Transcriptional Analysis and Subcellular Protein Localization Reveal Specific Features of the Essential WalKR System in Staphylococcus aureus

Abstract

The WalKR two-component system, controlling cell wall metabolism, is highly conserved among Bacilli and essential for cell viability. In Staphylococcus aureus, walR and walK are followed by three genes of unknown function: walH, walI and walJ. Sequence analysis and transcript mapping revealed a unique genetic structure for this locus in S. aureus: the last gene of the locus, walJ, is transcribed independently, whereas transcription of the tetra-cistronic walRKHI operon occurred from two independent promoters located upstream from walR. Protein topology analysis and protein-protein interactions in E. coli as well as subcellular localization in S. aureus allowed us to show that WalH and WalI are membrane-bound proteins, which associate with WalK to form a complex at the cell division septum. While these interactions suggest that WalH and WalI play a role in activity of the WalKR regulatory pathway, deletion of walH and/or walI did not have a major effect on genes whose expression is strongly dependent on WalKR or on associated phenotypes. No effect of WalH or WalI was seen on tightly controlled WalKR regulon genes such as sle1 or saouhsc_00773, which encodes a CHAP-domain amidase. Of the genes encoding the two major S. aureus autolysins, AtlA and Sle1, only transcription of atlA was increased in the ΔwalH or ΔwalI mutants. Likewise, bacterial autolysis was not increased in the absence of WalH and/or WalI and biofilm formation was lowered rather than increased. Our results suggest that contrary to their major role as WalK inhibitors in B. subtilis, the WalH and WalI proteins have evolved a different function in S. aureus, where they are more accessory. A phylogenomic analysis shows a striking conservation of the 5 gene wal cluster along the evolutionary history of Bacilli, supporting the key importance of this signal transduction system, and indicating that the walH and walI genes were lost in the ancestor of Streptococcaceae, leading to their atypical 3 wal gene cluster, walRKJ.

Introduction

Staphylococcus aureus, a highly versatile bacterium, is successful both as a commensal as well as a life-threatening pathogen. This dual lifestyle is facilitated by its sophisticated arsenal of virulence factors and the regulatory networks controlling their production and its general metabolism, allowing it to rapidly adapt to host infection stress. Several global regulation systems allowing coordinated expression of virulence factors have been described in S. aureus. Among these, the well-known AgrCA quorum sensing two-component system, active at high cell density, positively controls the expression of genes encoding secreted host cell and extracellular matrix degradation proteins and represses genes for surface proteins involved in adhesion and escape from the immune response [1]. This reciprocal regulation is thought to allow early bacterial colonization at low bacterial densities, and wide bacterial spread when higher cell densities induce the AgrCA system.

Another two-component system, WalKR, is essential for cell viability and also plays a role in S. aureus virulence. This system is mainly involved in regulation of cell wall metabolism [2–4], a cellular process crucial for bacterial fitness that can lead to NFκB-dependent induction of the innate immune inflammatory response through release of cell wall degradation products that bind to the Nod2 receptor in the host [5, 6]. Increased activity of the WalKR system also indirectly induces genes involved in degradation of the cellular and extracellular host matrix through the SaeRS two-component system [6].

Genes encoding the WalKR system are located within the wal genetic locus, encompassing five genes in S. aureus: walR, walK, walH, walI and walJ. The WalKR system is highly conserved among Firmicutes, and its characterization in several bacterial species (S. aureus, Bacillus subtilis, Streptococcus pneumoniae) has highlighted its crucial role in controlling cell wall metabolism [3, 4, 7, 8]. This regulation has been shown to be responsible for the essential nature of the WalKR system in S. pneumoniae and S. aureus [9, 10]. The conservation of the walR, walK, walH, walI and walJ gene order among Bacilli suggests functional interactions between the corresponding proteins [11], although this has only been partially characterized. In B. subtilis, the WalH and WalI membrane-bound proteins were shown to interact with the WalK histidine kinase, inhibiting its activity [12, 13]. The walJ gene, encoding a β-lactamase fold protein, was first described as affecting colony morphology in B. subtilis [12]. Deletion of walJ has also been reported to cause a DNA segregation defect in B. subtilis, leading to the hypothesis that WalJ could be involved in coordinating cell division and DNA replication [14]. In Bacillus anthracis, interruption of walJ leads to a spontaneous mutator phenotype, and it has recently been described as a novel 5'-3' double-stranded DNA exonuclease playing a role in DNA mismatch repair [15–17]. In Streptococcaceae, the wal locus is reduced to only three genes, walR, walK and walJ, with the walH and walI genes absent from the genome. WalJ is required for optimal growth in S. pneumoniae when cells are partially depleted for WalKR, suggesting a functional link between WalJ and the WalKR system [9]. In S. mutans, WalJ appears to have a pleiotropic effect on growth, natural competence, oxidative stress tolerance and biofilm formation [18].

In S. aureus, the roles of WalH, WalI and WalJ had never been studied. We carried out a transcriptional analysis of the entire wal locus in order to characterize its genetic structure. Protein topology, protein-protein interactions and subcellular localization of the Wal proteins allowed us to show that WalH and WalI are membrane-bound proteins which associate with WalK to form a complex at the cell division septum. Phenotypic investigation of the ΔwalH and ΔwalI mutants indicate that WalH and WalI do not act as major negative regulators of WalKR activity as they do in B. subtilis. A detailed phylogenomic study indicates coevolution of the wal genes in Firmicutes with a concomitant loss of the walH and walI genes in Streptococcaceae.

Materials and Methods

Bacterial strains and media

Escherichia coli K12 strain DH5α™ (Invitrogen, Thermo Fisher Scientific, Waltham, MA) was used for cloning experiments. Staphylococcus aureus strain HG001 was used for genetic and functional studies. Plasmids were first transformed into the restriction deficient S. aureus strain RN4220 before introduction into the HG001 strain. Strains and plasmids are listed in Table 1. E. coli was grown in LB medium with ampicillin (100 μg/ml) and kanamycin (50 μg/ml) added when required. S. aureus was grown in Trypticase Soy Broth (TSB; Difco; Becton, Dickinson and Co., Franklin Lakes, NJ) supplemented with erythromycin (1 μg/ml) or chloramphenicol (10 μg/ml) when required. Expression from the Pcad promoter was induced by adding cadmium chloride (CdCl₂) at a final concentration of 0.25 μM.

Table 1. Bacterial strains and plasmids used in this study.

Strain or plasmid	Description	Source, reference or constructiona
E. coli strains
DH5α	F- Φ80lacZΔM15 Δ(lacZYA-argF) U169 recA1 endA1 hsdR17 (rK–, mK+) phoA supE44 λ– thi-1 gyrA96 relA1	Invitrogen Life Technology
DHT1	F-glnV44(AS) recA1 endA1 gyrA96 (Nalr) thi-1 hsdR17 spoT1 rfbD1 cya-854 ilv-691::Tn10	[30]
S. aureus strains
RN4220	Restriction-deficient transformation recipient strain	[70]
HG001	NCTC 8325 rsbU+	[71]
ST1397	HG001ΔwalH	pMADwalH → HG001
ST1130	HG001ΔwalI	pMADwalI → HG001
ST1410	HG001ΔwalHI	pMADwalHI → HG001
ST1131	HG001ΔwalJ	pMADwalJ → HG001
ST1365	RN4220 PwalR(RI+RII)-lacZ	pSA14-PwalR(RI+RII) → RN4220
ST1366	RN4220 PwalR(RI)-lacZ	pSA14-PwalR(RII) → RN4220
ST1367	RN4220 PwalJ-lacZ	pSA14-PwalJ → RN4220
ST1398	RN4220 pSA14	pSA14→ RN4220
ST1301	HG001 walH-GFP	pOLSA-walH → HG001
ST1340	HG001 walK-GFP	pOLSA-walK → HG001
ST1341	HG001 walI-GFP	pOLSA-walI → HG001
ST1378	HG001 walJ-GFP	pOLSA-walJ → HG001
ST1415	HG001ΔwalH pMK4Pprot-walHI	pMK4Pprot-walHI → ST1397
ST1416	HG001ΔwalI pMK4Pprot-walHI	pMK4Pprot-walHI → ST1130
ST1417	HG001ΔwalHI pMK4Pprot-walHI	pMK4Pprot-walHI → ST1410
Plasmids
pMAD	Allelic exchange vector	[19]
pMADwalH	pMAD derivative for walH deletion	This study
pMADwalI	pMAD derivative for walI deletion	This study
pMADwalHI	pMAD derivative for simultaneous walH and walI deletion	This study
pMADwalJ	pMAD derivative for walJ deletion	This study
pMADwalK	pMAD derivative for walK deletion	This study
pSA14	Vector for constructing transcriptional lacZ fusions	[22]
pSA14-PwalR(RI+RII)	pSA14 derivative carrying the full-length walRKHI promoter region (promoters PI and PII)	This study
pSA14-PwalR(RII)	pSA14 derivative carrying the walRKHI PII promoter region	This study
pSA14-PwalJ	pSA14 derivative carrying the walJ promoter region	This study
pCN51	Vector for CdCl2-dependent gene expression	[31]
pOLSA	Vector for expression of GFP translational fusions	This study
pOLSA-walK	pOLSA derivative containing the entire walK coding sequence in frame with GFP	This study
pOLSA-walH	pOLSA derivative containing the entire walH coding sequence in frame with GFP	This study
pOLSAwalI	pOLSA derivative containing the entire walI coding sequence in frame with GFP	This study
pOLSAwalJ	pOLSA derivative containing the entire walJ coding sequence in frame with GFP	This study
pKTop	Vector expressing the dual reporter PhoA21-471/LacZ5-60, p15 ori	[26]
pKTop-walH	pKTop derivative for expression of WalH1-40/PhoA/LacZ protein fusion	This study
pKTop-walI	pKTop derivative for expression of WalI1-40/PhoA/LacZ protein fusion	This study
pKTop-walJ	pKTop derivative for expression of WalJ1-40/PhoA/LacZ protein fusion	This study
pKT25	BACTH vector designed to express a given polypeptide fused in frame at its N-terminal end with the CyaA T25 fragment; p15 ori	[29]
pKT25-Zip	Zip-Zip BACTH positive control vector	[29]
pKT25-walR	Full-length walR cloned into pKT25	This study
pKT25-walK	Full-length walK cloned into pKT25	This study
pKT25-walH	Full-length walH cloned into pKT25	This study
pKT25-walI	Full-length walI cloned into pKT25	This study
pKT25-walJ	Full-length walJ cloned into pKT25	This study
pUT18c	BACTH vector designed to express a given polypeptide fused in frame at its N-terminal end with the CyaA T18 fragment; ColE1 ori	[29]
pUT18c-Zip	Zip-Zip BACTH positive control vector	[29]
pUT18c-walR	Full-length walR cloned into pUT18c	This study
pUT18c-walK	Full-length walK cloned into pUT18c	This study
pUT18c-walH	Full-length walH cloned into pUT18c	This study
pUT18c-walI	Full-length walI cloned into pUT18c	This study
pUT18c-walJ	Full-length walJ cloned into pUT18c	This study
pMK4-Pprot	pMK4 derivative carrying a constitutive Gram-positive promoter for gene complementation	[21]
pMK4-PprotwalHI	pMK4-Pprot derivative carrying walHI	This study

