Signal Pathway in Salt-Activated Expression of the Salmonella Pathogenicity Island 1 Type III Secretion System in Salmonella enterica Serovar Typhimurium

Abstract

Salmonella enterica serovar Typhimurium secretes virulence factors for invasion called Sip proteins or Sips into its hosts through a type III secretion system (T3SS). In the absence of a host, S. enterica induces Sip secretion in response to sucrose or simple salts, such as NaCl. We analyzed induction of host-independent Sip secretion by monitoring protein secretion by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), assembly of needle complexes by electron microscopy, and transcription of virulence regulatory genes by quantitative reverse transcriptase PCR (real-time PCR). SDS-PAGE showed that addition of sucrose or simple salts, such as NaCl, to the growth medium induced Sip secretion without altering flagellar protein secretion, which requires a distinct T3SS. Electron microscopy confirmed that the amount of secreted Sips increased as the number of assembled needle complexes increased. Real-time PCR revealed that added sucrose or NaCl enhanced transcription of hilA, hilC, and hilD, which encode known regulators of Salmonella virulence. However, epistasis analysis implicated HilD and HilA, but not HilC, in the direct pathway from the salt stimulus to the Sip secretion response. Further analyses showed that the BarA/SirA two-component signal transduction pathway, but not the two-component sensor kinase EnvZ, directly activated hilD and hilA transcription and thus Sip secretion in response to either sucrose or NaCl. Finally, real-time PCR showed that salt does not influence transcription of the BarA/SirA-dependent csrB and csrC genes. A model is proposed for the major pathway in which sucrose or salt signals to enhance virulence gene expression.

Animal pathogens, such as Salmonella, pathogenic Escherichia coli, Shigella, or Yersinia spp., secrete virulence proteins into hosts using specialized protein secretion systems called type III secretion systems (T3SS) (21). Different species regulate T3SS differently. For instance, the level of type III secretion from Shigella spp. remains low until secretion is triggered by close contact with host tissues in vivo, a process that can be mimicked by low-speed centrifugation of a mixture of bacterial and host cells (50) or by addition of Congo red to bacterial cells (6, 36). On the other hand, type III secretion from Salmonella enterica serovar Typhimurium is triggered by interaction with a host or by manipulation of the growth conditions (14). Under physiological conditions, S. enterica serovar Typhimurium cells spontaneously secrete significant amounts of proteins into the medium. In one study, for example, 80% of the secreted proteins (called Sip proteins or Sips) were found in the infection medium (13). Among these secreted proteins, five flagellar proteins (flagellin, hook protein, and three hook-associated proteins) and five virulence factors (SipA, SipB, SipC, SipD, and InvJ) are abundant enough to stain with Coomassie brilliant blue (29). Secretion of these two groups of proteins is facilitated by a subset of the flagellar genes (2, 37) and by a subset of Salmonella pathogenicity island 1 (SPI1) genes (24), respectively. Each subset of genes encodes a distinct T3SS that has common structural features (1, 22).

The regulatory network responsible for ensuring proper expression, assembly, and function of the flagellar T3SS has been well established (2, 35). In contrast, the network responsible for controlling expression, assembly, and function of the SPI1 T3SS is still being actively investigated. It is quite clear that HilA is the proximal activator of SPI1 genes (4, 7, 8, 26, 33, 34, 43, 49) and that hilA transcription is controlled by two other regulators, HilD and HilC (3, 9, 10, 11, 19, 37, 39, 41, 42, 47, 52). These regulators, in turn, have been reported to be controlled by four two-component signal transduction pathways. PhoQ-PhoP and PhoR-PhoB both activate hilE, which encodes an inhibitor of hilD (9); EnvZ-OmpR activates hilD (32, 38); and BarA-SirA activates hilD, hilC, and hilA (31, 48). How the BarA-SirA pathway activates SPI1 is controversial; it has been proposed that BarA-SirA acts directly on hilC and hilA but indirectly through the Csr system to activate hilD (30) (Fig. 1).

FIG. 1.

Current model for the regulatory network of SPI1 gene expression. Boxes indicate two-component systems, ellipses indicate modulation proteins, italics indicate regulatory genes, solid arrows indicate direct influence, dashed arrows indicate indirect influence, and lines with a stop indicate repression.

