EzrA Contributes to the Regulation of Cell Size in Staphylococcus aureus

EzrA is a negative regulator of FtsZ in Bacillus subtilis, involved in the coordination between cell growth and cell division and in the control of the cell elongation–division cycle. We have now studied the role of the Staphylococcus aureus homologue of the B. subtilis EzrA protein and shown that it is not essential for cell viability. EzrA conditional and null mutants have an overall increase of the average cell size, compared to wild type strains. In the larger ezrA mutant S. aureus cells, cell division protein FtsZ and the cell wall synthesizing Penicillin Binding Proteins (PBPs) are not properly localized. This suggests that there may be a maximum cell diameter that allows formation of a Z-ring capable of recruiting the other components of the divisome and of driving cytokinesis. We propose that the major role of EzrA in S. aureus is in cell size homeostasis.


Introduction
Cell division in bacteria is a highly regulated process during which cells undergo a series of temporally and spatially controlled events that result in the generation of two identical daughter cells [1]. In nearly all bacteria, cell division initiates with the polymerization of a tubulin-like protein, FtsZ, into a ring structure located at the future division site [1,2]. This Z-ring serves as a scaffold for the recruitment of other proteins involved in cell division, resulting in the assembly of a multiprotein complex known as the divisome [1,3]. Constriction of the Z-ring and concomitant synthesis of new cell wall material initiate division septum formation, a process that eventually leads to the separation and release of two daughter cells.
In Bacillus subtilis, a model organism for the study of cell division, the concentration of FtsZ during the cell cycle remains constant [4]. This fact implies that timing of FtsZ assembly is not regulated via oscillations in the availability of FtsZ monomers, but primarily through cell cycle-dependent variations in FtsZ polymerization [4]. In agreement with this idea, previous studies have identified various septal proteins that modulate Z-ring assembly and dynamics: FtsA [5] and ZipA [6] (present in Gammaproteobacteria), which promote Z-ring formation and anchor FtsZ to the membrane; ZapA, a small protein which enhances both the assembly and the stability of the Z-ring [7]; ZapB (present in a subset of Gammaproteobacteria), which may promote the organization of shorter FtsZ protofilaments into a functional ring [8]; SepF (present in Gram-positive bacteria and in Cyanobacteria) which forms rings required for the regular arrangement of FtsZ filaments [9,10]; and EzrA (found in Gram-positive bacteria with low GC content), a negative regulator of Z-ring assembly [11].
EzrA (for extra Z-rings A) is an abundant protein with an estimated 10,000-20,000 molecules per B. subtilis cell, anchored to the membrane by an N-terminal transmembrane domain [12]. In B. subtilis, EzrA acts as an inhibitor that prevents formation of aberrant Z-rings at the poles of exponentially growing cells, ensuring the formation of only one Z-ring per cell cycle, at mid-cell [11]. The absence of EzrA lowers the critical concentration of FtsZ necessary to form a Z-ring, probably due to stabilization of the FtsZ ring [11,13]. EzrA interacts directly with the C-terminus of FtsZ, inhibiting FtsZ assembly in vitro [12,14] as this interaction not only reduces the GTP-binding affinity of FtsZ, but also increases the rate of GTP hydrolysis, both of which are unfavorable for FtsZ polymerization [15].
In B. subtilis, EzrA localizes to the cell membrane, where it is proposed to prevent FtsZ assembly at inappropriate locations other than mid-cell [11]. However, EzrA also localizes to the divisome at mid-cell [11], a localization which appears to be contradictory with EzrA's role as an FtsZ negative regulator. It seems that the presence of EzrA in the divisome reflects a second role of EzrA in maintaining proper FtsZ dynamics within the medial Z-ring [16] perhaps contributing to Z-ring remodeling by accelerating the disassembly of the Z-ring as cytokinesis progresses [7]. This latter role of EzrA may be important for the coordination between cell growth and cell division, as a strain encoding EzrA mutated in the conserved ''QNR patch'', required for EzrA localization to the medial Z-ring (but not for EzrA inhibition of FtsZ assembly at the poles) has an increased cell length [16]. More recently, EzrA together with the newly identified divisome component GpsB, has been proposed to have a third role, related to the control of the cell elongation-division cycle of B. subtilis, by modulating the recruitment of Penicillin-Binding Protein 1 (PBP1) to the divisome [17]. PBPs are enzymes responsible for the last stages of peptidoglycan biosynthesis, required for cell elongation as well as for the synthesis of the division septum in rod-shaped bacteria.
Cell division in general, and the regulation of the Z-ring in particular, has been studied mostly in the rod-shaped model bacteria Escherichia coli and B. subtilis. However, lessons learned from these organisms may not always be applicable to other bacteria, namely to bacteria with different morphologies. One interesting alternative model to study cell division is Staphylococcus aureus because it has spherical cells, and therefore, it is likely to have different mechanisms for placement of the division septum. Moreover, the bacterial cell division machinery constitutes a good potential target for the development of new antibiotics [18] and therefore it is important to understand this process in clinically relevant bacteria. S. aureus is a leading nosocomial pathogen due to its extraordinary capacity to acquire resistance to virtually all classes of available antibiotics. In fact, Methicillin-Resistant S. aureus strains (MRSA) currently cause more deaths in the USA than HIV/AIDS and tuberculosis combined [19].
