the link between the cytoplasmic MreB cytoskeleton and MreC. Quantification of the abundance of B. subtilis MreC by Western blotting showed that it is present at about 12 000 molecules (monomers) per cell (Fig. 5C), similar to the values reported previously for Mbl (12 000–14 000) and MreB (8000) (Jones et al., 2001). This high abundance suggests that MreC may have a structural, rather than an enzymatic role in cell wall synthesis, and in principle, it could make a structure on the outside of the cytoplasmic membrane that follows the MreB/Mbl helix along its entire length. This structure, together with MreD might act as a scaffold to which high-molecular-weight PBPs are recruited in order to spatially organize the synthesis of new cell wall material to maintain the rod shape of the cell.

Table 1.  MreC interactions.
ProteinGenePBP classaMreCb
  • a. 

    PBP, penicillin binding protein. HMW: high-molecular-weight PBPs are either class A with transpeptidase and transglycosylase activities or class B with transpeptidase activity. LMW: low-molecular-weight PBPs are either carboxypeptidases or endopeptidases.

  • b. 

    MreC: interaction with full-length MreC as determined in bacterial two-hybrid screen (scoring: –, no interaction; + to +++, weak to strong interaction, based on Fig. 6).

PBP1ponAHMW class A++
PBP4pbpDHMW class A++
PBP2cpbpFHMW class A+
PBP2dpbpGHMW class A+++
PBP2bpbpBHMW class B+++
PBP2apbpAHMW class B+++
PBP3pbpCHMW class B++
PbpHpbpHHMW class B+++
PBP4bpbpIHMW class B++
PBP4adacCLMW carboxypeptidase+
PBP4*pbpELMW endopeptidase
PBP5*dacBLMW carboxypeptidase
PbpXpbpXLMW endopeptidase
PBP5dacALMW carboxypeptidase
MreCmreC +++
MreDmreD +++

Experimental procedures

MreC protein expression and purification

The gene encoding MreC (NCBI Accession No. CAC99625; TrEMBL Q8Y6Y4) was amplified from L. monocotogenes genomic DNA (ATCC 19115D) by PCR starting from the codon for amino acid 50. The PCR product, generated with the forward primer (5′-CTTTAAGAAGGAGATATACATATGAATATAGTAGCTAAGCCTACTTCAT) and reverse primer (5′-AATGATGATGATGATGATGGGATCCTTGGCCTCCAGTCGTGTCTGAACTA), was used as a primer pair to insert the gene into vector pHis17 (B. Miroux, pers. comm.) according to a restriction-free cloning procedure as described previously (van den Ent and Löwe, 2006). The newly obtained plasmid, pFE164, contains the mreC gene downstream of the T7 promoter, encoding the periplasmic part of MreC with a His-tag at the C-terminus to facilitate purification. In the same way, the periplasmic domain of E. coli MreC and B. subtilis MreC were cloned, resulting in plasmids pFE147 (B. subtilis MreC starting from aa 50) and pFE150 (E. coli MreC starting from aa 57).