Although the flagellar and SPI1 regulatory networks are generally independent of each other, evidence showing some coordinated expression has accumulated. In S. enterica serovar Typhimurium, for example, expression of virulence genes is regulated by fliZ, which encodes a flagellar activator (25). In S. enterica serovar Typhi, SPI1 gene expression depends on the flagellar sigma factor FliA (16). Moreover, flagellar and SPI1 genes are reciprocally regulated by the rtsA and rtsB genes (17, 18). Fis, a DNA binding protein, is required for full expression of both SPI1 genes and flagellar genes (28, 52). Finally, the ATP-dependent protease Lon cleaves the SPI1 regulators HilC and HilD and the flagellar master regulator FlhC-FlhD, resulting in suppression of both systems (11, 12, 45, 46, 47).

Many environmental factors have been reported to influence SPI1 gene expression and/or Sip secretion, including low oxygen tension, neutral pH, acetate and other short-chain fatty acids, cationic peptides, and bile (4). To the best of our knowledge, only two reports have directly addressed the effect of osmolarity on SPI1 gene expression. Galan and Curtiss (23) explored the effect of high concentrations of salt (300 mM and higher) and concluded that the osmoinducibility of invA depends on changes in DNA supercoiling but not on the osmoregulator ompR, while Bajaj and coworkers (8) showed that osmolarity indirectly activates SPI1 gene expression and that the process is mediated by HilA.

In this paper, we carefully document enhanced Sip secretion in response to exposure to diverse salts and sucrose and a wide range of NaCl concentrations. In the process, we identify two distinct responses, one response that occurs upon exposure to high concentrations of salt and appears to involve DNA supercoiling and another response that occurs upon exposure to lower concentrations. We further demonstrate that the latter response requires several of the known SPI1 regulators, most notably HilA and HilD. We also show that this response requires the BarA-SirA two-component signal transduction pathway system. We propose a simple model that can serve as a firm foundation upon which to base further studies of SPI1 regulation.

MATERIALS AND METHODS

Strains and growth conditions.

The strains used in this study are listed in Table 1. Mutant strains were derived mainly from SJW1103, a wild-type strain of S. enterica serovar Typhimurium. DW269 (NK182 envZ::Cm) was a kind gift from Linda Kenny. The envZ, barA, and sirA mutations were transferred into SJW1103 by P1 transduction (44).

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant properties	Reference
Strains
SJW1103		53
χ3306	Virulence strain, gyrA	24
CS2724	ΔhilC in χ3306	47
CS2725	ΔhilD in χ3306	47
CS2802	ΔhilC ΔhilD in χ3306	47
CS3222	hilA::Km in χ3306	This study
AY001	barA::Km ΔbarA in SJW1103	This study
AY002	envZ::Cm ΔenvZ in SJW1103	This study
AY003	sirA::Km ΔsirA in SJW1103	This study
Plasmids
pTKY 748	hilA Apr	This study
pZA4 lacIq	lacIq Specr	This study

Except where indicated otherwise, cells were grown for 4 h at 37°C in 5 ml of salt-free TY medium (1% [wt/vol] tryptone and 0.5% [wt/vol] yeast extract) or TY medium supplemented with salts or sucrose at the indicated concentrations.

SDS-PAGE.

The proteins secreted into media were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). After the cells were removed by low-speed centrifugation (13,200 × g for 20 min), 1 ml of each supernatant was placed into an Eppendorf tube, mixed with prechilled 25% (wt/vol) trichloroacetic acid (TCA) (final concentration, 6%), chilled on ice for 15 min, and centrifuged at 10,000 × g for 10 min. The pellets were suspended with 0.3 ml of acetone and centrifuged at 10,000 × g for 10 min. Acetone washing was repeated twice to remove all of the TCA from the precipitates. Dry pellets were dissolved in SDS sample buffer. SDS-PAGE was carried out using a mini-gel kit from Bio-Rad Laboratories, Inc. The acrylamide concentration of the gels was 12.5%. Gels were stained with Coomassie brilliant blue.

Observation of osmotically shocked cells by electron microscopy.

Cells collected from 5 ml of culture were resuspended in 100 μl of a sucrose solution (20% [wt/vol] sucrose, 5 mM EDTA; pH 7). After 5 min of incubation at room temperature, the suspension was abruptly diluted with 10 ml prechilled water. Intact cells were removed by low-speed centrifugation, and the osmotically shocked cells that remained in the supernatant were collected by high-speed centrifugation (15,000 × g, 20 min). Samples were negatively stained with 2% (wt/vol) phosphotungstic acid (pH 7.0) and observed with a JEM-1200EXII electron microscope (JEOL, Tokyo, Japan). Micrographs were taken at an accelerating voltage of 80 kV.