In this work we studied the first steps of cell division in S. aureus. In particular, we focused upon the role of EzrA, which had been identified in a transposon-mediated differential hybridisation screen as a putative essential gene in S. aureus [20]. While this manuscript was being prepared, Steele and colleagues also reported that EzrA was essential in S.aureus by expressing it under the control of an inducible promoter [21]. Intriguingly, this result indicated that, contrary to what is observed in B. subtilis, EzrA could play an essential role in S. aureus, pointing to important differences between these two bacterial species. Using various approaches, we now show that EzrA is not essential in S. aureus but is required for maintaining correct cell size. In the absence of EzrA, there is a significant increase of the average diameter of S. aureus cells. Moreover, larger cells were found to have delocalized FtsZ and PBPs, suggesting that there may be a maximum cell diameter for proper assembly of the divisome in S. aureus.

EzrA is not essential in Staphylococcus aureus
Before we started the studies reported here, the gene encoding the S. aureus homologue of the B. subtilis cell division protein EzrA had been identified as an essential gene by Transposon-Mediated Differential Hybridisation screening [20]. As B. subtilis EzrA is not essential for survival, we hypothesized that its function could be more relevant in S. aureus. We therefore constructed conditional ezrA mutants, in order to confirm its essentiality and to study its role in cell division of Gram-positive cocci.
As a first approach we constructed strain BCBAJ031 in which the ezrA gene was placed under the control of the IPTG-inducible, LacI-repressible P spac promoter. For that purpose, the ezrA ribosomal binding site (RBS) together with the initial 533 bp of the ezrA gene were cloned downstream of the P spac promoter in the pMUTIN4 vector, which was then integrated into the ezrA locus of the S. aureus chromosome of strain NCTC8325-4. The plasmid pMGPII, containing a second copy of the lacI gene, was also introduced into this strain to enhance repression of ezrA transcription. The resulting strain BCBAJ031 was able to grow in the absence of the inducer IPTG (Fig. 1A) indicating that EzrA was not essential. EzrA was also placed under the control of the inducible P spac promoter in the background of strains RN4220 (BCBAJ036), COL (BCBAJ019), SH1000 (BCBAJ034) and Newman (BCBAJ035) which, similarly to BCBAJ031, were able to grow in the absence of IPTG both in solid (Fig. 1A) and in liquid medium (data not shown). To confirm that ezrA transcription was indeed repressed in the absence of inducer, we used quantitative real time PCR (RT-PCR) to quantify the amount of ezrA transcript in control strain BCBHV002 (NCTC8325-4 pMGPII) and in the ezrA inducible BCBAJ031 strain, grown in the presence and in the absence of IPTG. In the absence of inducer, the amount of ezrA transcript was 33 times lower than in the parental strain BCBHV002 while addition of IPTG (1 mM) led to recovery of nearly wild type levels of ezrA mRNA (Fig. 1B).
As a second approach to generate a conditional ezrA mutant, we constructed the replicative plasmid pBCBAJ006, encoding ezrA antisense RNA under the control of the xylose-inducible, glucoserepressible promoter pT5X. This plasmid was introduced into strains RN4220, SH1000, NCTC8325-4, Newman and COL. The empty pEPSA5 vector was introduced into the same strains and used as negative control. Growth of the resulting strains in the presence of 2% xylose (to induce antisense RNA expression and therefore repress ezrA expression) did not result in growth halt (data not shown), again indicating that EzrA is not essential in S. aureus. Depletion of EzrA due to ezrA antisense RNA expression was confirmed using strain BCBAJ024 expressing an EzrA-mCherry fusion protein from its native chromosomal locus. Quantification of EzrA-mCherry fluorescence in this strain grown in the presence of xylose or glucose showed that there was a 92% reduction in the amount of EzrA-mCherry protein produced when the ezrA antisense RNA was expressed (Fig. 1C).
Finally, in order to show that S. aureus cells were able to grow in the complete absence of EzrA, we constructed null mutants in the background of five S. aureus strains RN4220, NCTC8325-4, SH1000, Newman and COL by deleting the entire ezrA coding sequence, with the exception of its start and stop codons. For that purpose, up and downstream regions of the ezrA gene were cloned into the thermosensitive pMAD vector, which was then integrated into the S. aureus chromosome at 43uC. Growth in the absence of antibiotic selection allowed excision of the plasmid, which resulted in either wild type or null mutant strains. PCR screening of colonies in which plasmid excision had occurred resulted in the identification of an average of 46% of null mutants (the exact value depends on the background strain used), showing that EzrA is not essential in S. aureus. Growth curves of NCTC8325-4 and its ezrA null mutant BCBAJ030 show that absence of EzrA results in a small growth defect (Fig. 1D)

Absence of EzrA disturbs cell size homeostasis
In order to study the role of EzrA in S. aureus, we first examined the EzrA mutant strains by phase microscopy. Strains BCBAJ031 (in which ezrA is under the control of P spac in the NCTC8325-4 background and therefore is not expressed in the absence of IPTG), or BCBAJ030 (NCTC8325-4DezrA) showed an increased heterogeneity in cell size when compared to the corresponding controls. We therefore measured the diameter of around 1000 cells from each strain showing that lack of EzrA resulted in an overall increase of the average cell size (Fig. 2). We repeated our measurements using EzrA inducible and null mutant cells in RN4220, SH1000, Newman and COL backgrounds and again we observed an increase in average cell size of cells grown in the absence of EzrA when compared to the parental strains (Table S1).
