One of the main virulence factors of the pathogenic bacterium Streptococcus pneumoniae is the capsule, present at the bacterial surface, surrounding the entire cell. Virtually all the 90 different capsular serotypes of S. pneumoniae, which vary in their chemical composition, express two conserved proteins, Wzd and Wze, which regulate the rate of the synthesis of capsule. In this work, we show that Wzd, a membrane protein, and Wze, a cytoplasmic tyrosine kinase, localize at the bacterial division septum, when expressed together in pneumococcal cells, without requiring the presence of additional proteins encoded in the capsule operon. The interaction between the two proteins and their consequent septal localization was dependent on a functional ATP binding domain of Wze. In the absence of either Wzd or Wze, capsule was still produced, linked to the cell surface, but it was absent from the division septum. We propose that Wzd and Wze are spatial regulators of capsular polysaccharide synthesis and, in the presence of ATP, localize at the division site, ensuring that capsule is produced in co-ordination with cell wall synthesis, resulting in full encapsulation of the pneumococcal cells.
Streptococcus pneumoniae is associated with a variety of infections that can range in severity from otitis media to pneumonia or meningitis (Kadioglu et al., 2008). This Gram-positive bacterium is a common respiratory pathogen and a frequent cause of community-acquired pneumonia in developed countries. Despite the increased availability of new antibiotics, the mortality rate of patients presenting pneumococcal pneumonia and concurrent septicaemia has remained unchanged, with values higher than 20%. In developing countries pneumococcal septicaemia is a major cause of infant mortality, causing 25% of all preventable deaths in children under the age of 5 years (Kadioglu et al., 2008).
A capsular polysaccharide (capsule or CPS), surrounds pneumococcal bacteria, forming a thick layer of 200–400 nm (Sorensen et al., 1988). CPS may be linked to bacterial peptidoglycan (Sorensen et al., 1990), an heterogeneous polymer, which surrounds and protects bacterial cells, made of glycan chains of β-(1–4)-linked N-acetylglucosamine and N-acetylmuramic acid, and cross-linked by short peptides (Vollmer et al., 2008).
The pneumococcal CPS plays a major role in the colonization and persistence of S. pneumoniae in the infected host as non-encapsulated mutants of clinical pneumococcal isolates are non-virulent (Kadioglu et al., 2008). This is probably due to the ability of the CPS to form a shield that prevents antibodies and complement components, which bind to deeper cell structures such as teichoic acids and cell surface proteins, from interacting with their receptors on the host phagocytic cells (Winkelstein, 1981; Musher, 1992). The capsule also seems important to reduce the deposition of complement on the bacterial cell surface (Abeyta et al., 2003) and the trapping of bacteria in neutrophil extracellular traps (Wartha et al., 2007).
Although more than 90 different pneumococcal capsule serotypes are known, only a limited number seems to be associated with invasive pneumococcal disease (Babl et al., 2001) and these have been specifically targeted by the pneumococcal capsular polysaccharide conjugate vaccines recently developed (Pletz et al., 2008). The vaccination programmes in different countries seem to have led to a reduction of the total incidence of pneumococcal disease caused by the vaccines serotypes (Hicks et al., 2007). However, an increase of pneumococcal disease caused by serotypes that are not covered by the different vaccines, has been observed (Aguiar et al., 2010), as well as the alteration of the serotypes most frequently found in asymptomatic human carriers (Frazao et al., 2005; Huang et al., 2005). This worrying trend indicates that vaccines effective against most, if not all, serotypes would be of great importance.
The capsular biosynthetic loci of 90 different pneumococcal serotypes have been recently sequenced (Bentley et al., 2006). In virtually all serotypes, the cps operon includes serotype-specific genes, namely those that encode enzymes required for the synthesis of the specific sugar components of each capsular polysaccharide, as well as genes that are conserved among different serotypes (absent only in serotypes 3 and 37). These proteins are involved in the translocation of the capsular polysaccharide repeat unit to the outer surface of the plasma membrane (Wzx), in the polymerization of the capsule (Wzy), leading to the formation of high-molecular weight capsular polysaccharide, or in the regulation of CPS synthesis (Wzg, Wzh, Wzd and Wze, the products of the first four genes of the cps operon) (Bentley et al., 2006). In this work, we focused on the role of two of these proteins widely conserved in the different S. pneumoniae serotypes: Wzd and Wze.
Wze (also known as CpsD) is an autophosphorylating tyrosine kinase (Morona et al., 2000a). Some proteins that belong to the bacterial tyrosine kinase family function as co-polymerases in the biosynthesis of capsular and extracellular polysaccharides (Grangeasse et al., 2007; 2009). These proteins usually possess a transmembrane domain and an intracellular catalytic domain that, in Gram-negative bacteria, form a single polypeptide. On the contrary, in Gram-positive bacteria, the two domains occur in the form of two distinct polypeptide chains (Grangeasse et al., 2007; 2009). In the case of S. pneumoniae, Wzd (also known as CpsC), a membrane protein that belongs to the polysaccharide co-polymerase family (Morona et al., 2000b), constitutes the external domain, and Wze constitutes the catalytic domain. Because members of the bacterial tyrosine kinase family regulate the production and export of bacterial polysaccharides which play essential roles in the initial stages of infection by pathogenic bacteria, these proteins have raised attention as potential novel antibacterial targets (Cozzone, 2009). Most important, structural studies have shown that these bacterial tyrosine kinases possess unique features with no equivalent in eukaryotic kinases (Olivares-Illana et al., 2008).
As a bacterial tyrosine kinase, Wze possesses the conserved Walker A and B ATP binding motifs and the C-terminal tyrosine rich region, whose phosphorylation seems to regulate the rate of the synthesis of the capsular polysaccharide (Morona et al., 2000a). The binding of ATP to Wze is necessary for capsular polysaccharide synthesis, as Wze mutants carrying an inactive ATP binding site (G48A, K49A) seem to be unable to produce CPS, similarly to a wze null mutant (Morona et al., 2000a). On the other hand, tyrosine phosphorylation of Wze does not seem necessary for capsular polysaccharide production. In fact, S. pneumoniae expressing mutant forms of Wze, where the C-terminal tyrosines have been replaced by phenylalanines, making the protein unable to be phosphorylated, can still produce CPS (Morona et al., 2000a). More recently, an additional role of the phosphorylated Wze in the attachment of CPS to the pneumococcal cell wall, required for pneumococcal invasive disease, has been proposed (Morona et al., 2006).
Wzd is essential for the synthesis of CPS as no significant amounts of capsule, determined by colony phenotype or reaction with specific polyclonal antiserum, were found in wzd deletion mutants (Morona et al., 2000a; Bender et al., 2003). Wzd is thought to be involved in the regulation of the synthesis of the bacterial capsule through the promotion of the phosphorylation of the cytoplasmic Wze (Bender and Yother, 2001).
The current model for capsule synthesis states that Wzd and Wze interact in the presence of ATP to promote CPS synthesis. When Wzd induces Wze autophosphorylation at its C-terminal tyrosines, a reduction of the rate of CPS synthesis occurs, probably due to changes in interactions with Wzy (the capsule polymerase). This allows the CPS to be linked to the peptidoglycan by an unknown ligase. Finally, dephosphorylation of Wze enables the cycle to be repeated (Kadioglu et al., 2008).
