Isolation of a Membrane Protein Complex for Type VII Secretion in Staphylococcus aureus

ABSTRACT

The ESAT6-like secretion system (ESS) of Staphylococcus aureus promotes effector protein transport across the bacterial envelope. Genes in the ESS cluster are required for S. aureus establishment of persistent abscess lesions and the modulation of immune responses during bloodstream infections. However, the biochemical functions of most of the ESS gene products, specifically the identity of secretion machine components, are unknown. Earlier work demonstrated that deletion of essB, which encodes a membrane protein, abolishes S. aureus ESS secretion. Loss-of-function mutations truncating the essB gene product cause dominant-negative phenotypes on ESS secretion, suggesting that EssB is a central component of the secretion machinery. To test this prediction, we purified native and affinity-tagged EssB from staphylococcal membranes via dodecyl-maltoside extraction, identifying a complex assembled from five proteins, EsaA, EssA, EssB, EssD, and EsxA. All five proteins are essential for secretion, as knockout mutations in the corresponding genes abolish ESS transport. Biochemical and bacterial two-hybrid analyses revealed a direct interaction between EssB and EsaA that, by engaging a mobile machine component, EsxA, may also recruit EssA and EssD.

IMPORTANCE Type VII secretion systems support the lifestyle of Gram-positive bacteria, including important human pathogens such as Bacillus anthracis, Mycobacterium tuberculosis, and Staphylococcus aureus. Genes encoding SpoIIIE-FtsK-like ATPases and WXG100-secreted products are conserved features of type VII secretion pathways; however, most of the genes in T7SS clusters are not conserved between different bacterial species. Here, we isolate a complex of proteins from the membranes of S. aureus that appears to represent the core secretion machinery, designated ESS. These results suggest that three membrane proteins, EsaA, EssB, and EssA, form a secretion complex that associates with EssC, the SpoIIIE-FtsK-like ATPase, and with EsxA, a mobile machine component and member of the WXG100 protein family.

INTRODUCTION

Staphylococcus aureus colonizes the nasopharynx and intestinal tract of humans to cause skin and soft-tissue infections as well as invasive diseases (1, 2). A hallmark of S. aureus infection is the development of abscess lesions, where bacteria replicate in large communities while modifying innate and adaptive immune responses of infiltrating immune cells (3, 4). This virulence strategy of S. aureus enables persistent infections and dissemination to uninfected host tissues or transmission among individuals of a host population (3, 4). Genetic analyses identified the ESAT-6-like secretion system (ESS) as a key contributor for S. aureus pathogenesis and for the establishment of persistent abscess lesions (5–8).

The type VII secretion system (T7SS) pathway was originally identified in Mycobacterium tuberculosis as a gene cluster required for bacterial export of two WXG100 proteins, ESAT-6 and CFP-10 (9–12). The common factor between all T7SS pathways is an SpoIIIE-FtsK-like ATPase, EssC, that is thought to catalyze WXG100 transport (9). Members of the WXG100 protein family assemble as homo- or heterodimeric tetrahelical bundles, where each subunit contributes two α-helices separated by a Trp-Xaa-Gly (WXG) motif, introducing a 180° degree turn (13–17). Many additional genes are required for the proper functioning of T7SS pathways; however, these genes are not universally conserved among different bacterial species (18, 19). Further, each microbe appears to have evolved unique genes for effector proteins, thereby implementing species-specific virulence strategies (20–23). Bioinformatic analyses distinguished two broad classes of type VII secretion pathways: T7aSS of actinobacteria and T7bSS of firmicutes (18).

Bioinformatic analyses of the chromosomal region encoding the ESS (T7bSS) cluster from 153 clinical S. aureus isolates identified conserved and variable elements (24). Beginning with the 5′ end of the chromosomal plus strand boundary of ESS are five absolutely conserved genes designated module 1, esxA-esaA-essA-esaB-essB, where esxA encodes a WXG100 protein, esaA, essA, and essB encode membrane proteins, and the esaB gene product represents a cytoplasmic protein (24) (Fig. 1A and B). Module 2 is comprised of six variable genes, beginning with essC, which also encodes a membrane protein (Fig. 1B). While the essC gene is a conserved feature of all T7SS, in S. aureus four essC alleles (essC1, essC2, essC3, and essC4) have been distinguished based on sequence variations at the 3′ end of the open reading frame, in association with four distinct groups of genes located downstream of essC (24). Module 2 of the American epidemic clone for community-acquired, methicillin-resistant S. aureus (CA-MRSA) infections, strain USA300 LAC (CC8), includes essC1-esxC-esxB-essE-esxD-essD (24) (Fig. 1A). Similar to the product of esxA, the product of esxB, but not those of esxC and esxD, belongs to the WXG100 family of proteins that are prone to oligomerization (13–17). However, EsxA and EsxB of S. aureus do not assemble in EsxA/EsxB dimers or polymers (15, 22). Instead, EsxB forms a complex with EsxD (EsxB-EsxD) and EsxA assembles with EsxC or with itself (EsxA-EsxC and EsxA-EsxA, respectively) (22). The product of essD is a secreted effector with endonuclease activity (23, 25). In the staphylococcal cytoplasm, EssD associates with EssE, which functions as the secretion chaperone for EssD (8, 25). Module 3 is a complex arrangement of a variable number of genes encoding the domain of unknown function, DUF600, interspersed with genes encoding several predicted transmembrane proteins. In USA300 LAC, module 3 encompasses 9 genes, with DUF600 recently identified as the essI family, whose products associate with the C-terminal nuclease domain of EssD and prevent EssD-mediated degradation of the bacterial chromosome (Fig. 1A) (23, 25). Finally, module 4 contains two conserved genes whose products are predicted to be secreted via the Sec pathway or located in the bacterial cytoplasm; their contributions to ESS secretion or staphylococcal pathogenesis are not yet established (24).