While in rod-shaped B. subtilis inhibition of cell division results in longer cells [22], in spherical S. aureus cells inhibition of cell division, for example through depletion of FtsZ, results in larger cells which can have up to twice the diameter of a normal cell [23].
It is therefore possible that the larger size of S. aureus cells lacking EzrA results from a delay in cell division due to lack of proper coordination between cell growth and cell division.

EzrA localizes to the division septum in S. aureus dividing cells
In B. subtilis the proposed role of EzrA in the coordination between cell growth and cell division is dependent on its mid-cell localization. However, EzrA also localizes to the lateral membrane, which is required for its role in preventing FtsZ assembly at inappropriate locations different from the mid-cell. We have recently reported that EzrA-CFP localized to the division septa in S. aureus [24]. Analysis of EzrA localization pattern in the strain BCBAJ025 expressing EzrA-mCherry showed that EzrA was absent from the cell membrane that surrounds the staphylococcal cells undergoing division (Fig. 3). It therefore became interesting to determine if EzrA had any role in preventing the formation of extra Z-rings in S. aureus.
EzrA is required for correct formation of the FtsZ ring in S. aureus To study the localization of FtsZ in EzrA-depleted S. aureus cells we expressed FtsZ-CFP from the ectopic spa locus, under control of the IPTG-inducible P spac promoter in the NCTC8325-4 DezrA strain (BCBAJ032). We observed mislocalization of FtsZ in approximately 25% of the cells in the DezrA mutant compared to 4.5% in NCTC8325-4 cells expressing FtsZ-CFP (BCBHV011) (Fig. 4). FtsZ-CFP mislocalization patterns included protein present as dots around the membrane, protein delocalized all over the cell surface or the formation of incomplete, often asymmetrical, Z-rings (Fig. 4), as well as the formation of multiple complete FtsZ rings in a single cell in 7% of the cells. Interestingly, mislocalization of FtsZ-CFP was more pronounced in the subpopulation of very large EzrA mutant cells (diameter .1.75 mm). Results obtained by immunofluorescence using an anti-FtsZ polyclonal antibody were in agreement with these results (data not shown). Electron microscopy also confirmed the formation of extra incomplete septa in the ezrA inducible mutant BCBAJ031 grown in the absence of IPTG and in the null mutant BCBAJ030 (Fig. 5 C and B).
Absence of EzrA leads to mislocalization of cell wall synthetic proteins As stated above, lack of EzrA resulted in an increase of the average S. aureus cell size. Approximately 72% of cells had a cell diameter larger than 1.75 mm in the inducible mutant BCBAJ031 or 54% in the null mutant BCBAJ030, while only 7% of wild type NCTC8325-4 cells were this large. Given that S. aureus synthesizes cell wall mainly, if not only, at the septum [23] we wondered how cells were able to grow to such large sizes. To investigate this question, we tested if depletion of EzrA affected the localization of cell wall synthesis and cell wall synthetic proteins (PBPs) in S. aureus. For that purpose we labeled nascent cell wall using a fluorescent derivative of vancomycin (Van-FL) which binds the terminal D-Ala-D-Ala residues of the peptidoglycan muropeptides [23]. Cells of strain COL (parental strain) and BCBAJ014 (COL DezrA) were grown in the presence of an excess of D-serine, which leads to the replacement of the carboxyl-terminal D-alanine residue of the peptidoglycan precursor by a D-serine residue. Cells were then changed to growth medium without D-serine to allow incorporation of peptidoglycan precursors with D-Ala-D-Ala termini into the nascent cell wall. Labeling of the resulting cells with Van-FL showed that in larger cells from strain BCBAJ014, cell wall synthesis occurred around the cell periphery, in a dispersed manner (Fig. 6A). To study the localization of cell wall synthetic enzymes, we used a fluorescent derivative of penicillin, Bocillin-FL, which binds PBPs and therefore can be used to determine the localization of these enzymes. PBPs were labeled in NCTC8325-4 (parental strain) and BCBAJ030 (NCTC8325-4 DezrA) cells using Bocillin-FL at concentrations below the MIC (minimum inhibitory concentration), for 5 minutes, to minimize its effect on the metabolism of cell wall synthesis. Under these conditions, 48% of the total BCBAJ030 cells analyzed (n = 767) kept a normal mid cell localization of these proteins (class A in Fig. 6C), while 38% had displacement and irregular distribution of PBPs (class C in Fig. 6C). Furthermore, when these results were correlated with the size of each cell, we noticed that 77% of cells with a diameter over 1.75 mm had a delocalized Bocillin-FL signal, indicating that the cell wall synthetic machinery was delocalized in these cells (Fig. 6C).
