PmrA Represses Transcription of the SPI-2 ssaG Gene.

Given that the pmrA hypervirulence phenotype was observed in Nramp1^+/+ mice, we wondered whether the PmrA/PmrB system might regulate expression of SPI-2 genes. Genes located in the SPI-2 pathogenicity island are up-regulated in the spleens of Nramp1^+/+ mice relative to those of congenic Nramp1^−/− animals (25). Thus, we investigated expression of the SsrB-activated ssaG gene, which is the first gene of a large operon (from ssaG to ssaU) and encodes a component of the Spi/Ssa T3SS (22). We used a set of isogenic strains harboring a plasmid containing the ssaG promoter region fused to a promoterless gfp variant that specifies an unstable GFP (i.e., GFPaav) (26).

There was more fluorescence in the pmrA mutant than in the wild-type strain when bacteria were grown in minimal media at pH 4.6 with high Mg²⁺ (Fig. 2A), a condition reported to activate SPI-2 genes (27). This behavior was also true when bacteria were grown in minimal media at pH 4.6 with low Mg²⁺, although there was less of a difference between the two strains. By contrast, the expression was similar in wild-type and pmrA Salmonella in media containing neutral pH regardless of the Mg²⁺ concentration (Fig. 2A), which is a noninducing condition for SPI-2 genes (27, 28). Control experiments demonstrated that there was no detectable fluorescence in an ssrB mutant strain regardless of growth conditions (Fig. 2A), as anticipated by the established requirement for ssrB in ssaG transcription (21). Moreover, the elevated ssaG expression displayed by the pmrA mutant when grown in minimal media at pH 4.6 with high Mg²⁺ was restored to wild-type levels by a plasmid expressing the wild-type pmrA gene from a heterologous promoter but not by the plasmid vector (Fig. 2B).

Open in a separate window

Fig. 2.

PmrA decreases expression of the SPI-2 ssaG gene. (A) Fluorescence from a ssaG-gfp transcriptional fusion was determined in wild-type, pmrA, and ssrB Salmonella grown in N-minimal medium at pH 4.6 with 1 mM Mg²⁺ (LH) or 10 μM (LL), or pH 7.7 with 1 mM Mg²⁺ (HH) or 10 μM (HL) for 8 h. (B) Relative expression from a chromosomal ssaG-lacZ transcriptional fusion was determined in wild-type, pmrA, pmrA harboring the plasmid vector (pUHE21-2lacI^q), and pmrA harboring a plasmid expressing pmrA from a lac promoter derivative following growth in N-minimal medium at pH 4.6 with 1 mM Mg²⁺ (LH). Values are normalized relative to those displayed by the wild-type strain. (C) mRNA levels of ssaG produced by wild-type, pmrA, and ssrB Salmonella grown overnight in LB media (inoculum) and 4 h after infection of J774A.1 macrophages.

Because Fe³⁺ is a prime activator of the PmrA/PmrB system (6), we reasoned that it should be possible to modulate expression of SPI-2 genes in wild-type Salmonella by varying the concentrations of Fe³⁺ in the growth media. Indeed, ssaG expression was greatly reduced as the Fe³⁺ concentration increased (Fig. 3A). This effect was mediated by PmrA because ssaG expression was no longer responsive to the Fe³⁺ concentration in a pmrA mutant background (Fig. 3B).

Open in a separate window

Fig. 3.

Iron regulates expression of the SPI-2 ssaG gene in a PmrA-dependent manner. GFP activity from the ssaG-gfp transcriptional fusion was determined in wild-type Salmonella grown in N-minimal medium at pH 4.6 with 1 mM Mg²⁺ for 8 h with or without 100 μM Fe³⁺ (A); or in wild-type and pmrA Salmonella grown in N-minimal medium at pH 4.6 with 1 mM Mg²⁺ with various concentrations of iron (0.008, 0.04, 0.2, and 4 μM Fe³⁺) or without iron (B).