In this paper we report a previously uncharacterized role of Wzd and Wze in the co-ordination of the synthesis of the CPS with pneumococcal cell division, by showing that these two proteins localize at the division septum and that this localization is required for the presence of CPS at the septum. Pneumococcal Wzd and Wze null mutants were able to produce a capsule attached to the cell surface but this polysaccharide was absent from the division septum. We therefore propose that Wzd/Wze proteins may act as spatial regulators of capsule metabolism, ensuring the CPS is produced synchronously with the synthesis of peptidoglycan at the division septum.
Membrane-linked Wzd and cytoplasmic Wze are found at the division septa of encapsulated S. pneumoniae bacteria
The localization of the capsule synthetic machinery in S. pneumoniae has not been previously studied. In order to determine the localization of Wzd and Wze in the pneumococcal cell we have constructed fluorescent derivatives of these proteins in which their C-terminal end was fused to the fluorescent protein Citrine. Citrine is a variant of the Green Fluorescent Protein that fluoresces in the yellow range of the light spectrum and that has increased photostability and resistance to fluctuations of the environmental pH (Griesbeck et al., 2001). Genes encoding for Wzd-Citrine and Wze-Citrine proteins were individually introduced into the cps operon of the encapsulated ATCC6314 strain, substituting the native wzd and wze genes, in strains BCSMH003 and BCSMH004 respectively.
Wzd-Citrine and Wze-Citrine proteins were found at the division septa of pneumococcal bacteria and had similar localization patterns at different stages of the cell cycle (Fig. 1A), suggesting that the recruitment of Wzd and Wze to the division septum occurs in a similar manner.
To determine the temporal window of the pneumococcal cell cycle during which the localization of Wze takes place at the septum, we labelled an exponentially growing culture of strain BCSMH004, expressing Wze-Citrine, with DAPI, a fluorescent stain that labels DNA, and with Nile-Red, a dye that becomes fluorescent when incorporated into bacterial membranes. Using these dyes we were able to localize bacterial chromosomes and evaluate the stage of membrane invagination during the formation of the division septum. We then assigned labelled bacteria to five different classes, according to the progression of the cell cycle, from newborn bacteria with only one chromosome and no membrane at the division septum (class I, Fig. 1B) to bacteria that have nearly completed division, with fully segregated chromosomes and with a clear invagination of the division septum (class V, Fig. 1B). The same cells were also classified according to the localization of fluorescent Wze-Citrine protein (classes A to E, Fig. 1B). This analysis showed that in newborn bacteria, even before the onset of septum invagination (classes I and II), Wze is already localized at mid-cell, seen as two spots that indicate the formation of a ring (classes A and B). Later on the pneumococcal cell cycle, when a membrane septum is being formed and chromosome segregation has proceeded (class III), Wze-Citrine is observed as a line consistent with septal localization (class C). In a small percentage of cells (1.6%, n = 1254), we could observe that fluorescence was spread between the old and the new septa (class D). This is consistent with a migration of the Wze protein between the old and new septa along the membrane. Additionally, in a small percentage of the cells (2.1%, n = 1254), Wze is already found at the mid-cell of the newly born cells while still present at the old division septum (class E).
To test for the functionality of Wzd and Wze fluorescent derivatives expressed by BCSMH003 and BCSMH004, respectively, we performed a Quellung reaction. For that purpose, these strains were incubated with rabbit-serum raised against pneumococcal capsular polysaccharide serotype 14, which resulted in a positive reaction of agglutination. Agglutination reactions were also observed with the parental encapsulated ATCC6314 strain, with the Wzd null mutant expressing constitutively untagged Wzd (strain BCSMH025) and with the Wze null mutant expressing constitutively untagged Wze (strain BCSMH026) (Fig. S1). This indicated that both Wzd-Citrine and Wze-Citrine protein fusions were functional, as Wzd and Wze null mutants (strains BCSMH001 and BCSMH002 respectively) seemed impaired in their ability to produce capsule as they failed to agglutinate (Fig. S1).
We further confirmed that the fluorescent derivatives of Wzd and Wze were functional by looking for phosphorylated Wze. It was previously shown that Wzd is required for Wze autophosphorylation at its C-terminal tyrosines (Bender and Yother, 2001). Therefore, if the Wze fluorescent derivative retained its ability to autophosphorylate, and the Wzd fluorescent derivative its capacity to induce Wze autophosphorylation, this would further indicate the protein fusions were functional. We checked for the presence of the phosphorylated Wze fluorescence derivatives in unencapsulated strains BCSMH006 and BCSMH011, expressing only Wze-mCherry, and in strains BCSMH009 and BCSMH014, expressing Wze-mCherry together with Wzd-Citrine, by Western blot using an antibody against phosphorylated tyrosines. Wze-mCherry was phosphorylated when Wzd-Citrine was also expressed but not in its absence (Fig. S2A), indicating that both fusions were functional. We did not detect additional bands that could result from phosphorylated untagged Wze, originated from cleavage of the fusion proteins. The phosphorylated Wze protein was only observed in the encapsulated ATCC6314 control strain (Fig. S2A).
Importantly for experiments described below, we observed that expressing Wzd-Citrine or Wze-Citrine in the encapsulated strains BCSMH019 and BCSMH016, respectively, under the control of a constitutive promoter, also resulted in a fluorescent signal localized at the division septum (see below). Expression of Wze-Citrine fluorescent protein from the replicative plasmid pBCSMH004 in strain BCSMH016 occurred at similar levels as when this protein was produced from the coding sequence inserted into the chromosome, under the control of the chromosomal cps promoter, in strain BCSMH004 (Fig. S2B). The possibility that Wzd and Wze fluorescent derivatives could be being cleaved, which would result in functional and untagged Wzd and Wze proteins was ruled out because BCSMH004 and BCSMH003 strains produced fluorescent proteins of the expected molecular weight of the fusion proteins. This was determined by Western blotting with a specific antibody against the fluorescent domain of the protein fusions (Fig. S2C).
Coexpression of Wzd and Wze is sufficient for septal localization of both proteins
Having determined that fluorescent derivatives of Wzd and Wze localized at the division septum of encapsulated pneumococcal bacteria, we then asked whether this localization was dependent on the ability of bacteria to produce a capsule. To answer this question, plasmids pBCSMH007 and pBCSMH004 were transformed into unencapsulated strains R36A (unencapsulated laboratory strain) and BCSMC001 (unencapsulated strain derived from ATCC6314, in which the entire cps operon was deleted). Constitutive expression of either Wzd-Citrine in unencapsulated BCSMH008 and BCSMH013 strains or Wze-Citrine in unencapsulated BCSMH007 and BCSMH012 strains, did not result in localization of the fluorescent signal at the division septum (Fig. 2A): Wzd-Citrine was distributed all over the membrane and Wze-Citrine was dispersed throughout the entire cytoplasm. We further confirmed the cellular localization of Wzd-Citrine and Wze-Citrine by cell fractionation analysis, which showed that Wzd-Citrine was found at the membrane while Wze-Citrine was found in the cytoplasm (Fig. S2D). These results are in accordance with the fact that Wzd carries two hydrophobic regions, responsible for its insertion in the membrane, while Wze is predicted to be a cytoplasmic protein.
