Programmed Delay of a Virulence Circuit Promotes Salmonella Pathogenicity

To investigate the expression of EIIA^Ntr, we first investigated its transcription levels by measuring β-galactosidase activity produced by a p_rpoN-lacZ fusion given that the ptsN gene is located downstream of the rpoN gene, forming an operon (21). Although EIIA^Ntr is a component of a nitrogen-metabolic PTS (21, 22), transcription levels of rpoN remained unaltered by 100-fold changes in the concentration of a nitrogen source (Fig. 1A). We next examined rpoN expression at different pH values or concentrations of Mg²⁺, representing environmental conditions that Salmonella might encounter during infection (6, 14). However, none of those changes modified the expression of rpoN (Fig. 1B and C).

Despite the absence of notable changes in p_rpoN-lacZ expression, we examined EIIA^Ntr protein amounts under these conditions. Similar to rpoN expression, EIIA^Ntr amounts were not responsive to changes in nitrogen source (Fig. 1D). Surprisingly, however, EIIA^Ntr abundance was significantly reduced when Salmonella was exposed to low-Mg²⁺ or acidic-pH conditions (Fig. 1E and F).

This rpoN-independent alteration of EIIA^Ntr levels (Fig. 1A to F) raised the possibility that transcription of the ptsN gene might not just be from the rpoN promoter. Indeed, primer extension analysis indicated the presence of a transcriptional start site 67 nucleotides upstream of the EIIA^Ntr start codon (Fig. 1G). Therefore, we investigated the expression of a p_ptsN-lacZ transcriptional fusion in bacteria grown under the above-described conditions. However, none of those conditions altered the expression of p_ptsN-lacZ (Fig. 1H to J). Taken together, these results suggest that Salmonella probably controls EIIA^Ntr levels through posttranscriptional regulatory mechanisms.

Given that an acidic pH and low Mg²⁺ are signals activating the sensor PhoQ (12–14), we hypothesized that the PhoP/PhoQ system might be involved in altering EIIA^Ntr amounts. To test this hypothesis, we examined the transcription and translation levels of EIIA^Ntr in isogenic wild-type and the phoP mutant Salmonella strains. Consistent with the expression of the p_ptsN-lacZ fusion gene from a plasmid (Fig. 1H to J), the chromosomal ptsN-lacZ fusion also showed similar β-galactosidase activities under acidic and neutral pH conditions (see Fig. S1A in the supplemental material). Moreover, mutation of the phoP gene did not alter ptsN expression (Fig. S1A). In contrast, EIIA^Ntr amounts were significantly higher in the phoP null mutant than in the wild type when grown under acidic conditions (Fig. 2A), indicating that PhoP reduces EIIA^Ntr amounts independent of its transcription. The increased EIIA^Ntr abundance in the phoP mutant was restored by a plasmid expressing PhoP from a heterologous promoter (Fig. 2A). The absence of the cognate sensor kinase PhoQ also coordinated with a higher abundance of EIIA^Ntr (Fig. S1B) as in the phoP mutant (Fig. 2A), indicating that PhoP’s action in altering EIIA^Ntr abundance is dependent on PhoP’s phosphorylation. Furthermore, the lack of PhoP increased EIIA^Ntr abundance (Fig. 2B) even when transcription of ptsN was induced by isopropyl-β-d-thiogalactopyranoside (IPTG), further supporting the notion that PhoP modulates the abundance of EIIA^Ntr posttranscriptionally.

We next investigated how PhoP controls EIIA^Ntr levels posttranscriptionally. PhoP may decrease EIIA^Ntr levels by reducing the stabilities of ptsN mRNA and/or EIIA^Ntr protein. ptsN mRNA showed similar levels of decay in both the wild type and the isogenic phoP mutant upon addition of rifampin to stop transcription (Fig. 2C). In contrast, the amount of EIIA^Ntr decreased in the wild type after inhibition of protein synthesis with chloramphenicol treatment (half-life [t_1/2] < 60 min), whereas it remained constant in the strain lacking PhoP (t_1/2 > 120 min) (Fig. 2D). Furthermore, this EIIA^Ntr degradation was detected when Salmonella was grown at an acidic pH but not at a neutral pH (Fig. S2). These results suggest that PhoP boosts the degradation of EIIA^Ntr protein in an acidic environment.