FIG 1.

Solubilization of EssB from Staphylococcus aureus membranes. (A) Schematic representation of the ESS gene cluster of S. aureus strain USA300 LAC with its four modules (1 to 4). Transmembrane segments were deduced using the topology prediction software TMHMM Server v. 2.0 (32). (B) Proposed topology and domain architecture of four proteins (EsaA, EssA, EssB, and EssC) with predicted transmembrane segments. Structural data are not available for EsaA and EssA; the EssB homodimer is drawn as reported previously (27). The two forkhead-associated and three ATPase-type domains of EssC are shown in light and dark orange, respectively (42). Numbers indicate positions of amino acids in the sequences of the proteins. (C) S. aureus USA300 LAC was grown on tryptic soy agar supplemented with 0.2% horse serum. Staphylococci were suspended in buffer and washed, and cell wall peptidoglycan was digested with lysostaphin. Cell lysate was obtained by removing unbroken cells by centrifugation at 5,000 × g for 15 min. Cell lysate in the supernatant then was subjected to ultracentrifugation at 100,000 × g for 2 h, separating membrane proteins in the sediment (UP1, ultracentrifugation pellet 1) from soluble cytoplasmic proteins (US1, ultracentrifugation supernatant 1). The UP1 sample was suspended in buffer with 2% n-dodecyl-β-d-maltoside (DDM) and again centrifuged at 100,000 × g for 2 h, generating samples UP2 and US2. Proteins in all fractions were separated by 15% SDS-PAGE and visualized with Coomassie blue staining or transferred to PVDF membranes for immunoblotting with rabbit antibodies specific for sortase A (anti-SrtA), EssB (anti-EssB), EsxA (anti-EsxA), and the cytoplasmic ribosomal protein L6 (anti-L6). Numbers indicate the migration of molecular mass markers in kilodaltons.

We hypothesize that the products of module 1 genes, in conjunction with the module 2 EssC ATPase, function as a membrane-embedded machinery for the secretion of WXG100 proteins and EssD. EssB is a 444-amino-acid protein with an M_r of 52,023 and central membrane-spanning domain, i.e., a predicted transmembrane helix derived from hydrophobic amino acids at residues 229 to 251 (Fig. 1B). Earlier studies generated mutations in essB that prematurely truncate EssB polypeptide at variable positions; the resulting EssB variants fractionate with staphylococcal membranes in a manner dependent on the hydrophobic domain (residues 229 to 251) and cannot complement the ESS secretion defect of the essB null mutant (26). Further, EssB truncation variants exert a dominant-negative ESS secretion phenotype when expressed in wild-type staphylococci, suggesting that EssB is a key factor of the secretion machinery (26).

Geobacillus thermodenitrificans EssB has a membrane topology similar to that of S. aureus EssB (27). The crystal structures of purified recombinant, soluble N-terminal (EssB_N; located in the bacterial cytoplasm) and C-terminal (EssB_C; positioned on the bacterial surface) domains of G. thermodenitrificans EssB have been determined (27). EssB_N assumes a fold related to protein kinases with two globular domains separated by a cleft (27). The helical fold of EssB_C is thought to extend over the trans side of the membrane, forming a cradle-like structure (27). X-ray crystallography of staphylococcal EssB_N confirmed the structural analysis of G. thermodenitrificans EssB_N (28). Analyzing staphylococcal EssB_N and purified recombinant proteins as well as bacterial two-hybrid encounters with cloned EsxA, EsxB, EsaB, or EsxC did not identify specific interactions with EssB_N (29). Further, in vivo cross-linking experiments followed by blue native PAGE analyses did not identify heteromeric interactions between EssB and other ESS components, including EsaA, EssA, or EssC (29). Cross-linking agents were found to capture homomeric complexes of each protein, and it was concluded that EsaA, EssB, and EssC are unable to interact with each other. However, the use of strains expressing esaA, essB, and essC from a plasmid with an inducible promoter may have prevented the assembly of supramolecular complexes in favor of homooligomer formation (29).