In B. subtilis, EzrA has recently been proposed to have a role in the recruitment of PBP1 to the septum [17]. PBP1 is the major transglycosylase/transpeptidase bifunctional PBP in B. subtilis, similarly to PBP2 in S. aureus. To investigate if EzrA interacted with PBPs in S. aureus, we performed a bacterial-two hybrid assay using the BATCH system based on a cyclic AMP signaling cascade in E. coli [25]. This indicated that EzrA can interact with itself, with PBP1 and PBP2 (Fig. 7).
However, if EzrA does indeed interact with PBPs in S. aureus, these interactions (at least with PBP2) may occur only during a specific time during the cell cycle, as EzrA and PBP2 are not recruited to the septum at the same time (Fig. 8). This was seen in strain BCBAJ033 (NCTC8325-4 background) and BCBAJ017 (COL background) expressing both ezrA-mCherry and gfp-pbp2  fusions from their respective native loci. Although both EzrA and PBP2 were visible at mid cell, as expected, co-localization occurred in 51% (n = 802) or 34% (n = 1420) of the cells for NCTC8325-4 and COL backgrounds, respectively (Fig. 8B). In the remaining cells the two proteins did not colocalize, as EzrA was seen in early forming septa (visible as two spots at the membrane) while PBP2 was not yet present at mid cell (class a, Fig. 8B); EzrA was found localized over the entire septum (seen as a line across the cell) while PBP2 was still present in a ring around the cell (visible as two spots at the membrane) (class b, Fig. 8B); or, when cells started to prepare the next round of division, EzrA could already be found at the next division site, whereas PBP2 was still present at the mid cell of the two daughter cells which were splitting (class d, Fig. 8B).

Discussion
In order to study the role of EzrA in S. aureus, we have used three different approaches to deplete S. aureus cells of EzrA: placing the gene under the control of the inducible promoter P spac , expressing antisense RNA and constructing null mutants in which the complete ezrA gene was deleted. This was performed in the background of five different strains -RN4220, NCTC8325-4, SH1000, Newman and COL, generating a total of 15 strains which are able to grow in the absence of EzrA, indicating that EzrA is not an essential protein in S. aureus. These results are not in agreement with those recently reported by Steele and colleagues, using a SH1000 strain expressing ezrA under the control of the same P spac inducible promoter [21]. One hypothesis for the discrepancy between our results and those from Steele and colleagues would be the presence of suppressor mutations in the strains reported here, which would allow them to grow in the absence of EzrA. However, we think that is unlikely because null mutants were obtained at rates up to 70%. If a suppressor mutation was required for the viability of the null mutants, we would expect a strong predominance of wild type colonies during the excision of the plasmid pBCBAJ003 used to construct the null mutants (this excision can result either in wild type or DezrA cells, see materials and methods). A second reason to discard the hypothesis of suppressor mutations is that is does not explain the viability of the ezrA inducible mutants we have constructed. These strains, in which ezrA expression in under the control of IPTGinducible P spac promoter, were maintained in the constant presence of high (1 mM) concentrations of IPTG, i.e., in the constant presence of EzrA, except during the depletion experiments. Therefore these conditions should not favor selection of suppressor mutations for EzrA absence.
One possible explanation for the lack of growth, in the absence of inducer, of the ezrA inducible strain VF79 reported by Steele and colleagues could be the presence of a mutation synthetic lethal with ezrA, that could have been transduced from the background of RN4220 (a mutagenized strain) into SH1000 during the construction of VF79. This hypothesis is based on the fact that mutations in at least five genes have been shown to be synthetic  lethal with the absence of ezrA in B. subtilis: zapA [7], gpsB [17], sepF [9], noc [26] and ftsL [27].
When EzrA was initially described in B. subtilis, the first obvious phenotype caused by lack of this protein was the formation of extra Z-rings near the poles [11]. However, later studies by Haeusser and colleagues suggested that EzrA has a second function in contributing to the maintenance of proper FtsZ assembly dynamics in the medial Z-ring [16]. This function would render the Z-ring sensitive to factors responsible for coordinating cell growth and cell division. As a consequence, impairing EzrA's ability to localize to the medial Z-ring (without impairing its ability to inhibit FtsZ assembly at the cell poles), resulted in cells with stabilized FtsZ assembly at midcell, which were significantly longer than wild type cells [16]. Importantly, these two functions of EzrA -preventing Z-ring formation at the cell poles and coordinating cell growth and cell division -are directly related to the two localization patterns observed for EzrA in B. subtilis cells -at the membrane and at the nascent septal site, respectively.
