|Home | About | Journals | Submit | Contact Us | Français|
The type III secretion system (T3SS) encoded by Salmonella pathogenicity island 2 (SPI-2) is involved in systemic infection and intracellular replication of Salmonella enterica serovar Typhimurium. In this study, we investigated the function of SsaE, a small cytoplasmic protein encoded within the SPI-2 locus, which shows structural similarity to the T3SS class V chaperones. An S. enterica serovar Typhimurium ssaE mutant failed to secrete SPI-2 translocator SseB and SPI-2-dependent effector PipB proteins. Coimmunoprecipitation and mass spectrometry analyses using an SsaE-FLAG fusion protein indicated that SsaE interacts with SseB and a putative T3SS-associated ATPase, SsaN. A series of deleted and point-mutated SsaE-FLAG fusion proteins revealed that the C-terminal coiled-coil domain of SsaE is critical for protein-protein interactions. Although SseA was reported to be a chaperone for SseB and to be required for its secretion and stability in the bacterial cytoplasm, an sseA deletion mutant was able to secrete the SseB in vitro when plasmid-derived SseB was overexpressed. In contrast, ssaE mutant strains could not transport SseB extracellularly under the same assay conditions. In addition, an ssaE(I55G) point-mutated strain that expresses the SsaE derivative lacking the ability to form a C-terminal coiled-coil structure showed attenuated virulence comparable to that of an SPI-2 T3SS null mutant, suggesting that the coiled-coil interaction of SsaE is absolutely essential for the functional SPI-2 T3SS and for Salmonella virulence. Based on these findings, we propose that SsaE recognizes translocator SseB and controls its secretion via SPI-2 type III secretion machinery.
A number of gram-negative pathogenic bacteria use a type III secretion system (T3SS) to interact with eukaryotic host cells. T3SS delivers bacterial effectors through the needle-like structure extending across the inner and outer membranes of the bacterium and into the cytosol of eukaryotic cells (28). Salmonella enterica serovar Typhimurium is an enteropathogenic bacterium that causes gastroenteritis in humans and typhoid-like fever in mice. Salmonella possesses two different T3SSs encoded by Salmonella pathogenicity island 1 (SPI-1) and SPI-2. Upon entry into host cells, S. enterica serovar Typhimurium resides in a special membrane-bound compartment termed the Salmonella-containing vacuole (23). Expression of SPI-2 is induced within the vacuole (8) and is essential for intracellular replication and virulence of S. enterica serovar Typhimurium (26, 44).
Functional SPI-2 genes are clustered within six large transcriptional units. Thirty-one potential open reading frames on the SPI-2 region encode proteins that are directly involved in the assembly and regulation of the T3SS (50). The Ssa proteins are involved in the assembly of the syringe-like type III secretion injectisome, the so-called nanomachine (41). A set of nine Ssa proteins conserved among T3SSs forms the injectisome core. Transport of some effectors through the injectisome is facilitated by formation of a complex between an effector and a chaperone encoded in SPI-2. For example, SscB, a protein encoded immediately upstream of sseF, acts as a chaperone for SseF (14). In addition to effectors, the translocators SseB (a component of the oligomeric filament structure of the type III secretion apparatus), SseC, and SseD (the pore-forming translocator complex) are also secreted through the SPI-2-encoded type III secretion machinery. Efficient secretion of SseC and SseD is required for the presence of SseB, but SseB secretion is independent of these two translocators (31, 42). Furthermore, stable expression of these translocators requires the chaperone SseA for SseB and SseD and a putative chaperone SscA for SseC (49, 62, 63).
Several different types of chaperones have been classified on the basis of structural and functional analyses (57). Class I chaperones are small, acidic (pI ~4 to 5), usually dimeric proteins that interact with a single T3SS effector or with two or three effectors and mediate targeting of their cognate binding partners to the type III secretion apparatus (35). Class II chaperones bind to translocator proteins of the T3SS. They possess tetratricopeptide repeats, which form a curved layer of α-helices. These chaperones play a regulatory role in type III secretion (6, 20). Class III is represented by flagellar chaperones (such as FliS), which have an entirely different structure. FliS prevents polymerization of flagellin protein FliC in the cytoplasm prior to secretion (19). The CesA protein of enteropathogenic Escherichia coli (EPEC), which acts as a chaperone for EspA, is structurally different from other classes of chaperones. Therefore, it has been designated a class IV chaperone. The aggregation of EspA in the bacterial cytosol is prevented through binding with CesA (59). Class V chaperones are a heterogeneous group of proteins that interact with the needle subunit of the type III secretion apparatus. These include Yersinia YscE and YscG and Pseudomonas aeruginosa PscE and PscG. Polymerization of needle components Yersinia YscF and P. aeruginosa PscF in the bacterial cytoplasm is prevented by binding to these chaperones, YscE/YscG and PscE/PscG, respectively (47).