Next, we examined whether the enhanced ssaG expression resulting from inactivation of the pmrA gene was also observed when Salmonella was inside macrophages. This test was important because expression of SPI-2 genes greatly increases upon Salmonella phagocytosis (29, 30), and the PmrA/PmrB system is active inside macrophages (11). We found ssaG expression to be ~30-fold higher at 4 h after internalization by J774A.1 macrophages relative to expression displayed by bacteria grown in LB (Fig. 2C), which is in agreement with previous results (22, 29). Inactivation of the pmrA gene further increased ssaG expression (Fig. 2C), as observed during growth in minimal media (Fig. 2A). We note that the increased ssaG expression does not depend on the presence of a functional Nramp1 protein because J774A.1 macrophages originated from an Nramp1^−/− animal (31). Taken together, these results indicate that PmrA functions as a negative regulator of ssaG and potentially of other SPI-2 genes.

PmrA Represses ssaG Transcription Indirectly, by Hindering Expression of the ssaG Activator SsrB.

In principle, PmrA could repress ssaG transcription directly, by binding to the ssaG promoter, or, indirectly, by hindering transcription of an ssaG activator(s) or by furthering expression of an ssaG repressor(s). Sequences resembling a PmrA box (i.e., a direct repeat of the sequence CTTAAT separated by 5 nt present in PmrA-activated promoters; refs. 32–34) were not found in the ssaG promoter region. By contrast, we identified a perfect match to the consensus PmrA box downstream of the mapped transcriptional start site of the ssrB gene (Fig. 4A), which encodes an activator necessary for ssaG transcription (22). This analysis suggested that PmrA might bind to the ssrB promoter hampering ssrB transcription.

Open in a separate window

Fig. 4.

PmrA binds to the promoter of the ssrB gene. (A) DNA sequence of the promoter region of the ssrB gene. A +1 corresponds to the reported transcriptional start site, the boxed sequences indicate those matching the PmrA-binding consensus sequence, and the underlined sequence indicates the region protected by the PmrA protein in the ssrB promoter shown in B. Mutated nucleotides at the PmrA binding site in the ssrB promoter mutant strain are indicated beneath the sequence. (B) DNase I footprinting analysis of the ssrB promoter using purified PmrA-His6 protein (0, 100, 200, and 300 pmoles) was performed as described in Materials and Methods. Lane A+G corresponds to dideoxy chain-termination sequences for the ssrB promoter DNA. The black bar indicates the region protected by the PmrA-His6 protein. (C) In vivo occupancy of the ssrB, pmrC, and ssaG promoter regions by the PmrA-HA protein when Salmonella strains expressing PmrA-HA from its original chromosomal location and promoter were switched from noninducing conditions (N-minimal medium at pH 7.7 with 1 mM Mg²⁺) into noninducing or inducing conditions (N-minimal medium at pH 7.7 with 10 μM Mg²⁺ and 100 μM Fe³⁺). (D) β-galactosidase activity expressed by wild-type, ssrB, and pmrA Salmonella harboring a plasmid with a p_lac1–6-driven transcriptional fusion of the 5′ untranslated region (5′UTR) of ssrB or spiR to a promoterless lacZ gene. Bacteria were grown in N-minimal medium at pH 7.7 with 10 μM Mg²⁺. Values are normalized by that of the wild-type strain. (E) Fluorescence originating from a plasmid-linked ssaG-gfp transcriptional fusion was determined in wild-type, pmrA, ssrB, and ssrB pmrA Salmonella grown in N-minimal medium at pH 4.6 with 1 mM Mg²⁺ (LH) for 8 h.