MreC was expressed in Bl21AI cells (Invitrogen) at 25°C and purified using Histrap HP (Amersham) affinity purification at pH 6.0, essentially as described before (van den Ent and Löwe, 2000). After the protein was eluted from the column it was concentrated by ultrafiltration using vivaspin 10 k MWCO (Vivascience) and loaded onto a size-exclusion column Sephacryl S200 in TEN7.0 (20 mM Tris 7.0, 1 mM EDTA, 1 mM NaAzide). Typically, yields were around 18 mg from 1 l of culture for L. monocytogenes MreC. Selenomethionine-containing MreC was expressed in Bl21AI cells, which is non-auxotrophic for methionine. By using specific growth conditions methionine biosynthesis was inhibited (Van Duyne et al., 1993). A pre-culture, grown in 2× TY, was used in a 1:1000 dilution to inoculate 200 ml of minimal medium containing 1× M9 supplemented with 0.4% glucose, 2 mM MgSO4, 100 μg ml−1 ampicillin, 1 μg ml−1 of each of the vitamins (riboflavin, niacinamide, pyrodoxine, thiamine), and 1:100 diluted Trace elements (1 l Trace elements contains 50 g of EDTA, 0.8 g of FeCl3·6H2O, 0.05 g of ZnCl2, 0.01 g of CuCl2, 0.01 g of CoCl2·6H2O, 0.01 g of H3BO3, 1.6 g of MnCl2, 0.01 g of NiSO4, 0.01 g of molybdic acid, adjusted to pH 7.0). The overnight culture was diluted 1:50 into 10 l of minimal medium and grown for 4 h to optical density at 600 nm (OD600) of 0.30 at 36°C, when the temperature was dropped to 25°C. Once the OD600 reached a value of 0.5 a mixture of l-amino acids was added as solids [per litre of culture: 50 mg of selenomethionine (Acros), 50 mg of leucine, isoleucine, valine and 100 mg of lysine, threonine and phenylalanine (all from Fluka)]. After 15 min, protein expression was induced by the addition of 0.2% arabinose and the culture was grown for an additional 9 h. Protein purification was performed as described for the non-substituted protein, except that 5 mM β-mercaptoethanol was added to the buffers for the Ni2+ column and 5 mM dithiothreitol (DTT) was used in all other buffers. A typical yield was 2.5 mg of protein per litre of cells. Electrospray mass spectrometry measurements of the non-substituted protein and the selenomethionine-containing protein was used to check selenomethionine incorporation, indicating that all five methionines were substituted by selenomethionine (MreC observed 27 587.0 Da, calculated 27 589.0 Da; SeMet MreC observed 27 822.0 Da; calculated 27 822.0 Da).

Crystallization and structure determination

MreC was crystallized at 86 mg ml−1 by sitting-drop vapour diffusion in 17.7% ethanol, 100 mM Bis-Tris-Propane, pH 7.1 and the crystals were flash frozen in liquid nitrogen in the presence of 18.7% ethanol, 15% glycerol and 100 mM Bis-Tris-Propane 7.1. Selenomethionine substituted MreC crystallized in 14.4% ethanol and 100 mM ADA 6.0 and the crystals were flash frozen in 17% ethanol, 15% glycerol and 100 mM ADA 6.0. MreC crystals belong to space group P622 with two molecules in the asymetric unit. Cell constants were a = 162.95 Å, b = 162.95 Å, c = 95.39 Å (Table 2). Data sets were collected at beamlines SRS 10.1 (Cianci et al., 2005) (selenomethionine containing crystals) and ESRF ID29 (native data set).

Table 2.  Crystallographic data.
P622:
Crystal
a = 162.95 Å, b = 162.95 Å, c = 95.39 Å
l (Å)Resol. (Å)I/sIaRmbMultipl.c
  • a. 

    Signal to noise ratio of intensities, highest resolution bin in brackets.

  • b. 

    Rm: ShSiII(h, i) − I(h)I/ShSi I(h, i) where I(h, i) are symmetry-related intensities and I(h) is the mean intensity of the reflection with unique index h.

  • c. 

    Multiplicity for unique reflections, for MAD data sets I(+) and I(–) are kept separate.

PEAK0.98023.022.9 (6.3)0.08910.2
INFL0.98043.017.3 (3.9)0.12810.2
HREM0.97663.021.0 (5.2)0.10110.1
NATI0.97922.59.8 (1.9)0.1003.4

All data were indexed and integrated with mosflm (Leslie, 1991) and further processed using the ccp4 package (Collaborative Computing Project, 1994). An initial electron density map was generated with the MAD data using solve (Terwilliger and Berendzen, 1999) and phases were calculated with sharp (de la Fortelle and Bricogne, 1997). The model was built manually with the program main (Turk, 1992) and refined using cns (Brünger et al., 1998). Details of the refined model are shown in Table 3.

Table 3.  Refinement statistics.
 NATI
  • a. 

    Five per cent of reflections were randomly selected for determination of the free R-factor, prior to any refinement.

  • b. 

    Temperature factors averaged for all atoms.

  • c. 

    RMS deviations from ideal geometry for bond lengths and restraint angles (Engh and Huber, 1991).

  • d. 

    Percentage of residues in the ‘most favoured region’ of the Ramachandran plot and percentage of outliers (procheck; Laskowski et al., 1993).

  • e. 