The most obvious phenotype in all of the ezrA mutants we have constructed was a difference in cell size, with ezrA mutants having on average larger cells than wild type strains. Given that lack of EzrA results in the presence of extra Z-rings in B. subtilis [11], we looked for the formation of additional Z-rings in S. aureus strains depleted of EzrA and expressing a CFP fluorescent derivative of FtsZ. Although we did observe mislocalization patterns for FtsZ in approximately one quarter of the cells analyzed, two complete Zrings were found in less than 7% of the cells. This is in sharp contrast with B. subtilis ezrA mutants in which 54% of the cells have more than one FtsZ ring [11]. We have also observed that EzrA localizes to the division site in cells undergoing division, and is essentially absent from the cell membrane in these cells. Putting this information together, we propose that the major role for EzrA in S. aureus is to maintain proper FtsZ assembly dynamics at mid cell, contributing to the coordination of cell growth and cell division, and not to prevent the assembly of Z-rings at inappropriate locations. One hypothesis for the role of EzrA at mid cell in B. subtilis is that it contributes to Z-ring remodeling by accelerating the disassembly of the Z-ring as cytokinesis progresses [7]. Lack of EzrA therefore leads to stabilized Z-rings, which may exist for longer periods during the bacterial cell cycle. In S. aureus, the septum is the only place where cell wall synthesis occurs [23]. Therefore, it is conceivable that in the absence of EzrA, there is more time to synthesize cell wall at the septum and therefore additional cell wall material may be incorporated into the septum. Splitting of the staphylococcal division septum generates one hemisphere of each new daughter cell and, in principle, expansion of the flat septum to generate the inflated hemispheres could occur by limited cell hydrolysis only [28]. Therefore it is possible that a septum containing more cell wall material would lead to larger hemispheres in the daughter cells, and consequently to larger daughter cells. These cells may be able to further increase in size due to dispersed cell wall synthesis around the periphery of the  EzrA mutant cells. These mechanisms could explain the larger cell sizes that we have observed in the S. aureus ezrA mutants.
Approximately 71% and 54% of EzrA inducible and null NCTC8325-4 mutant cells have diameters over 1.75 mm. Both FtsZ and PBPs are mostly delocalized in these large cells, raising the possibility that in the absence of EzrA, FtsZ dynamics is altered so that the Z-ring can no longer be assembled/maintained. An alternative hypothesis would be the existence of a maximum cell diameter in S. aureus which allowed the formation of a productive Z-ring, capable of recruiting the other protein components of the divisome and of driving cytokinesis. Above that maximum diameter, a functional Z-ring would not be stable or would not be able to form and consequently, proteins such as cell wall synthetic enzymes would not be recruited to the divisome and/or would become delocalized. This hypothesis is in agreement with previous observations made in a S. aureus mutant in which FtsZ was placed under the control of the IPTG-inducible P spac promoter. FtsZ-depleted cells increased in size up to twice their initial diameter [23]. This increase in size was not reversible even when IPTG was added to restore FtsZ production, indicating that a Z-ring cannot be formed in staphylococcal cells with very large diameters [23].
Lastly, we have obtained some evidence that EzrA may be part of a protein complex that includes PBPs, as EzrA is able to interact with PBP1 and PBP2 in a bacterial two-hybrid assay. These results are in agreement with the most recent role attributed to EzrA in B. subtilis, in promoting the recruitment of the major transglycosylase/transpeptidase PBP1 to the division site. However, if an interaction between EzrA and PBP2 (the major transglycosylase/ transpeptidase of S. aureus) does occur in live staphylococcal cells, it is probably temporary, as although EzrA and PBP2 colocalize at the septum during certain stages of the cell cycle (see Fig. 8), microscopy images indicate that EzrA is recruited to the division septum before PBP2.
In summary, our work suggests that the major role of EzrA in S. aureus is to contribute to the coordination of cell growth and cell division and consequently to the maintenance of the correct size of the staphylococcal cells. Furthermore, our work reinforces the importance of studying the role of conserved cell division proteins in different model organisms, namely in organisms with different morphologies.

Bacterial strains, plasmids and growth conditions
All plasmids and strains used in this study are listed in Tables 1 and 2. The sequences of the primers used are listed in Table 3. E. coli strain DH5a was grown on Luria-Bertani agar (LA) or in Luria-Bertani broth (LB, Difco), supplemented with ampicillin (100 mg/ml) as required. S. aureus strains were grown on tryptic soy agar (TSA, Difco) at 37uC or in tryptic soy broth (TSB, Difco) at 37uC with aeration, unless otherwise stated. The medium was supplemented, when necessary, with 10 mg/ml of erythromycin (Sigma), 100 mg/ml of 5-bromo-4-chloro-3-indolyl-b-D-galactopyranoside (X-Gal, Apollo Scientific) or 0.1 mM (for induction of FtsZ-CFP expression) or 1 mM (for induction of EzrA expression) Conditional mutants were always grown in the presence of 1 mM IPTG except during depletion assays. Growth was monitored by the increase in optical density at 600 nm (OD 600 nm ).

General procedures
DNA manipulations and E. coli transformations were carried out using standard methods [29]. Restriction enzymes were purchased from New England Biolabs. Polymerase Chain Reaction (PCR) was performed using Phusion high-fidelity DNA polymerase (Finnzymes). Sequencing reactions were carried out at Macrogen.
S. aureus RN4220 cells were transformed by electroporation as previously described [30] and plasmids were moved into different S. aureus strains by transduction using the phage 80a [31].