Quantitative reverse transcriptase PCR (real-time PCR) and calculation of relative expression levels.

All PCRs were performed using an Mx3000P quantitative PCR system (Stratagene) by following the manufacturer's instructions. For each PCR, the reaction mixture (total volume, 25 μl) was prepared using FullVelocity SYBR green quantitative reverse transcriptase PCR master mixture (Stratagene) with 0.1 μg/μl total RNA, 10 pmol of the PCR forward primer, 10 pmol of the PCR reverse primer, and the supplied ROX reference dye (Stratagene). The PCR cycling conditions were an initial denaturation step of 30 s at 95°C, followed by 40 PCR cycles of denaturation for 10 s at 95°C and annealing/extension for 30 s at 60°C. The relative amplification for expression of the genes in the presence of salt compared to that in a salt-free culture was automatically calculated.

The following primers were used: for the 16S rRNA gene, forward primer TGTAGCGGTGAAATGCGTAG and reverse primer CAAGGGCACAACCTCCAAG (predicted length of the transcript, 161 bp); for hilA, forward primer CATGGCTGGTCAGTTGGAG and reverse primer CGTAATTCATCGCCTAAACG (predicted length of the transcript, 150 bp); for hilC, forward primer GGACTTGTTGCCAGGGATG and reverse primer TGACCATTTGCGGGTGAG (predicted length of the transcript, 241 bp); for hilD, forward primer ACTCGAGATACCGACGCAAC and reverse primer CTTCTGGCAGGAAAGTCAGG (predicted length of the transcript, 129 bp); for hilE, forward primer CAGAGACACCAACGAAATGG and reverse primer AAACCTTTGATCCGGCTTTC (predicted length of the transcript, 159 bp); for sipC, forward primer CTGTGGCTTTCAGTGGTCAG and reverse primer TGCGTTGTCCGGTAGTATTTC (predicted length of the transcript, 150 bp); for csrA, forward primer CTGGACTGCTGGGATTTTTC and reverse primer CATGATTGGCGATGAGGTC (predicted length of the transcript, 144 bp); for csrB, forward primer GCGTTAAAGGACACCTCCAG and reverse primer ACCTTACGGCCTGTTCATCC (predicted length of the transcript, 146 bp); and for csrC, forward primer GAAGACAAACGTCCGGAGAC and reverse primer CCTTAACGGGTTCCACCATC (predicted length of the transcript, 116 bp).

RESULTS

Rationale for in vitro assay of T3SS.

The complex interaction between bacterial cells and host cells may involve both unknown chemicals and variable physical stimuli. This makes it difficult to dissect the signaling networks that control interaction-dependent behavior, such as the biogenesis of the T3SS, which bacteria use to secrete virulence proteins (Sips) into host cells. It is known, however, that cells of S. enterica serovar Typhimurium grown in LB medium secrete Sips even in the absence of host cells (14, 29). Thus, we can study the regulation of secretion in a host-independent manner. In such a system, we can control the nature and amount of chemicals and/or stimuli added exogenously. We propose that results from such in vitro experiments can reveal fundamental properties of the secretion system.

In this study, we investigated the effect of osmolarity on the expression of the SPI1 T3SS and dissected the pathway through which osmolarity exerts this effect. In practice, we first analyzed secretion proteins by SDS-PAGE. If Sip proteins were secreted as expected, we assumed that SPI1 genes were expressed normally and that the T3SS had formed successfully. When Sip secretion was abnormal, we analyzed SPI1 gene expression by real-time PCR and observed T3SS by electron microscopy.

Increased osmolarity induces Sip secretion. (i) Sip secretion is NaCl dependent.

We routinely recovered secreted proteins from cell-free spent media that had been harvested following 4 h of incubation by TCA precipitation and analyzed these proteins by SDS-PAGE. Protein bands were blotted on a polyvinylidene fluoride membrane and analyzed by N-terminal amino acid sequencing (29). When wild-type cells were grown in TY medium supplemented with 100 mM NaCl, the major proteins in each band were either flagellar proteins (FlgK, FliC, FlgE, and FlgL) or Sip proteins (SipA, SipB, SipC, SipD, and InvJ) (Fig. 2A). In contrast, when cells were grown in the absence of salt, they secreted only the flagellar proteins (Fig. 2A). The bands containing FlgE or SipC and the bands containing InvJ or SipD overlapped; thus, band assignments were made cautiously. The amount of flagellar protein secreted into the media varied from one protein to another and was not necessarily proportional to the amount required for formation of a structure. We concluded that NaCl induces Sip secretion.