    Protein Data Bank identifiers for co-ordinates and structure factors respectively.

Residues1: 74–283; 2: 77–277
Resolution2.5 Å
R-factor, R-freea0.245, 0.266
B averageb75.6 Å2
Geometry bonds/anglesc0.009 Å, 1.469°
Ramachandrand86.9%/0.0%
PDB IDe2j5u, r2j5usf

Construction of a strain of B. subtilis expressing MreCTS229AA

The Thr229 and Ser230 codons of the B. subtilis mreC gene were mutated into alanines. As MreC is an essential gene (Lee and Stewart, 2003; Leaver and Errington, 2005) these residues were mutated at the chromosomal locus in a strain with an ectopic copy of mreC at the amyE locus under the control of the Pxyl promoter. The mutation was introduced by linkage to a neo cassette that conveys resistance to kanamycin. To do this we utilized strain 3427 (trpC2Ωneo3427; Leaver and Errington, 2005), which contains a neomycin cassette upstream the promoter of mreB. Chromosomal DNA of this strain was used to PCR amplify segments of DNA that contained the neo cassette and the mreBCD operon with the desired mutation in mreC. Two primers (USF5, TCCTGGTATGCAGGCCTGTT and USF3, TGCTTCAATAAGATCGCCTTTTT mutation indicated) were designed to amplify a fragment of DNA stretching from 3 kb upstream of the neomycin marker to the codon for the threonine residue at position 229 which was mutated to an alanine codon by the 3′ primer USF3. Two more primers (DSF3, CCAGCACCCGAGCAGCAGGC and DSF5, GCAGGGACAGGCGGTGTTTT CCC) were designed to amplify a fragment stretching from the serine residue at position 230, which was mutated to code for an alanine by the 5′ primer DSF5, and extending 3 kb downstream of the mreC stop codon. Primers USF3 and DSF5 were phosphorylated with polynucleotide kinase (Roche) and used with primers USF5 and DSF3 to PCR amplify fragments of DNA from chromosomal DNA of strain 3427. These two fragments where then purified, ligated together and used to transform competent cells of strain 3437, which contains an ectopic copy of mreC at the amyE locus. The resulting strain, 3473, had the mutated form of mreC at the chromosomal locus (designated allele mreC3473), which was complemented by the wild-type copy of mreC at amyE. The genotype of this strain was confirmed by PCR and sequencing of the whole mreBCD operon. In order to separate the mutant mreC allele from the ectopic copy, wild-type B. subtilis was transformed with chromosomal DNA of strain 3473 and selected for kanamycin. Strains were isolated that were kanamycin resistant and spectinomycin sensitive, indicating that they had not inherited the ectopic copy of mreC. Analysis by PCR and digestion showed that four out of five strains had the mreC3473 allele. The presence of the mutation was verified by sequence analysis (strain 3474).

Growth and maintenance of B. subtilis strains

For growth experiments, B. subtilis strains were grown at 37°C in CH media (Partridge and Errington, 1993), supplemented with or without xylose. B. subtilis cells were transformed by the method of Kunst and Rapoport (1995), then plated onto nutrient agar (Oxoid) supplemented with 5 μg ml−1 kanamycin with 0.5% xylose. To screen transformants, colonies were patched onto nutrient agar containing 50 μg ml−1 spectinomycin with 0.5% xylose.

Quantitative Western

The number of MreC molecules per cell was determined by quantitative Western analysis. The quantity of protein from total cell extract was compared with a dilution series of purified protein. Rabbit antiserum was raised against MreC from B. subtilis (periplasmic part aa 50–290) using a commercial service provided by Eurogentec. Wild-type B. subtilis was grown in CH media to an OD600 of approximately 0.8 and the number of cells per OD unit was determined using a haemocytometer. Two dilutions of total cell lysate and a dilution series of pure MreC protein were separated by electrophoresis on a 4–12% Bis-Tris midi gel (Invitrogen) before being transferred to polyvinylidene difluoride membrane for 3.5 h at 20 mAmps using a wet transfer apparatus (TransBlotTMCell, Bio-Rad). The membrane was blocked with 4% milk powder in PBS before being incubated with a 1:10 000 dilution of anti-MreC antiserum then a 1:10 000 dilution of anti-rabbit IgG conjugated to horseradish peroxidase (Sigma). The antibody was detected using an ECL+ kit (Amersham Biosciences) and the membrane was scanned on an FLA-5000 fluoroimager (Fuji). The protein concentration of the total cell lysate was estimated by comparing the fluorescence intensities of bands on the scanned membrane to those of the standard protein using the Aida Image Analyser program (Raytest).