To test our hypothesis, here we used a biochemical approach, purifying native and affinity-tagged EssB from staphylococcal membranes via dodecyl-maltoside extraction. This approach identified a complex assembled from five ESS secretion factors: EsaA, EssA, EssB, EssD, and EsxA. Biochemical and bacterial two-hybrid analyses revealed direct interactions between EssB and EsaA that, by engaging the mobile machine component, EsxA, also may be involved in recruiting EssA and EssD.

RESULTS

Solubilization of EssB and EsxA from staphylococcal membranes.

To purify EssB and other components of the ESS pathway, secretion machinery proteins must first be solubilized from staphylococcal membranes. Here, we asked whether EssB and EsxA can be solubilized from membranes by extraction with n-dodecyl-β-d-maltoside (DDM). Wild-type (WT) CA-MRSA strain USA300 LAC was grown on plates, bacteria were washed and suspended in phosphate buffer, and bacterial peptidoglycan was digested with lysostaphin. Crude bacterial lysate was ultracentrifuged, and the resulting membrane sediment (UP1, for ultracentrifugation pellet 1) was separated from soluble cytoplasmic proteins (US1, for ultracentrifugation supernatant 1). The UP1 fraction was suspended in 50 mM Bis-Tris (ACA) buffer with 2% DDM for 2 h at 4°C. The extract was ultracentrifuged again to generate two new fractions, UP2 and US2. Fraction UP2 contained proteins refractory to DDM treatment, whereas US2 contained proteins that were solubilized by DDM. Proteins in US1, UP1, US2, and UP2 were separated by 15% SDS-PAGE and visualized by staining with Coomassie brilliant blue (Fig. 1C). The results demonstrated that DDM extraction solubilized some, but certainly not all, membrane proteins of S. aureus in UP1 into US2 (Fig. 1C). Immunoblot analysis was used to analyze DDM extraction of EssB and sortase A (SrtA), a membrane-anchored transpeptidase (30). Data shown in Fig. 1C demonstrate that DDM extraction of staphylococcal membranes solubilized both EssB and SrtA. As expected, the cytosolic ribosomal protein L6 did not sediment following the first centrifugation and was found exclusively in fraction US1 (Fig. 1C). Of note, the WXG100 protein EsxA fractionated equally between US1 and UP1, suggesting that the protein reversibly associates with staphylococcal membranes. EsxA in the UP1 fraction was effectively solubilized by DDM (Fig. 1C).

Size exclusion chromatography of DDM-solubilized ESS complexes.

When analyzed by size exclusion chromatography on Superdex 200, DDM-solubilized proteins in the US2 sample were separated into large peaks of absorbance at 215 nm (Fig. 2A). Proteins in the first peak (peak A) eluted with the void volume of the column (∼46 ml), which corresponds to assemblies with molecular masses of ≥900 kDa. The asymmetric peak A displays a prominent right shoulder, suggesting the presence of multiple different complexes with smaller mass (≤800 kDa). A second, broad peak (peak B) with an elution volume of 64 ml includes protein assemblies with a molecular mass of ∼220 kDa. Immunoblot analysis of eluted fractions revealed the presence of EssB in peak A and peak B fractions (Fig. 2B). Proteins eluting in peaks A and B were separated by SDS-PAGE and stained with Coomassie, and groups of protein bands were excised for trypsin cleavage and peptide identification via mass spectrometry and bioinformatic analysis with predicted tryptic peptides derived from the translated genome of S. aureus USA300 LAC (31). This approach identified full-length EssB and EsaA in peak A and EssB and EsxA in peak B (Fig. 2C). EsaA is a 1,009-amino-acid-long protein with an M_r of 114,825. The protein contains six hydrophobic segments with a large extracytoplasmic loop between residues 28 and 825 (32) (Fig. 1B). EsxA is a 97-amino-acid WXG protein with an M_r of 11,036. The presence of EssB and EsaA in peak A as well as EssB and EsxA in eluted fractions of peaks A and B, respectively, was confirmed by immunoblotting using polyclonal rabbit antibodies raised against purified recombinant proteins (Fig. 2B).

FIG 2.

Size exclusion chromatography of DDM-solubilized ESS machine components. (A) DDM-solubilized US2 sample (∼10 mg protein) from S. aureus USA300 LAC was subjected to Superdex 200 16/60 HiLoad preparation-grade (pg) gel filtration chromatography at 4°C in the presence of 0.03% DDM while measuring absorbance at 215 nm (mAU) per ml elution volume. Eluted peaks (A, A′, and B) are indicated with the corresponding molecular masses derived by calibration of the column with established protein standards. (B) Aliquots of representative elution fractions were analyzed by 15% SDS-PAGE and immunoblotting with polyclonal rabbit antibodies anti-EssB, anti-EsaA, and anti-EsxA. Numbers indicate the migration of molecular mass markers (in kilodaltons). (C) The amino acid sequences of tryptic peptides (red) from EssB and EsaA were identified by mass spectrometry in the 46-ml fraction of peak A. Tryptic peptides of EsxA were identified in the 65-ml fraction of peak B. Numbers inserted in the amino acid sequence of EsaA, 337//796, depict a segment of the protein for which tryptic peptides were not identified, and the amino acid sequence for this segment is not shown.

esaA is required for ESS secretion in S. aureus USA300 LAC.