^a Arrows indicate plasmid introduction by electroporation

Mutant construction and complementation

Oligonucleotides used in this study were synthesized by Eurofins Genomics (Ebersberg, Germany) and their sequences are listed in Table 2. Gene deletions were performed using the thermosensitive replication vector pMAD carrying DNA fragments of approximately 800 bp corresponding to the upstream and downstream regions flanking the gene of interest [19]. The cloned DNA fragments were amplified by PCR using chromosomal DNA from strain HG001 and the oligonucleotide pairs listed in Table 2. Given the close proximity/overlap of the walK, walH and walI genes and their likely translational coupling, oligonucleotides for gene deletions were carefully designed so as to preserve translation initiation signals and avoid any polar effects on the distal genes. Oligonucleotide pairs used for amplifying DNA fragments for gene deletions were as follows: ΔwalH (OP238/OSA528 and OSA529/OP241); ΔwalI (OP242/OP243 and OP244/OP245); ΔwalHI (OP238/OSA528 and OP244/OP245); ΔwalJ (OP201/OP202 and OP203/OP204); ΔwalK (in the ΔwalHI mutant, OP369/OP370 and OP244/OP245). The resulting plasmids were then introduced into S. aureus at a permissive temperature (30°C) and integration/excision cycles were induced by temperature shifts (30°C-37°C) as previously described [19]. Candidate mutants were then tested by PCR.

Table 2. Oligonucleotides used in this study.

Name	Sequence	Description
Construction of pMAD derivatives
OP238	TGCAGGATCCAGGATTTGTAACTGGTTATA	walH upstream fragment (BamHI/EcoRI)
OSA528	ATGGAATTCATTATTCATCCCAATCACCGTCTTC
OSA529	TATGAATTCGGGAGGCTTGAATAAATGAACTGG	walH downstream fragment (EcoRI/NcoI)
OP241	ATTCTCCATGGGATTCATATTAATGATTAATAATTTTAG
OP242	TGCAGGATCCATGCAACCATACACAGATATCATCACA	walI upstream fragment (BamHI/EcoRI)
OP243	AGCCGAATTCTTATTCAAGCCTCCCATCGTTATAAAC
OP244	AGCCGAATTCTATGAATCGTAATAAGCTAGTATTGCA	walI downstream fragment (EcoRI/NcoI)
OP245	ATTCTCCATGGTGCGGATCTATTGCATCATGTGAC
OP201	ATTCTGGATCCGTAAAATACGAACAAACGT	walJ upstream fragment (BamHI/EcoRI)
OP202	AGCCGAATTCCCTTTCTCTTTAAACAGTCA
OP203	ATGCATGAATTCAAATGAGAGTCACCCTAT	walJ downstream fragment (EcoRI/BglII)
OP204	GTGGTTAGATCTTGGAAGGCGTCTCCTGCG
OP369:	TGCAGGATCCCCGATTGCTGATATTTTAGAAT	walK upstream fragment
OP370	AGCCGAATTCTCGTTTCGACCTCTACTCATGT	(BamHI/EcoRI)
Complementation of wal mutants
OP277	TGCAGGATCCTGAAGTCATTGAAGACGGTGATT	walHI coding sequences
OSA542	CTCTGCAGTGATGAGCTTGCAATGCTAGCTTATTACG	(BamH1/PstI)
Construction of transcriptional lacZ fusions
OSA477	AATCTGCAGTAGTAATATATAAGTTTATATTGG	walR upstream promoter region, PI + PII, forward (PstI)
OSA478	AATCTGCAGAATAGTAAGCGACATCCTGTG	walR upstream promoter region, PII, forward (PstI)
OSA473	TCTGGATCCTTGCATAAACCTCTTTTCTTAAATC	walR upstream region, reverse (BamHI)
OSA479	TTACTGCAGATGTTTTTGCGTCTCCAACGTG	walJ upstream region, forward (PstI)
OSA475	GCGGGATCCCCTTTCTCTTTAAACAGTCAATC	walJ upstream region, reverse (BamHI)
Construction of pKTop derivatives
OMA16	CGGATCCATGTGAAGTCATTGAAGACGGTGATTGGG	WalH (AA 1 to 40) coding sequence (BamHI/KpnI)
OMA17	TCGGTACCCTATCTGTATTGTCGACATTTGCAATATCAGG
OMA12	TGGATCCTTTATAACGATGGGAGGCTTGAATAAATG	WalI (AA 1 to 40) coding sequence (BamHI/KpnI)
OMA13	ACGGTACCTTGTTACTCTCGACTTCATTAATGTGTGAGCG
OMA14	CGGATCCATAGAGAGATTGACTGTTTAAAGAGAAAGG	WalJ (AA 1 to 40) coding sequence (BamHI/KpnI)
OMA15	AAGGTACCTCTTCCATTTTCTTTCCAGTCAAACCAACATC
Construction of pOLSA and derivatives
OSA338	CCCCCGGGTCAGGGTCAGGGTCAAAAGGAGAAGAATTATTTACAGGGG	Peptide linker and GFPopt coding sequence (XmaI/KpnI)
OSA339	AAGGTACCTTACTTATATAATTCATCCATTCCG
OSA340	AAGGATCCGAGTAGAGGTCGAAACGAATGAAGTGGC	WalK coding sequence (BamHI/XmaI)
OSA341	CCCCCGGGTTCATCCCAATCACCGTCTTCAATGACTTCAC
OSA357	TGGGATCCTTGAAGACGGTGATTGGGATGAATAATAAGG	WalH coding sequence (BamHI/XmaI)
OSA358	TTCCCGGGTTCAAGCCTCCCATCGTTATAAACATACC
OMA12	TGGATCCTTTATAACGATGGGAGGCTTGAATAAATG	WalI coding sequence (BamHI/XmaI)
OMA34	TCCCCGGGATGATTAATAATTTTAGGGTTATTATTTGTCG
OMA14	CGGATCCATAGAGAGATTGACTGTTTAAAGAGAAAGG	WalJ coding sequence (BamHI/XmaI)
OMA35	CTCCCGGGTATTGTATATATTGGCGTTGGAATAGC
Construction of BACTH plasmids
OSA361	TTGGATCCCATGGCTAGAAAAGTTGTTGTAGTTGATGATG	WalR coding sequence (BamHI/KpnI)
OSA362	TTGGTACCTTTCGACCTCTACTCATGTTGTTGG
OSA363	TCGGATCCCATGAAGTGGCTAAAACAACTACAATCCC	WalK coding sequence (BamHI/KpnI)
OSA364	AGGGTACCTATGCTCCTTATTATTCATCCC
OSA365	GTGGATCCCATGAATAATAAGGAGCATATTAAATCTG	WalH coding sequence (BamHI/KpnI)
OSA366	TGGGTACCTCAGTTTCCAGTTCATTTATTCAAGCC
OSA367	CTGGATCCCATGAACTGGAAACTGACAAAGACAC	WalI coding sequence (BamHI/KpnI)
OSA368	TTGGTACCATTCATATTAATGATTAATAATTTTAGGG
OSA369	AGGGATCCCATGAGCCGCTTGATACGCATGAGTG	WalJ coding sequence (BamHI/KpnI)
OSA370	TTGGTACCATGACTCTCATTTATATTGTATATATTGGC
Reverse transcription PCR analysis
OJU20	TATTGTGACGCGTAGAGGCGTTGG	3' part of walR, forward
OJU21	CTTGACGGTTGGCATACTCACT	5' part of walK, reverse
OJU22	CGCACAATGGTCGTATTTGGGC	3' part of walK, forward
OJU23	CGAATGAATAATCTGAAATGGC	5' part of walH, reverse
OJU24	GAAGTGCAGATTAACAGTGAACTCG	3' part of walH, forward
OJU25	TCGACTTCATTAATGTGTGAGCG	5' part of walI, reverse
OJU1	CGGTTACTACTCAGTCGTGAATGAAACG	3' part of walI, forward
OP227	CTTGAATATTACGGTCAATTTGACT	5' part of walJ, reverse
OJU17	TTCATTGCACGTGATTATC	Intergenic walI-walJ region, forward
Primer extension experiments
OSA433	CTTATAAATGGTAAATTATATAATAGG	walR upstream region, forward
OSA434	GTTAAATTCTAAAATATCAGCAATCGG	walR coding sequence, reverse
OSA435	TTTGCTGGTACGGTTTATTATGGC	walJ upstream region, forward
OSA437	TAATACACTCATGCGTATCAAGCG	walJ coding sequence, reverse
qRT-PCR experiments
OSA161	ACGTGGATAACCTACCTATAAGACTGGGAT	16S rRNA
OSA162	TACCTTACCAACTAGCTAATGCAGCG
OSA138	GTGTACTGTGCATACGATGGTAATGATGC	walR
OSA139	CGTTACATAGTCATCTGCACCTAGTTCTA
OSA203	AACAGCACCAACGGATTAC	atlA
OSA204	CATAGTCAGCATAGTTATTCATTG
OSA209	AATTATATTCATACAATCCTGGTG	saouhsc-00773
OSA210	GGTGCTTGCTTAACTACTAC
OSA224	AAGTATCTGGCTCAAGTAATTCTAC	sle1
OSA225	TGATGGACGGCTACTATTGC

Plasmid pMK4Pprot, a derivative of shuttle vector pMK4 [20] carrying a constitutively expressed Gram-positive promoter sequence [21] was used for gene complementation experiments. Complementation of the ST1397 (ΔwalH), ST1130 (ΔwalI) and ST1410 (ΔwalHI) strains was carried out using a 2204 bp DNA fragment corresponding to the complete walHI coding sequences, amplified with oligonucleotides OP277/OSA542, generating BamHI/PstI restriction sites at the extremities, and cloned in the replicative plasmid pMK4-Pprot, to give plasmid pMK4Pprot-walHI.