In order to confirm that the localizations observed in the unencapsulated strains are indeed different from the ones observed in the encapsulated strains we calculated the fluorescence ratio (FR) between the intensity of the fluorescent signal at the division septum, of cells that had initiated division (class III), and the fluorescent signal measured at the lateral cell wall of the same cells. When a fluorescent protein is evenly distributed all over the membrane, the intensity of the fluorescence at the septum, which contains two membranes, should be approximately twice the fluorescence at the lateral wall. On the other hand, when a protein is localized at the septum the fluorescence ratio should be higher than two (Pereira et al., 2007; Atilano et al., 2010). Determination of the FR in encapsulated strains BCSMH019 and BCSMH016, clearly indicated that Wzd and Wze, respectively, were localized at the septum (Fig. 2A). This is different from the result obtained for the unencapsulated strains expressing Wzd (BCSMH008 and BCSMH013) and Wze (BCSMH007 and BCSMH012) fluorescent derivatives. The FR determined for these strains was consistent with an even distribution of the proteins over the cells (Fig. 2A).
We ruled out the hypothesis that the fluorescent domain of the protein fusions was preferably cleaved in the unencapsulated strains, which would result in the presence of the fluorescent signal throughout the cytoplasm of the cells, because expression of Wzd fluorescent derivative in both encapsulated BCSMH019 and in the unencapsulated BCSMH008 and BCSMH013 strains produced protein products of similar molecular weight and at similar levels (Fig. 2B, left panels). Similar results were obtained with expression of Wze fluorescent derivative in both encapsulated BCSMH016 and in the unencapsulated BCSMH007 and BCSMH012 strains (Fig. 2B, right panels). Wzd and Wze fluorescent derivatives were not cleaved as shown by Western blotting with a specific antibody that recognizes Citrine protein (Fig. 2B, lower panels).
The observation that fluorescent derivatives of Wzd and Wze could localize at the division septum in bacteria capable of producing a capsular polysaccharide, but not in unencapsulated strains, suggested that there was an additional protein, encoded by the cps operon, involved in recruitment of Wzd and Wze to the septa. Alternatively, a protein complex containing Wzd and Wze, could be able to find the division septum on its own, or by interacting with a topological marker present at the pneumococcal division septum. We therefore decided to coexpress, in unencapsulated pneumococcal bacteria, Wze, fused to the fluorescent mCherry protein, and Wzd-Citrine. Expression of Wze-mCherry alone, in strains BCSMH006 and BCSMH011, resulted in a dispersed fluorescent signal throughout the entire cytoplasm. Expression of Wzd-Citrine alone, in strains BCSMH008 and BCSMH013, resulted in dispersal of the fluorescent signal all over the membrane (Figs 2C and S3A), similarly to what was described in Fig. 2A. However, when Wze-mCherry and Wzd-Citrine were coexpressed in the unencapsulated strains, BCSMH009 and BCSMH014, both proteins localized at the septum (Figs 2C and S3A). Similar results were obtained when Wze-mCherry was coexpressed with Wzd in the unencapsulated BCSMH010 strain. We confirmed that in all strains there was no cleavage of the fluorescent domains of the Wzd and Wze derivatives, and that the proteins were being produced at similar levels (Fig. S3B). Furthermore, we confirmed that, as expected, untagged mCherry (expressed in the unencapsulated BCSLF001 strain and in the encapsulated BCSLF002 strain) was found dispersed throughout the entire cytoplasm (Fig. S3C).
These results show that the pair Wzd/Wze is recruited to the division septa in the absence of any additional protein encoded by the cps operon.
Wzd and Wze interact and this interaction requires the Wze ATP binding domain
As coexpression of proteins Wzd and Wze is necessary and sufficient for their localization at the division septum, it seemed likely that these two proteins interacted. In order to test this hypothesis, we used a bacterial two-hybrid system (Karimova et al., 1998), based on the expression, in an Escherichia coli strain deficient in endogenous adenylate cyclase, of two inactive fragments of the catalytic domain of Bordetella pertussis adenylate cyclase, T25 and T18, which, when fused to interacting polypeptides, can activate the synthesis of cAMP and consequently restore the ability of the E. coli mutant strain to ferment lactose or maltose. This approach showed that Wzd was indeed capable of interacting with Wze, as extracts from E. coli cells expressing the proteins T25-Wzd and T18-Wze had a β-galactosidase activity 30 times higher than that resulting from the sole expression of the untagged protein fragments T25 and T18 (Fig. 3A).
A functional Walker A motif and the enriched Tyr cluster in Wze were previously shown to be important for the regulation of the synthesis of the capsule in S. pneumoniae (Morona et al., 2000a). We therefore asked if these motifs were also important for the ability of Wze to interact directly with Wzd, and consequently for the localization of the pair Wzd/Wze at the division septum. We used again the bacterial two-hybrid system to test the interaction between Wzd and Wze(WA), a previously described mutated form of Wze in which the Walker A motif is inactivated by the mutations G48A and K49A (Morona et al., 2000a). In this case, the β-galactosidase activity was similar to the negative control (Fig. 3A). Similar results were obtained with a mutated form of Wze, named Wze(WA2), in which the Walker A motif was inactivated by the mutation K49M (data not shown). This mutation results in an inactive P-loop in the Wze homologue of Staphylococcus aureus, CapB, without significant alterations of the protein structure (Olivares-Illana et al., 2008). On the contrary, when the mutant Wze(Y) protein, with the tyrosines of the C-terminus cluster mutated to phenylalanines, was tested in the bacterial two-hybrid system, it resulted in β-galactosidase activity 16 times higher than the negative control (Fig. 3A).
These results show that a functional Walker A motif is essential for Wze to interact with Wzd, while the ability of Wze to phosphorylate its C-terminal tyrosines is not.
The ATP binding domain of Wze is required for recruitment of the pair Wzd/Wze at the division septum
Because mutations in the conserved Walker A and C-terminal tyrosine cluster motifs of Wze had different effects on its interaction with Wzd, we tested the effect of those mutations on the localization of Wze. In order to do that, we transformed the encapsulated ATCC6314 strain with plasmids pBCSMH005 and pBCSMH006, encoding the Walker A, Wze(WA), and the tyrosine cluster, Wze(Y), mutant forms of Wze, respectively, fused to fluorescent protein mCherry, obtaining strains BCSMH017 and BCSMH018.
Microscope visualization of the encapsulated BCSMH017 strain, expressing Wze(WA)-mCherry showed that the protein did not localize at the division septa (Fig. 3B). Similar results were obtained with encapsulated BCSMH024 strain, expressing Wze(WA2)-mCherry (data not shown). On the other hand, expression of Wze(Y)-mCherry in the encapsulated BCSMH018 strain, resulted in septal localization (Fig. 3B). The three forms of Wze fluorescent derivatives had similar sizes and were produced at similar levels (Fig. 3C).