To date, some proteins encoded within SPI-2 remain uncharacterized. One of these is SsaE. The ssaE open reading frame of 243 bp is predicted to encode a protein of 80 amino acids containing three predicted α-helical regions. SsaE has homology to EscE (Orf2) of EPEC and Citrobacter rodentium, YscE of Yersinia pestis, PscE of P. aeruginosa, and AscE of Aeromonas hydrophila. Yersinia yscE mutants are unable to export their effectors (16). In P. aeruginosa, PscE is required for a functional T3SS and cytotoxicity by controlling the needle complex biogenesis of the T3SS (47). While the core type III components are generally conserved, allowing the function and/or cellular locations of some of the proteins to be inferred (4, 28), these small proteins are not broadly conserved among other T3SSs or flagellar export systems. For example, no homologue of PscG is present in S. enterica serovar Typhimurium.
In this study, we have characterized the role of SsaE in secretion and translocation via SPI-2 T3SS. A Salmonella mutant that lacks SsaE failed to secrete SseB (a translocator) and PipB (an SPI-2 effector). Using pull-down assays, we showed that SsaE directly interacts with SseB and a putative ATPase, SsaN. Furthermore, deletion and site-directed mutagenesis of SsaE identified a C-terminal coiled-coil domain involved in protein-protein interactions of SseB. We finally found that Salmonella expressing the point-mutated SsaE(I55G), which is unable to form a C-terminal coiled-coil domain, has dramatic defects in virulence, comparable to those of the SPI-2 null mutant. These data suggest that the chaperone-like small molecule SsaE plays a crucial role in SPI-2 secretion by interacting with SseB via the C-terminal coiled-coil domain.
The Salmonella strains and plasmids used in this study are listed in Table Table1.1. In this study, S. enterica serovar Typhimurium strain SL1344 was used as the wild-type strain, and strains harboring an ssaV mutation and an spiA(C133S) point mutation were used as general SPI-2 T3SS component mutants. E. coli strains DH5α (Gibco BRL) and MC1061 (7) were used for molecular cloning and expression of recombinant proteins. E. coli strain S17.1λpir (38) was used for propagation of π-dependent plasmids and for conjugation. Bacteria were routinely grown in LB broth (Sigma) at 37°C overnight with aeration. Ampicillin (100 μg/ml), chloramphenicol (25 μg/ml), kanamycin (25 μg/ml), and streptomycin (25 μg/ml) were used when required.
Deletion mutants of strain SL1344 were constructed using a λ Red disruption system (15). S. enterica serovar Typhimurium wild-type strain SL1344 derivatives with chromosomally encoded FLAG and tandem hemagglutinin (HA) fusion proteins were constructed using the integrational plasmid pLDΩKm2 by conjugation (37) and a λ Red disruption system (15), respectively. Double mutant strains were created by phage P22-mediated transduction. To consider the effect of downstream gene expression, a Salmonella strain containing an SsaE initiation codon mutation [ssaE(mut)] and a strain expressing a point-mutated SsaE(I55G) were constructed by allele exchange using the temperature- and sucrose-sensitive suicide vector pCACTUS containing ssaE(mut) or ssaE(I55G) as described previously (37). All constructs were verified by PCR and DNA sequencing.
For construction of the complementing plasmids pSsaE and pACYC-SseB, ssaE and sseB genes were amplified from the genomic DNA of strain SL1344 with primers ssaE-SacI-FW and ssaE-SphI-RV, and sseB-FW and sseB-RV, and were ligated into pMW118 and pACYC184, respectively.
To construct the plasmid encoding the N-terminal glutathione S-transferase (GST)-tagged SseB fusion protein, sseB genes were amplified from the genomic DNA of strain SL1344 with primers sseB-FW-BamHI and sseB-XhoI-RV and cloned into pGEX-6P-1 (GE Healthcare).
To construct the plasmids encoding the C-terminal FLAG-tagged fusion proteins, the DNA fragments containing the respective genes were amplified by PCR with specific primers and cloned into pFLAG-CTC (Sigma). To obtain the plasmids encoding truncated SsaE proteins fused with FLAG, inverse PCRs were performed with primers ssaE-StuI-R3 and ssaE-StuI-R4 (for pSsaEΔ1-FLAG), ssaE-StuI-R1 and ssaE-StuI-R2 (for pSsaEΔ2-FLAG), or ssaE-StuI-R7 and ssaE-StuI-R8 (for pSsaEΔ3-FLAG). The point mutation on pSsaE-FLAG was created by a QuikChange site-directed mutagenesis kit (Stratagene) using the primers sense-ssaEI55G and antisense-ssaEI55G to replace isoleucine with glycine at position 55 in SsaE. The mutation was confirmed by DNA sequencing.