Nucleotide sequencing of plasmid constructs was carried out by Beckman Coulter Genomics (Danvers, MA).

β-Galactosidase assays

Plasmid pSA14 [22] is a derivative of shuttle vector pMK4 [20], carrying a promoterless E. coli lacZ gene and was used to construct transcriptional lacZ reporter fusions. Promoter regions of the walRKHI operon and the walJ gene were amplified by PCR using oligonucleotides introducing PstI/BamHI restriction sites (see Table 2). The DNA fragments were cloned between the corresponding restriction sites of the pSA14 vector, yielding plasmids pSA14-PwalR(RI+RII), pSA14-PwalR(RII) and pSA14-PwalJ (Table 1).

For β-galactosidase assays in S. aureus, cells carrying lacZ fusions were grown in TSB until OD_{600 nm} ≈ 3 and harvested by centrifuging 2 ml culture samples (2 min; 5,400 x g). Assays were performed as previously described [23] and β-galactosidase specific activities expressed as Miller units mg⁻¹ protein [24]. Protein concentrations were determined using the Bio-Rad protein assay (BioRad, Hercules, CA) [25].

Protein topology analysis

To study the topology of the WalH, WalI and WalJ proteins, we used the pKTop plasmid carrying the dual pho-lac reporter gene, which encodes the E. coli alkaline phosphatase fragment PhoA_21-471, fused in-frame with the α-peptide of E. coli β-galactosidase, LacZ_4-60, allowing protein topology determination in vivo [26]. Extracellular or membrane localization of the reporter leads to high alkaline phosphatase activity and low β-galactosidase activity, whereas a cytosolic localization results in high β-galactosidase activity and low alkaline phosphatase activity. A strain carrying the pKTOP vector alone, in the absence of any inserts, will exhibit β-galactosidase activity, but no alkaline phosphatase activity, as there are no secretion signals for the ‘PhoA’-‘LacZ’ fusion [26]. Transmembrane domain predictions were carried out using the Phobius web server (http://phobius.sbc.su.se; [27]). Sequences encoding the first 40 amino acids, encompassing the predicted transmembrane domain, were amplified by PCR using oligonucleotides OMA16/OMA17 for walH and OMA12/OMA13 for walI (Table 2). The WalJ protein, predicted to be cytosolic, was also included in this study, and a DNA fragment corresponding to the first 40 codons was amplified using oligonucleotides OMA14/OMA15 (Table 2). These DNA fragments were cloned between the BamHI and KpnI restriction sites of pKTop, resulting in translational fusions with the dual pho-lac reporter genes.

Protein topology assays were carried out using E. coli strain DH5α on dual-indicator LB agar plates containing isopropyl-β-D-thiogalactopyranoside (IPTG; 1 mM); 6-chloro-3-indolyl-β-D-galactopyranoside (Red-Gal; Sigma-Aldrich, St. Louis, MO) at 100 μg/ml (for β-galactosidase activity) and 5-bromo-4-chloro-3-indolyl phosphate disodium salt (X-Pho; Sigma-Aldrich) at 80 μg/ml (for alkaline phosphatase activity), as previously described [28].

BACTH protein interaction assays

For testing protein interactions using the Bacterial Adenylate Cyclase Two-Hybrid System (BACTH), genes encoding the proteins of interest were cloned into plasmids pKT25 and pUT18c leading to translational fusions with the T25 and the T18 domains of the Bordetella pertussis adenylate cyclase [29]. DNA fragments were amplified by PCR using chromosomal DNA from strain HG001 and specific oligonucleotides (see Table 2) and digested with BamHI and KpnI for introduction into pKT25 and pUT18c. The resulting plasmids were then co-transformed into E. coli strain DHT1 [30] to detect protein interactions and transformants were selected on kanamycin (50 μg/ml) for pKT25 derivatives and ampicillin (100 μg/ml) for pUT18c derivatives.

E. coli strain DHT1 transformants carrying combinations of pKT25 and pUT18c derivatives were tested for cyclic AMP-dependent activation of lacZ expression. Strains were grown in LB medium supplemented with ampicillin (100 μg/ml) and kanamycin (50 μg/ml). Overnight cultures were then spotted on LB-agar plates containing isopropyl β-D-1-thiogalactopyranoside (IPTG; 0.5 mM), ampicillin (100 μg/ml), kanamycin (50 μg/ml), and 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-Gal; 100 μg/ml). Plates were incubated for 24 H at 30°C and examined for the characteristic blue color indicative of β-galactosidase activity through X-Gal hydrolysis. Quantitative β-galactosidase assays were performed on exponentially growing E. coli liquid cultures (30°C, OD_{600 nm} ≈ 1), on SDS/chloroform permeabilized cells as previously described [24]. Zip-Zip associations were used as positive controls [29].

Subcellular localization of fluorescent fusion proteins

The pOLSA plasmid was used to produce fluorescent GFP fusion proteins and is derived from the pCN51 vector carrying the Pcad-cadC promoter module (cadmium chloride-inducible promoter and CadC repressor gene) [31]. A DNA fragment encoding an optimized GFP (GFPopt) preceded by a six amino acid peptide linker (PGSGSG) was cloned downstream from the Pcad promoter in order to obtain a vector allowing production of a fluorescent fusion protein under cadmium-dependent expression. The GFPopt gene was amplified by PCR from the pTetONGFPopt plasmid [32] using oligonucleotides OSA338 and OSA339. Oligonucleotide OSA338 has a 5' extension encoding a PGSGSG peptide linker and allowing in frame translation with the downstream GFPopt. The amplified fragment was cloned into the pCN51 vector between the XmaI and the KpnI restriction sites. Translational fusions of the WalK, WalH, WalI and WalJ proteins with GFPopt were constructed by PCR amplification using HG001 chromosomal DNA and the OSA340/OSA341, OSA357/OSA358, OMA12/OMA34, and OMA14/OMA35 oligonucleotide pairs, respectively (Table 2). Amplicons were cloned into pOLSA between the BamHI and XmaI restriction sites, allowing transcription from the Pcad promoter and production of fusion proteins composed of the protein of interest, the peptide linker, and GFPopt.

Subcellular protein localizations were performed in HG001 S. aureus strains transformed with pOLSA derivatives producing GFP fusion proteins of interest. Fluorescence microscopy was carried out on cells grown in liquid cultures in TSB supplemented with CdCl₂ (0.25 μM) to induce gene fusion expression. When cells reached OD_600nm ≈ 1.5 (exponential growth phase), they were harvested and concentrated 20 times in PBS. Cell suspensions were mixed with Vectashield mounting media (Vector Laboratories, Burlingame, CA) and used for microscopic observations with a Nikon Eclipse E600. Images were acquired with a Nikon Digital Camera DXM1200F. ImageJ software was used for quantifying fluorescence (http://imagej.nih.gov/ij/index.html; [33]). Fluorescence ratios were calculated by measuring fluorescence at the division septa divided by the fluorescence at the lateral wall after subtracting background fluorescence. Quantification was performed for 22 cells (11 septa) for each strain and the results were statistically analyzed by the Wilcoxon signed-rank test using GraphPad Prism 5.0d software (GraphPad Software, San Diego, CA; http://www.graphpad.com) with a P value < 0.05 considered significant.

Bacterial autolysis assays

Bacteria were grown in TSB at 37°C with shaking until OD_{600 nm} ≈ 1, pelleted (10 min; 5,400 x g), resuspended in phosphate buffered saline (PBS) with Triton X-100 (0.1%), and incubated at 37°C with shaking. Lysis was determined as the decrease in OD_{600 nm} over time and indicated as a percentage of the initial OD (measured OD_{600 nm} / initial OD_{600 nm}).

Biofilm formation assays

Strains were grown overnight in TSB, diluted (1/1000) in TSB with glucose (0.75%), NaCl (3.5%) and distributed in PVC microtiter plates (200 μl per well). After 24 h incubation at 37°C, adherent biomass was rinsed twice with PBS, stained with crystal violet, resuspended in ethanol-acetone (80:20) and quantified by measuring OD_{595 nm}, normalized to the OD_{600 nm} of each culture.

Total RNA extraction

Bacteria were grown in TSB at 37°C with shaking and harvested once they reached OD_{600 nm} = 1 (2 min, 5,400 x g). RNA extractions were then performed as previously described [34], followed by a DNase I treatment with the TURBO DNA-free reagent (Ambion, Austin, TX) to eliminate residual genomic DNA.

cDNA synthesis and quantitative real time PCR (qRT-PCR)

Complementary DNAs (cDNAs) were synthesized using the iScript cDNA synthesis kit (Bio-Rad, Hercules, CA), in a 20 μl reaction volume containing 1 μg total RNA. Oligonucleotides were designed for 100–200 bp amplicons using BEACON Designer 4.02 software (Premier Biosoft International, Palo Alto, CA) (see Table 2). qRT-PCRs, critical threshold cycles (CT) and n-fold changes in transcript levels were performed and determined as previously described [3] using the SsoFast^TM EvaGreen Supermix (Bio-Rad, Hercules, CA) and normalized with respect to 16S rRNA whose levels did not vary under our experimental conditions. All assays were performed using quadruplicate technical replicates, and repeated with three independent biological samples, and the data are presented as the mean and standard deviation.