These observations indicated that the ability of Wze to bind ATP, but not to autophosphorylate its C-terminal tyrosines, is necessary for its interaction with Wzd and consequent recruitment to the division septa.
In order to determine if ATP binding was required only for the interaction of Wze with Wzd or also for the actual recruitment of the Wzd/Wze complex to the division septum, we linked both proteins through an 11 amino acid flexible peptide linker (L11), which should overcome the requirement of ATP binding for Wze to interact with Wzd and allow us to evaluate its role solely on the localization of the pair of proteins.
Strain BCSTR001, expressing the chimera protein Wzd-L11-Wze-Citrine in the background of unencapsulated R36A, showed a fluorescent signal localized at the division septum, similarly to what was observed in strain BCSMH009, when Wzd-Citrine was coexpressed together with Wze-mCherry (Fig. 4). On the contrary, strain BCSTR002, unencapsulated R36A strain expressing the protein Wzd-L11-Wze(WA)-Citrine, which has the inactivated Walker A motif, presented a fluorescent signal, which was no longer localized at the septum (Fig. 4). The results obtained here showed that ATP binding is required not only for the ability of Wze to interact with Wzd but also for the correct localization of these two proteins at the division septa.
The chimera protein Wzd-L11-Wze-Citrine seems functional, because when this protein was expressed in strain BCSTR003, lacking Wzd, or in strain BCSTR004, lacking Wze it could restore capsule production as determined by Quellung reaction (Fig. S1). Furthermore, detection of phosphorylated tyrosines in the unencapsulated R36A expressing Wzd-L11-Wze-Citrine (strain BCSTR001), in a protein with the correct size for the fusion protein, indicated that this was not cleaved and that both Wzd and Wze domains were functional (Fig. S2E), as it has been shown that Wze only phosphorylates its C-terminal tyrosines in the presence of Wzd (Bender and Yother, 2001).
Capsular polysaccharide is produced in S. pneumoniae Wzd or Wze null mutants, but it is absent from the division septa
In order to study the effect of lack of Wzd or Wze on CPS production we deleted the wzd and wze genes from the chromosome of the encapsulated ATCC6314 strain, obtaining the BCSMH001 and BCSMH002 strains respectively. No resistance marker was left in the CPS operon to minimize polar effects in downstream genes. The last 50 nt of each deleted gene were kept in the chromosome of the mutant strains because we found that they were required for the expression of the protein encoded by the gene located downstream in the cps operon (data not shown). Both BCSMH001 and BCSMH002 strains seemed to have an impaired ability to synthesize the capsular polysaccharide, compared with the parental ATCC6314 strain, as assayed by Quellung reaction, in which the mutant strains showed a decreased ability to agglutinate in the presence of serum against the type 14 capsular polysaccharide (Figs 5A and S1).
We performed a dot-blot with purified commercially available serum, which recognizes capsular polysaccharide 14, to determine if capsule was present in wzd and wze null mutant cells, in their purified cell walls and in samples of purified peptidoglycan. As seen in Fig. 5B, we found that capsule was being produced at lower levels in the wzd and wze null mutant cells than in the wild-type encapsulated ATCC6314 cells and it was absent from the unencapsulated BCSMC001 cells. Moreover, the capsule was present in samples of purified cell walls from the wzd and wze null mutants, similarly to the parental encapsulated strain, and absent in samples of the purified peptidoglycan (Fig. 5B). These results suggest that the capsule was covalently attached to the peptidoglycan in both wzd and wze null mutants, identical to what was observed in the parental strain, and it was absent when hydrofluoric acid was used to remove material linked to peptidoglycan through phosphodiester bonds, a step used in the purification of bacterial peptidoglycan. Similar amounts of cell walls and peptidoglycan were present in the dot-blot (see Experimental procedures and Fig. 5B).
In order to determine not only if, but also where, capsule was present, we labelled the capsular polysaccharide using again the purified rabbit-serum raised against pneumococcal capsular polysaccharide serotype 14 and showed by immunofluorescence microscopy that in BCSMH001 and BCSMH002, the ATCC6314 wzd and wze null mutants strains, respectively, the capsular polysaccharide was absent from the division septum (Fig. 5C). This is the site where Wzd and Wze proteins can localize if coexpressed, and where new cell wall is synthesized (see below). However, the capsule was still present in the mature cell wall, judged by the observation of an immunofluorescence signal. This is in accordance with what Morona and colleagues have reported previously (Morona et al., 2000a). As expected, the capsule was present over the entire surface of the pneumococci cells in the parental ATCC6314 strain, and no signal was detected in BCSMC001, the cps null mutant of ATCC6314 (Fig. 5C).
To determine where the new cell wall was being synthesized and if it corresponded to the same place where Wze and Wzd localized, we labelled encapsulated bacteria BCSMH015, expressing Wze-mCherry, with Van-FL, a fluorescent derivative of vancomycin, an antibiotic that binds to the recently translocated peptidoglycan precursors that still carry an intact D-Ala-D-Ala carboxyl terminus (Daniel and Errington, 2003). Fluorescent signals of Van-FL and Wze-mCherry colocalized at the division septum (Fig. 5D), indicating that Wze localizes where the new peptidoglycan is being synthesized.
The observation that Wzd and Wze were required for the presence of the capsular polysaccharide at the division septum suggests that these proteins are involved in the spatial-temporal regulation of the capsule metabolism, ensuring the capsule is present at the division septum, where the new cell wall is being synthesized.
Protein localization studies in S. pneumoniae have been done mainly by immunofluorescence, probably due to the fact that these bacteria are microaerophiles. Low levels of oxygen could prevent post-translational oxidation of GFP fusion proteins with the consequent absence of fluorescent signal, when expressed in pneumococci. However, Eberhardt and colleagues have recently described the localization in live cells of fluorescent derivatives of proteins involved in the essential pathway of choline metabolism in S. pneumoniae (Eberhardt et al., 2009).
In this work we have studied the localization of the regulators of capsule synthesis in S. pneumoniae, Wzd and Wze. We substituted wzd and wze in the cps operon by genes encoding the fluorescent derivatives of Wzd and Wze. This strategy increases the likelihood that fluorescent proteins are produced at the same level as the native unlabeled proteins, with minimal alterations in the expression of the other proteins encoded in the cps operon.
Wzd and Wze localized at the septum of live and dividing encapsulated pneumococcal cells. Both proteins were recruited to the septum early in the cell cycle, as the corresponding fluorescent signal was observed at mid-cell in newborn bacteria, before the onset of invagination.
Surprisingly, we found that when the localization of the fluorescent derivatives of Wzd and Wze was determined in unencapsulated bacteria both Wzd and Wze proteins were no longer localized at the septum, but had a dispersed localization consistent with their amino acid composition: Wzd was found all over the cell membrane, while Wze was dispersed throughout the cell cytoplasm.