To construct the plasmids encoding the C-terminal 2HA-tagged fusion protein, plasmids p2HA-CTC, pACHS-2HA, and pACPJ-2HA were used. For construction of p2HA-CTC, inverse PCR was performed using pFLAG-CTC circular DNA as the template with primers 2HA-StuI-1 and 2HA-StuI-2. For construction of pACHS-2HA, the DNA fragment containing a 2-HA-tag sequence amplified from p2HA-CTC with primers FLAG-SphI-FW and FLAG-BamHI-RV was cloned into pACYC184. To construct pACPJ-2HA that contains the gene encoding the 2-HA-tagged protein under the control of the sseJ promoter, DNA fragments containing the sseJ promoter region amplified from the genomic DNA of the SL1344 strain with primers ProsseJ-SalI-FW and ProsseJ-SphI-RV and a 2-HA-tag sequence amplified from p2HA-CTC with primers FLAG-SphI+ATG-FW and FLAG-BamHI-RV were cloned into pACYC184. To generate the plasmids encoding the C-terminal CyaA-2HA fusion protein, a DNA fragment encoding the catalytic domain of CyaA was amplified from pMS109 with primers cyaA-BglII-FW and cyaA-XhoI-RV and cloned into pACPJ-2HA. The primers used for the construction of all constructs are listed in Table Table22.
Rabbit anti-Salmonella O4 polyclonal antibody (Denka Seiken) was used at a dilution of 1:5,000. Mouse polyclonal anti-SseB and rabbit polyclonal anti-PagC antibodies were used as described previously (37, 43). The mouse monoclonal antibodies, anti-FLAG (1:20,000) (Sigma), anti-GST (1:2,000) (Upstate, Lake Placid, NY), and anti-DnaK (1:2,000) (Calbiochem) were used as primary antibodies for immunoblotting. The mouse monoclonal anti-HA epitope tag HA.11 (Covance) was used at a dilution of 1:2,000 for immunofluorescence microscopy and immunoblot analysis. Alexa 488-conjugated goat anti-mouse immunoglobulin G (IgG) and Alexa 594-conjugated goat anti-rabbit IgG secondary antibodies (dilution of 1:500) were obtained from Molecular Probes. Alkaline phosphatase-conjugated goat anti-mouse and anti-rabbit IgG antibodies were purchased from Sigma and were used at a dilution of 1:10,000.
HeLa cells were grown in minimal essential medium (Sigma) supplemented with 10% fetal bovine serum and were cultured in the presence of gentamicin (100 μg/ml) and kanamycin (60 μg/ml) at 37°C in a 5% CO2 atmosphere. Bacterial infections of HeLa cells were performed as described previously (37). For immunofluorescence labeling, cells were fixed, permeabilized, and probed with various antibodies as described previously (37). Labeled cells were analyzed by a Zeiss confocal laser scanning microscope (LSM510 Meta).
Translocation of a CyaA fusion protein by SPI-2 T3SS into infected host cells was measured using a cyclic AMP (cAMP) enzyme immunoassay system (Amersham Biosciences) to quantify intracellular levels of cAMP. HeLa cells were infected with Salmonella strains carrying the plasmids encoding CyaA fusion proteins for 22 h. After infection, cells were lysed and processed according to the manufacturer's instructions. The data are the means ± standard deviations from triplicate determinants.
Salmonella strains grown in minimal medium low in phosphate and magnesium (LPM) (pH 5.8) were fractionated. The culture supernatant fraction was obtained as described previously (10). Bacterial fractionation was based on the method described by Gauthier et al. (24). Bacteria grown in 100 ml of LPM (pH 5.8) for 16 h at 37°C were used for the isolation of cytoplasmic, periplasmic, and membrane fractions. Samples were run on 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels and transferred to polyvinylidene difluoride membranes (Immobilon; Millipore). The gels were hybridized with various antibodies and developed using a Sigma Fast 5-bromo-4-chloro-3-indolyl phosphate (BCIP)/Nitro Blue Tetrazolium detection system as described previously (37).
Immunoprecipitation of the FLAG-tagged protein complexes from Salmonella and E. coli lysates was performed by using FLAG beads conjugated with anti-FLAG M2 antibody (Sigma) as previously described (55). Lysates were clarified by centrifugation and incubated with FLAG beads for 2 h at 4°C. Beads were washed seven times with ice-cold phosphate-buffered saline (PBS) containing 1 mM phenylmethylsulfonyl fluoride (PBS-P), and bound proteins were competitively eluted from the beads using FLAG peptide (Sigma) at a final concentration of 90 μg/ml. The eluted protein fraction was then analyzed by SDS-PAGE and Western blotting.
GST and GST-SseB fusion proteins were purified by affinity chromatography with glutathione-Sepharose 4B beads according to the manufacturer's instructions (GE Healthcare). For GST pull-down assays, glutathione-Sepharose 4B beads immobilized with GST or GST-SseB were incubated with cleared extracts from the Salmonella ssaE(mut) mutant or E. coli DH5α strains expressing SsaE-FLAG for 2 h at 4°C. Beads were washed five times with ice-cold PBS-P, and the bound proteins were eluted from the beads using 10 mM glutathione. The eluted protein fraction was then analyzed by SDS-PAGE and Western blotting.