We asked whether the EsaA membrane protein is required for ESS secretion in S. aureus USA300 LAC. Using allelic replacement, we deleted the esaA gene in USA300 LAC, thereby generating the ΔesaA variant. To examine the contribution of esaA to ESS secretion, ΔesaA variant cultures were centrifuged to separate bacterial cells and the extracellular medium. Bacterial cells were lysed by lysostaphin digestion. Proteins in the cell and medium compartments were separated by SDS-PAGE and analyzed by immunoblotting. This experiment revealed that in wild-type staphylococci, EsaA is located in bacterial cells, whereas EsxA and EsxC are found in both the extracellular medium and in bacterial cells (Fig. 3). Deletion of esaA abolished the expression of esaA but not the production of EsxA and EsxC. However, deletion of esaA abrogated the secretion of EsxA and EsxC into the extracellular medium (Fig. 3). Multiple immune-reactive species specific for EsaA could be detected in wild-type but not esaA mutant extracts, suggesting that EsaA is physiologically processed or degraded during sample preparation. Plasmid-borne expression of wild-type esaA in the ΔesaA variant restored both EsaA production, with a similar pattern of immune-reactive species, and the secretion of EsxA and EsxC (Fig. 3). While a faint reactive band of a smaller EsaA species was detected in strains overproducing EsaA, the cytosolic ribosomal protein L6 was not found in the medium fraction, ruling out the possibility of extensive cell lysis during sample preparation. We conclude that esaA expression is essential for ESS secretion in S. aureus USA300 LAC.

FIG 3.

esaA gene is essential for ESS secretion of S. aureus USA300 LAC. Cultures of S. aureus USA300 LAC wild type (WT), its ΔesaA isogenic esaA deletion mutant, or a ΔesaA mutant carrying a plasmid for expression of wild-type esaA, the ΔesaA(pesaA) mutant, were grown to an A₆₀₀ of 1.0. Cultures were centrifuged to separate the bacterial sediment (cell) from the culture medium (medium). Bacterial cells were suspended in buffer and peptidoglycan digested with lysostaphin, and cellular proteins were precipitated with trichloroacetic acid (cell). Proteins in cell and medium fractions were subjected to immunoblotting with antibodies anti-EsaA, anti-EsxA, anti-EsxC, and anti-L6 to rule out cell lysis. Numbers indicate the migration of molecular mass markers (in kilodaltons).

Size exclusion chromatography of DDM-solubilized ESS from essB and esaA mutant staphylococci.

Using the extraction procedure described above for wild-type staphylococci, DDM-solubilized membrane proteins from ΔessB or ΔesaA mutant S. aureus USA300 LAC were subjected to gel filtration chromatography on a Superdex 200 column. Peak A (≥900 kDa; 44 to 46 ml) of the ΔessB mutant sample displayed a diminished asymmetrical shoulder at 47 ml, and the abundance of peak B (∼200 kDa; 58 to 68 ml) was reduced (Fig. 4A). A new, broad peak B′ (∼80 kDa) eluted at 68 to 78 ml. Immunoblot analysis confirmed the absence of EssB in all fractions (Fig. 4B). Proteins eluting in the 65- and 73-ml fractions of the ΔessB mutant sample were subjected to trypsin digestion and mass spectrometry. EssB and EsaA were not detected; however, EsxA peptides were identified in both the 65- and the 73-ml fractions. Mass spectrometry data were corroborated by immunoblotting of eluate samples with antibodies specific for EsxA, EssB, and EsaA. EsaA protein was detected in peak A, whereas EsxA was detected in peaks B and B′ (Fig. 4B).

FIG 4.

Size exclusion chromatography of DDM-solubilized ESS machine components from ΔessB and ΔesaA mutants of S. aureus. (A) DDM-solubilized US2 samples (∼10 mg protein) from S. aureus USA300 LAC (WT) or the ΔessB or ΔesaA mutant were subjected to Superdex 200 16/60 HiLoad pg gel filtration chromatography at 4°C in the presence of 0.03% DDM while measuring absorbance at 215 nm (mAU) per ml elution volume. Eluted peaks (A, A′, B, and B′) are indicated with the corresponding molecular masses derived by calibration of the column with established protein standards. Aliquots of representative elution fractions were analyzed by 15% SDS-PAGE and immunoblotting with polyclonal rabbit antibodies anti-EssB, anti-EsaA, and anti-EsxA in US2 samples from ΔessB (B) or ΔesaA (C) mutant staphylococci.