Cytoplasmic proteases, including ClpXP and Lon, are involved in the proteolysis of cytosolic proteins in Gram-negative bacteria (29), and EIIA^Ntr is a cytoplasmic protein. As PhoP counteracts ClpXP-mediated proteolysis of RpoS via IraP (30), we first investigated the potential role of ClpXP in controlling EIIA^Ntr abundance. A Salmonella strain lacking ClpXP produced amounts of EIIA^Ntr comparable to those of the wild type, unlike the phoP mutant (Fig. 2E). However, a lack of Lon increased the abundance of EIIA^Ntr protein compared with that of the wild type, similar to the case with the phoP mutant strain (Fig. 2E). If Lon is responsible for the degradation of EIIA^Ntr, the lon mutant should make EIIA^Ntr stable. Like the phoP mutant (Fig. 2D), the lon mutant displayed sustained abundance of EIIA^Ntr after chloramphenicol treatment (t_1/2 > 120 min), whereas EIIA^Ntr levels dwindled in the wild type (t_1/2 < 60 min) (Fig. 2F). Furthermore, double deletion of the phoP and lon genes resulted in amounts of EIIA^Ntr comparable to those in the phoP or lon single-deletion mutants (Fig. S3). These results suggest that PhoP favors Lon protease-mediated degradation of EIIA^Ntr.

We next wondered why Salmonella curtails EIIA^Ntr amounts when it encounters an acidic pH, a PhoP-inducing condition inside macrophage phagosomes (31). To understand the role of EIIA^Ntr, we investigated genes that are regulated by EIIA^Ntr using a DNA microarray experiments with wild-type and isogenic ptsN mutant strains grown in acidified minimal medium. We found 768 differentially expressed genes in the ptsN mutant compared with the wild type (>2-fold): 371 upregulated genes and 397 downregulated genes (Fig. S4). Consistent with a previous report (28), SPI-2 genes were more highly expressed in the ptsN mutant than in the wild type (Table S1). Interestingly, we found that transcript levels of PhoP-regulated genes were higher in the strain lacking EIIA^Ntr than in the wild type (Table S1). We further verified the EIIA^Ntr’s regulatory effects on PhoP-activated genes using quantitative reverse transcription-PCR (qRT-PCR): the ptsN mutant displayed 3- to ∼7-fold-higher transcript levels of PhoP-regulated genes than the wild type (Fig. 3A). Moreover, the elevated expression of those genes in the ptsN mutant was restored to wild-type levels by a plasmid expressing the ptsN gene from a heterologous promoter but not by the plasmid vector (Fig. 3A). Interestingly, plasmid-driven heterologous expression of an unphosphorylatable variant of EIIA^Ntr (H73A) or a variant of EIIA^Ntr mimicking the phosphorylated form (H73E) (25) was also able to rescue the expression of PhoP-regulated genes similarly to wild-type EIIA^Ntr (Fig. 3A). These results suggest that EIIA^Ntr modulates the expression of PhoP target genes independent of EIIA^Ntr’s phosphorylation status and PhoP transcription.

We next wondered whether EIIA^Ntr controls expression of PhoP to regulate PhoP regulon. If EIIA^Ntr directly controls PhoP-activated genes, EIIA^Ntr should be able to regulate those genes in the absence of PhoP. However, the absence of PhoP abrogated the regulatory effects of EIIA^Ntr on the expression of PhoP-activated genes (Fig. 3B), indicating that control of the PhoP regulon by EIIA^Ntr requires PhoP. If the regulation of PhoP-regulated genes by EIIA^Ntr is due to altered phoP transcription, heterologous expression of phoP from a plasmid should abolish the effect of EIIA^Ntr on the expression of those genes. A lack of EIIA^Ntr increased expression levels of PhoP-regulated genes, even when PhoP was produced from a heterologous promoter (Fig. 3C). These results indicate that EIIA^Ntr regulates PhoP regulon in a PhoP-dependent manner.

Given that EIIA^Ntr regulates other regulatory systems via protein-protein interaction (23, 24, 28, 32), we next investigated whether EIIA^Ntr interacts with the PhoP protein. We used the bacterial two-hybrid system, in which β-galactosidase levels are dependent on the proximity of fused proteins to fragments (i.e., T25 and T18) of the Bordetella pertussis adenylate cyclase in an Escherichia coli strain lacking its own adenylate cyclase (33). Coexpression of T25-EIIA^Ntr and T18-PhoP resulted in approximately 141-fold-higher levels of β-galactosidase activity than in strains expressing T25–EIIA^Ntr and T18 fragment or empty vectors (Fig. 4A), indicating that EIIA^Ntr interacts with PhoP. This activity was comparable to that from the positive-control strain harboring T25 and T18 fragments fused to the leucine zipper of the transcription factor GCN4 (Fig. 4A). Consistent with the observation that unphosphorylatable EIIA^Ntr (H73A) functions like the wild type in controlling the PhoP regulon (Fig. 3A), EIIA^Ntr (H73A) displayed an interaction with PhoP similar to that of the wild-type protein (Fig. 4A).