Construction of plasmids and strains
To express ezrA from the P spac inducible promoter, a DNA fragment including the putative RBS and a truncated 59 region of ezrA (533 bp) was amplified by PCR using primers EzrAP22 and EzrAP7. The PCR fragment was digested with EcoRI and BamHI and cloned downstream of the IPTG-inducible P spac promoter in the pMUTIN4 vector [32], resulting in the plasmid pBCBAJ009. The plasmid was electroporated into strain BCBAJ001 (which carries pMGPII, a multi-copy plasmid encoding an extra lacI gene [33]) where it integrated into the ezrA chromosomal locus, resulting in strain BCBAJ036. This construct was transduced into the S. aureus strains NCTC8325-4, SH1000, Newman and COL, resulting in strains BCBAJ031, BCBAJ034, BCBAJ035 and BCBAJ019, respectively. All procedures were performed in the presence of IPTG 1 mM. Insertion at the correct chromosomal locus was confirmed by PCR.
A second strategy for EzrA depletion in S. aureus involved the use of antisense RNA. For that purpose, a 39end of the ezrA gene (566 bp) was amplified from S. aureus COL genomic DNA, using primers BTHEzrAP2 and EzrAP8. The resulting PCR product was digested with EcoRI and KpnI and cloned into the pEPSA5 vector [34], resulting in pBCBAJ006, in which the ezrA gene fragment is transcribed in the opposite direction compared to the ezrA chromosomal copy. This replicative plasmid was introduced into S. aureus RN4220 by electroporation and transferred to S. aureus strains NCTC8325-4, SH1000, Newman, COL and BCBAJ012 by transduction, selecting with chloramphenicol 20 mg/mL and glucose 0.2% (w/v). Final strains were named BCBAJ006, BCBAJ046, BCBAJ048, BCBAJ050, BCBAJ016 and BCBAJ024, respectively. The empty vector pEPSA5 was also introduced into the same strains which were used as controls and named BCBAJ005, BCBAJ045, BCBAJ047, BCBAJ049, BCBAJ015 and BCBAJ023.
To delete the entire ezrA gene from the chromosome of S. aureus RN4220 leaving only its start and stop codons (ATGTAA), a 1102 bp fragment upstream of the ezrA gene and a 1107 bp fragment downstream of the ezrA gene were amplified by PCR from S. aureus COL genomic DNA using primer pairs upEzrAP1/downEzrAP3 and upEzrAP2/EzrAP2, respectively. The two fragments were joined by overlap PCR using primers upEzrAP1/EzrAP2, gel purified, digested and cloned into the EcoRI/BamHI sites of  pMAD vector, yielding pBCBAJ003. The resulting plasmid was sequenced, introduced into RN4220 cells by electroporation and transduced to S. aureus NCTC8325-4, SH1000, Newman and COL using phage 80a. Blue colonies of BCBAJ002, BCBAJ037, BCBAJ038, BCBAJ039 and BCBAJ011 grown at 30uC on TSA with erythromycin and supplemented with X-Gal, were selected. Growth at 43uC with erythromycin allowed selection of colonies in which pBCBAJ003 had integrated at the ezrA locus, which was confirmed by PCR. Plasmid excision occurred after a second homologous recombination event, in cells growing in the absence of antibiotic. White, erythromycin sensitive colonies were selected, and ezrA gene deletion was confirmed by PCR. The resulting strains were named BCBAJ004, BCBAJ030, BCBAJ040, BCBAJ041 and BCBAJ014. For localization of EzrA, a C-terminal fusion of mCherry to EzrA was expressed in S. aureus from its native locus. For that purpose, two PCR fragments containing the truncated 39end of ezrA gene (1099 bp) without its stop codon and the 1107 bp sequence downstream of ezrA gene, were amplified from S. aureus COL genomic DNA using primer pairs EzrAP4/EzrAP5 and EzrAP1/EzrAP2, respectively. The two fragments were joined by overlap PCR using primers EzrAP4 and EzrAP2 which resulted in the introduction of a 5 codon linker at the 39 end of the ezrA gene. The resulting fragment was digested with EcoRI/BamHI and cloned into the thermosensitive pMAD vector [35], generating the plasmid pBCBAJ001. The mCherry coding sequence (711 bp) was amplified from the plasmid pBCB4-ChE [24] using primers mChP3 and mChP4 and cloned into pBCBAJ001 using enzymes NheI and XhoI, in frame with ezrA gene, generating plasmid pBCBAJ002. This plasmid was introduced into RN4220 cells by electroporation, resulting in strain BCBAJ028, and was transduced from this strain to NCTC8325-4 and COL using phage 80a, generating strains BCBAJ042 and BCBAJ013. These strains were incubated at 43uC in the presence of erythromycin for selection of colonies in which plasmid pBCBAJ002 was integrated into the ezrA locus. These colonies were then grown without antibiotic selection. After a second recombination event leading to plasmid excision, cells in which ezrA was replaced by an ezrA-mCherry fusion were   identified by PCR and the resulting strains were named BCBAJ025 and BCBAJ012. For localization studies of FtsZ we constructed strains where FtsZ-CFP fusion is ectopically expressed, under the control of the P spac promoter, from the spa locus of S. aureus chromosome. To construct these strains, phage 80a was used to transduce pBCBHV003 plasmid [36] into the null mutant strain BCBAJ030. The resulting strain BCBAJ043 was incubated at 43uC in the presence of erythromycin to allow the integration of the pBCBHV003 plasmid. Incubation in the absence of antibiotic selection, allowed the selection of white colonies in which the vector had been excised and the spa gene was exchanged by ftsZcfp generating strain BCBAJ044.