The observation that Wzd and Wze localized at the division septum only in encapsulated bacteria suggested that one or more proteins, encoded in the cps operon, could recruit Wzd and Wze to the division septum. An alternative hypothesis was that Wzd and Wze, assembled in a protein complex and were able to find the septum on their own, or through the interaction with a topological marker present at the pneumococcal division septum. The second scenario seems more likely, as we observed that coexpression of fluorescent derivatives of both Wzd and Wze in unencapsulated strains resulted in a clear septal localization. We have not identified the topological marker that determines Wzd/Wze septal localization. Interestingly, Wze has homology with proteins from the large P-loop NTPase superfamily that include members such as the MinD ATPase, a constituent of the MinCDE complex that contributes to the determination of the place where the division septum is formed in different bacteria (Shih and Rothfield, 2006). In Gram-positive bacteria, such as Bacillus subtilis, MinD can interact with the bacterial membrane and accumulate at cell division sites and cell poles (Marston et al., 1998). This results in the localization of MinC, an inhibitor of the polymerization of FtsZ, at these sites, which ensures that bacterial division is temporally and spatially regulated (Marston et al., 1998).
Bacterial two-hybrid assays showed that Wzd and Wze interact and that a functional ATP binding domain in Wze is required for their interactions. This is in agreement with recent data obtained with Wzd and Wze homologues from Streptococcus thermophilus using a yeast two-hybrid assay (Cefalo et al., 2011). Wze contains a divergent Walker A motif (A/G-X-X-X-X-G-K-S/T, where X can be any amino acid); and a Walker B ATP binding motif (h-h-h-h-D, where h represents a hydrophobic amino acid) (Leipe et al., 2002). Besides these two motifs, Wze also has an additional motif, termed Walker A′ (D-X-D), generally found in cell division regulatory proteins of the MinD/Mrp family and present in different bacterial tyrosine kinases involved with the synthesis of capsular polysaccharides (Cozzone, 2009). Morona and colleagues have described that a mutation of the Walker A motif, which inactivates the ATP binding ability of the protein, resulted in the elimination of capsule synthesis (Morona et al., 2000a). On the contrary, mutation of the tyrosines at the C-terminus of the protein had no apparent effect on capsule production. We now found that a functional Walker A motif in Wze is also essential for its interaction with Wzd, while the tyrosines at the C-terminus of Wze are not. These results suggest that the reason why an ATP binding Wze mutant is impaired in its ability to produce capsule, may be a consequence of its inability to interact with Wzd and properly localize.
Wze belongs to a bacterial tyrosine kinase protein family that include members reported to act as co-polymerases in the biosynthesis of capsular and extracellular polysaccharides (Grangeasse et al., 2007; 2009). In Gram-negative bacteria, the tyrosine kinase domain is linked to a transmembrane domain, functional homologous to Wzd, forming a single polypeptide. We have reproduced the Gram-negative organization of bacterial tyrosine kinases in S. pneumoniae by constructing a fluorescent derivative of the chimera Wzd-L11-Wze protein, in which Wzd was linked to Wze through an 11-amino acid flexible linker. This strain allowed us to test if ATP binding activity was required only for Wzd/Wze interactions (as showed in the bacterial two-hybrid assays), or also for the septal localization of these proteins, as the need of ATP binding for the interaction between Wzd and Wze would be by-passed by the physical linkage of the two proteins. The fact that expression of Wzd-L11-Wze-Citrine, but not Wzd-L11-Wze(WA)-Citrine (containing a Walker A mutant form of Wze), in the unencapsulated strain R36A resulted in localization at the division septum, showed that the ability of Wze to bind ATP is required not only for its interaction with Wzd, but also for the correct localization of the proteins. However, at this moment, we cannot rule out the possibility that the absence of a functional ATP binding domain may alter the structure of Wze(WA) in such a way that Wzd and Wze(WA) domains can not interact properly in the Wzd-L11-Wze(WA)-Citrine protein. If this were the case, it would also result in the inability of the fusion protein to localize at the division septum.
In order to clarify the role of Wzd and Wze in the production of capsular polysaccharide, we constructed mutants of the encapsulated strain ATCC6314 where the genes encoding these proteins were deleted. The synthesis of the capsule in these wzd and wze null mutants seemed to be affected, as seen by their reduced ability to agglutinate when incubated with serotype specific serum in a Quellung reaction. These results are in accordance with previous reports describing that Wzd or Wze mutants were not able to produce capsule, as judged by Quellung reactions and ELISA experiments (Morona et al., 2000a; 2004; Bender and Yother, 2001).
However, we have now determined that wzd and wze mutants still produce capsule but in a different manner from the parental ATCC6314 strain. In the encapsulated parental strain the capsule was distributed all around the surface of the cells, as expected. However, in the wzd and wze null mutants, capsule was produced at lower levels than those found in the encapsulated ATCC6314 parental strain, as determined by quantitative detection of capsule bound to cells in a dot-blot assay. More interestingly, the capsular polysaccharide was absent from the division septum of the wzd and wze null mutant cells, but was still present in the mature cell wall as determined by a immunofluorescence microscopy assay using a purified serum against serotype 14 capsular polysaccharide. A similar observation has been made by Morona and colleagues (Morona et al., 2000a), with deletion mutants constructed in an encapsulated serotype 19 S. pneumoniae strain. The authors explained this observation as resulting from impaired synthesis of capsule, where this polysaccharide was synthesized, exported and attached to the cell wall at very low basal rate (Morona et al., 2000a).
In theory, absence of the capsule from the division septum of wzd and wze null mutants could result in the inability of these mutants to attach capsule to the newly assembled peptidoglycan. In fact, phosphorylated Wze has been previously proposed to have a role in the attachment of capsule to the cell wall (Morona et al., 2006). However, we do not favour this hypothesis, as we observed that the capsule produced by the wzd and wze null mutants was as tightly associated to peptidoglycan as in the parental strain (Fig. 5B). The capsular polysaccharide was removed from the peptidoglycan only after incubation with hydrofluoric acid, which is known to cleave phosphodiester bonds and thus remove teichoic acids and polysaccharides covalently bound to the bacterial peptidoglycan, similarly to what has been previously described (Sorensen et al., 1990). It should be emphasized that the proposed role for Wzd/Wze in attaching the capsule to the cell wall has been made based on the characterization of the amount of capsule produced by mutants strains constructed in the background of the encapsulated serotype 2 S. pneumoniae D39 strain. As different serotypes are known to produce variable amounts of capsule polysaccharide, we cannot exclude that the absence of Wzd/Wze may affect differently the synthesis of the capsule in different serotypes.
We propose that Wzd/Wze function as spatial regulators of the capsular polysaccharide synthesis, ensuring that it occurs at the proper place, the division septum, and at the proper time, possibly to ensure the concealment of the newly synthesized cell wall. According to this model, wild-type Wzd and Wze interact and migrate to the middle of the dividing cells, events that are dependent on a functional ATP binding Walker A motif of Wze (Fig. 6). At the septum, the complex Wzd/Wze is probably subjected to cycles of phosphorylation/dephosphorylation events at the C-terminal tyrosine cluster of Wze, that may affect other components of the polysaccharide synthesis and assembly complex through conformational changes, as previously proposed from the studies of the S. aureus Wzd/Wze homologues, CapA and CapB (Olivares-Illana et al., 2008).