FIG. 2.

SDS-PAGE gel pattern of secreted proteins under various salt conditions. (A) Typical SDS-PAGE gel pattern of secreted proteins in the absence and presence of NaCl or sucrose for S. enterica serovar Typhimurium. Secreted proteins were recovered from cell-free spent culture media by TCA precipitation. Cells were grown in TY medium with no NaCl (None), TY medium supplemented with 100 mM NaCl, or TY medium supplemented with 200 mM sucrose. Each protein band was assigned to a flagellar protein (left) or a virulence protein (right). The molecular mass of each protein is indicated in parentheses. (B) An overnight culture in TY medium was transferred into fresh TY media containing various concentrations of NaCl, including 0, 100, 200, 300, 400, 500, 600, 700, 800, and 900 mM. NaCl concentrations up to 300 mM had no effect on the growth rate or maximum cell density. To standardize loading, the optical densities at 660 nm of the cultures at the time of harvest were 1.651 (0 mM), 1.677 (100 mM), 1.658 (200 mM), 1.590 (300 mM), 1.563 (400 mM), 1.566 (500 mM), 1.600 (600 mM), 1.561 (700 mM), 1.551 (800 mM), and 1.450 (900 mM).

To optimize conditions for the study of NaCl-induced Sip secretion, we performed a dose dependence experiment. The effect of NaCl was biphasic with respect to concentration (Fig. 2B). In the absence of NaCl, low levels of Sips were secreted. As the NaCl concentration was increased up to 100 mM, the amounts of secreted Sips gradually increased (Fig. 2B and data not shown). At an NaCl concentration of 400 mM or higher, Sip secretion decreased gradually. In contrast, the amounts of secreted flagellar proteins remained relatively constant at NaCl concentrations between 0 and 500 mM. These results are consistent with the hypothesis that in S. enterica serovar Typhimurium the flagellum-specific and SPI1-specific T3SS systems are regulated independently, at least with regard to salt concentration.

In the presence of an NaCl concentration of 600 mM or higher, the intensities of the bands of both the Sip and flagellar proteins decreased, while many minor background bands appeared. These results are consistent with global damage to secretion induced by high salt concentrations. Such concentrations clearly inhibited growth (data not shown), and microscopic examination revealed that the cells were motile but elongated like snakes. We also observed similar abnormal secretion when cells were grown in the presence of 100 mM NaCl if a DNA gyrase inhibitor (novobiocin or coumermycin A1) was also added; growth was impaired, microscopic examination revealed elongated cells (data not shown), and SDS-PAGE analysis revealed many minor background bands (Fig. 3). The presence of 20 μg/ml novobiocin resulted in a secreted protein profile similar to that observed with cells grown in the presence of 700 mM NaCl (compare Fig. 3A and 2B). Coumermycin A1 was more effective, resulting in a similar secreted protein profile at concentrations as low as 2 μg/ml (Fig. 3B). These results are consistent with the hypothesis that high concentrations of NaCl hamper the normal function of DNA gyrase and thus reduce DNA superhelicity, as suggested previously (43).

FIG. 3.

SDS-PAGE gel pattern of secreted proteins in the presence of gyrase inhibitors. (A) Secreted proteins from TY media supplemented with 100 mM NaCl and various concentrations of novobiocin, including 0, 1, 5, 10, 20, and 50 μg/ml. (B) Secreted proteins from TY media containing 100 mM NaCl and various concentrations of coumermycin A1, including 0, 0.1, 0.2, 0.5, 1, 2, 5, and 10 μg/ml.

(ii) Simple salts induce Sip secretion.

We next surveyed the effects of various salts on Sip secretion. Cells were precultured in TY medium and inoculated into TY media containing the following salts at a concentration of 100 mM: NaCl, KCl, KBr, NH₄Cl, CaCl₂, MgCl₂, MgSO₄, Na₂SO₄, and (NH₄)₂SO₄. With one exception, these salts yielded a secreted protein profile that resembled the pattern produced in the presence of NaCl (Fig. 4). The intensity of the bands produced in the presence of KCl, KBr, or MgCl₂ was generally similar to the intensity of the bands produced in the presence of NaCl. The profile was generally less intense in the presence of NH₄Cl, MgSO₄, Na₂SO₄, or (NH₄)₂SO₄. In contrast, CaCl₂ did not induce Sip secretion. Thus, it appears that with the exception of CaCl₂, the halogenated salts exerted a stronger effect than the sulfated salts. For all further studies, we employed NaCl.

FIG. 4.