PhoP promotes transcription of the PhoP regulon by binding to DNA (34) when PhoP is activated by PhoQ-mediated phosphorylation under inducing conditions (35) or by reducing acetylation of PhoP (36). Thus, the interaction of EIIA^Ntr with PhoP could decrease PhoP activity by inhibiting the interaction of PhoP with the cognate kinase PhoQ (i.e., reducing phosphorylation), by promoting acetylation of PhoP, or by reducing deacetylation of PhoP. Alternatively, EIIA^Ntr could interfere with PhoP binding to DNA.

If EIIA^Ntr hampers PhoP phosphorylation by PhoQ, the lack of PhoQ should abolish the regulatory effects of EIIA^Ntr on the PhoP regulon. Because PhoP is not active in the absence of PhoQ, we investigated the function of EIIA^Ntr in a phoP* phoQ strain lacking PhoQ and expressing a PhoP variant that autophosphorylates from acetyl phosphate (35). EIIA^Ntr reduced pagD expression even in the absence of PhoQ (Fig. S5), indicating that the regulatory action of EIIA^Ntr is independent of PhoQ. We next examined whether acetylation of PhoP is responsible for EIIA^Ntr-mediated regulation of PhoP target genes by mutating known acetylase (Pat) or deacetylase (CobB) (36). However, EIIA^Ntr displayed similar regulatory effects on pagD expression in the absence of Pat or CobB (Fig. S5).

To test whether EIIA^Ntr inhibits PhoP’s DNA binding ability, a gel shift assay was conducted using purified PhoP and EIIA^Ntr proteins with the phoP-activated pagD promoter. Purified PhoP bound to the pagD promoter DNA and formed a complex with the probe DNA in vitro (Fig. 4B). EIIA^Ntr prevented PhoP from binding to the target DNA: the PhoP-DNA complex decreased to generate the unbound pagD promoter DNA when amounts of EIIA^Ntr increased (although an excess of EIIA^Ntr could not fully dissociate PhoP from DNA), and EIIA^Ntr alone did not form a complex with the DNA (Fig. 4B). However, addition of an EIIA^Ntr paralogue, EIIA^Glc, did not alter binding of PhoP to DNA, indicating that it is specific to EIIA^Ntr (Fig. S6A). This inhibitory function of EIIA^Ntr is specific, as another regulatory protein, PmrA, bound to the pbgP promoter DNA regardless of EIIA^Ntr (Fig. S6B). Taken together, these data suggest that EIIA^Ntr reduces expression of the PhoP regulon by inhibiting PhoP binding to its target promoter DNA.

We next examined if EIIA^Ntr could control the expression of PhoP-activated genes during infection. To evaluate this, macrophages were infected with wild-type and mutant Salmonella strains harboring a gfp fusion with the promoter of the PhoP-activated gene pagD, and fluorescence was measured. Activation of pagD expression inside macrophages was completely dependent on PhoP, because the phoP mutant was unable to produce any fluorescence, in contrast to the wild type (Fig. 5A). The ptsN mutant showed higher and earlier activation of pagD expression than the wild type (Fig. 5A), indicating that EIIA^Ntr inhibits PhoP activation inside macrophages. These results suggest that Salmonella delays expression of PhoP-activated genes inside macrophages via EIIA^Ntr. The ptsN mutant also displayed accelerated activation of the pagD gene compared to that of the wild type in acidic pH (Fig. S7A). Consistent with a previous report (28), the lack of EIIA^Ntr rendered Salmonella virulence attenuated in mice inoculated via the intraperitoneal route (Fig. 5B, left). Defective virulence of the ptsN mutant Salmonella was also observed when mice were inoculated via the oral route (Fig. 5B, right). These results are in agreement that EIIA^Ntr controls the expression of various virulence genes, including the PhoP and SsrB regulons (Fig. 3) (28). And it is possible that the delayed virulence gene expression by EIIA^Ntr might be critical for Salmonella pathogenicity.