The S. enterica serovar Typhimurium strain TM312 that expresses SsaE-FLAG was resuspended in PBS containing lysozyme and protease inhibitor cocktails (Roche), disrupted by sonication, and cleared by centrifugation. The sample coprecipitated with SsaE-FLAG using FLAG beads was subjected to trypsin digestion. The digested peptides were analyzed by nano-liquid chromatography-electrospray ionization mass spectrometry (nano-LC-ESI-MS/MS) using a DiNa nano-LC system (KYA Technologies) with an L-column 2 octyldecyl silane (0.05 mm by 100 mm, 3 μm; CERI, Japan) coupled to a QStar Elite hybrid LC tandem MS (LC/MS/MS) system (Applied Biosystems). Peptide and protein identification were performed using Protein Pilot version 2.0 software (Applied Biosystems) with default parameters. Each MS/MS spectrum was searched for species of S. enterica serovar Typhimurium against the NCBI database.
Female BALB/c mice (5 to 6 weeks old) were used for the mixed infection assay and were housed at Kitasato University according to the standard Laboratory Animal Care Advisory Committee guidelines. At least three mice were inoculated intraperitoneally with a mixture of two strains comprising 5 × 104 CFU of each strain in physiological saline, and the number of viable bacteria from infected spleens was determined at 48 h after infection as described previously (37). The competitive index (CI) was calculated by the formula CI = output (mutant strain/wild-type strain)/inoculum (mutant strain/wild-type strain). In the case of two tested strains that have same virulence, the CI is 1.0. In contrast, in the case of the mutated gene that contributes to the virulence, the CI is <1.0. Each CI value is the mean from at least three independent infections ± standard deviation. CI data were analyzed by Mann-Whitney U test for statistical significance. P values of 0.05 or less were considered significant.
In the SPI-2 region of S. enterica serovar Typhimurium, ssaE is flanked by spiB (also referred to as ssaD) and sseA (Fig. (Fig.1A).1A). SsaE shares some characteristics with other type III chaperone proteins (reviewed in references 2, 11, and 56). SsaE has a predicted molecular mass of 9.4 kDa and a predicted pI of 4.9, both of which are common characteristics of T3SS chaperones. In addition, SsaE was predicted to have three α-helices by PSIPRED (30). Among them, α3 could be predicted as a perfect amphipathic helix in the C terminus. We searched for similarities between SsaE and other T3SS chaperones using SKE-CHIMERA (54) and the fold recognition algorithm server SPARKS2 (61) (Fig. (Fig.1B).1B). The model of the three-dimensional structure of the SsaE was built with the FAMS program (45) on the basis of this alignment using a Y. pestis T3SS-specific chaperone, YscE (Protein Data Bank identification no. 1ZW0), as template (Fig. 1C and D). These observations indicated that SsaE encodes a protein as the T3SS-specific chaperone.
The subcellular localization of SsaE was determined by fractionation of a Salmonella strain, followed by SDS-PAGE and immunoblotting. An S. enterica serovar Typhimurium strain encoded chromosomally by an ssaE-FLAG fusion gene was grown in LPM medium (pH 5.8), and the cell components were separated into cytoplasm, periplasm, and membrane fractions. SsaE (SsaE-FLAG) was not secreted and was localized to the cytoplasmic fraction (Fig. 2A and B). We further attempted to detect translocation of SsaE into the host cell cytosol by confocal immunofluorescence microscopy. As an HA epitope-tagged derivative of effector proteins has been used to detect SPI-2-dependent translocation in host cells (29, 32), a plasmid (pSsaE-2HA) expressing SsaE-2HA was generated and transformed into Salmonella wild-type SL1344. In HeLa cells infected with the wild-type strain expressing SsaE-2HA, SsaE was not detected in the host cell cytosol (data not shown). Furthermore, a lack of SsaE translocation into host cells was confirmed quantitatively by a cAMP assay using the wild-type strain expressing the SsaE-CyaA-2HA fusion protein. While SsaE-CyaA-2HA was determined to be stable in Salmonella (data not shown), background levels of cAMP were detected in cells infected with the wild-type strain harboring either pACPJ-CyaA-2HA (15.6 ± 0.5 fmol cAMP/μg protein) or pSsaE-CyaA-2HA (15.5 ± 0.6 fmol cAMP/μg protein). In the same assay, the SPI-2 effector SseF-CyaA-2HA fusion protein was used as a positive control for SPI-2 T3SS translocation, and intracellular cAMP was clearly elevated by SseF-CyaA-2HA of the wild-type strain (335.8 ± 107.6 fmol cAMP/μg protein) relative to levels for the SPI-2 T3SS component mutant (ΔssaV) harboring pSseF-CyaA-2HA (10.9 ± 3.0 fmol cAMP/μg protein). These results suggest that SsaE is a cytoplasmic protein but not an effector translocated by SPI-2 T3SS.
To investigate the role of SsaE in SPI-2 T3SS, we constructed a site-directed mutation in the initiation codon of ssaE [ssaE(mut); ATG to CTT] and then examined the ability of the ssaE(mut) mutant strain to secrete the translocator SseB into the culture supernatant. As shown in Fig. Fig.3A,3A, SseB in the Salmonella mutant strain carrying an ssaE(mut) mutation accumulated within bacterial cells but was not secreted into the supernatant when bacteria were grown under conditions that induced SseB expression and secretion. Complementation of the ssaE(mut) mutant with a wild-type ssaE allele restored the secretion of SseB to a level comparable to that in the wild-type strain. The mutation could be also complemented by ssaE-FLAG or ssaE-2HA introduced on a plasmid (data not shown). These results strongly suggest that SsaE is required for secretion of SseB. In addition, Salmonella strains containing sseAB::lacZ transcriptional fusion on the chromosome in both ssaE+ and ssaE mutant [ssaE(mut)] backgrounds were grown under SPI-2-inducing conditions, and β-galactosidase activity was measured. Consistent with the above results, transcription of sseAB was not affected by loss of SsaE (data not shown).