SDS-PAGE gel pattern of secreted proteins from TY culture media supplemented with various salts, including no salts, 100 mM NaCl, 100 mM KCl, 100 mM KBr, 100 mM NH₄Cl, 100 mM CaCl₂, 100 mM MgCl₂, 100 mM MgSO₄, 100 mM Na₂SO₄, and 100 mM (NH₄)₂SO₄.

(iii) Sucrose induces Sip secretion.

To determine if the salt effect results from changes in medium osmolarity, we examined Sip secretion in the presence of 200 mM sucrose. The secretion profile (Fig. 2A) was similar but not identical to the profiles observed with NaCl (Fig. 2A). Whereas sucrose induced secretion of SipA and SipD/InvJ at levels that were similar to those observed in the presence of NaCl, the secretion of SipB and SipC was not nearly as strong. We concluded that increased osmolarity induces SPI1-dependent secretion.

NaCl induces needle complex assembly.

The SPI1 T3SS secretes Sips through a needle complex (24, 27) that can be seen by electron microscopy. We therefore employed electron microscopy to determine if salt influences the assembly of needle complexes. Wild-type cells were grown in TY medium supplemented or not supplemented with 100 mM NaCl; the final cell densities under the two conditions were similar, and microscopic examination revealed no obvious morphological defects. For electron microscopy, the cells were harvested from culture media, suspended in a sucrose solution, and shocked osmotically. When cells were grown in the presence of 100 mM NaCl, about 70% of the cells had about 50 to 100 needle complexes on the cell surface (Fig. 5A). In contrast, when they were grown in the absence of NaCl, the majority of the cells displayed only a few needle complexes (Fig. 5B). These data are consistent with the hypothesis that the amount of secreted Sips correlates with the number of needle complexes per cell.

FIG. 5.

Electron micrographs of osmotically shocked cells. Cells grown in (A) TY medium without salt or (B) TY medium supplemented with 150 mM NaCl were osmotically shocked, stained with 2% phosphotungstic acid (pH 7), and observed by electron microscopy. Only the cell surfaces are shown. The arrowheads indicate needle complexes. Bars = 200 nm.

Increased osmolarity induces transcription of SPI1 regulators.

To identify the part of the proposed regulatory network (Fig. 1) that controls needle complex assembly and Sip secretion in response to increased osmolarity, we used quantitative reverse transcriptase PCR (real-time PCR) to monitor transcription of known regulators of SPI1 genes. We first examined transcription of the hilA gene, which directly regulates SPI1 gene expression (7, 8, 33, 34), and transcription of the hilC and hilD genes, which are known to positively regulate hilA (8, 33).

The presence of 100 mM NaCl increased the expression of hilA, hilC, and hilD nearly sixfold, threefold, and nearly fivefold, respectively. In contrast, salt had little effect on hilE transcription (Table 2). Similar results were observed when 100 mM NaCl was replaced with 200 mM sucrose (data not shown). These results are consistent with previous reports indicating that HilA, HilC, and HilD regulate SPI1 gene transcription and that HilC and HilD activate hilA transcription equally (3, 7, 8, 9, 10, 11, 19, 33, 34, 37, 40, 41, 47, 49). To determine if the salt response required HilC and HilD equally, we performed an epistasis analysis. We compared the wild-type Sip secretion profile to the secretion profile of mutant cells defective for hilA, defective for either hilC or hilD, or defective for both hiC and hilD (Fig. 6). The hilA, hilD, and hilC hilD mutants did not secrete Sips. In contrast, the hilC mutant secreted Sip proteins about as well as its wild-type parent. Accordingly, hilA, hilC, and hilD transcripts were barely detectable in hilD and hilC hilD mutants even in the presence of NaCl (Table 3). In contrast, salt-activated transcription of the hilA and hilD genes was observed in the hilC mutant, although the levels were lower than those in the wild-type parent. Electron microscopy showed that there were many needle complexes in the hilC mutant but few needle complexes in the hilA, hilD, and hilC hilD mutants (data not shown). We therefore propose that the salt signal passes through HilD to HilA and that HilC exerts a modulating effect on hilD and hilA transcription, but this modulating effect is not enough to influence the Sip secretion profile.

TABLE 2.