Reverse transcription PCR (RT-PCR)

We used RT-PCR to characterize the wal locus transcripts. Specific wal transcript cDNAs were synthesized from DNase I treated RNA samples (20 μg of total RNA) using reverse oligonucleotides OJU21 (walK), OJU23 (walH), OJU25 (walI), and OP227 (walJ) and AMV Reverse Transcriptase (Roche, Basel, Switzerland) according to the manufacturer’s recommendations. Specific cDNAs were then used as templates to test for co-transcription of wal genes by PCR amplification using Taq DNA polymerase (MP Biomedicals, Santa Ana, CA) with the following oligonucleotide pairs: OJU20/OJU21 to amplify the walR-walK junction, OJU22/OJU23 for the walK-walH junction, OJU24/OJU25 for the walH-walI junction; OJU1/OP227 for the walI-walJ junction, and OJU17/OP227 as a control to amplify the walJ 5' region (see Table 2 for oligonucleotide sequences). Absence of genomic DNA in RNA preparations was verified in control reactions omitting the reverse transcription step. Positive controls were obtained by using HG001 genomic DNA as the template. RT-PCR products were visualized after electrophoresis in 1.5% agarose gels with ethidium bromide staining.

Primer extension analysis

Oligonucleotide primers (OSA434 for walR, and OSA437 for walJ) were 5’end-labeled with [γ-³²P] ATP (Perkin Elmer, Waltham, MA) using T4 polynucleotide kinase (New England Biolabs, Ipswich, MA) and used in primer extension reactions with AMV reverse transcriptase (Roche). Briefly, 20 μg of total RNA and 1 pmol of labeled oligonucleotide were hybridized by heating to 65°C for 3 min and cooling to room temperature. The hybridized oligonucleotide was then extended with 10 U of AMV reverse transcriptase for 30 min at 42°C in the presence of all four dNTPs. The reaction was stopped by addition of formamide stop solution and subjected to electrophoresis on DNA sequencing gels alongside the corresponding dideoxy chain termination sequencing reactions with the same oligonucleotide. DNA sequencing was performed with the Sequenase PCR Product Sequencing Kit (USB, Cleveland, OH) using DNA fragments corresponding to the upstream regions of the genes, generated by PCR using oligonucleotide pairs OSA433/OSA434 for walR and OSA435/OSA437 for walJ.

Phylogenomic analysis

A reference species tree was built from a concatenation of 47 widely distributed ribosomal proteins from 119 representatives of the Bacilli class plus 24 representatives of the Clostridia class as an outgroup. Exhaustive homology searches of the complete set of ribosomal proteins from Firmicutes were performed by BlastP, TBlastN, and different seeds on a local database of complete Firmicute genomes obtained from the NCBI. Each protein data set was aligned using Muscle [35] with default parameters, and unambiguously aligned positions were automatically selected using the BMGE software for multiple-alignment trimming [36] with a BLOSUM30 substitution matrix. Trimmed alignments were then concatenated by allowing a maximum of 17 missing proteins per data set, giving a final character supermatrix of 47 ribosomal proteins and 5,945 amino acid positions for phylogenetic analysis. A maximum likelihood tree was calculated by RAxML [37], with the PROTCATLGF model and four categories of evolutionary rates, as suggested by the Akaike information criterion (AIC) implemented in Treefinder [38]. Support at nodes was calculated by nonparametric bootstrap from 1000 resamplings of the original alignment.

A search for homologs of the WalR-K-H-I-J proteins followed the same strategy as described above. Single gene trees were obtained by RAxML and showed nearly identical topologies, justifying concatenation. Assembly of a concatenated data set followed the same strategy detailed above, by allowing a maximum of 17 missing proteins per data set giving a final character supermatrix of 1,102 amino acid positions. A maximum likelihood tree was calculated by RAxML as described above. Wal protein accession numbers are listed in S1 Table.

Results

The S. aureus wal locus consists of two separate transcription units

The wal locus contains 5 genes in S. aureus, walR and walK, encoding the essential WalKR two-component system (TCS), followed by walH, walI and walJ. Sequence inspection of the S. aureus wal locus indicates that the walK ATG is only 12 bp downstream from the walR amber stop codon, which overlaps the walK ribosome-binding site. Likewise, the walK and walH coding sequence have an 8 bp overlap with the walK ochre codon located within the beginning of the walH coding sequence (ATGaaTAA), whereas the walH ochre codon is immediately followed by the walI ATG codon. We noted that the intergenic region between walI and walJ is particularly long in S. aureus, 389 bp, compared to only 21 bp in B. subtilis. These observations suggest that walJ may have evolved to be transcribed independently from the other wal genes in S. aureus, and that walRKHI are cotranscribed as an operon, with translational coupling between all four genes leading to stoichiometric production of WalR, WalK, WalH and WalI.

This prompted us to carry out a transcriptional analysis of the wal locus in S. aureus. Total RNA was extracted from mid-exponential phase cultures (OD_{600 nm} = 1) in rich medium (TSB) and used for RT-PCR experiments with specific primers designed to amplify DNA fragments encompassing the junctions between each of the wal locus genes (see Table 2 and Fig 1A). As shown in Fig 1A, using wal transcript cDNAs as a template, we were able to amplify DNA fragments corresponding to the junctions of walR/walK, walK/walH and walH/walI, indicating that these four genes are co-transcribed as a tetra-cistronic operon (Fig 1B). However, there was no amplification from cDNA using primers designed to amplify the walI/walJ junction, in agreement with our sequence based prediction that walJ is transcribed as an independent transcription unit. Positive PCR controls using chromosomal DNA as the matrix led to a DNA fragment corresponding to the walI/walJ intergenic region, indicating that the primers hybridize efficiently. Using the wal transcript cDNAs as the matrix, we were able to amplify a DNA fragment using oligonucleotides hybridizing within the walJ 5’ UTR and the beginning of the walJ coding sequence (Fig 1A, last panel), confirming that walJ mRNA was not degraded in the RNA preparation.

Fig 1. The S. aureus wal locus consists of two separate transcription units.

(A) S. aureus total RNA was used to synthesize cDNAs and amplify intergenic regions with oligonucleotides hybridizing specifically within the upstream and downstream genes (indicated by arrows, not to scale). DNA fragments were separated by electrophoresis on ethidium bromide stained 1.5% agarose gels and their size is indicated. Lanes 1: positive PCR control using HG001 genomic DNA as the template; lanes 2: PCR using cDNA as the template; lanes 3: control PCR reaction using total RNA as the template (without reverse transcriptase treatment). M: 100 bp molecular mass ladder (Eurogentec, Angers, France). (B) Representation of the wal locus and the corresponding transcripts (not to scale).

Taken together, our results indicate that walJ is transcribed as a monocistronic transcription unit whereas walRKHI are transcribed as an operon (Fig 1B). Indeed, sequence analysis indicates that the walI ochre codon lies immediately upstream from a probable rho-independent transcription terminator inverted repeat sequence followed by a poly(T) stretch (AAGCTAGCAttGCaagctcatcatatgtgagaaGCggTGCTAGCTT; ΔG = -98 kJ mol^-1). Likewise, a potential rho-independent transcription terminator followed by a poly(T) stretch is located downstream from the walJ ochre codon (CAtCCGATAAAGTtccgcattgctgtgagacgACTTTATCGGgTG; ΔG = -85.3 kJ mol^-1).

The walRKHI operon and the walJ gene are transcribed from σ^A-dependent promoters

To further characterize transcription of the wal locus, we determined the transcriptional start sites for the walRKHI operon and the walJ gene. Primer extension experiments were performed with specific oligonucleotides hybridizing within the walR or walJ coding sequences. As shown in Fig 2A, with the walR oligonucleotide, two cDNAs were synthesized by primer extension (RI and RII). The 3’ end of the longer cDNA (RI) corresponds to a distal transcription start site located 119 bp upstream from the walR translational start site. It is preceded by appropriately spaced potential -10 (TATAAG) and -35 (TATATA) regions, sharing similarities with the consensus sequences of promoters recognized by the vegetative form of RNA polymerase holoenzyme, Eσ^A (Fig 2B). The shorter cDNA, whose 3’ end is 28 bp upstream from the walR translation initiation codon, could correspond either to mRNA processing or to transcription initiation from a secondary promoter. In front of this potential proximal transcriptional start site, there is a well-conserved -10 box (TAAAAT) preceded by a possible -35 sequence (TAAACG), and we noted that the -10 hexamer is directly preceded by a TG dinucleotide (Fig 2B) suggesting that this promoter belongs to the extended -10 promoter family, which are active even in the absence of a consensus -35 sequence [39]. These sequence elements suggest that the shorter transcript could be synthesized by transcription from a secondary promoter (Fig 2B).

Fig 2. The walRKHI operon and the walJ gene are transcribed from σ^A promoters.

(A) Primer extension analysis of the walR (left panel) and walJ (right panel) mRNA. Sanger dideoxy chain termination sequencing reactions (GATC) were carried out on PCR-generated DNA fragments corresponding to the respective upstream regions. (B) DNA sequence of the walR and walJ upstream regions with the identified transcription start sites indicated by arrows and consensus -10 and -35 sequences and extended -10 promoter TG dinucleotides shown by grey boxes. Translation and transcription initiation sites are indicated in bold letters. (C) Expression of the walRKHI operon and the walJ gene was followed using lacZ transcriptional fusions in S. aureus strains ST1365 (PwalR RI + RII), ST1366 (PwalR RII) and ST1367 (PwalJ) and control strain ST1398 carrying plasmid pSA14 alone. β-Galactosidase assays were performed as described in Materials and Methods and measured during late exponential growth at 37°C in TSB. Cultures were spotted on TSB plates containing X-Gal as shown above the histogram bars. (D) Schematic representation of the wal locus genetic structure. Transcription start sites, nucleotide distances or overlap between coding sequences and transcription terminators downstream from walI and walJ are shown.

A single transcriptional start site was identified 181 bp upstream of the walJ translational start site (Fig 2A). This site is preceded by a perfectly conserved -10 box (TATAAT) and a potential -35 box (TTTAGT). The presence of a TG dinucleotide just upstream from the -10 hexamer indicates that the walJ promoter could also belong to the extended -10 promoter family (Fig 2B) [39].