The PmrA protein appears to repress ssrB directly because: First, the purified PmrA protein protected the −45 to −16 region from ATG of the ssrB gene in vitro (Fig. 4B), which contains the predicted PmrA binding site in the ssrB promoter (Fig. 4A). Second, chromatin immunoprecipitation experiments revealed that an HA-tagged PmrA expressed from its normal promoter and chromosomal location bound to the ssrB promoter in vivo but not to the ssaG promoter (Fig. 4C). Importantly, the enrichment in ssrB promoter DNA was comparable to that of other known PmrA-regulated genes, including pmrC (Fig. 4C). Third, substitution of four conserved nucleotides in the predicted PmrA box of the ssrB promoter (Fig. 4A) reduced PmrA binding to ssrB promoter DNA in vitro (Fig. 5A) and increased ssaG expression in vivo (Fig. 5B). Notably, the ssaG expression levels exhibited by the strain with the mutation of the PmrA-binding site in the ssrB promoter were similar to those displayed by the pmrA null mutant (Fig. 5B). Furthermore, nucleotide substitution of the PmrA-binding site in the ssrB promoter increased expression of both ssrB and ssaG within the J774A.1 macrophage-like cell line (Fig. 5C) and in bone marrow-derived macrophages prepared from C3H/HeN mice (Fig. S1). We note that the expression levels of the ssrB and ssaG genes in the ssrB promoter mutant were comparable to those in the pmrA mutant (Fig. 5C and Fig. S1) and that the ssrB promoter mutation had no effect on the expression of the PmrA-activated pmrC gene (Fig. 5C).

Open in a separate window

Fig. 5.

Binding of PmrA to the ssrB promoter is required for PmrA-mediated repression of the SPI-2 ssaG gene. (A) Electrophoretic mobility shift assays of wild-type or mutated ssrB promoter region DNA fragments using purified PmrA-His6 protein (0, 75, 150, 300, 600, and 1,200 pmoles) were carried out as described in Materials and Methods. The shifted bands can be competed out by the corresponding cold probe (unlabeled DNA) in the presence of the maximum amount of PmrA-His6 protein used. (B) Fluorescence from a plasmid-linked ssaG-gfp transcriptional fusion was determined in wild-type, pmrA mutant, the ssrB promoter mutant (p_ssrBmu), and pmrA ssrB promoter double mutant (pmrA p_ssrBmu) Salmonella grown in N-minimal medium at pH 4.6 with 1 mM Mg²⁺. (C) J774A.1 macrophages were infected with wild-type, pmrA mutant, ssrB mutant, and the ssrB promoter mutant (p_ssrBmu) strains. mRNA levels of the ssrB, ssaG, and pmrC genes were determined at the indicated times after infection.

PmrA’s action at the ssrB promoter is not simply to antagonize activation of ssrB transcription by the SsrB protein because the levels of β-galactosidase activity originating from a plasmid-linked ssrB-lacZ transcriptional fusion were higher in the pmrA mutant than in the wild-type strain and the ssrB mutant (Fig. 4D) even though this fusion was driven by the heterologous lac promoter derivative p_lac1–6 (35). The only Salmonella sequences present in the plasmid-linked ssrB-lacZ fusion were those located between the ssrB transcription start site and the ssrB start codon, which include the PmrA binding site. By contrast, inactivation of pmrA had no effect on the expression of a similar transcriptional fusion that included the leader region of the spiR gene instead of that of ssrB (Fig. 4D). In addition, similarly low fluorescence levels were displayed by isogenic ssrB and ssrB pmrA strains harboring a plasmid carrying a fusion between the ssaG promoter and a promoterless gfp (aav) gene (Fig. 4E). And inactivation of pmrA in the ssrB promoter mutant background did not alter the levels of ssaG expression (Fig. 5B). These results demonstrated that SsrB is necessary to activate the ssrB promoter even in the absence of PmrA repression. Finally, PmrA’s role as a repressor of ssrB transcription is not simply to prevent the PhoP and OmpR proteins from activating ssrB transcription (21, 36) because there were similarly low levels of ssaG expression in phoP, phoP pmrA, ompR, and ompR pmrA mutants (Fig. S2). Taken together, the results presented in this section indicate that PmrA’s effects on ssaG expression result from direct repression of ssrB transcription.

PmrA Delays Salmonella-Induced Macrophage Death.