When DDM membrane extracts of the ΔesaA strain were chromatographed over the Superdex 200 column, peaks A and B were not observed. Instead, a new peak, B′, eluting at 68 ml to 78 ml (∼80 kDa), was observed (Fig. 4A). Immunoblotting did not identify EssB in eluate fractions otherwise corresponding to peak A (fractions of 43 to 52 ml) and peak B (fractions of 58 to 68 ml) (Fig. 4A). EsxA was detected in the ∼80-kDa peak B′, although EsxA abundance was diminished compared to that for peak B′ from ΔessB DDM-extracted membranes (Fig. 4B).

Together, the results show that recruitment of EsxA to a protein complex eluting in peak B requires both intact EsaA and EssB. Further, in the absence of EsaA, EssB is no longer observed and the retention time of EsxA is drastically increased, suggesting a complete loss of DDM-extractable complexes.

Affinity chromatography of DDM-solubilized EssB_His.

Earlier work demonstrated that essB expression is essential for ESS secretion in S. aureus USA300 LAC. The ESS secretion defect of the ΔessB mutant can be complemented by plasmid-borne expression of wild-type essB and by expression of an essB variant extended at its 3′ end with six codons specifying a six-histidyl tag (essB_His). Here, we asked whether assembly of EssB_His into the ESS secretion machinery can be detected via DDM extraction of staphylococcal membranes and affinity chromatography on nickel-nitrilotriacetic acid (Ni-NTA)–Sepharose. To address this question, S. aureus USA300 ΔessB(pessB) and ΔessB(pessB_His) mutant cultures were centrifuged, bacterial cells lysed, and membranes sedimented at 100,000 × g. Membranes (UP1) were extracted with 2% DDM and again centrifuged; US2 supernatant was subjected to Ni-NTA–Sepharose chromatography, columns were washed, and bound proteins were eluted with 250 mM imidazole buffer (eluate). Separation of eluate from Ni-NTA–Sepharose chromatography of DDM-solubilized membrane proteins from S. aureus USA300 ΔessB(pessB) and ΔessB(pessB_His) mutants on Coomassie-stained SDS-PAGE gel revealed many different protein species in the EssB_His sample but not in the EssB control (Fig. 5). The identity of ESS secretion machine components was revealed by immunoblotting with specific antibodies. As expected, EssB was detected in the eluate of the EssB_His affinity chromatography sample but not in the EssB control. Similarly, EsaA, EssA (152-residue membrane protein, M_r of 17,392), EssC (1,479-residue SpoIIIE-FtsK-like ATPase of the ESS pathway, M_r of 170,930), EsxA, and EssD (614-residue nuclease effector, M_r of 68,318) all coeluted with EssB_His. As a control, sortase A (SrtA), the DDM-solubilized transpeptidase, did not coelute with EssB_His following Ni-NTA affinity chromatography. Thus, Ni-NTA affinity chromatography of DDM-solubilized membrane proteins reveals the incorporation of EssB_His into assemblies of the ESS secretion machinery encompassing at least six different polypeptides: EssB, EsaA, EssA, EssC, EsxA, and EssD.

FIG 5.

Affinity chromatography of EssB_His from DDM-solubilized staphylococcal membranes. S. aureus USA300 LAC ΔessB(pessB) (lanes 1) and ΔessB(pessB_His) (lanes 2) variants were grown on tryptic soy agar supplemented with 0.2% horse serum, and US2 samples were prepared via DDM extraction of staphylococcal membranes and ultracentrifugation. DDM-solubilized membrane proteins (∼10 mg protein in US2 samples) were subjected to FPLC-operated Ni-NTA affinity chromatography at 4°C in the presence of 0.03% DDM. Columns were washed extensively with buffer containing 50 mM imidazole prior to elution with 250 mM imidazole. Eluate samples were separated by 15% SDS-PAGE and proteins stained with Coomassie blue. Eluate samples were also subjected to immunoblotting with polyclonal rabbit antibodies anti-EssB, anti-EsaA, anti-EssA, anti-EssC, anti-EsxA, anti-EssD, and anti-SrtA. Numbers indicate the migration of molecular mass markers (in kilodaltons).

Bacterial two-hybrid analysis of protein interactions in the staphylococcal ESS pathway.

We wondered whether ESS machine components interact with other factors of this pathway to ensure their functional assemblies. To test this, we generated hybrids between ESS pathway genes and fragments of adenylate cyclase, monitoring protein interactions as increased adenylate cyclase activity and expression of β-galactosidase activity on agar plates with 5-bromo-4-chloro-indolyl-β-d-galactopyranoside (X-Gal) as an indicator or in liquid cultures with o-nitrophenyl-β-d-galactopyranoside (ONPG) substrate (33, 34). This experimental plan revealed productive interactions between EssB:EssB, EssB:EsaA, and EsxA:EsxA, results that are consistent with biochemical studies reporting homotypic interactions among EsxA and EssB. Strains used for analyzing EsxA:EssB, EsxA:EsaA, and EsaA:EsaA hybrids yielded either little to no adenylate cyclase activity, as monitored on X-Gal agar plates, or ONPG β-galactosidase activity in culture (Fig. 6A and B). Taken together, these data indicate that EssB self-assembles into a homopolymeric structure that also recruits EsaA.

FIG 6.