As mentioned above, two potential promoters were identified upstream from the walRKHI operon. In order to verify that the shorter cDNA identified by primer extension with the walR oligonucleotide corresponds to a bona fide transcription initiation site and compare the relative expression levels of the walRKHI operon and the walJ gene, transcriptional lacZ fusions were constructed using the pSA14 vector. DNA fragments were generated by PCR, corresponding to the upstream regions of the walJ gene (oligonucleotides OSA479/OSA475; pSA14-PwalJ) or the walRKHI operon, encompassing either both the PI and PII promoters (oligonucleotides OSA477/OSA473; pSA14-PwalRRI+RII) or the isolated proximal PII promoter region (oligonucleotides OSA478/OSA473; pSA14-PwalRRII) and cloned between the PstI and BamHI sites of plasmid pSA14. The resulting plasmids were introduced into S. aureus strain RN4220 and expression was monitored during growth at 37°C. As shown in Fig 2C, the walR'-lacZ fusion containing both the PI and PII promoters was strongly expressed (approximately 2500 units mg^-1 protein; Fig 2C) whereas expression from the walRKHI operon PII promoter alone was five-fold lower (approximately 460 Units mg^-1 protein; Fig 2C), indicating that the shorter cDNA did not result from mRNA processing but is due to transcription from the secondary promoter. The walJ promoter lacZ fusion was expressed at approximately 540 units mg^-1 protein whereas the background value for the promoterless pSA14 plasmid alone was 0.65 units mg^-1 protein (Fig 2C).

Thus, as shown in Fig 2D, the S. aureus wal locus, unlike that of B. subtilis, is composed of two distinct σ^A-dependent transcription units, one encompassing walRKHI and a second monocistronic unit corresponding to walJ. Furthermore, transcription of the walRKHI operon involves dual σ^A-dependent promoters with different expression levels.

WalH and WalI are membrane-anchored extracellular proteins

Analysis of the primary structures of the Wal proteins using the Phobius web server (http://phobius.sbc.su.se; [27]) suggests that WalR (233 aa) and WalJ (266 aa) are likely cytoplasmic. In contrast, WalH (444 aa) and WalI (262 aa) have predicted single transmembrane domains (residues 7–27 and 9–26, respectively), whereas WalK (608 aa) has a typical histidine kinase domain structure with two amino-terminal transmembrane domains (residues 12–33 and 183–202) flanking a 148 amino acid extracellular loop. In order to determine the membrane topology of WalH and WalI, we used the pho-lac dual reporter system in E. coli that allows distinction between extracellular (alkaline phosphatase activity), and intracellular protein localization (β-galactosidase activity) [40]. Carboxy-terminal in-frame fusions of the ‘phoA’-‘lacZ’ reporter gene with DNA fragments encoding the first 40 amino acids of WalJ (fusion point residue E40) and the predicted transmembrane domain coding sequences of WalH (D40) and WalI (N40) were created with the pKTop plasmid [26]. The resulting plasmids were introduced into E. coli and the strains tested for enzymatic activities on LB plates containing X-Pho as a reporter for alkaline phosphatase activity and Red-Gal for β-galactosidase activity.

As shown in Fig 3A, strong alkaline phosphatase activity is detected with ‘phoA’-‘lacZ’ fusions to the first 40 AA of WalH or WalI. These results indicate that WalH and WalI are anchored to the cell membrane through their amino-terminal transmembrane domain, and are consistent with the remainder of the proteins being located extracellularly (residues 28 to 444 for WalH and 27 to 262 for WalI) (Fig 3B). The WalJ_1-40-‘PhoA’-‘LacZ’ fusion leads to β-galactosidase activity, consistent with the predicted cytosolic localization for WalJ (Fig 3A and 3B).

Fig 3. WalH and WalI are membrane-anchored extracellular proteins.

(A) E. coli DH5α cells producing ‘PhoA’-‘LacZ’ fusion proteins (WalH_1-40-PhoA_21-471-LacZ_4-60, WalI_1-40-PhoA_21-471-LacZ_4-60, or WalJ_1-40-PhoA_21-471-LacZ_4-60) were plated on indicator medium with two chromogenic substrates, Red-Gal (for β-galactosidase activity) and X-Pho (for phosphatase activity). Blue coloring of the colonies (high phosphatase activity) indicates a membrane or extracellular localization of the fusion point. Red coloring of the colonies (high β-galactosidase activity) indicates cytosolic localization of the fusion point. (B) Schematic representation of Wal protein localization and topology with respect to the cell membrane. In the case of WalK and WalR, topology and localization were deduced from primary sequence analysis using the Phobius Hidden Markov Model (indicated by stars).

WalH interacts with both WalI and WalK

The conserved gene order of the wal locus in Bacilli and the translational coupling of the walRKHI operon strongly suggest that the Wal proteins interact. In order to test possible interactions between the Wal proteins, we used the bacterial adenylate cyclase two-hybrid system (BACTH). We fused the full-length WalK, WalR, WalH, WalI and WalJ proteins to the C-terminal domain of either the T18 or T25 subunits of Bordetella pertussis adenylate cyclase using the pKT25 and pUT18c plasmids [29]. To probe putative interactions, E. coli strain DHT1 was co-transformed with combinations of pKT25 and pUT18c derivatives carrying translational fusions to the different wal genes. Upon protein-protein interactions, the proximity between the T18 and T25 subunits restores adenylate cyclase activity, leading to cAMP synthesis and activation of the lactose operon. Interactions were tested both by spotting the resulting strains on LB plates containing X-Gal and by measuring β-galactosidase activity (Fig 4A and 4B). To determine pair-wise interactions, we chose a β-galactosidase activity cut-off value of 250 Miller Units to indicate a positive interaction between the protein fusions.

Fig 4. WalH interacts with both WalI and WalK.

(A) The five Wal proteins (WalR, WalK, WalH, WalI, WalJ) were systematically tested for pairwise interactions in E. coli using the BACTH complementation assay by fusing the corresponding full length coding sequences of each of the wal genes to those of the T25 or T18 adenylate cyclase domains. DHT1 E. coli strains co-transformed with each possible combination of plasmids were spotted on LB agar plates with X-Gal as a chromogenic reporter of β-galactosidase activity (see Materials and Methods). Empty pUT18c and pKT25 vectors were also co-transformed in DHT1 as a negative interaction control, and pUT18-zip and pKT25-zip as a strong positive interaction control (boxed spots). Arrows show β-galactosidase positive cells indicating protein-protein interactions. (B) Quantitative β-galactosidase activity assays for each strain shown in panel A (expressed in Miller Units). Shaded cells indicate significant β-galactosidase activities resulting from positive protein-protein interactions.

Strong β-galactosidase activity was observed with the pUT18c-walH/pKT25-walH and pUT18c-walI/pKT25-walI combinations, indicating that both WalH and WalI self-associate, and the appropriate plasmid pairs allowed us to show that the WalH and WalI proteins also interact with each other (Fig 4A and 4B). Although it is well known that two-component system histidine kinases and response regulators homodimerize, we did not detect self-interaction of WalK or WalR, likely due to instability of the complexes. Concerning interactions between the accessory Wal proteins and the WalKR TCS, the results show that WalH forms a complex with the WalK histidine kinase, but no association between WalI and WalK was detected (Fig 4A and 4B). Thus, the three membrane proteins encoded by the wal locus, i.e. the WalK histidine kinase and the WalH and WalI accessory proteins, associate through interactions with WalH, which appears to be the keystone of the complex. We did not detect any interactions of WalK, WalH or WalI with the cytoplasmic proteins of the system, WalR and WalJ, suggesting either that they do not associate with the other Wal components or that the interaction is transient and that the complex is not sufficiently stable to be detected using this technique.

WalK and WalH are localized at the division septum

We have shown that WalK, WalH and WalI form a complex. To confirm the results of the BACTH experiments and determine the subcellular localization of this complex in S. aureus cells, we examined strains expressing translational GFP fusions by fluorescence microscopy. A dedicated plasmid was constructed, pOLSA (see Materials and Methods) allowing translational fusions to GFP. This plasmid is derived from pCN51 in which the Pcad cadmium-inducible promoter drives expression of the downstream gene [31]. We added an adapter sequence encoding a flexible hexapeptide linker (P-G-S-G-S-G) to avoid steric hindrance between the protein of interest and the fluorescent marker, followed by the GFPopt gene, a codon optimized GFP gene for low-GC% Gram-positive bacteria derived from plasmid pTetONGFPopt [32]. The resulting pOLSA vector allows inducible production of fluorescent protein fusions. Plasmids pOLSA-walK, pOLSA-walH, pOLSA-walI and pOLSA-walJ were constructed, fusing the full length coding sequences of the wal genes to GFP, and introduced into S. aureus strain HG001. The resulting strains were grown in TSB until OD_{600 nm} = 0.2, and CdCl₂ (0.25 μM) was added to the cultures in order to induce expression of the gene fusions. Cells were then harvested at OD_{600 nm} ≈ 1.5 and prepared for fluorescence microscopy (see Materials and Methods).

As shown in Fig 5A, the localization patterns of WalK and WalH were identical, with a strong concentration of the fusion proteins at the division septa (white arrows and zoomed in views), whereas WalI was found to be evenly distributed in the cell membrane and WalJ was located throughout the cytoplasm, in agreement with the ‘phoA’-‘lacZ’ analysis.

Fig 5. WalK and WalH are localized at the cell division septum.

(A) S. aureus HG001 strains producing fluorescent Wal protein fusions (WalK-GFP, WalH-GFP, WalI-GFP, WalJ-GFP) were grown in TSB and observed in mid-exponential phase by fluorescence microscopy. White arrows and zoomed views indicate septal enrichment. (B) Fluorescence ratios (septum/lateral membrane) were quantified for strains producing membrane protein fusions using ImageJ software and plotted using GraphPad Prism. Horizontal lines correspond to average fluorescence ratios with values greater than 2, indicating preferential septal localization (see Materials and Methods). ** P<0.002, *** P<0.001 as determined by the Wilcoxon signed-rank test.

In order to confirm these observations, fluorescence ratios were calculated by measuring fluorescence at the septum versus that at the lateral membrane. Indeed, since the septum contains two membranes, if a fluorescent protein is homogeneously distributed throughout the cell membrane, the intensity of the fluorescent signal at the septum will be approximately twice that at the lateral membrane, whereas the ratio should be greater than 2 if the protein is specifically accumulated at the division septum. As shown in Fig 5B, fluorescence ratios (septum/lateral membrane) were significantly higher than 2 for both WalK and WalH, confirming that they are preferentially localized at the division septum, whereas WalI was evenly distributed throughout the cell membrane.