Construction of a strain of E. coli expressing MreCTS222AA

The Thr222 and Ser223 codons of the E. coli mreC gene were mutated to encode Ala222 and Ala223 using the Quickchange site-directed mutagenesis method (Stratagene). Briefly, plasmid pFB211[lacIqPlac::mreC-rfp] was used as a template for temperature-cycled amplification with primers designed to introduce mutations into both codons. Amplification was carried out with KOD polymerase (Novagen) and primers 5′-GGTGATGTGCTGGTTGCGGCCGGTCTGGGCGGTCGTTTCC-3′ and 5′-GGAAA CGACCGCCCAGACCGGCCGCAACCAGCACATCACC-3′ (alanine codons underlined). The product was treated with DpnI to digest unamplified DNA and transformed into competent DH5α cells. The resulting plasmid (pFB218) contained the desired mutations as verified by sequence analysis.

Construction of a truncated MreC protein encoding amino acids 1 through 104 fused to RFP was made as follows. The mreC reading frame from E. coli MG1655 chromosomal DNA was amplified by PCR with primers 5′-CTAGTCTAGAATACGAGAATACGCATAACTT-3′ and 5′-CCGCTCGAGCAGCGGGGAACCCAGCAGCTCG-3′. The product was digested with XbaI and XhoI (sites underlined). Unexpectedly, the product contained an additional XhoI site near the 3′ primer annealing site (beginning of site is in italics) leading to a slightly shorter fragment than anticipated. The 339 bp fragment was used to replace the 1143 XbaI–XhoI fragment of pFB211, yielding pFB223[Plac::mreC(1–104)-rfp].

Construction of plasmids pFB112, pFB128 and pFB211, and of strain FB15 will be detailed elsewhere (F. Bendezu and P. de Boer, in preparation).

Bacterial two hybrid

We used the bacterial two-hybrid system developed by Karimova et al. (1998). Full-length and truncated versions of the mreC gene were amplified by PCR and ligated into the pKT25 vector. All the other constructs carrying MreD and PBP fusions were derived from plasmid pUT18c (M. Leaver, R. Daniel and J. Errington, in preparation). Half microlitre of pUT18c derivative and 0.5 μl of pKT25 derivative was added to a pre-chilled 0.5 ml tube with 20 μl of E. coli BTH101 competent cells and incubated at 4°C for 30 min. Cells were then heat shocked at 42°C for 90 s. One hundred and eighty microlitres of 2× TY supplemented with 20 mM glucose was added and cells were allowed to recover at 30°C for 45 min before being spotted onto minimal media agar plates with 0.5 mM IPTG. Minimal medium was made up of A+B (3.6 μM FeCl3·6H2O, 40 μM MgCl2·6H2O, 0.1 mM MnCl2·4H2O, 10 mM NH4Cl, 75 μM Na2SO4, 0.5 mM KH2PO4, 1.2 mM NH4NO3) supplemented with 1 mM MgSO4, 0.8% glucose, 0.0001% thiamine, 0.2% casine hydrolysate (Oxoid), 0.008% X-gal [form a 4% stock in dimethylformamide (BDH)], 100 μg ml−1 ampicillin, 25 μg ml−1 kanamycin, and solidified with 1.5% bacteriological agar (agar number 1, Oxoid). Plates were incubated at 30°C for 24–48 h, and any colour change was noted.

Acknowledgements

This work was supported by a Human Frontier Science Program Award (RGP0001-C103) to J.L., P.d.B. and J.E. F.B. was also supported by NIH NRSA Institutional Training Grant T32GM08056. ML was supported by an MRC Graduate Studentship. We would like to thank Dr R. Daniel for his help and stimulating discussions during the course of this work and are grateful to Dr M. Cianci for his support while beamline SRS 10.1 (Daresbury, UK) was commissioned.

Ancillary