Interaction between ESS components examined with bacterial two-hybrid analysis. (A) Fusions of recombinant esaA, essB, or esxA with the fragmented adenylate cyclase gene were cloned into pKT25 or pUT18C plasmid and transformed into E. coli strain BTH101, and recombinant plasmids were selected on agar containing kanamycin and ampicillin. Bacterial cultures probing homotypic (EssB:EssB, EsaA:EsaA, and EsxA:EsxA) as well as heterotypic (EssB:EsaA, EssB:EsxA, and EsaA:EsxA) interactions were spotted on LB agar containing X-Gal and incubated at 37°C for 24 h. White color indicates failure to reconstitute active adenylate cyclase, as shown with the control strain transformed with empty pKT25 and pUT18C plasmids (p25:p18). Blue color indicates X-Gal hydrolysis and efficient complementation between active hybrids as shown with the control strain transformed with plasmids pKT25-zip and pUT18C-zip (Zip:Zip). (B) E. coli cultures described for panel A were grown to an A₆₀₀ of 0.5 to 0.7, bacteria were lysed, and β-galactosidase activity was measured via the addition of o-nitrophenyl-β-d-galactopyranoside. Production of o-nitrophenol was recorded in a spectrophotometer at 420 nm and calculated in Miller units to account for variations in cell density between samples. Data represent averages from at least three independent experiments. E. coli expressing Zip:Zip plasmids was used as a positive control.

DISCUSSION

Type VII secretion systems (T7SS) support the pathogenic strategies of several Gram-positive microbes, including B. anthracis, M. tuberculosis, and S. aureus. Genes encoding SpoIIIE-FtsK-like ATPases and WXG100-secreted products are the only conserved features of T7SS; however, each pathway encompasses a cluster of 20 to 24 genes whose functions are largely unknown (24). We have studied the T7bSS of S. aureus, designated ESS, which is required for the establishment of persistent abscess lesions and the modulation of immune responses during staphylococcal bloodstream infections (5–8). The identity of the ESS secretion machinery and its biochemical attributes are unknown. Earlier work showed that deletion of essB abolishes S. aureus ESS secretion and that truncating mutations in essB display a dominant-negative phenotype (5, 26). We therefore asked whether or not EssB functions as a central assembly component of the T7bSS pathway in S. aureus and may be exploited to reveal the identity of the ESS machinery.

Our biochemical and genetic studies revealed that EssB assembles into a polymeric structure and associates directly with EsaA, a polytopic membrane protein of the ESS pathway. Large EssB:EsaA complexes could be isolated from DDM-extracted staphylococcal membranes either with or without additional components of the ESS pathway, including EssD and EsxA. When probed by affinity chromatography of fully functional, six-histidyl-tagged EssB_His, complexes comprising six different proteins could be isolated: EsaA, EssA, EssB, EssC, EssD, and EsxA. We note that some of these proteins, specifically EssC and EssD, were in part degraded, which precluded determination of stoichiometric relationships between individual ESS machinery components. It is interesting that substrates that travel the ESS pathway, mobile elements such as EsxA and EssD, can be found associated with membrane complexes. Perhaps substrates accumulate at the membrane until the machinery is complete or environmental cues are perceived for the opening of the secretory pore. Alternatively, secreted substrates may be a functional part of the machinery, suggesting a self-secretion process. We interpret the isolation of complexes between the conserved T7SS SpoIIIE-FtsK-like machine component (EssC), the conserved WXG100 protein (EsxA), three membrane proteins (EsaA, EssA, and EssB), and the T7bSS effector EssD to mean that DDM extraction solubilized functional ESS complexes assembled by ATPase (EssC) engagement of substrates (EsxA and EssD) and of the rudimentary ESS machine (EsaA:EssB).

Recent work demonstrated the isolation of the ESX-5 T7aSS machine from Mycobacterium xenopi using a streptavidin tag (Strep-tag II) at the carboxy terminus of EccC₅ (the functional equivalent of EssC in S. aureus) and Strep-Tactin affinity chromatography of DDM-solubilized membrane proteins (35). Isolated T7aSS complexes encompassed four proteins, EccB₅, EccC₅, EccD₅, and EccE₅, revealing structural assemblies with 6-fold rotational symmetry and a central cavity (35). With the sole exception of EccC₅, the T7aSS ATPase of ESX-5, the other membrane proteins of ESX-5 (EccB₅, EccD₅, and EccE₅) do not display sequence homology to S. aureus ESS components. Nonetheless, we propose that EsaA, EssA, EssB, and EssC represent functional homologs of mycobacterial EccB₅, EccC₅, EccD₅, and EccE₅ in S. aureus. In this regard, it will be interesting to determine both the stoichiometry and the structure of DDM-solubilized EsaA, EssA, EssB, EssC, EssD, and EsxA complexes.