In order to co-localize EzrA and PBP2 in the same cell, a phage lysate of the RNpPBP2-31 strain [37] was used for transduction into BCBAJ025 and BCBAJ012. Colonies expressing GFP-PBP2 were further screened using a 473 nm laser of a FUJI FLA5100 imager and the resulting strains expressing GFP-PBP2 and EzrA-mCherry were named BCBAJ033 and BCBAJ017.

Growth conditions for EzrA depletion experiments
To evaluate the essentiality of EzrA for S. aureus growth, inducible strains BCBAJ036 (RN4220 background), BCBAJ031 (NCTC8325-4 background), BCBAJ034 (SH1000 background), BCBAJ035 (Newman background) and BCBAJ019 (COL background) and the corresponding parental strains containing the plasmid pMGPII, with an extra lacI gene BCBAJ001, BCBHV002, BCBAJ051, BCBAJ052 and BCBAJ010, were grown overnight in TSB with erythromycin, chloramphenicol and IPTG 1 mM. The following day, the cultures were diluted 1/200 in TSB supplemented with erythromycin, chloramphenicol and IPTG and grown until midexponential phase so that they could recover from stationary phase. At this time, the cultures were washed twice with TSB and diluted in pre-warmed TSB (with or without antibiotics) to an OD 600 nm of 0.001. These cultures were split in two and 1 mM IPTG was added to one of the aliquots. To modulate expression of EzrA using antisense RNA, overnight cultures of strains BCBAJ006, BCBAJ046, BCBAJ048, BCBAJ050, BCBAJ016 and BCBAJ024 grown with 20 mg/mL of chloramphenicol and glucose 0.2% (w/v) were diluted 1/500 into TSB (without antibiotics) supplemented with 2% xylose (w/v), to induce the expression of the 39-end ezrA antisense RNA. Control strains transformed with the empty pEPSA5 vector, BCBAJ005, BCBAJ045, BCBAJ047, BCBAJ049, BCBAJ015 and BCBAJ023 were also grown under the same conditions.
To confirm depletion of EzrA upon antisense RNA expression, strain BCBAJ024 was grown in TSB containing 2% xylose (w/v) or 0.2% glucose (w/v) and strain BCBAJ012 was grown in TSB. Cells were harvested from 20 mL of culture, washed with phosphate buffered saline (PBS) and resuspended in 0.2 mL of PBS. Cell lysates were prepared by disrupting cells using a Fast Prep FP120 (Thermo) in 3 cycles of 45 seconds, at maximum speed, with 10 minutes incubation on ice between cycles. Lysates were clarified by centrifugation and 56 SDS-loading buffer (500 mM Dithiothreitol; 10% SDS; 250 mM Tris?HCL, pH 6.8; 30% glycerol; 0.02% bromophenol blue) was added to the supernatant. Samples were heated to 30uC for 10 minutes and analyzed by SDS-PAGE (10%). After running, the gel was washed with MilliQ water and EzrA-mCherry protein fluorescence was detected using a 532 nm laser in a FUJI FLA-5100 reader. The integrated density of the bands was calculated using Image J 1.42q software.
Quantitative Real-Time PCR S. aureus strains BCBHV002 and BCBAJ030 were grown in TSB until early-exponential phase for RNA extraction. The inducible mutant strain BCBAJ031 was grown with 1 mM IPTG and without IPTG inducer, as indicated above, until early-exponential phase. Prior to harvesting, RNAprotect Bacteria Reagent (twice the culture volume, Qiagen) was added to the culture and the mixture was immediately vortexed for 10 sec. The cells were harvested, the pellet was quickly frozen in liquid N 2 and stored at 280uC overnight. The next day, the pellet was resuspended with 1 mL of FastRNA blue reagent (MP Biomedicals) and cells were disrupted using silica beads and a FastPrep FP120 apparatus (Thermo). RNA was extracted with 200 mL chloroform and recovered by precipitation with isopropyl alcohol, washed with 80% ethanol, and resuspended in milli-Q water. Integrity of total RNA was analysed in an agarose gel under denaturing conditions (0.25 M formaldehyde) and quantified using a Nanodrop Spectrophotometer ND-100.
cDNA was generated from 300 ng of each RNA sample using TaqMan RT Reagents (Applied Biosystems, Foster City, CA). The reaction mix included 5.5 mM MgCl 2 , 500 mM dNTPs, 2.5 mM random hexamers, 16 RT Buffer, 0.8 U/ml RNase Inhibitor and 1.25 U/ml MultiScribe RT in a final volume of 25 ml. The Reverse Transcription conditions were 10 min at 25uC, 15 min at 42uC and 5 min at 99uC.