The Salmonella enterica serovar Typhimurium strains used in this study were derived from strain SL1344. The strains and plasmids used in this study are listed in Table S2A. Phage P22-mediated transduction was performed as described previously (52). All Salmonella strains were grown aerobically at 30 or 37°C in Luria-Bertani (LB) or M9 minimal medium at the desired pH and Mg²⁺ concentrations to mid- to late log phase unless specified. Antimicrobial peptide C18G was treated at 5 µg/ml for an hour. Antibiotics were used at the following concentrations: ampicillin, 50 μg/ml; chloramphenicol, 25 μg/ml; and kanamycin, 50 μg/ml. Primers used for the construction of bacterial strains and plasmids are listed in Table S2B.

To generate a ptsN-FLAG strain, a cat cassette was introduced in the 3′ end of the ptsN gene as follows: the cat fragment was amplified from pKD3 using primers ptsN-FLAG-F/ptsN-FLAG-R and then introduced into wild-type Salmonella (SL1344) harboring plasmid pKD46 as previously described (53). The cat cassette was removed with plasmid pCP20 (53).

To generate a ptsP mutant, a cat fragment was amplified from pKD3 using primers ptsP-Red-F/ptsP-Red-R and then introduced into wild-type Salmonella harboring plasmid pKD46 (53).

To generate a phoP mutant strain, a kan fragment was amplified from pKD13 using primers phoP-Red-F/phoP-Red-R and then introduced into wild-type Salmonella harboring plasmid pKD46 (53). Next, the kan cassette was removed with plasmid pCP20 (53).

To generate a phoQ mutant strain, a kan fragment was amplified from pKD13 using primers phoQ-Red-F/phoQ-Red-R and then introduced into wild-type Salmonella harboring plasmid pKD46 (53). Next, the kan cassette was removed with plasmid pCP20 (53).

To generate a cobB mutant strain, a kan fragment was amplified from pKD13 using primers cobB-P1-F-kan/cobB-P4-R-kan and then introduced into wild-type Salmonella harboring plasmid pKD46 (53).

To generate a pat mutant strain, a kan fragment was amplified from pKD13 using primers pat-P1-F-kan/pat-P4-R-kan and then introduced into wild-type Salmonella harboring plasmid pKD46 (53).

To generate a clpXP mutant, a cat fragment was amplified from pKD3 using primers clpP-Red-F/clpX-Red-R2 and then introduced into wild-type Salmonella harboring plasmid pKD46 (53). Next, the kan cassette was removed with plasmid pCP20 (53).

To generate a lon mutant, a cat fragment was amplified from pKD3 using primers lon-Red-F/lon-Red-R and then introduced into wild-type Salmonella harboring plasmid pKD46 (53). Next, the kan cassette was removed with plasmid pCP20 (53).

To generate a strain with the p_ptsN-lacZ fusion in the normal chromosomal location, pCP20 was introduced into SR3203 (the ptsN mutant). Next, the lacZ fusion was generated with plasmid pCE70 (54).

To generate a strain with the p_pagD-lacZ fusion in the normal chromosomal location, a cat fragment was amplified from pKD3 using primers pagD-1/pagD-2 and then introduced into wild-type Salmonella (SL1344) harboring plasmid pKD46 as previously described (53). The cat cassette was removed with plasmid pCP20 (53). Next, the lacZ fusion was generated with plasmid pCE70 (54).

A plasmid expressing ptsN-FLAG was constructed as follows: the ptsN-FLAG fragment was amplified from Salmonella expressing ptsN-FLAG (SR4045) using primers ptsN-pF2/ptsN-pR2 and then introduced between the EcoRI and BamHI sites of pUHE21-2lacI^q (55).

A plasmid expressing the phoP gene was constructed as follows: the phoP coding region was amplified from wild-type Salmonella (SL1344) using primers phoP-com-F2/phoP-com-R2 and then introduced between the EcoRI and BamHI sites of pUHE21-2lacI^q (55).

A plasmid expressing phoQ gene was constructed as follows: the phoQ coding region was amplified from wild-type Salmonella (SL1344) using primers phoQ-com-F/phoQ-com-R and then introduced between the EcoRI and BamHI sites of pUHE21-2lacI^q (55).