Bioinformatic analysis revealed that the S. aureus ESS pathway, while expressed in all clinical isolates, represents a variable trait that is associated with increases in staphylococcal virulence (7, 24). The variant genetic traits are located in module 2, including 3′ sequences of essC and genes representing the effectors of the ESS pathway, for example, essD (23, 25) (Fig. 1A). Variant genetic traits are also found in module 3, which encompasses genes whose products fulfill chaperone functions for EssD or that prohibit nuclease activity of the effector in the bacterial cytoplasm (EssI and its homologs) (Fig. 1A) (8, 23, 25). Of note, genes whose products are proposed to represent the ESS secretion machinery (EsaA, EssA, EssB, and the N-terminal domains of EssC) are conserved among the ESS gene clusters of all S. aureus clinical isolates, in agreement with their proposed catalytic function.

MATERIALS AND METHODS

Bacterial cultures.

S. aureus cultures were grown in tryptic soy broth (TSB) or agar (TSA). Chloramphenicol was added to a final concentration of 20 μg/ml for plasmid selection. Anhydrotetracycline was used at 50 ng/ml for pKOR1-mediated allelic replacements of target genes. To assay for protein production and secretion, staphylococcal cultures were grown with vigorous shaking at 37°C in TSB supplemented with 0.2% horse serum until the desired density was reached. Escherichia coli cultures were grown in Luria-Bertani medium at 37°C. For plasmid selection in E. coli, ampicillin and kanamycin were used at 100 and 50 μg/ml, respectively. Where indicated, 0.5 mM isopropyl-β-d-1-thiogalactopyranoside (IPTG) and 40 μg/ml 5-bromo-4-chloro-indolyl-β-d-galactopyranoside (X-Gal) were used to monitor β-galactosidase (LacZ) activity.

Bacterial strains and plasmids.

S. aureus RN4220 was used to passage plasmid DNA. USA300 LAC, a clone of the American community-acquired, methicillin-resistant S. aureus (CA-MRSA) epidemic (36), was used as the wild-type S. aureus strain. E. coli DH5α and JM109 were used for cloning experiments. E. coli BTH101 (Euromedex, France) was used for the bacterial two-hybrid study. The USA300 LAC variant lacking essB (ΔessB) or carrying the complementing plasmid pessB (ΔessB/pessB) has been described earlier (26). Like pessB, plasmid pessB_His is a derivative of pWWW412, whereby essB_His is cloned under the control of the constitutive hprK promoter (37) to produce a variant of EssB with a C-terminal 6-histidyl tag. Plasmid pessB_His was assembled by cloning a PCR product amplified with primer sequences (5′-AACTCGAGATGGTTAAAAATCATAACCCTAAAAATG-3′ and 5′-AAAAGGATCCCTAGTGATGGTGATGGTGATGTTTTTTTCTTTCAGCTTCTTG-3′). The USA300 ΔesaA variant was generated by allelic replacement with plasmid pKOR1 to fuse the first 15 and last 15 codons of esaA. Briefly, two 1-kbp DNA sequences flanking esaA were amplified from USA300 template DNA with primer pairs with sequences 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTCAGGAGGTTTCTAGTTATGGCAATG-3′ and 5′-AAAGATCTACTGAACATGTTTATAAAACAC-3′ (amplification of upstream fragment) as well as 5′-AAAGATCTTAATTAAAGTGACAATTAATGCATAAATCC-3′ and 5′-GGGGACCACTTTGTACAAGAAAGCTGGGTGGTGTATGATTGTCATTAATGTC-3′ (amplification of downstream fragment). Ligated PCR products were cloned into pKOR1 using the Gateway BP Clonase II enzyme from Invitrogen (no. 11789-020). Procedures for cloning and allelic replacement were described earlier (38). The complementing plasmid pesaA was assembled by PCR amplification of esaA coding sequence using USA300 template and primers (5′-AACATATGAAAAAGAAAAATTGGATTTATGC-3′ and 5′-AACTCGAGTTAGATTAATCTCTCTTTCTTAAAGTG-3′). The PCR product was cloned in plasmid pWWW412, placing esaA under the transcriptional control of the constitutive hprK promoter (37). For bacterial two-hybrid analyses, the full-length genes esaA, essB, and esxA were amplified using the genomic DNA of strain USA300 and primer pairs with the following sequences: esaA, 5′-CGCAGTCTAGACATGAAAAAGAAAAATTGGATTTATGCA-3′ and 5′-CGCAGGGTACCTTAGATTAATCTCTCTTTCTTAAAGTGT-3′; essB, 5′-CGCAGTCTAGACATGGTTAAAAATCATAACCCTAAAAAT-3′ and 5′-CGCAGGGTACCCTATTTTTTTCTTTCAGCTTCTTGGCGT-3′; and esxA, 5′-CGCAGTCTAGACATGGCAATGATTAAGATGAGTCCAGAG-3′ and 5′-CGCAGGGTACCTTATTGCAAACCGAAATTATTAGAAAGT-3′. PCR products were cloned into the pCRII-Blunt-TOPO vector (Invitrogen) and verified by DNA sequencing prior to subcloning into plasmids pKT25 and pUT18C (Euromedex, France), resulting in plasmids pT18esaA, pT25esaA, pT18essB, pT25essB, pT18esxA, and pT25esxA.