The absence of WalH and/or WalI leads to increased expression of atlA and decreased biofilm formation but not to increased autolysis

The walRK genes are always followed by walH, walI and walJ in the Bacilli class of bacteria, except for Streptococcaceae, where walH and walI are absent from the genome and the locus is reduced to walRKJ [4]. The strong conservation of the gene order suggests functional interactions between WalH, WalI and WalJ and the WalKR TCS [11]. In order to identify a functional role for WalH, WalI and WalJ, we generated S. aureus mutant strains lacking either walH, walI, walJ or both walH and walI. Growth of the resulting mutant strains was identical to that of the parental HG001 strain (S1 Fig), indicating that contrary to the WalKR system, which is essential for cell viability, WalH, WalI and WalJ do not play crucial roles in S. aureus physiology. Indeed, although the ΔwalH and ΔwalHI mutants displayed a slight lag during the first hour post inoculation, the growth rate and OD_600nm from then on were not significantly different from those of the other strains including the parental strain (S1 Fig). Doubling times (http://www.doubling-time.com/compute.php) were calculated during the exponential growth phase (56 min to 214 min) and gave identical values of 36 min for each strain. This is quite different from the situation in Bacillus subtilis where the absence of either WalH or WalI leads to a cell growth defect [12].

The ΔwalH and ΔwalI mutants of B. subtilis also display a cell wall defect and increased autolysis, due to increased WalKR activity [12]. We have shown that increased WalR activity in S. aureus also leads to increased autolysis [6]. To test whether the absence of WalH and/or WalI is associated with increased susceptibility to cell lysis, we tested the effect of a non-ionic detergent, Triton X-100, on autolysis. Triton X-100 is thought to trigger autolysis by removing lipoteichoic acids, which act as inhibitors of endogenous autolysins [41]. Cells were grown in TSB until approximately OD_600nm = 1. Cells were then harvested and resuspended in PBS containing 0.1% Triton X-100 (See Materials and Methods). Autolysis rates were measured by following the decrease in OD_{600 nm} over time. Autolysis of the ΔwalH mutant was found to be slower, as shown in Fig 6. Simultaneous deletion of both walH and walI led to the same autolysis resistance phenotype as the walH deletion, and complementation of the ΔwalH or ΔwalHI mutants restored autolysis rates to those of the parental HG001 parental strain (Fig 6). Deletion of walI alone had no effect on cell lysis, which was identical to that of the HG001 parental strain under the same conditions (S2 Fig). This contrasts with the situation in B. subtilis where the ΔwalH and ΔwalI mutants display increased autolysis, suggesting that WalKR activity in S. aureus is not higher in the absence of WalH or WalI.

Fig 6. Triton-induced autolysis is decreased in the absence of WalH.

Bacteria were grown in TSB at 37°C with shaking until OD_{600 nm} ≈ 1, pelleted (10 min; 5,400 x g), resuspended in phosphate buffered saline (PBS) with Triton X-100 (0.1%), and incubated at 37°C with shaking. Lysis was determined as the decrease in OD_{600 nm} over time and indicated as a percentage of the initial OD (measured OD_{600 nm} / initial OD_{600 nm}). Results are shown as the mean and standard deviation of three independent experiments. Strains: HG001 (■); ST1397 ΔwalH (○); ST1410 ΔwalHI (△); ST1415 ΔwalH pMK4Pprot-walHI (●); ST1417 ΔwalHI pMK4Pprot-walHI (▲).

Studies in S. aureus, B. subtilis and Streptococcus pneumoniae have shown that cell wall metabolism is the main function regulated by the WalKR system [3, 7–9]. As shown in this work, WalH is a membrane-anchored protein interacting with both WalI and the WalK histidine kinase, leading us to test whether WalH could affect WalKR-dependent regulation of cell wall hydrolase genes. AtlA and Sle1 are the two major autolysins of S. aureus, playing a crucial role in daughter cell separation by splitting the septum during the division process [42, 43]. We have previously shown that expression of sle1 and atlA is positively controlled by WalKR [3, 6]. The WalKR-activated saouhsc_00773 gene encodes a protein harboring a LysM cell wall binding domain and a CHAP domain suggesting a role in cell wall degradation [3]. We compared expression of atlA, sle1 and saouhsc_00773 in the parental HG001 strain and in the ΔwalH, ΔwalI, and ΔwalHI mutant strains by quantitative real time PCR (qRT-PCR). As shown in Fig 7, expression of atlA is increased five- to six-fold in the ΔwalH, ΔwalI and ΔwalHI mutants, however expression of sle1 and saouhsc_00773 was not significantly different in any of the mutants as compared to the parental strain (S3 Fig). Complementation of the ΔwalH, ΔwalI and ΔwalHI mutants with the pMK4-PprotwalHI plasmid fully restored atlA expression levels to those of the parental HG001 strain (Fig 7). Since expression of sle1 and saouhsc_00773 is strongly regulated by WalKR [3, 6], this suggests that WalH and WalI are not involved in WalKR-dependent regulation of these genes under our conditions, and that their effect on atlA expression may be indirect, ie that they are not acting by increasing WalKR activity.

Fig 7. Expression of atlA is increased in the ΔwalH, ΔwalI and ΔwalHI mutants.

The HG001 parental strain and the ΔwalH, ΔwalI and ΔwalHI mutant strains were grown in TSB rich medium until OD_{600 nm} = 1. Total RNA was extracted and quantitative real time PCR was used to compare gene expression. Expression levels were normalized using 16S rRNA as an internal standard and are indicated as the n-fold change with respect to the HG001 parental strain, expressed as means and standard deviations. Dark grey bars indicate expression levels in the parental and mutant strains and light grey bars correspond to values in the complemented strains carrying pMK4-PprotwalHI. * P<0.05 as determined using Student’s t-test.

In order to verify that the absence of WalH and WalI does not lead to a cell wall defect that could be linked to either increased or decreased WalKR activity, strains HG001 and ST1410 (ΔwalHI) were grown overnight in TSB and embedded in thin sections for ultrastructure examination by transmission electron microscopy. No significant differences in cell division, septum placement or cell wall thickness were seen between the parental strain and the ΔwalHI mutant (S4 Fig). We also examined the effect of WalH and WalI on lysostaphin sensitivity. Lysostaphin is a glycyl-glycine endopeptidase that specifically cleaves the pentaglycine cross-bridges of the staphylococcal cell wall, leading to rapid lysis of the bacteria. Strains were grown in TSB until mid-exponential phase (OD_600nm = 1). Cells were then pelleted, washed and resuspended in PBS. Lysostaphin-induced lysis was then followed on the nongrowing cells in the presence of lysostaphin (200 ng/ml) at 37°C and measured as the decline in OD_600nm over time. No difference in lysostaphin sensitivity was observed in the absence of WalH and/or WalI as compared to the parental HG001 strain (S5 Fig).

We previously showed that S. aureus biofilm formation is enhanced when walRK expression is increased or when WalR activity is higher [3, 6]. We therefore tested biofilm formation for the ΔwalH, ΔwalI and ΔwalHI mutants. As shown in Fig 8, biofilm formation was strongly decreased in the absence of WalH and/or WalI. Complementation of the ΔwalH, ΔwalI and ΔwalHI mutants with the pMK4-PprotwalHI plasmid restored biofilm formation to levels comparable to those of the parental HG001 strain (Fig 8). This again suggests that WalKR activity is not increased in the ΔwalH, ΔwalI and ΔwalHI mutants, unlike the situation in B. subtilis.

Fig 8. Biofilm formation is decreased in the ΔwalH, ΔwalI and ΔwalHI mutants.

Biofilm assays were performed in microtiter plates after growth at 37°C for 24 h. Adherent biomass was quantified, normalized to the OD_{600 nm} of each cell culture and represented as n-fold variation compared to the parental strain. Dark grey bars indicate biomass levels in the parental and mutant strains and light grey bars correspond to values in the complemented strains carrying pMK4-PprotwalHI. Experiments were carried out in quadruplicate and standard deviations are indicated. ** P<0.01 as determined using Student’s t-test.

Evolution of wal operon structure and function among Firmicutes

The unique nature of the wal operon structure in Staphylococcus aureus prompted us to carry out a phylogenomic analysis of this locus. The walRK genes are highly conserved in Bacilli, displaying striking synteny with the walHIJ genes in the analyzed genomes, strongly suggesting functional linkage (S1 Table). As previously reported, the walH and walI genes are absent from the genomes of Streptococcaceae [4, 44], which nevertheless maintain synteny between walRK and walJ. Interestingly, we observed that walJ is missing from all Leuconostoc genomes (S1 Table).

We sought to understand if this conservation is due to reshuffling through horizontal gene transfer among Bacilli or a consequence of vertical inheritance. We therefore carried out a phylogenomic analysis of the 5 gene wal locus in 119 representatives of the Bacilli class plus 24 representatives of the Clostridia class as an outgroup. We built a reference species tree by concatenation of 47 widely distributed ribosomal proteins (Fig 9A, see Materials and Methods for details). In parallel, we analyzed the phylogenies of all 5 Wal proteins. Albeit generally poorly resolved, due to the well known lack of a strong phylogenetic signal in single marker genes, these Wal trees were globally consistent with the reference species tree, allowing them to be concatenated into a single character supermatrix of over a thousand positions. The resulting tree (Fig 9B) is highly consistent with the reference species tree, indicating that the ancestor of Bacilli already harbored a 5 gene wal cluster, and that this system was inherited in a largely vertical fashion through speciation events (mapped on Fig 9A). This is unusual for gene clusters, which should be easy to exchange among bacteria. Either the essential nature of the Wal system is such that all attempts at homologous replacement are strongly counterselected, or the system works with additional components that lie elsewhere in the genomes, preventing transfer of a fully functional system. Moreover, our analysis unequivocally shows that the lack of some wal genes in specific genomes is due to gene loss. This is the case for Streptococcaceae, where walH and walI appear to have been neatly excised from the cluster, as well as for Leuconostoc species where walJ was lost (indicated by red crosses in Fig 9A).

Fig 9. Phylogenetic relationships within the Bacilli class and inferred losses of wal genes.