Recently, it has been demonstrated that the SPI-2-encoded regulatory molecules SsaL, SsaM, and SpiC, the latter two of which occur in complex, are required for translocator protein secretion (10, 60). However, mutations of ssaL and ssaM resulted in enhanced secretion of SPI-2 effectors encoded outside of SPI-2 (10, 60). Therefore, to further determine the role of SsaE in the secretion of SPI-2 effectors, we constructed ssaL mutants with ssaE+ and ssaE mutant backgrounds and tested the secretion of PipB. As expected, no detectable SseB was secreted by any of the Salmonella mutants tested. In contrast, the PipB-2HA fusion protein was detected in secreted protein fractions from the ΔssaL single mutant but not the double mutants carrying mutations in ΔssaL and ssaE(mut) or ΔssaL and spiA(C133S), a point mutation in the outer membrane secretin SpiA (37) (Fig. (Fig.3B),3B), suggesting that SsaE is required for secretion of PipB in wild-type cells and in cells lacking SsaL. Similar results were obtained from other SPI-2 effectors, SopD2 (SopD2-2HA) and SspH2 (SspH2-2HA), or when the Salmonella ΔssaM mutant was used (data not shown). Thus, phenotypically, a mutant lacking SsaE is different from a mutant lacking SsaL or SsaM.
Several physical features and predicted structural characteristics of SsaE indicated that SsaE functions as a chaperone. Therefore, to identify the SsaE partner proteins encoded by SPI-2 in Salmonella, we next performed coimmunoprecipitation assays using lysates prepared from the wild-type strain expressing SsaE-FLAG grown under SPI-2-inducing conditions. The bound proteins were eluted and subjected to MS analyses. Seven SPI-2 proteins, i.e., SsaH, SsaK, SsaN, SsaQ, SseA, SseB, and SseC, were identified as SsaE-interacting proteins. In general, translocator proteins of T3SSs are known to require chaperones before secretion by the T3SS pathway. Thus, we further characterized the SsaE interaction with SPI-2 T3SS translocator proteins SseB and SseC.
To examine the binding specificity of SsaE with SseB, we performed FLAG pull-down assays using lysates prepared from the wild-type strain SL1344 carrying a plasmid (pSsaE-FLAG) expressing SsaE-FLAG. The bound protein was analyzed by SDS-PAGE and immunoblotting with mouse anti-SseB antibody. Controls for binding specificity included binding to bacterial alkaline phosphatase (BAP)-FLAG and blotting with anti-PagC antibody, as PagC is an SPI-2-unrelated outer membrane protein. The wild-type strain harboring a plasmid (pSsaE) expressing untagged SsaE was also used as a control for nonspecific binding to anti-FLAG beads. The results showed that SseB was coprecipitated by SsaE-FLAG but not BAP-FLAG or untagged SsaE (Fig. 4A and B and data not shown). In contrast, PagC was not coprecipitated by either SsaE-FLAG, untagged SsaE, or BAP-FLAG, suggesting that SsaE is capable of interacting with SseB (Fig. 4A and B and data not shown). This interaction was also confirmed by reverse experiments using Salmonella extracts from the wild-type strain expressing SseB-FLAG and SsaE-2HA grown under the same conditions (data not shown).
Next, to clarify whether the interaction between SsaE and SseB is direct or requires other SPI-2 proteins bridging two proteins, we performed GST pull-down assays using SsaE-FLAG expressed in E. coli DH5α that does not encode the Salmonella SPI-2 proteins. Cell lysate was incubated with glutathione-Sepharose beads conjugated with GST-SseB or GST, and the bound proteins were analyzed by SDS-PAGE and immunoblotting with anti-FLAG antibody. SsaE-FLAG was coeluted with GST-SseB, whereas GST alone did not bind to the SsaE-FLAG (Fig. (Fig.4C).4C). These results strongly suggest that SsaE interacts with SseB and that none of the additional proteins encoded by SPI-2 are required for this interaction.
Under the same assay conditions, when SsaE-FLAG and SseC-2HA were coexpressed from plasmids (pSsaE-FLAG and p2HA-SseC) in S. enterica serovar Typhimurium, SseC-2HA coimmunoprecipitated with SsaE-FLAG onto FLAG beads (data not shown). However, SseC-2HA did not coelute with SsaE-FLAG-containing beads when SseC-2HA was expressed in E. coli (data not shown), indicating that the interaction between SsaE and SseC is indirect binding.