Real-time PCR for expression of T3SS-related regulatory genes: hilA, hilC, hilD, and hilE gene expression in the wild-type strain in the presence or absence of 100 mM NaCl

Presence of NaCl	Expression levels of genesa
	hilA	hilC	hilD	hilE
−	1.00 ± 0.10	1.00 ± 0.09	1.00 ± 0.07	1.00 ± 0.03
+	5.70 ± 0.67	3.00 ± 0.43	4.69 ± 0.54	1.16 ± 0.10

^amRNA levels of the genes from cultures grown in TY medium supplemented with 100 mM NaCl or with no added salt are indicated. The values for the relative levels obtained under the two conditions were determined by defining the level in TY medium with no added salt as 1.

FIG. 6.

Protein secretion from hilA, hilC, hilD, and hilC hilD deletion mutants: SDS-PAGE gel pattern of secreted proteins from wild-type strain χ3306 (WT), a hilA deletion mutant (ΔhilA), a hilC deletion mutant (ΔhilC), a hilD deletion mutant (ΔhilD), and a hilC hilD double-deletion mutant (ΔhilC,D) grown in TY medium (−) or TY medium supplemented with 100 mM NaCl (+). Note that cells of wild-type strain χ3306 and cells of its derivatives can produce either the FliC or FljB flagellin. Under these conditions the hilA mutant secreted FliC flagellin, while the wild-type parent and the hilC, hilD, and hilC hilD mutants secreted the FljB flagellin. The underlying mechanism remains unknown.

TABLE 3.

Real-time PCR for expression of T3SS-related regulatory genes: hilA, hilC, and hilD gene expression in hilC, hilD, and hilC hilD deletion mutants

Strain	Presence of NaCl	Expression levels of genesa
		hilA	hilC	hilD
Wild type	−	1.00 ± 0.08	1.00 ± 0.16	1.00 ± 0.03
	+	5.27 ± 0.37	5.97 ± 1.09	3.56 ± 0.13
ΔhilC	−	0.14 ± 0.04	0.00 ± 0.00	0.24 ± 0.02
	+	1.40 ± 0.28	0.00 ± 0.00	1.36 ± 0.10
ΔhilD	−	0.00 ± 0.00	0.03 ± 0.01	0.00 ± 0.00
	+	0.00 ± 0.00	0.06 ± 0.01	0.00 ± 0.00
ΔhilC ΔhilD	−	0.00 ± 0.00	0.00 ± 0.00	0.00 ± 0.00
	+	0.00 ± 0.00	0.00 ± 0.00	0.00 ± 0.00

^aSee Table 2, footnote a.

To test this model, we transformed a hilD mutant with a multicopy plasmid that carries the wild-type allele of hilA. As a control, we also transformed a hilA mutant. We then exposed the strains to salt and monitored hilA, hilD, and sipC transcription. Relative to the wild type, the presence of the hilA plasmid increased hilA transcription 40- and 70-fold and sipC transcription 20- and 30-fold in the hilA and hilD mutants, respectively, even in the absence of salt (Table 4). Overexpression of HilA enhanced hilD transcription, as reported previously (15, 49). In the hilA mutant, which expressed HilD, exposure to salt increased hilA, hilD, and sipC transcription. In the hilD mutant, exposure to salt had no effect on sipC transcription and caused a significant decrease in hilA transcription. On the basis of these observations, we concluded that the salt signal acts on hilA transcription and thus transcription of SPI1 genes (e.g., sipC) via HilD.

TABLE 4.

Real-time PCR for expression of T3SS-related regulatory genes: hilA, hilD, and sipC gene expression in a hilA mutant and a hilD mutant transformed with a plasmid expressing the hilA gene

Strain	Presence of NaCl	Expression levels of genesa
		hilA	hilD	sipC
Wild type	−	0.06 ± 0.01	0.28 ± 0.02	0.04 ± 0.00
	+	1.00 ± 0.18	1.00 ± 0.05	1.00 ± 0.08
ΔhilA/philA	−	41.39 ± 6.57	1.83 ± 0.13	19.06 ± 0.00
	+	51.78 ± 0.13	4.23 ± 0.26	31.22 ± 0.07
ΔhilD/philA	−	68.94 ± 9.98	0.00 ± 0.00	27.00 ± 1.75
	+	44.64 ± 0.34	0.00 ± 0.00	25.46 ± 0.09

^aSee Table 2, footnote a.

The BarA-SirA two-component system mediates the response to high osmolarity.