(A) Maximum likelihood phylogeny of Bacilli based on a concatenation of 47 ribosomal proteins comprising 5,945 amino acid positions. The conservation of wal genes in each genome is indicated by circles: walR (red), walK (green), walH (blue), walI (yellow), walJ (purple). Red crosses indicate the loss of walH and walI in Streptococcaceae and of walJ in Leuconostoc. (B) Maximum likelihood phylogeny based on a concatenation of WalR, WalK, WalH, WalI and WalJ protein sequences comprising 1,102 amino acid positions. For both analyses, values at nodes represent bootstrap proportions calculated on 1,000 resamplings of the original data set. For clarity, only the values corresponding to monophyly of families and their evolutionary relationships are shown. The scale bar represents the average number of substitutions per site. For details on analyses, see Materials and Methods.

Streptococcaceae differ from other Firmicutes with respect to the WalKR regulatory pathway. Indeed, whereas the genes encoding the WalR response regulator (RR) and the WalK histidine kinase (HK) are both essential for cell viability in other Firmicutes, only the walR gene is essential in Streptococcaceae, suggesting that the RR can be phosphorylated either by a non-cognate HK or by small phosphate donors such as acetyl phosphate. Non-essentiality of walK in Streptococcaceae is always associated with the absence of walH and walI. Although Streptococci and Staphylococci differ in physiology, this observation prompted us to test whether walK was still essential in a S. aureus strain that lacks both walH and walI. We used a deletion strategy with the pMAD allelic replacement vector [19]. The region upstream from walK was fused to the region downstream of walI in the plasmid construct, pMADwalK (S6 Fig). When cells are grown at the non-permissive temperature for pMAD replication, integration of the entire circular plasmid occurs by a Campbell-type single crossover event, with concomitant duplication of the walR and walJ genes (S6 Fig). Depending on the recombination site, this could then result either in a separation of walRK from their upstream promoter, with one copy of walR alone expressed from its promoter (S6 Fig, event A), or conservation of the walRK locus transcribed from its own promoter (S6 Fig, event B). We determined the integration site for 8 clones, showing that event B, which leaves transcription of walRK intact, occurred in 100% of the integrants. Growth temperature shifts were then performed in order to favor plasmid excision (second crossover event). Depending on the recombination regions, this second crossover could result either in deletion of the walK gene (S6 Fig, event C) or conservation of the wal locus structure in the original ΔwalHI recipient strain (S6 Fig, event D). Two independent experiments were performed and among the 107 tested clones, all retained the walK gene. These results strongly suggest that in S. aureus, the walK gene remains essential, even in the absence of WalH and WalI. They also suggest a global evolution of the WalKR system in Streptococcaceae that has led to non-essentiality of the walK gene.

Discussion

Two-component systems have evolved in free-living Eubacteria as highly conserved mechanisms for adaptation to environmental changes. Among these, the WalKR TCS, ubiquitous among low G+C% bacteria and specific to the Firmicutes phylum, is one of the few that have been shown to be essential for cell viability. This has been linked to its major role in controlling cell wall hydrolases, both in S. pneumoniae and S. aureus [9, 10]. Studies of the WalKR system have essentially focused on the walR and walK genes, yet they are part of a conserved locus encompassing five genes (walR, walK, walH, walI, and walJ) except in Streptococcaceae, which lack the walH and walI genes and have a tricistronic walRKJ operon [4, 44, 45]. In B. subtilis, the wal locus contains an additional gene, yycK (aka yyxA), located downstream from walJ and encoding an HtrA protease orthologue, and it has been shown that walR, K, H, I, J and yycK are all co-transcribed [46]. However, YycK does not appear to play a role in WalKR-dependent regulation [47]. The B. anthracis and Listeria monocytogenes wal loci are also followed by a gene encoding a HtrA-like serine protease [4]. In L. monocytogenes, expression of htrA is controlled by the LisK/LisR TCS, and HtrA plays a role in stress response and pathogenesis [48, 49].

The co-occurence and synteny of the wal genes is a strong indication of a functional link between the Wal proteins. Studies in B. subtilis have shown that WalH and WalI interact with the WalK histidine kinase to negatively control its activity [12, 13, 47, 50]. Although no direct links between WalJ and the WalKR two-component system have been clearly demonstrated, WalJ has also been suggested to be involved in regulation of cell wall metabolism. Indeed, in B. subtilis and S. pneumoniae, deletion of walJ leads to increased sensitivity to some cephalosporin antibiotics and walJ is essential in both bacteria when walRK expression is reduced [9, 14].

WalK, WalR and WalJ are well conserved in Firmicutes (46%, 75% and 59% amino acid sequence identity between B. subtilis and S. aureus, respectively). In contrast, WalH and WalI share no significant similarities, either between different Firmicutes (17% and 21% amino acid sequence identity between B. subtilis and S. aureus, respectively) or with each other (16% identity in S. aureus). Despite their poor conservation, the crystal structures of WalH and WalI from B. subtilis are remarkably similar, with a common fold [51, 52].

In S. aureus, nothing was known about the walH, walI and walJ genes. We carried out a detailed analysis of wal locus transcription and showed that the first gene of the locus, walR, is preceded by two independent σ^A-type promoters driving expression of walR, walK, walH and walI. We note that these experimentally determined promoters are different from the potential promoter previously proposed by in silico analysis [53]. A recent study also identified two promoters upstream from the S. aureus walRK genes, however their position is slightly different from the ones identified here, due to the technique that was used [54]. In that study, a mutation affecting the -10 of the PII promoter identified here led to lowered transcription, in agreement with our results. Based on our established consensus sequence [2], no WalR-binding sites could be identified in the walRKHI operon promoter region, consistent with our previous results indicating that expression of the walRK genes is not autoregulated [6]. As shown in Fig 2C, the fact that transcription of walRKHI occurs from two promoters with different strengths could suggest that one is involved in maintaining a basal level of expression whereas transcription from the other may be regulated in response to specific conditions.

Interestingly, walJ was found to be independently transcribed in S. aureus, whereas in B. subtilis and S. pneumoniae, walJ is co-transcribed with the preceding wal genes [46, 55]. In S. aureus, the intergenic region between walI and walJ is unusually large, with a conserved length of 389 bp. A careful survey of the walI walJ region in Firmicutes indicates that this large intergenic sequence is particularly well conserved and specific to S. aureus strains. The walJ transcription start site is located 182 bp upstream from the start of the walJ coding sequence. This transcribed but untranslated region is highly structured (Pascale Romby, CNRS Strasbourg, personal communication), suggesting it could be involved in post-transcriptional regulation of walJ expression. A transcript corresponding to the walJ upstream sequence, Teg107, was previously identified by high-throughput sequencing of the S. aureus transcriptome [56] and our results show that it corresponds to the walJ 5'-UTR region. In agreement with our findings, an in silico study also predicted that walJ is transcribed independently in S. aureus [57]. Interestingly, walJ has been lost in Leuconostoc species (Fig 9), where the gene immediately following the walRKHI operon encodes an HtrA-like protease, as in B. subtilis and L. monocytogenes.

In an effort to understand the roles of the WalH, WalI and WalJ accessory proteins, we constructed ΔwalH, ΔwalI, ΔwalHI and ΔwalJ mutant strains in S. aureus. The mutants did not display any distinguishing phenotypes, and cell morphology and growth were not affected. Although a number of reports link the S. aureus WalKR system and WalH with vancomycin-intermediate resistance (VISA phenotype) [54, 58–63], the ΔwalH, ΔwalI, ΔwalHI and ΔwalJ mutants were not significantly affected in sensitivity to vancomycin. Indeed, only a very small change was seen for the ΔwalH and ΔwalHI strains (MIC increase from 3 μg/ml to 4 μg/ml) which could be fully complemented (S7 Fig), whereas no difference was observed for the ΔwalI and ΔwalJ mutants. Furthermore, it was recently reported that walR overexpression did not affect vancomycin MICs in Bacillus anthracis [64].

Deletion of walJ in other bacteria leads to decreased resistance to oxidative stress and cell wall targeting antibiotics, and a DNA partition default during cell division, and more recently was associated with a spontaneous mutator phenotype in B. anthracis [14–18]. We tested several of these phenotypes with the ΔwalJ mutant of S. aureus and saw no difference in chromosomal segregation during bacterial division (S8 Fig) and in sensitivity to paraquat-induced oxidative stress (S9 Fig) or cell wall targeting antibiotics (fosfomycin, penicillin, cefotaxime, cefalotin) (data not shown).

In B. subtilis the WalK kinase is localized at the division septum in an FtsZ-dependent manner, interacting with the DivIB, Pbp2B and FtsL division proteins, whereas the WalH and WalI proteins are not localized at the septum but around the cell [50, 65]. In non-dividing B. subtilis cells, in the absence of the divisome, WalH and WalI associate with WalK through their transmembrane helices to inhibit its activity [12, 13, 47]. Thus, in the B. subtilis model, WalK is localized at the division septum in actively dividing cells, where it can no longer interact with WalH and WalI, and the active WalKR system can then coordinate cell wall plasticity with cell division [4, 50].

Our study has highlighted several major differences between the WalKR systems of S. aureus and B. subtilis, with features that appear to be specific to Staphylococci. We show that WalH and WalI are membrane-anchored extracellular proteins, that WalH and WalK interact using the BACTH system in E. coli, and that in contrast to the situation in B. subtilis, WalH and WalK are both localized at the division septum. We also showed that WalH and WalI interact with each other, suggesting that WalH, WalI and WalK form a ternary complex. However, although we could not detect any interactions between WalI and WalK, WalI is also found at the septum, even if it does not localize there preferentially, such that all three proteins can interact there. However, we cannot exclude that fusing GFP to WalI may prevent it from localizing efficiently to the septum. We have also shown that the walJ gene is transcribed independently from the other wal genes in S. aureus, unlike the wal operon of B. subtilis.

Furthermore, unlike the situation in B. subtilis, no increased autolysis was observed for the ΔwalH, ΔwalI and ΔwalHI mutant strains (see Fig 6 and S2 Fig), nor did the mutants present growth or cell wall defects (S1 and S4 Figs). On the contrary, Triton X-100 induced autolysis was actually decreased in the ΔwalH and ΔwalHI mutant strains, and this phenotype could be fully reversed upon complementation (Fig 6).