To examine the physiological consequences of PmrA repression of ssrB transcription, we examined the kinetics of SPI-2–promoted macrophage death, which is critical for Salmonella virulence (37). We infected macrophages with various Salmonella strains and evaluated cell death by measuring lactate dehydrogenase released from dead or damaged cells. Cell death took place earlier in macrophages infected with the pmrA mutant than with the wild-type strain (Fig. 6). The faster cell death kinetics was also displayed by macrophages infected with the ssrB promoter mutant (Fig. 6). Control experiments showed reduced cell death in macrophages infected with the ssrB mutant (Fig. 6), in agreement with previous reports (38). These data indicate that the PmrA protein delays macrophage death induced by wild-type Salmonella.

Open in a separate window

Fig. 6.

PmrA delays SPI-2–promoted macrophage death. J774A.1 macrophages were infected with wild-type, pmrA, ssrB, and the ssrB promoter mutant (p_ssrBmu) Salmonella. The percentage of macrophage death was determined by detecting the amount of lactate dehydrogenase released into the supernatants at the indicated times after infection.

How PmrA Decreases Salmonella Virulence.

The antivirulence function of PmrA can be ascribed to its role as transcriptional repressor of the regulatory gene ssrB, which is essential for expression of genes required for proliferation inside macrophages and systemic infection in mice (23, 41). This antivirulence function is supported by the following: First, the hypervirulence phenotype of the pmrA null mutant could be phenocopied by mutating the PmrA binding site in the ssrB promoter (Fig. 1A). Second, a pmrA ssrB promoter double mutant exhibited the same hypervirulence phenotype of either single mutant (Fig. 1A). And third, SPI-2–dependent macrophage killing was accelerated in both the pmrA mutant and the ssrB promoter mutant (Fig. 6).

The SsrB protein is necessary for transcription of the genes specifying the T3SS encoded in the SPI-2 pathogenicity island as well as its secreted effectors (19, 20). Inactivation of the SPI-2 T3SS enhances Salmonella’s growth inside phagocytes during infection of an animal (42). These results suggest that PmrA repression of ssrB might enable Salmonella to control bacterial proliferation within a cell and to favor dissemination to neighboring cells. Alternatively or in addition, PmrA may enable proper repression of SPI-2 in the intestine, which is a PmrA-inducing environment (43, 44).

What is the identity of the SsrB-regulated gene(s) that must be tightly controlled for a normal course of Salmonella infection? We considered the possibility of the SsrB-repressed sciS being such a gene because its inactivation was reported to increase Salmonella virulence in BALB/c mice infected orally (4). However, we established that an sciS mutant is attenuated for virulence in C3H/HeN mice inoculated i.p. (Fig. S3), which is in contrast to the hypervirulence phenotype exhibited by the pmrA null mutant in the same mouse strain and route of inoculation (Fig. 1A).

Repression of ssrB, the master regulator of SPI-2 genes, constitutes a unique and unexpected function for the PmrA protein, whose best-described role was as the major activator of LPS modification genes (6, 16, 33). Whereas PmrA-dependent LPS modifications can reduce the risk of Salmonella detection by the Toll-like receptor-4 in a mammalian host (15, 45), its role in ssrB repression may reduce expression of Spi/Ssa-secreted effectors that dampen the host immune response (46, 47).

GFP Assay Salmonella.

Cells expressing a gfp variant from the ssaG promoter were grown in minimal media. The expression of the GFP was measured by using multidetector (VICTOR3; PerkinElmer) with OD₆₀₀ values. The measured values for GFP expression were divided by 1,000 times the OD₆₀₀ values.

We thank Tammy Latifi for the DNase I footprinting experiment; Ephraim Fass for constructing plasmid pFPV25AAV-p_ssaG; and Eunjin Lee, John May, and Nathan Schwalm for comments on the manuscript. This research was supported, in part, by National Institutes of Health Grant AI042236 (to E.A.G.). E.A.G. is an Investigator of the Howard Hughes Medical Institute.