Extraction and purification of detergent-solubilized proteins from S. aureus.

Membrane isolation and detergent extraction experiments were conducted as previously described, with some modifications (39–41). Briefly, overnight cultures of S. aureus were diluted in TSB to a final absorbance at 600 nm (A₆₀₀) of 0.1 and plated on TSA supplemented with 0.2% horse serum and antibiotics where necessary, and plates were incubated at 37°C for 16 h. Cells were scraped off agar plates, suspended, and washed twice in 50 mM sodium potassium phosphate (NaKP) buffer (pH 5.5) and lysed in 20 ml NaKP buffer containing lysostaphin (200 μg/ml). Lysates were cleared by centrifugation at 5,000 × g for 15 min, and supernatant fractions were subjected to ultracentrifugation at 100,000 × g for 2 h at 4°C, yielding the ultracentrifugation supernatant 1 (USP1) and pellet 1 (UP1) fractions, respectively. Fraction UP1 was suspended in 1 ml of 750 mM 6-amino-caproic acid, 50 mM Bis-Tris (ACA) buffer (pH 7.0) containing 2% n-dodecyl-β-d-maltoside (DDM), and incubated for 2 h at 4°C. Detergent-solubilized proteins were recovered as the supernatant fraction 2 (US2) following ultracentrifugation at 100,000 × g for 2 h at 4°C. For size exclusion chromatography, 1 ml of US2 samples was loaded on an ÄKTA purifier-operated HiLoad 16/60 Superdex 200 preparation-grade (pg) column (GE Healthcare) at a flow rate of 0.5 ml min⁻¹, and 0.5-ml fractions were collected. Chromatograms of eluted proteins were recorded as milli-absorbance units (mAU) at 215 nm, and elution volumes were calibrated with reference proteins, including thyroglobulin (669 kDa), ferritin (440 kDa), catalase (232 kDa), lactate dehydrogenase (140 kDa), and bovine serum albumin (67 kDa) (GE Healthcare). For affinity chromatography, US2 samples containing EssB_His were passed over a fast protein liquid chromatography (FPLC)-operated 5-ml HisTrap HP nickel affinity column (GE Healthcare) at 4°C. The column was washed with buffer A (20 mM imidazole in 750 mM ACA buffer [pH 7.0], 0.03% DDM) followed by gradual elution using buffer B (500 mM imidazole in ACA buffer [pH 7.0], 0.03% DDM). EssB_His eluted at 275 mM imidazole in a single fraction. When necessary, eluted fractions were concentrated using Amicon protein concentrators (Millipore) with a 30-kDa molecular mass cutoff. Proteins in the eluate were separated by SDS-PAGE and identified by mass spectrometry or electrotransferred to a polyvinylidene difluoride (PVDF) membrane for immunoblot analyses using specific polyclonal antibodies. Immunoreactive products were revealed by chemiluminescent detection after incubation with an anti-rabbit horseradish peroxidase-conjugated secondary antibody (Cell Signaling Technology).

Protein secretion assay.

To assess protein secretion, S. aureus cultures were grown to an A₆₀₀ of 1.0 and centrifuged (9,000 × g for 10 min), and the extracellular medium (MD) containing secreted proteins was transferred to a new tube. Bacteria in the sediment were washed once and subsequently lysed in 50 ml 50 mM Tris-HCl (pH 7.0) containing 20 μg/ml lysostaphin at 37°C for 60 min to yield cell lysates. Proteins in each fraction were precipitated with a final concentration of 10% trichloroacetic acid (TCA) and solubilized in sample buffer (0.5 M Tris-HCl [pH 8.0], 4% SDS, 10% glycerol, 5% 2-mercaptoethanol, 0.01% bromophenol blue) for separation by SDS-PAGE followed by electrotransfer to a PVDF membrane for immunoblot analyses.

Protein interactions examined with the bacterial two-hybrid system.

Genes esaA, essB, and esxA cloned in pKT25 or pUT18C were transformed into E. coli strain BTH101 on agar containing kanamycin and ampicillin. For control experiments, BTH101 was transformed with pKT25 and pUT18C (negative control) and the plasmid pair pKT25-zip and pUT18C-zip (Euromedex, France), in which the leucine zipper of GCN4 is genetically fused to the complementary T25 and T18 fragments of Bordetella pertussis adenylate cyclase (positive control). Isolated colonies were used to inoculate LB medium supplemented with kanamycin and ampicillin and grown for 16 h at 37°C. Cultures were diluted to an A₆₀₀ of 0.5, and 10 μl was plated on solid medium supplemented with kanamycin, ampicillin, IPTG, and X-Gal. The plates were incubated at 37°C for 24 h to visualize color changes from white to blue as a measure of protein-protein interaction. Quantitative measurement of β-galactosidase activity was performed using bacterial cultures and the synthetic compound ONPG as described by Miller (33).