The following four two-component signal transduction systems could affect hil gene expression: PhoQ-PhoP, PhoR-PhoB, EnvZ-OmpR, and BarA-SirA. We could eliminate the PhoQ-PhoP pathway because these proteins activate hilE (9) and the presence of either 100 mM NaCl or 200 mM sucrose had no effect on the expression of hilE (Table 2 and data not shown). To test the involvement of the EnvZ-OmpR and BarA-SirA pathways, we compared the secretion profiles of barA and envZ mutants to the profile of their isogenic wild-type parent. The barA and sirA mutants did not secrete Sips even in the presence of 100 mM NaCl (Fig. 7) or 200 mM sucrose (data not shown). In contrast, the envZ mutant exhibited a secretion profile indistinguishable from that of the wild-type parent (Fig. 7 and data not shown). Consistent with these data, 100 mM NaCl induced hilA and hilD transcription in the envZ mutant but not in the barA and sirA mutants (Table 5). Furthermore, electron microscopy revealed numerous needle complexes on the surface of the envZ mutant but not on the surface of the barA mutant (data not shown). These data are consistent with the hypothesis that the BarA/SirA pathway, but not EnvZ, is intimately involved in salt-activated Sip secretion.

FIG. 7.

Protein secretion from mutants deficient in either the envZ/ompR or barA/sirA two-component system. Wild-type strain SJW1103 (WT), a barA deletion mutant (ΔbarA), a sirA deletion mutant (ΔsirA), and an envZ deletion mutant (ΔenvZ) were grown in TY medium with no added salt (−) or with 100 mM NaCl (+). SDS-PAGE gel patterns of secreted proteins from wild-type strain SJW1103 and the barA, sirA, and envZ deletion mutants are shown.

TABLE 5.

Real-time PCR for expression of T3SS-related regulatory genes: hilA and hilD gene expression in mutants deficient for barA, sirA, or envZ

Strain	Presence of NaCl	Expression levels of genesa
		hilA	hilD
Wild type	−	1.00 ± 0.24	1.00 ± 0.04
	+	3.74 ± 1.24	2.72 ± 0.12
ΔbarA	−	0.01 ± 0.00	0.06 ± 0.01
	+	0.16 ± 0.05	0.32 ± 0.01
ΔsirA	−	0.01 ± 0.00	0.06 ± 0.00
	+	0.27 ± 0.06	0.27 ± 0.02
ΔenvZ	−	0.42 ± 0.07	0.46 ± 0.02
	+	4.80 ± 0.88	2.32 ± 0.16

^aSee Table 2, footnote a.

It is known that the BarA-SirA pathway activates transcription of the small csrB and csrC RNAs. These small RNAs bind the protein CsrA (5, 30). CsrA has been shown to inhibit hilD expression (30). We therefore tested if salt influences the transcription of csrA, csrB, and csrC (Table 6). In contrast to hilA transcription, which exhibited a BarA-SirA-dependent response to 100 mM NaCl, csrA, csrB, and csrC transcription did not respond to salt even though transcription of csrB and csrC did depend on the BarA-SirA pathway, as reported previously (20, 30). We concluded that the salt stimulus acts via the BarA-SirA pathway but not through the Csr system.

TABLE 6.

Real-time PCR for expression of T3SS-related regulatory genes: hilA, csrA, csrB, and csrC gene expression in mutants deficient for barA, sirA, or envZ

Strain	Presence of NaCl	Expression levels of genesa
		hilA	csrA	csrB	csrC
Wild type	−	1.00 ± 0.22	1.00 ± 0.14	1.00 ± 0.05	1.00 ± 0.59
	+	8.36 ± 1.61	1.11 ± 0.12	1.13 ± 0.05	0.52 ± 0.28
ΔbarA	−	0.02 ± 0.00	0.95 ± 0.11	0.14 ± 0.01	0.19 ± 0.09
	+	0.29 ± 0.06	0.64 ± 0.07	0.30 ± 0.01	0.16 ± 0.10
ΔsirA	−	0.02 ± 0.01	0.94 ± 0.11	0.13 ± 0.01	0.18 ± 0.08
	+	0.13 ± 0.02	0.74 ± 0.08	0.17 ± 0.04	0.13 ± 0.05
ΔenvZ	−	0.83 ± 0.20	1.14 ± 0.12	0.90 ± 0.09	0.63 ± 0.32
	+	8.87 ± 1.59	0.87 ± 0.10	1.31 ± 0.06	0.41 ± 0.17

^aSee Table 2, footnote a.

Second effect of salt.