SsaE is predicted to have three α-helical regions (the amino acids spanning regions 5 to 11, 24 to 32, and 52 to 68) (Fig. 1B and C). Thus, to identify the SsaE domain involved in interaction with partner proteins, we constructed a series of SsaE-FLAG derivatives and subjected the resulting proteins, pSsaEΔ1-FLAG (SsaE(21-80)-FLAG), pSsaEΔ2-FLAG (SsaE(Δ21-41)-FLAG), and pSsaEΔ3-FLAG (SsaE(Δ52-68)-FLAG), to FLAG pull-down assays using bacterial lysates prepared from the Salmonella wild-type strain (Fig. (Fig.5A).5A). Similar to the full-length SsaE-FLAG, SseB was precipitated with SsaEΔ1-FLAG and SsaEΔ2-FLAG, but not with SsaEΔ3-FLAG (Fig. (Fig.5A),5A), showing that SseB interacts with the C-terminal portion of SsaE encompassing amino acid residues 52 to 68.
Coiled-coil sequences in proteins consist of seven residues, termed the heptad repeat (a-b-c-d-e-f-g)n, which includes the nonpolar hydrophobic residues at positions a and d (25). Residues in positions a and d typically form a core in the center of the α-helical bundle, providing the driving force for protein-protein interaction (25, 36). The Ile-55 of the SsaE protein within the C-terminal coiled-coil sequence was determined to be a nonpolar hydrophobic residue at position a by the COILS program (http://www.ch.embnet.org/software/COILS_form.html), which is a tool for predicting coiled-coil domains (Fig. (Fig.5B).5B). Thus, to further confirm the importance of the C-terminal coiled-coil domain of SsaE for protein-protein interaction, site-directed mutagenesis was performed by using the plasmid pSsaE-FLAG as a template and replacing the Ile-55 with glycine. While the resulting SsaE(I55G)-FLAG protein lacking the C-terminal coiled-coil structure was stable in Salmonella, a replacement of Ile-55 with glycine in SsaE resulted in a loss of the ability to interact with SseB (Fig. (Fig.5A5A).
Next, to investigate the functional importance of the C-terminal coiled-coil domain of SsaE, we constructed a Salmonella mutant strain expressing SsaE(I55G) (chromosome mutant) and examined its SseB secretion. This mutant strain was unable to secrete SseB into the supernatant in LPM medium at pH 5.8, and the SseB secretion could be restored by introduction of the wild-type ssaE gene on the plasmid into the SsaE(I55G)-expressing Salmonella strain (Fig. (Fig.6A).6A). These results suggest that the interaction between SseB and SsaE is essential for SseB secretion. In addition, we further examined the secretion of SPI-2 effector PipB in an ssaE(I55G) mutant strain. Similar to SseB, less PipB-2HA fusion protein was secreted into the supernatant from the ssaE(I55G) mutant strain (Fig. (Fig.6B),6B), suggesting that the C-terminal coiled-coil domain of SsaE is important for secretion of the SPI-2 effectors.
It has been reported that SseA functions as a chaperone for SseB and is required for stabilization of SseB in the bacterial cytosol and for SseB export to the bacterial surface (9, 49, 62, 63). Therefore, to demonstrate the presence of the SseA-SseB-SsaE protein complex in the Salmonella cytoplasm, the immunoprecipitation by FLAG pull-down assays using lysates from the Salmonella strain expressing SseA-2HA (chromosomal mutation) harboring either pSsaE-FLAG or pBAP-FLAG were carried out. The pulled-down proteins were detected by immunoblotting with anti-FLAG (for SsaE-FLAG and BAP-FLAG), anti-HA (for SseA-2HA), and anti-SseB antibodies. As shown in Fig. Fig.6C,6C, SseA-2HA and SseB were coprecipitated by SsaE-FLAG but not BAP-FLAG (data not shown). Thus, both SseA and SsaE were able to associate with SseB in the Salmonella cytoplasm.
We next confirmed that SseA plays a role in SseB secretion by using a Salmonella sseA mutant strain. As previously reported, the sseA mutant strain exhibited a dramatic decrease in the intracellular accumulation of SseB compared to that of the wild-type strain, and secreted SseB was not detectable in the supernatant fraction by Western blot analysis (data not shown). To rule out the possibility that reduced intracellular production of SseB was responsible for the defective secretion in the sseA mutant, we constructed a plasmid (pACYC-SseB) expressing SseB constitutively and introduced it into the wild-type strain and mutant strains deficient in sseA and ssaE genes. In addition, Salmonella strains that lack SseB and SsaV harboring pACYC-SseB were used as positive and negative controls for SseB secretion. SseB was found in the cytoplasmic fraction at levels similar to those for the wild-type strain when mutant strains harbored a plasmid expressing SseB constitutively (Fig. (Fig.6D).6D). However, no detectable SseB was found by Western blot analysis of the supernatant fraction in the ssaE(mut) and ssaE(I55G) mutant strains harboring the plasmid pACYC-SseB (Fig. (Fig.6D6D and data not shown). In contrast to a previously reported study (49), SseB was easily detected in the supernatant of the sseA mutant strain harboring pACYC-SseB (Fig. (Fig.6D).6D). The same results were obtained when detached fractions containing bacterial surface proteins from these strains were analyzed by Western blotting using anti-SseB antibody (data not shown). Our results suggest that, in addition to SseA, SsaE plays an important role in SseB secretion.