The formation of capsule at the division septum of the pneumococcal cells seems to be regulated by Wzd/Wze proteins. If ATP is not able to bind to Wze, or in the absence of Wzd or Wze, capsule is absent from the septum, where the new cell wall is synthesized, resulting in the presence of capsule at the surface of cells only in regions of mature cell wall. This raises the hypothesis that there are two modes to synthesize the capsular polysaccharide in pneumococcal bacteria (septal and non-septal), similarly to what has been proposed for peptidoglycan synthesis (Zapun et al., 2008). The absence of capsule from the septum could then be explained by the existence of two capsular polysaccharide synthetic machineries or alternatively by the activation of the synthetic machinery at the lateral cell wall or at the division septum, the latter dependent on the pair Wzd/Wze.
Why bacteria require such fine-tuned regulation of the synthesis of the capsular polysaccharide is unknown. One possibility is that the newly synthesized cell wall, located at the septum, has to be concealed or covered by the capsular polysaccharide to avoid bacteria recognition by the host complement system, or prevent detection by the host innate immune system. The pair Wzd/Wze would therefore assure that the capsule polysaccharide is produced or exported where it is most needed: the division septum of bacteria.
Bacterial strains and growth conditions
Bacterial strains and plasmids used in this study are listed in Table 1. E. coli was routinely grown in Luria–Bertani (LB) medium at 37°C unless otherwise indicated. When needed, antibiotics were used at the following concentrations: ampicillin 100 µg ml−1, kanamycin 50 µg ml−1 and erythromycin 100 µg ml−1. Isopropyl-β-d-thiogalactopyranoside (IPTG, Apollo Scientific) was used at 0.5 mM and 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (Xgal, Apollo Scientific) at 40 µg ml−1. S. pneumoniae was grown in C + Y liquid medium (Lacks and Hotchkiss, 1960) at 37°C, without aeration, or in trypic soy agar (TSA, Difco) plates supplemented with 5% sheep blood. Tetracycline (Sigma-Aldrich) and erythromycin (Apollo Scientific) were added to the media when appropriate to make a final concentration of 1 and 0.25 µg ml−1 respectively. For white/blue colony selection, Xgal was used at 120 µg ml−1. Lactococcus lactis was grown in M17 broth (Difco), supplemented with glucose (0.5% w/v), at 30°C. Erythromycin was used at 100 µg ml−1.
Table 1. Bacterial strains and plasmids.
supE thi (lacproAB) (F′traD36 proAB lacIq ZΔM15), repA from pWV01 integrated in the chromosome
pORI280 containing up- and downstream regions of cps operon, Emr, LacZ+
pORI280 containing the 3′ of wzd, the entire citrine sequence, and the downstream region of wzd, Emr, LacZ+
pORI280 containing the 3′ of wze, the entire citrine sequence, and the downstream region of wze, Emr, LacZ+
pORI280 containing up- and downstream regions of wzd, Emr, LacZ+
pORI280 containing up- and downstream regions of wze, Emr, LacZ+
DNA manipulation procedures
Escherichia coli competent cells were prepared and transformed as described (Sambrook et al., 1989). S. pneumoniae competent cells preparation and transformation was executed as previously described (Martin et al., 1995). PCR products and plasmid DNA were purified with kits Wizard SV Gel and PCR Clean-up System and Wizard Plus SV Minipreps respectively (Promega). PCR fragments were amplified using Phusion high-fidelity DNA polymerase (Finnzymes). Restriction enzymes were from New England Biolabs. Primers used in this study are listed in Table 2.
Construction of null mutants and substitution of capsule wild-type genes for fluorescent derivatives
In order to delete the entire cps gene cluster from the chromosome of S. pneumoniae ATCC6314 strain, plasmid pORI280 was used (Kloosterman et al., 2006). Fragments of DNA corresponding to the upstream and downstream regions of the cps operon were amplified with primers 1 and 2 and primers 3 and 4 respectively. Primers 2 and 3 have a 23 base pair overlapping region. After this first amplification, the two fragments were joined by overlapping PCR using primers 1 and 4. The product of this reaction was restricted with BamHI and PstI, cloned into pORI280, resulting in plasmid pBCSMC005, and transformed to E. coli strain EC101. The encapsulated ATCC6314 S. pneumoniae strain (serotype 14) was transformed with pBCSMC005. The cps null mutant, strain BCSMC001, was obtained by excision of the plasmid as previously described (Kloosterman et al., 2006).
Construction of wzd and wze null mutants (strains BCSMH001 and BCSMH002, respectively) was done in the same way, using primers 5 to 8 and 9 to 12 respectively. These primer pairs were used to clone the upstream and downstream regions of wzd and wze, resulting in plasmids pBCSMH012 and pBCSMH013 respectively. In the construction of these mutants ∼ 50 bp of the 3′ end of the deleted gene were left intact, to guarantee that the expression of the following gene on the operon was not disturbed.
Plasmid pORI280 was also used to substitute wzd and wze in the S. pneumoniae genome by genes encoding the correspondent Citrine fluorescent derivatives. For the construction of strain BCSMH003 (ATCC6314wzd::wzd_citrine), primers 13 and 14 were used to amplify the 3′-part of wzd and the entire citrine coding sequence, using plasmid pBCSMH007 as template DNA. Primers 15 and 16 were used to amplify the downstream region of wzd, using ATCC6314 chromosomal DNA as template. Primers 14 and 15 possess an overlapping region, allowing the above fragments to be joined by overlap PCR. The resulting fragment was restricted with enzymes BamHI and EcoRI and cloned in pORI280, resulting in pBCSMH010. For the construction of strain BCSMH004 (ATCC6314wze::wze_citrine), a fragment comprising the 3′-part of wze and the entire citrine sequence was amplified with primers 17 and 18 from plasmid pBCSMH004. The downstream region of wze was amplified with primers 11 and 19, from the chromosomal DNA of strain ATCC6314. Primers 11 and 18 contain an overlapping region so that the two fragments could be joined by overlap PCR. The resulting fragment was restricted with BamHI and EcoRI and cloned in pORI280, giving pBCSMH011. As for the construction of the null mutants, a duplicated fragment of ∼ 50 bp of the 3′ end of each gene was left on the chromosome, after the citrine sequence, so that the expression of the following gene on the operon was maintained at normal levels.
Plasmids were routinely propagated in E. coli EC101 or L. lactis LL108 and purified before being transformed into S. pneumoniae bacteria.
Construction of plasmids for protein expression in S. pneumoniae
Plasmid pBCSLF001 was constructed by amplification of plasmid pLS578 with primers 20 and 21, and subsequent auto ligation. This eliminated the CAT reporter gene, which encodes a chloramphenicol acetyltransferase, from pLS578 and added NheI, NcoI and BglII restriction sites.