Quantification of ezrA expression was achieved using the ABI7000SDS (Applied Biosystems), SYBR Green chemistry, and the standard curve method for relative quantification. The PCR reagents consisted of: 16 SYBR Green PCR Master Mix (Applied Biosystems), 400 nM of each primer, and 5 ml of sample cDNA, in a final volume of 25 ml. The thermocycling profile was: 10 min at 95uC followed by 40 cycles of 15 s at 95uC and 1 min at 60uC. 16SrRNA was used as the endogenous control, and the uniformity of the 16SrRNA expression levels throughout the heterogeneous samples was confirmed before proceeding with subsequent qPCR comparisons. Briefly, cDNA from all strains subjected to different experimental conditions were generated from exactly the same amount of RNA and were subjected to qPCR analysis as described above. The obtained Threshold cycle (Ct) values varied less than 0.2 between samples from the same strain, and less than 0.7 between samples from different strains (data not shown), unequivocally validating the use of the 16SrRNA gene as an endogenous control. qPCR primers for both ezrA (EzrAP29 and EzrAP30) and 16SrRNA (16SrRNAP3 and 16SrRNAP4) ( Table 3) were designed using ABI7000SDS -specific software, Primer Express (Applied Biosystems).
For each sample, the amount of ezrA and 16SrRNA was determined from the respective standard curve by conversion of the mean threshold cycle values, and normalization was obtained by dividing the quantity of ezrA by the quantity of 16SrRNA. The specificity of the amplified products was verified by analysis of the dissociation curves generated by the ABI 7000 software based on the specific melting temperature for each amplicon.

Fluorescence Microscopy
S. aureus strains were grown overnight in TSB at 37uC with appropriate selection. After reaching the desired OD, 1 ml of cells were pelleted and resuspended in 20 ml of PBS from which 1 ml was placed on a thin layer of 1% (w/v) agarose in PBS, mounted on a slide. Microscopy was performed using a Zeiss Axio Observer.Z1 microscope and images were taken with a Photometrics CoolSNAP HQ2 camera (Roper Scientific) using Metamorph 7.5 software (Molecular Devices).
Labeling of nascent cell wall with Vancomycin-FL was done as previously described [23].
Bocillin-FL (Molecular Probes), a BODIPY-fluorescent derivative of penicillin, was used to visualize PBPs. For that purpose, 1 mL of culture of NCTC8325-4 and BCBAJ030 (NCTC8325-4 DezrA) cells was incubated with 0.1 mM Bocillin-FL, for 5 minutes, protected from light. Cells were visualized as described above.

Bacterial Two Hybrid Assays
The ezrA gene was amplified from S. aureus COL genomic DNA. To construct a fusion of the ezrA gene in frame with the 59 end of the cya gene we used primer pair BTHEzrAP1/BTHEzrAP2 and cloned the resulting PCR product into pKNT25 vector [25], resulting in plasmid pBCBAJ010. The ezrA gene was also cloned in frame with the 59end of the cya gene into pUT18 vector [25], using primers BTHEzrAP1/BTHEzrAP2, resulting in pBCBAJ011 plasmid. Both constructs were confirmed by sequencing. To test for the existence of putative interactions between EzrA and PBPs, plasmids with the genes coding for each PBP cloned in frame with the cya gene were used as previously described [38]. Briefly, all possible combinations of those plasmids were co-transformed into the reporter E. coli strain BTH101. The empty vectors pUT18 and pKNT25 were used as negative control, and plasmids p18Zip and p25Zip [25], which express each a domain of a leucine zipper, were used as a positive control.
To qualitatively evaluate putative interactions, cultures of E. coli BTH101 cells co-transformed with BTH plasmid pairs were diluted 10 22 , 10 24 to 10 25 and 10 mL of each dilution was plated on LA supplemented with X-Gal 40 mg/mL, IPTG 0.5 mM, ampicillin 100 mg/mL and kanamycin 50 mg/mL.

Electron Microscopy
S. aureus strains NCTC8325-4 and BCBAJ030 were grown in TSB until OD600 nm 0.7-0.8. EzrA conditional mutant strain, BCBAJ031, was grown with 1 mM IPTG until mid-exponential phase. The culture was washed twice with TSB, diluted in pre-warmed media to an initial OD 600 nm of 0.002 and grown with or without 1 mM IPTG until OD 600 nm of 0.8. All samples were washed with 0.1 M sodium-cacodylate buffer (pH 7.4) and fixed with 10 times their pellet volume of ice-cold 2.5% (v/v) glutaraldehyde in 0.1 M sodium-cacodylate buffer, pH 7.4. The cells were then centrifuged, washed with the same buffer, and post fixed with 2% osmium tetroxide for 2 h. After a brief rinse with buffer, the cells were dehydrated using a graded acetone series and embedded in Spurr medium [39]. Thin sections were stained with uranyl acetate and lead citrate and viewed in a Philips BioTwin CM120 electron microscope at 100 kV. Individual micrographs were recorded on Kodak 4489 film at a nominal magnification of 11.0006. At least 2000 cells of each strain were analyzed.

Supporting Information
Table S1 Frequency of cells (%) with diameter below and above 1.75 mm for different strains. (DOC)