The type III secretion apparatus is associated with an ATPase that presumably provides the energy for the secretion process. Interestingly, mass spectrometer analysis revealed that SsaE coeluted the predicted SPI-2 T3SS ATPase SsaN. We thus examined whether the binding of SsaE to SsaN could be mediated by direct SsaE-SsaN interaction or bridging interactions with other SPI-2 proteins. To confirm this, SsaE was expressed in E. coli as a FLAG-tagged fusion protein, and SsaN was expressed as a 2-HA-tagged fusion protein. SsaE-FLAG-fusion protein prepared from E. coli lysate was immobilized on FLAG beads, and then the SsaN-2HA fusion protein prepared from E. coli lysate was incubated with FLAG fusion protein immobilized beads. While the BAP-FLAG control protein did not interact with SsaN-2HA (data not shown), SsaE-FLAG precipitated the SsaN-2HA (Fig. (Fig.7A).7A). This binding required the C-terminal coiled-coil domain of SsaE, since SsaE(I55G)-FLAG and SsaEΔ3 did not pull down SsaN-2HA (Fig. (Fig.7B7B and data not shown).
Furthermore, we investigated whether the SPI-2 T3SS chaperone SseA interacts with SsaN. To do this, we performed FLAG pull-down assays using the E. coli lysates expressing SseA-FLAG and SsaN-2HA fusion proteins. As shown in Fig. Fig.7C,7C, there was no detectable interaction between SsaN-2HA and SseA-FLAG. These results suggest that SsaE, but not SseA, interacts directly with SsaN, and this interaction requires the C-terminal coiled-coil domain of SsaE.
Finally, to examine the virulence function of ssaE in vivo, we performed mixed infections in mice. Control experiments with the wild-type strain and the wild-type strain harboring pMW118 showed a CI of 1.13 ± 0.19. The CI of the wild-type strain versus the ssaE(I55G) mutant strain harboring pMW118 was significantly decreased to 0.047 ± 0.013 (P < 0.01), compared to that of the SPI-2 mutant strain (the CI of the wild-type strain versus the ΔssaV mutant harboring pMW118 was 0.017 ± 0.026). In addition, the CI of the wild-type strain versus the ssaE(I55G) mutant strain expressing intact SsaE from a plasmid was 1.16 ± 0.21 (no significant difference from the CI of the wild-type strain), showing that the replication defect of the ssaE point-mutant strain was due to the loss of SsaE function. These results clearly demonstrate that ssaE contributes to Salmonella virulence in the mouse model of systemic infection.
Bacterial type III secretion machines are involved in the transport of virulence effectors directly into the cytoplasm of target cells (22). In the initial stages of the assembly of the type III secretion apparatus, the T3SS is involved in protein secretion into the extracellular space, while efficient translocation of effectors to the host cells takes place when the T3SS is completely developed. In this study, we have characterized SsaE, which is encoded by an operon for components of the type III secretion apparatus within the SPI-2 locus. We show that SsaE is localized in the bacterial cytoplasm, indicating that SsaE functions as neither an SPI-2 T3SS core component nor an effector. The structural model of SsaE indicates that it is a protein similar to class V chaperones of the YscE family. Similar to generalized SPI-2 core component mutants, the S. enterica serovar Typhimurium strain lacking SsaE failed to secrete translocator and effector proteins, suggesting that ssaE is required for complete SPI-2 T3SS function. In addition, we demonstrated that SsaE was involved in Salmonella pathogenesis in a mouse model of infection.
Generally, five different classes of T3SS proteins, i.e., proteins for apparatus, translocators, effectors, chaperones, and regulators, are encoded by pathogenicity islands. In addition to these proteins, a new class of type III secretion proteins has recently been identified. This class of proteins is intracellularly localized and is necessary for the ordered secretion of translocator and effector proteins through the T3SS pathway. In Salmonella, InvE of SPI-1 (34) and SsaL and SsaM/SpiC of SPI-2 (10, 60) are suggested to be involved in the recognition of translocator complexes and to establish the ordered secretion. Since some studies suggest that chaperones play a role in setting the secretion hierarchy (3, 5), it is possible to speculate that SsaE is a protein that can distinguish translocator protein SseB from other secreted proteins and enable the export of SseB at the appropriate time. In this study, however, we showed that while Salmonella mutant strains carrying the SsaE, SsaL, or SsaM mutation could not secrete SseB into the culture supernatant, SsaL and SsaM mutants, but not the SsaE mutant, were able to transport the SPI-2-dependent effectors, including PipB, SopD2, and SspH2, extracellularly. Thus, SsaE is functionally different from SsaL and SsaM/SpiC. From these results, although we could not provide evidence concerning the effect of ssaE mutation on the SPI-2 type III secretion apparatus, it is possible to note that, as is the case for other T3SS class V chaperones, the inability of ssaE mutant strains to secrete SPI-2 effectors is due to failure to assemble the needle complex. Further analyses of secretion and/or translocation of SPI-2 effectors by ssaE-deficient mutant strains are needed to clarify SsaE function with SPI-2 type III secretion apparatus assembly.