The data presented thus far are consistent with a simple linear pathway in which BarA senses the salt/sucrose stimulus and transduces the information to SirA, which informs HilD, which activates hilA transcription, which permits HilA-dependent activation of SPI1 genes, including the T3SS and its Sips. Alternatively, however, salt/sucrose may directly affect the conformation of intermediate regulatory proteins and thus modify their function. Because plasmid-borne hilA in the hilA mutant induced hilA, hilD, and sipC transcription independent of salt (Table 4), we reasoned that this strain could allow us to observe any transcription-independent salt effects. Indeed, Sip secretion, which was completely inhibited in the hilA mutant, recovered in the presence of plasmid-borne HilA (Fig. 8). Intriguingly, in the absence of salt, the secretion profile of this transformant resembled that of wild-type cells exposed to 200 mM sucrose; the amount of secreted SipB and SipC was reduced in the absence of NaCl relative to the amount in the presence of NaCl (Fig. 8). We propose that NaCl impacts Sip secretion in two distinct ways: (i) like sucrose, it induces SPI1 transcription via a pathway that includes BarA, SirA, HilD, and HilA; and (ii) unlike sucrose, it enhances secretion of SipB and SipC.

FIG. 8.

SDS-PAGE gel patterns of secreted proteins from wild-type strain χ3306 (WT), a hilA mutant (ΔhilA), and the hilA mutant transformed with the hilA plasmid (ΔhilA/philA⁺) in the absence (−) and presence (+) of 100 mM NaCl.

DISCUSSION

In this study, we showed that exposure to simple salts (100 mM) and sucrose (200 mM) specifically induces expression and assembly of the SPI1-encoded T3SS and its associated needle complex and expression and secretion of SPI1-encoded Sips. Because Sip secretion occurred in response to a variety of salts and to sucrose, we propose that elevated osmolarity is the relevant environmental stimulus. This response may explain why Salmonella infection of animal tissues is most effective in saline solutions (23).

Osmolarity response pathway: a simple model.

To understand how cells respond to relatively low concentrations of salt (100 mM) and sucrose (200 mM), we used genetic analyses to track the pathway through which salt activates Sip secretion. On the basis of these studies, we propose the following model (Fig. 9). In response to salt, the sensor kinase BarA autophosphorylates and donates its phosphoryl group to its cognate response regulator, SirA. Phosphorylated SirA then activates transcription of a factor X, which is not the small csr RNAs. Factor X activates hilD, which activates hilA. The evidence for this simple model is as follows. Exposure to salt or sucrose activated hilD, hilC, and hilA transcription (Table 2). hilD, hilC, and hilA mutants did not secrete Sips in response to salt (Fig. 6). This response depended on hilD and hilA but not on hilC (Table 3), supporting the hypothesis that HilC has an accessory role. The barA and sirA mutants did not secrete Sips in response to either salt or sucrose, and hilA and hilD transcription depended on barA (Fig. 7 and Table 5). In contrast, the envZ mutant secreted Sips and transcribed hilA and hilD as well as its wild-type parent; thus, EnvZ plays a minor role in this pathway (Fig. 7 and Table 5). Because exposure to salt or sucrose had no effect on hilE transcription (Table 2), it is unlikely that the salt/sucrose response involves the PhoQP and PhoRB pathways.

FIG. 9.

Model for the major pathway of the salt signal for SPI1 gene expression. Boxes indicate two-component systems, ellipses indicate modulation proteins, italics indicate regulatory genes, solid arrows indicate direct influence, dotted arrows indicate indirect influence, and the line with a stop indicates repression.

Second effect of salt.

Although sucrose and salt induced Sip secretion through the same pathway, the secretion profiles induced by sucrose and salt were not identical. Although exposure to sucrose and exposure to salt resulted in detection of similar amounts of SipA and SipD, only exposure to salt resulted in detection of significant amounts of SipB and SipC (Fig. 2A). A possible reason for this discrepancy became apparent when we overexpressed HilA in a hilA mutant (Table 4). This genetic manipulation bypassed the salt requirement for Sip secretion with two notable exceptions: SipB and SipC. In the absence of salt, we detected small amounts of these two proteins. In contrast, we detected substantial amounts of both proteins upon exposure to salt. Since all four proteins are encoded in a single operon in the order sipBCDA, it is quite likely that they are expressed at similar levels. Indeed, regardless of treatment, all four Sip proteins could be detected in the medium when they were harvested at earlier time points (data not shown). We therefore propose that SipB and SipC are expressed and secreted but are degraded during prolonged exposure to the medium. Via a presently unknown mechanism, 100 mM NaCl, but not 200 mM sucrose, appears to protect SipB and SipC from this degradation. One difference between the two pairs of Sips is that SipB and SipC are escorted by the specific chaperone SicA, while SipA and SipD are not (51).