The mCherry coding sequence was amplified from plasmid pROD17 with primers 22 and 23 and cloned in plasmid pBCSLF001. In this process, KpnI and NotI restriction sites were added upstream and downstream of the mCherry sequence, respectively, and plasmid pBCSLF002 was obtained. This plasmid was again amplified with primers 24 and 25, restricted with SpeI and ligated. This procedure mutated two guanine nucleotides to adenine, eliminating an internal ribosome binding site present on the 5′ end of the mCherry sequence. The resulting plasmid was named pBCSMH001. The Citrine gene was amplified from plasmid pcDNA3_Citrine with primers 28 and 29 and ligated to the product of amplification of pBCSMH001 with primers 26 and 27, producing plasmid pBCSMH002. The coding sequence for Wze was amplified with primers 30 and 31 and cloned in pBCSMH001, upstream of the mCherry sequence, giving plasmid pBCSMH003. Cloning of the same DNA in pBCSMH002 produced pBCSMH004. In order to mutate the Walker A ATP binding motif of wze, the gene was amplified in two different fragments with primer pairs 30 and 32 and 33 and 31. The two fragments were joined by overlapping PCR using primers 30 and 31 and the resulting product was cloned in pBCSMH001, producing plasmid pBCSMH005. Primers 32 and 33 encode for the following mutations: G48 → A; K49 → A. A second mutation of the Walker A ATP motif was done by the same procedure, using primers 46 and 47 instead of primers 32 and 33 respectively. The resulting plasmid is pBCSMH015, in which Wze carries a K49M mutation. Substitution of the four tyrosines (residues 218, 221, 224 and 227) for phenylalanines on the C-terminal tyrosine cluster of Wze was achieved through a single PCR step with primers 30 and 34. Cloning of this DNA fragment in pBCSMH001 produced pBCSMH006.
The wzd gene was amplified with primers 35 and 36, and cloned on plasmid pBCSMH002 generating the plasmid pBCSMH007.
In order to coexpress the Wze and Wzd fluorescent derivatives in the same cell, plasmid pBCSMH008 was constructed. This was done by amplification of the wzd_citrine sequence from plasmid pBCSMH007 with primers 37 and 38, and cloning on pBCSMH003. Plasmid pBCSMH009 was also constructed, through amplification of wzd with primers 37 and 39, restriction and ligation into pBCSMH003, to allow coexpression of Wze-mCherry and untagged Wzd.
Plasmids for the constitutive expression of untagged Wzd and Wze were also constructed. For this purpose, plasmids pBCSMH007 and pBCSMH004 were used as templates in a PCR reaction with primers 26 and 48 and 26 and 49 respectively. The resulting DNA fragments were restricted with BglII and autoligated, giving plasmids pBCSMH016 (Wzd expression) and pBCSMH017 (Wze expression).
In order to determine the localization of Wzd linked to Wze, fragments corresponding to both genes were amplified with primers 35 and 40 and 41 and 31, respectively, and joined by overlap PCR with primers 35 and 31. The resulting fragment that encode for Wzd linked to Wze through the linker NH2-AERGSVKTIKG-COOH was cloned into pBCSMH002 and originated the plasmid pBCSTR001. The same procedure was done for the Wze(WA) mutant, giving plasmid pBCSTR002.
The nucleotide sequences of the modified regions of the constructed plasmids were confirmed by sequencing.
To test in vitro the interaction between proteins Wzd and Wze, a bacterial two-hybrid activity assay was employed. Plasmid pBCSMC001, encoding the fusion protein T25-Wzd, was constructed by cloning the DNA fragment amplified with primers 42 and 43 from ATCC6314 chromosomal DNA into plasmid pKT25. Plasmid pBCSMC002, encoding the fusion protein T18-Wze, was constructed by cloning the DNA fragment amplified with primers 44 and 45 from ATCC6314 chromosomal DNA into plasmid pUT18C. Cloning of the wze mutated forms in plasmid pUT18C was done using primers 44 and 45 and plasmids pBCSMH005, pBCSMH015 and pBCSMH006 as templates. The resulting plasmids are pBCSMC003 [encoding T18-Wze(WA)], pBCSMH014 [encoding T18-Wze(WA2)] and pBCSMC004 [encoding T18-Wze(Y)]. The nucleotide sequences of the modified regions of the constructed plasmids were confirmed by sequencing.
Interactions between proteins Wzd and Wze were first tested by transformation of BTH101 cells with the constructed bacterial two-hybrid plasmids. Plates were incubated at 30°C and screened for blue/white colonies, in which blue indicated a positive interaction. Single colonies were grown at 30°C in the presence of 0.5 mM of IPTG and the interactions confirmed by β-galactosidase activity measurements, as described (Karimova et al., 2005).
Streptococcus pneumoniae strains were grown until early exponential phase and observed by fluorescence microscopy on a thin layer of 1% agarose in PreC medium (Lacks and Hotchkiss, 1960). Membranes were labelled with Nile Red (10 µg ml−1, Molecular Probes) and DNA with Hoechst 33342 (0.2 µg ml−1, Molecular Probes). To determine the site of cell wall synthesis, cells were labelled with a 1:1 mixture of vancomycin (Sigma) and the fluorescent BODIPY FL conjugate of vancomycin (VanFL; Molecular Probes) at a final concentration of 0.6 µg ml−1, during 5 min on ice. Images were obtained using a Zeiss Axio Observer. Z1 microscope equipped with a Photometrics CoolSNAP HQ2 camera (Roper Scientific). After acquisition, these images were analysed, adjusted and cropped using Metamorph software (Meta Imaging series 7.5) and Image J software (Abràmoff et al., 2004).
Quellung reactions were performed using 1 ml aliquots of liquid cultures (OD600 ∼ 0.5). Cells were washed 3 × with fresh C + Y medium at 37°C and resuspended in a final volume of 50 µl of C + Y medium. A volume of 2 µl of this suspension was mixed with 2 µl of CPS14 pneumococcal antisera (SSI Diagnostica) and the resulting reaction was observed under the microscope.
In vivo detection of the capsule present at the surface of S. pneumoniae cells was achieved using an antibody specific for the capsular polysaccharide of serotype 14 (SSI Diagnostica), purified against the nonencapsulated strain BCSMC001, as previously described (Xayarath and Yother, 2007). Briefly, 300 ml of a culture of the null mutant strain at OD ∼ 0.3 was heat-killed during 45 min at 56°C, followed by centrifugation and washing with PBS. Cells were suspended in 300 µl of Anti-CPS14 diluted 1/100 in PBS and incubated o/n at 4°C with gentle agitation. On the next day, the sample was centrifuged, the supernatant filtered (0.2 µm pore, Millipore) and stored at 4°C until further use. Immunofluorescence of live pneumo cells was done by incubating, on ice during 5 min, 1 ml aliquots of early exponential phase cultures with 50 µl of purified antibody. Samples were washed 2 × with C + Y at 37°C, and incubated with 0.4 µg ml−1 of Anti-rabbit Alexa Fluor 488 (Invitrogen), on ice during 5 min. Cells were washed twice with fresh C + Y medium, suspended in 50 µl and 2–3 µl loaded on PreC + 1% agarose slides.
Fluorescence ratio (FR) calculation was done similarly to a previously described method (Pereira et al., 2007; Atilano et al., 2010). FR was determined by quantifying the fluorescence at the septum of cells initiating division (cells that belong to class III as defined in Fig. 1) divided by the average of the fluorescence at the lateral cell wall of the two future daughter cells. Average background fluorescence was subtracted from every value. Quantification was performed for at least 100 cells of each strain.