The homologues of SsaE are conserved in several bacterial pathogens, including PscE of P. aeruginosa (16%), YscE of Y. pestis (25%), EscE of EPEC (27%), and AscE of A. hydrophila (33%), which function as T3SS chaperones (class V). While Salmonella SsaE shares some characteristics with these T3SS class V chaperones, the role of SsaE as a chaperone for SseB appears to be unique. In P. aeruginosa, PscE forms a 1:1:1 ternary heteromolecular complex composed of a putative chaperone, PscG, and a needle component protein, PscF, which is necessary to maintain a monomeric state of PscF (47). Recently, the crystal structures of the PscE/PscF55-85/PscG complex (48), the YscE/YscF/YscG complex (53), and the AscE/AscF/AscG complex (39) have resolved the detailed protein-protein interactions between the chaperone-needle subunit complex of T3SS. However, Salmonella harbors none of the PscG/YscG/AscG homologues. In addition, SsaE is not associated with the stability of binding partner SseB, since the ssaE mutant strain could express SseB at the same level as the wild-type strain, and SseA has been shown to be required for stable expression of SseB as a chaperone (49, 62). These findings provide evidence that Salmonella SsaE should be categorized structurally as a class V chaperone, although it has a function that is distinct from that of other class V chaperones.
Although SseA has previously been found to be essential for export of SseB to the bacterial surface in addition its principal role in stabilization of SseB within the bacterial cytoplasm (49), our results showed that SseA is dispensable for SseB secretion into culture supernatant. This apparent inconsistency is probably due to the difference in sample preparation procedures between the present study and the previous ones. We adopted the in vitro secretion assays established by Coombes et al. (10) to detect SPI-2-dependent secreted proteins using minimal medium previously reported to induce the expression of SPI-2 genes (42) and an antibody specific for SseB (37). This procedure is optimized for isolation of SPI-2 translocator proteins without the need to extract the bacterial surface with n-hexadecane (49) or mechanical shearing (42), which had previously been required to detect sufficient levels of SPI-2 translocator proteins. Under the same assay conditions, the ssaE mutant strain did not secrete SseB into the supernatant, suggesting that SsaE, rather than SseA, is essential for export of SseB.
Multiple chaperones have been reported to act with T3SS translocators for their complete secretion. For example, EspA of EPEC is one of the T3SS substrates that forms a needle structure on the bacterial membrane surface, and efficient secretion of EspA requires two distinct chaperones, CesAB and CesA2. These chaperones increase the stability of EspA and show direct EspA-binding activity (12, 13, 52). Recently, it has been reported that EscL, in addition to CesAB and CesA2, interacts directly with EspA, enhances the stability of intracellular EspA, and is also essential for the complete secretion of EspA (33). Similarly, we found that two small T3SS molecules, SsaE and SseA, function as chaperones for the Salmonella SPI-2 translocator SseB, an EspA homologue. The two chaperones have some common chaperone properties, which include a low molecular mass (<15 kDa), an acidic pI, and a C-terminal amphipathic helix (2, 11, 56). However, these chaperones are functionally different, since SsaE associates with the putative T3SS-associated ATPase SsaN, while SseA does not. Recent evidence strongly suggests that the T3SS-associated ATPase promotes the initial docking of T3SS substrates to the secretion apparatus, and the subsequent translocation of proteins via T3SS depends on the protein motive force (1, 39, 46, 58). Thus, it might be expected that SseB is efficiently escorted by SsaE to the type III secretion apparatus via specific binding to SsaN.
In conclusion, the SPI-2-encoded small protein SsaE was shown to interact with the T3SS translocator SseB and the putative ATPase SsaN. The C-terminal coiled-coil domain of SsaE is critical for these protein-protein interactions. Furthermore, a point mutation of Ile55 to glycine in SsaE completely abolishes the function of SsaE, and a Salmonella strain expressing the point-mutated SsaE exhibits attenuated virulence in mice, much like the SPI-2 component null mutant, suggesting that the coiled-coil-mediated protein-protein interactions of SsaE are absolutely essential for the functional expression of SPI-2 T3SS. Since no T3SS component has been found to have a function identical to that of SsaE, further studies will be needed to identify T3SS proteins that play a role equivalent to that of SsaE.
We are grateful to Kazuhiko Kanou and Hideaki Umeyama for the protein modeling and for helpful input. We also thank Seisuke Hattori for advice on mass spectrometry experiments.
This work was supported in part by a Grant-in-Aid for Young Scientists (B) (17790292) from the Japanese Ministry of Education, Culture, Sports, Science, and Technology and by a Grant-in-Aid for Scientific Research (C) (21590490) from the Japan Society for the Promotion of Science. Support (T.M.) was also received in the form of a Kitasato University Research Grant for Young Researchers (2009).
Published ahead of print on 18 September 2009.