Plasmids expressing EIIA^Ntr variants [EIIA^Ntr (H73A) and EIIA^Ntr (H73E)] were constructed as follows: the pJJ14 plasmid was mutated using a QuikChange Lightning site-directed mutagenesis kit (Agilent Technologies) with primers ptsN_H73A-F/ptsN_H73A-R for H73A substitution and ptsN_H73A-F/ptsN_H73A-R for H73E substitution.

A plasmid expressing His₆-tagged PhoP was constructed as follows: the phoP coding region was amplified from wild-type Salmonella (SL1344) using primers pPhoP-F/pPhoP-His6-R and then introduced between the BamHI and HindIII sites of pUHE21-2lacI^q (55).

A plasmid expressing His₆-tagged PmrA was constructed as follows: the pmrA coding region was amplified from wild-type Salmonella (SL1344) using primers pPmrA-F/pPmrA-His6-R and then introduced between the BamHI and HindIII sites of pUHE21-2lacI^q (55).

Plasmids containing promoter fusions were constructed as follows: the rpoN and ptsN promoter regions were amplified from wild-type Salmonella (SL1344) using primers rpoN-pF1/rpoN-pR1 and ptsN-pF4/ptsN-pR4, respectively. Next, they were introduced between the EcoRI and BamHI sites of pRS415 (56). The pagD promoter region was amplified from wild-type Salmonella (SL1344) using primers PpagD-gfp-F/PpagD-gfp-R and then introduced between the EcoRI and BamHI sites of pFPV25 (57).

A plasmid expressing T18-PhoP fusion protein was constructed as follows: the phoP gene was amplified from wild-type Salmonella (SL1344) using primers phoP-F2/phoP-R1 and then introduced between the BamHI and EcoRI sites of pUT18C (58).

Salmonella strains expressing the EIIA^Ntr-FLAG protein from its normal chromosomal location or under the control of a heterologous promoter were grown as described in “Bacterial strains, plasmids, and growth conditions,” above. Bacteria were collected by centrifugation, and cell lysates were prepared using B-PER solution (Pierce). Cell lysates were separated by 12% SDS-PAGE, and EIIA^Ntr and DnaK were detected using anti-FLAG (Sigma) and anti-DnaK (Abcam) antibodies, respectively. Blots were developed using anti-mouse IgG horseradish peroxidase-linked antibody with the ECL detection system (Amersham Biosciences).

β-Galactosidase assays were carried out in triplicate, and the activity was determined as described previously (59).

Salmonella strains were grown as described above, and total RNA was isolated using an RNeasy minikit (Qiagen). After DNase treatment of the isolated RNA, cDNA was synthesized using Omnitranscript reverse transcription reagents (Qiagen) and random hexamers (Invitrogen). Quantification of the cDNA was carried out using 2× iQ SYBR Green Supermix (Bio-Rad), and real-time amplification of the PCR products was performed using the iCycler iQ real-time detection system (Bio-Rad). The primers used for detection of the gene transcripts are listed in Table S2B . Data were normalized to the abundance of 16S rRNA expression levels.

To test RNA stability, bacterial cultures were treated with 0.1 mg/ml of rifampin to stop transcription, and samples were collected at the desired time points. Total RNA was isolated, and mRNA levels were determined by qRT-PCR as described above. To test the protein stability, bacterial cultures were treated with 0.2 mg/ml of chloramphenicol to stop protein synthesis, and samples were collected at the desired time points. Protein levels were analyzed using Western blot analysis.

E. coli BTH101 organisms harboring derivatives of plasmids pUT18 and pKT25 were grown overnight in LB broth containing ampicillin (100 µg/ml) and kanamycin (50 µg/ml), adding a 1:100 dilution to 1 ml of the same fresh medium containing 0.5 mM IPTG and employing shaking at 30°C overnight as previously described (33, 58).

His₆-tagged PhoP, His₆-tagged PmrA and His₆-tagged EIIA^Ntr were expressed in E. coli BL21(DE3). Bacterial cells were grown in LB medium at 37°C until the optical density at 600 nm (OD₆₀₀) reached 0.5, and the expression of those proteins was induced by addition of IPTG (0.5 M) followed by growth at 30°C for 5 h. Cells were harvested, washed, and suspended in buffer A (20 mM Tris [pH 8.0], 150 mM NaCl, and 20 mM imidazole). Then the cells were disrupted by sonication, and cell debris was removed by centrifugation at 20,000 × g at 4°C for 30 min. The supernatant was applied to a 1.5-ml nickel- nitrilotriacetic acid (Ni-NTA) agarose column equilibrated in buffer A, washed with a 25-column volume of the same buffer, and eluted using a gradient of buffer A and buffer B (20 mM Tris [pH 8.0], 150 mM NaCl, and 250 mM imidazole). The fractions were then collected and analyzed by SDS-PAGE, and selected fractions were dialyzed against buffer C (20 mM Tris [pH 8.0], 150 mM NaCl, and 10% glycerol).