Biofilm formation, which we have shown to be directly correlated with WalR activity, was lowered in the ΔwalH, ΔwalI, ΔwalHI mutant strains (Fig 8), and WalKR-dependent expression of sle1 and saouhsc_00773 was not affected by the absence of WalH and WalI (S3 Fig). Taken together, these results indicate that WalH and WalI do not negatively regulate WalKR activity in S. aureus as they do in B. subtilis, consistent with their divergent sequences. In agreement with our results, a report where full length WalH, WalI, WalK and WalR proteins of S. aureus were purified and combined in vitro showed that WalH and WalI had no significant effect on WalK autophosphorylation or phosphotransfer to WalR [66]. A two-fold increase in autolysis and biofilm formation was recently reported for a walI mutant of S. aureus, however the mutant was a gene disruption, leaving much of the protein produced, and the strain used was a clinical isolate carrying uncharacterized mutations leading to increased biofilm formation and Agr activity [67].

Both walH and walI have been lost from Streptococcaceae as shown by our phylogenomic analyses (Fig 9). This was suggested to be linked to a lack of capacity for electron transport [44]. However, we saw no difference in expression of walK, walR or atlA during aerobic or anaerobic growth of S. aureus indicating neither expression levels nor activity of the system are affected under these conditions (S10 Fig). WalH and WalI are absent in Streptococcaceae where the WalK kinase lacks an extracytoplasmic domain and is not essential [4, 45]. By measuring resolution frequencies for tandem chromosomal duplications following integration of plasmids designed to delete walK, we showed that WalK remains essential in S. aureus even in the absence of WalH and WalI (S6 Fig). This is consistent with the fact that in S. pneumoniae, where WalK is not essential, the WalR response regulator can be phosphorylated by small molecule phosphodonors such as acetyl phosphate [68], whereas purified WalR of B. subtilis or S. aureus could not be phosphorylated using radiolabeled acetyl phosphate (S. Dubrac, unpublished results). A recent report has shown that in Listeria monocytogenes, WalK also appears as not essential for cell viability, as in Streptococcaceae [69]. However, sequence analysis of the wal locus in Listeria monocytogenes reveals a very different structure from that of S. aureus or B. subtilis: the walR gene in L. monocytogenes does not appear to be co-transcribed with the remaining walKHIJ genes. Indeed, the intergenic region between walR and walK is 184 bp versus 12 in S. aureus, with a likely Rho-independent transcription terminator located downstream from walR. This suggests that WalR in L. monocytogenes may have evolved so as to be active even in the absence of its cognate kinase, much like in S. pneumoniae.

Cell wall biosynthesis and degradation require fine-tuning in order to ensure proper cell division and avoid cell lysis. In B. subtilis WalH and WalI control WalKR activity whereas they are absent in Streptococcaceae. Our results are consistent with the idea that the WalKR system of Staphylococci may represent an evolutionary intermediate between those of Bacillus and Streptococci, with a progressive loss of importance of WalH and WalI. In Staphylococci, WalH and WalI are still present, but no longer play an essential role in negatively controlling WalKR activity as in B. subtilis. This suggests that other components could be involved in fine-tuning WalKR activity, and this hypothesis is currently being investigated in our laboratory.

Supporting Information

S1 Fig. Growth curves of S. aureus HG001 and wal mutant strains.

Bacterial cultures were grown overnight, inoculated in TSB at a calculated OD_600nm of 0.05 and incubated at 37°C with shaking. Optical densities were followed over a 7.5 hour period. Results are shown as the mean and standard deviation of three independent growth curves. Doubling times (http://www.doubling-time.com/compute.php) were calculated during the exponential growth phase (56 min to 214 min) and gave identical values of 36 min for each strain. Strains: HG001 (■); ST1397 ΔwalH (◆); ST1130 ΔwalI (▲); ST1410 ΔwalHI (△); ST1131 ΔwalJ (○).

(TIF)

S2 Fig. WalI does not affect Triton-induced autolysis.

Bacteria were grown in TSB at 37°C with shaking until OD_{600 nm} ≈ 1, pelleted (10 min; 5,400 x g), resuspended in phosphate buffered saline (PBS) with Triton X-100 (0.1%), and incubated at 37°C with shaking. Lysis was determined as the decrease in OD_{600 nm} over time and indicated as a percentage of the initial OD (measured OD_{600 nm} / initial OD_{600 nm}). Results are shown as the mean and standard deviation of three independent experiments. Strains: HG001 (■); ST1130 ΔwalI (▲).

(TIF)

S3 Fig. WalH and WalI have no effect on expression of the sle1 and saouhsc_00773 WalKR regulon autolysin genes.

The HG001 parental strain and the ΔwalH, ΔwalI and ΔwalHI mutant strains were grown in TSB rich medium until OD_{600 nm} = 1. Total RNA was extracted and quantitative real time PCR was used to compare gene expression in the strains. Expression levels were normalized using 16S rRNA as an internal standard and are indicated as the n-fold change with respect to the parental HG001 strain, expressed as means and standard deviations.

(TIF)

S4 Fig. The ΔwalHI mutant does not display any cell wall defects.

Strains HG001 and ST1410 (ΔwalHI) were grown overnight in TSB and embedded in thin sections for ultrastructure examination by transmission electron microscopy. Bacteria were fixed with 2,5% glutaraldehyde for 2 hrs at RT, postfixed with 1% osmium tetroxide for 1 hr at RT, dehydrated in a graded series of ethanol baths, and embedded in Epon. Thin sections were cut with a Leica Ultramicrotome Reichert Ultracut S, and stained with uranyl acetate and lead citrate. Images were taken with a FEI Tecnai Bio-Twin Transmission Electron Microscope at 120kV. Bars represent 0.5 μm.

(TIF)

S5 Fig. Lysostaphin sensitivity of wal mutant strains.

Bacteria were grown in TSB at 37°C with shaking until mid-exponential phase (OD_{600 nm} ≈ 1). Cells were then pelleted, washed, and resuspended in PBS. Lysostaphin-induced lysis was then followed on the nongrowing cells in the presence of lysostaphin (200 ng/ml) at 37°C, measured as the decline in OD_600nm over time and indicated as a percentage of the initial OD (measured OD_{600 nm} / initial OD_{600 nm}). Results are shown as the mean and standard deviation of three independent experiments. Strains: HG001 (▰); ST1397 ΔwalH (●); ST1130 ΔwalI (▲); ST1410 ΔwalHI (◆).

(TIF)

S6 Fig. The walK gene remains essential in S. aureus, even in the absence of walH and walI.

Schematic representation of the wal locus in strain HG001 and the ΔwalHI mutant and the recombinant pMADwalK plasmid designed to delete the walK gene in the walHI mutant (not to scale). A, B, C, and D indicate the possible recombination events and their frequencies as tested by PCR screening with appropriate primers.

(TIF)

S7 Fig. Vancomycin MICs are not significantly affected by the ΔwalH, ΔwalI, ΔwalHI and ΔwalJ mutations.

Bacterial strains were grown overnight at 37°C with shaking in TSB (supplemented with 10 μg/ml chloramphenicol for the complemented strains). Overnight cultures were diluted in TSB to reach an optical density around 0.1 (0.5 McFarland) and flooded on BHI agar plates. The E-tests (Biomérieux) were then applied on dried plates and incubated for 24 hours at 37°C before observation.

(TIF)

S8 Fig. Chromosome segregation is not affected in the S. aureus ΔwalJ mutant.

The HG001 and ST1131 (ΔwalJ) strains were grown in TSB until OD_600nm ≈1. Half of each culture was treated with novobiocin (0.5 mg/ml) in order to inhibit chromosomal replication (this has been shown to amplify the DNA segregation phenotype of a ΔwalJ mutant strain in B. subtilis), and incubation was pursued. Cells were harvested after one hour and resuspended in PBS. Cell membranes were labeled with FM1-43FX (red) and DNA with DAPI (blue). Suspensions were immediately mounted with Vectashield and observed with a Nikon Eclipse E600 microscope. Images were acquired with a Nikon Digital Camera DXM1200F. No differences were observed between cultures with or without novobiocin treatment, and the images shown are from the cultures treated with novobiocin. Panels A: membrane labeling (FM1-43FX), B: DNA labeling (DAPI), C: merged. No defects in DNA segregation were observed for the ΔwalJ mutant as compared to the HG001 parental strain.

(TIF)

S9 Fig. The ΔwalJ mutant strain does not show increased sensitivity to paraquat-induced oxidative stress.

S. aureus strains HG001 and ST1131 (ΔwalJ) were grown overnight, and inoculated at a calculated OD_600nm of 0.05 in TSB supplemented with the indicated concentrations of paraquat (methyl viologen hydrate). Growth was followed by measuring OD_600nm over a six hour period. Growth of the HG001 parental strain (closed symbols) and the ΔwalJ strain (open symbols) was not affected by paraquat concentrations up to 10 mM, whereas growth of both strains was strongly inhibited in the presence of 20 mM paraquat. (circles: 5 mM paraquat; squares: 10 mM; diamonds: 20 mM).

(TIF)

S10 Fig. Expression of walRK and the atlA WalKR-regulated autolysin gene is not affected by anaerobiosis.

S. aureus strain HG001 wild type strain was grown overnight, inoculated in TSB at a calculated OD_600nm of 0.05 and half of the culture was incubated at 37°C under agitation whereas the other half was incubated at 37°C in an anaerobic jar. At an OD_600nm ≈ 1, cells were harvested and treated for RNA extraction, cDNA synthesis and qRT-PCR analysis (see Materials and Methods section). Data are represented as n-fold change during growth under anaerobiosis compared to aerobiosis, expressed as means and standard deviations.

(TIF)

S1 Table. Sequence accession numbers for Wal proteins used in the phylogenomic analysis.

(PDF)

Data Availability

All relevant data are within the paper and its Supporting Information files.

Funding Statement

Support was provided by Institut Pasteur Transversal Research Program No. 336 [http://www.pasteur.fr]; Agence Nationale de la Recherche, Grant ANR-08-ALIA-0011 NABAB [http://www.agence-nationale-recherche.fr/]; and Agence Nationale de la Recherche, Grant ANR-09-MIEN-0010 GRABIRON [http://www.agence-nationale-recherche.fr/]. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.