Fluorescent protein analysis
For the preparation of S. pneumoniae total cell extracts, cells were harvested from mid-exponential phase cultures, resuspended in PBS and disrupted in a FastPrep FP120 bead beater (Thermo Electron Corporation). The cell breakage programme used was three cycles of 90 s at 6.0 r.p.m., followed by 5 min intervals of incubation on ice. Total cell extracts were obtained after centrifugation at 14 000 r.p.m., during 20 min at 4°C (Eppendorf 5430R), to remove unbroken cells and insoluble material. The BCA Protein Assay Kit from Pierce (Thermo Scientific) was used to quantify the total amount of protein.
For the separation of the membrane and cytoplasmic fractions, total cell extracts of nonencapsulated bacteria constitutively expressing Wzd-Citrine (strain BCSMH008) and Wze-Citrine (strain BCSMH007) were prepared as described above. These total cell extracts were then ultracentrifuged for 20 min at 20 000 r.p.m. (Beckman TL 100). The supernatants resulting from this centrifugation were ultracentrifuged again, this time at 80 000 r.p.m. during 45 min. The pellet obtained in this step contained membrane bound proteins, such as Wzd-Citrine, while the supernatant contained Wze-Citrine and other cytoplasmic proteins.
Protein samples were incubated with solubilization buffer (200 mM Tris_HCl pH 8.8, 20% glycerol, 5 mM EDTA pH 8.0, 0.02% bromophenol blue, 4% SDS, 0.05 M DDT) (Drew et al., 2006) at 37°C during 5 min and separated on SDS-PAGE gels. Gel images were acquired on a Fuji FLA 5100 laser scanner (Fuji Photo Film) with 635 nm excitation and > 665 nm band pass emission filter for protein molecular weight marker detection, 532 nm excitation and > 575 nm band pass emission filter for mCherry fusions detection and 473 nm excitation and > 510 nm band pass emission filter for Citrine fusions detection.
Bacterial cell lysates of 1 ml of culture (OD600 ∼ 0.3) were harvested and boiled during 3 min before being separated on a 10% SDS-PAGE gel. Proteins were transferred into a Hybond PVDF Membrane (Amersham) and probed with mouse anti-phosphotyrosine monoclonal antibody cocktail clone 4G10 Platinum (Millipore) used at 1:2000, followed by 1:5000 of goat anti-mouse antibody conjugated to horseradish peroxidase (Amersham). Detection was done with ECL Plus Western Blotting Detection Reagents (Amersham). For the detection of capsule proteins fluorescent derivatives, the cultures were harvested at OD600 ∼ 0.5. Proteins were transferred and detected as above. Probing was done with Living Colors Av. Peptide Antibody (Clontech) used at 1:500, followed by 1:100000 of goat anti-rabbit antibody conjugated to horseradish peroxidase.
Cell samples were prepared by harvesting cells at early exponential growth-phase (OD600 ∼ 0.3), and resuspending them in water. After adjusting the samples to same cell density, cells were lysed with deoxycholate and boiled for 3 min before use. Samples of purified cell walls and peptidoglycan were prepared as previously described (Severin and Tomasz, 1996). Briefly, cells were boiled into sodium dodecyl sulfate (SDS, final concentration, 4%) for 30 min to inactivate any enzyme that could modify the bacteria cell wall. After removal of SDS, cell walls were mechanically broken by shaking with an equal volume of acid-washed glass beads with a FastPrep FP120 apparatus. Cell walls were digested with Dnase and Rnase (for 3 h at 37°C), and trypsin (overnight at 37°C), which were inactivated by boiling in 1% (final concentration) SDS. Cell walls were washed twice with water, once with 8 M LiCl and then with 100 mM EDTA. Before lyophilization, broken cell walls (CW) were washed three times with water. For purification of peptidoglycan, the cell walls (5 mg) were treated with 2 ml of 49% hydrofluoric acid (HF) for 48 h at 4°C, which is known to cleave phosphodiester bonds and thus remove teichoic acids and polysaccharides covalently bound to the bacterial peptidoglycan. The peptidoglycan (PGN) was recovered and washed as previously described (Severin et al., 1997).
In order to load in the dot-blot similar amounts of cell walls and peptidoglycan, the content of muramic acid was determined in each purified sample, by fluorescence HPLC analysis after an hydrolysis process, as described (Appuhn et al., 2004). Twenty microlitres of peptidoglycan (10 mg ml−1) and 50 µl of cell wall (20 mg ml−1) were mixed with 37.25 µl of 12 M HCL and incubated at 95°C for 2 h. The mixture was lyophilized, washed with 0.5 ml of water, lyophilized again and the final pellet resuspended in 100 µl of water. Twenty microlitres of sample was mixed with 80 µl of OPA reagent, which was prepared by dissolving 68 mg of o-phtaldialdehyde (Fluka) in 50 µl of 2-mercaptoethanol and 1.4 ml methanol, followed by completion of the volume to 10 ml with borate buffer (0.4 M H3BO3, pH 10.4). Samples were injected onto Hypersil ODS column C18 (Thermo Electron Corporation). The flow rate of the mobile phase (0.05 M sodium citrate, 0.05 M sodium acetate, methanol and tetrahydrofuran at a ratio of 90:8.5:0.75:0.75 in vol %) was 1.5 ml min−1 at 35°C for 80 min, followed by a 20 min gradient from 0 to 80% of a cleaning buffer (water and methanol at a ratio of 50:50 in vol %) and 20 min mobile phase again for re-conditioning of the column. Fluorometric emission of the amino sugar derivatives was measured at a wavelength of 445 nm with 340 nm as the excitation wavelength (Shimadzu RF-10Axl).
Samples were loaded into Hybond PVDF (Amersham) membranes, pre-equilibrated in PBS and placed on top of PBS-soaked Hybond Blotting Paper. The membranes were allowed to air-dry for 30 min and then blocked during 1 h in Blocking Buffer (5% non-fat dried milk in PBS). Membranes were washed in PBS-T (PBS + 0.05% Tween 20) and incubated overnight at 4°C with primary antibodies Anti-CPS14, purified as described above, and Anti-CWPS (SSI Diagnostica) diluted 1/1000 in PBS-T. The latter serum, Anti-CWPS, was used to verify that the samples were retained in the membrane of the dot-blot, as it recognizes the phosphorylcholine residues of the Cell Wall Polysaccharide, or pneumococcal teichoic acids. After washing with PBS-T, membranes were incubated during 1 h at room temperature with secondary antibody Anti-Rabbit IgG peroxidase linked diluted 1/100000 in PBS-T. Membranes were again washed with PBS-T and detected using the ECL Plus Western Blotting Detection Reagents (Amersham).
We thank Dr Mariana Pinho for critical reading of the manuscript and Magda Atilano for helping us in performing the analysis of the different samples by HPLC. This work was funded by ‘Fundação para a Ciência e Tecnologia’ through research Grants PTDC/SAU-MII/75696/2006 and PTDC/BIA-MIC/100747/2008 (to S.R.F.) and fellowships SFRH/BD/43797/2008 (to M.H.).