DNA fragments containing the promoter region of the pagD or pbgP gene were amplified by PCR using the primers EMSA-pagD-F/EMSA-pagD-R and EMSA-pbgP-F/EMSA-pbgP-R, respectively. Purified promoter DNA (80 fmol) was incubated with the desireded concentrations of purified PhoP-His₆ or PmrA-His₆ with EIIA^Ntr-His₆ or EIIA^Glc-His₆ at room temperature for 20 min in 15 μl of binding buffer (10 mM Tris [pH 7.5], 0.5 mM EDTA, 1 mM MgCl₂, 0.5 mM dithiothreitol [DTT], and 50 mM NaCl) containing 5 ng/µl of poly(dI-dC). Samples were prepared by addition of 3 μl of 6× electrophoretic mobility shift assay (EMSA) gel loading solution and separated by electrophoresis using a 6% nondenaturing polyacrylamide gel. DNA staining was performed according to the manufacturer’s instructions (EMSA kit; E33075; Thermo Fisher Scientific).

The murine-derived macrophage line RAW264.7 was cultured in Dulbecco’s modified Eagle’s medium (DMEM; Life Technologies) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Life Technologies) at 37°C with 5% CO₂. Macrophages were seeded in 24-well tissue culture plates at 5 × 10⁵ per well 1 day before infection with Salmonella. Confluent monolayers were inoculated with bacterial cells that had been grown overnight in LB broth, washed with phosphate-buffered saline (PBS), and resuspended in 0.1 ml of prewarmed DMEM at a multiplicity of infection of 20. Following a 30-min incubation, the wells were washed three times with prewarmed PBS to remove extracellular bacteria and then incubated with prewarmed medium supplemented with 100 µg/ml of gentamicin for 1 h to kill extracellular bacteria. Next, the wells were washed three times with prewarmed PBS and incubated with prewarmed medium supplemented with 10 µg/ml of gentamicin. At the desired time points, the cells were washed three times with prewarmed PBS and subjected to the following procedures. For green fluorescent protein (GFP) assessment, washed cells were scraped with 200 µl of PBS and subjected to fluorescence measurements at 510 nm. For CFU measurements, washed cells were lysed with PBS containing 1% Triton X-100 and plated on LB agar plate at the proper dilutions.

Six-week-old female BALB/c mice were purchased from the Institute of Laboratory Animal Resources at Seoul National University. Five mice in each group were infected intraperitoneally or orally with 0.1 ml of PBS containing approximately 10² or 10⁶ Salmonella cells grown in LB broth overnight, respectively. All animals were housed in temperature- and humidity-controlled rooms and maintained on a 12-h light/12-h dark cycle. All procedures complied with the regulations of the Institutional Animal Care and Use Committee of Seoul National University.

RNA labeling, hybridization to the microarrays, scanning, and data analysis were performed at Macrogen. Triplicates of total RNAs from wild-type and ptsN mutant strains grown in acidified M9 medium (pH 5.8) were purified as described above and subjected to microarray using a CombiMatrix chip for the Salmonella Typhimurium SL1344 genome (12,396 probes covering 4,441 genes). Arrays were scanned using the Axon GenePix 4000B scanner (Molecular Devices LLC). Image analysis and feature extraction were performed using Axon GenePix Pro software (Molecular Devices). The data were analyzed using Avadis Prophetic software version 3.3 (Strand Genomics). Fold changes were calculated by comparing averaged normalized signal intensities in wild-type versus ptsN mutant Salmonella. The t test was performed in parallel with the use of a false-discovery rate correction for multiple testing (60). A P value of <0.05 was used to pinpoint significantly different expression levels of genes. A cutoff of a 2-fold change for up- or downregulated expression was chosen to define genes that were differentially expressed.

Reverse transcription was conducted using ptsN-P1 and Superscript II (Invitrogen). The ladder was generated with a template DNA that was amplified using primers ptsN-PE-F/ptsN-PE-R and the genomic DNA of SL1344.