The LysR-Type Transcriptional Regulator LeuO Controls Expression of Several Genes in Salmonella enterica Serovar Typhi

The bacterial strains and plasmids used are listed in Table S1 in the supplemental material. Salmonella serovar Typhi was grown in LB medium (10 g of tryptone per liter, 5 g of yeast extract per liter, 10 g of NaCl per liter) or in liquid MA medium (7 g of nutrient broth per liter, 1 g of yeast extract per liter, 2 g of glycerol per liter, 3.75 g of K₂HPO₄ per liter, 1.3 g of KH₂PO₄ per liter) (21). When required, the following antibiotics were added: kanamycin (30 μg/ml), tetracycline (12 μg/ml), and ampicillin (100 μg/ml). Salmonella serovar Typhi and E. coli strains were grown aerobically at 37°C.

Plasmid isolation and genomic DNA isolation were performed using previously described protocols (47). Primers for PCR amplification were provided by the oligonucleotide synthesis facility at our institute (see Table S2 in the supplemental material). Restriction enzymes, ligase, kinases, nucleotides, and polymerases were obtained from New England Biolabs or Gibco BRL. For sequencing, double-stranded DNA was purified with a High Pure plasmid isolation kit (Boehringer, Mannheim, Germany), and sequencing was performed with an automatic Perkin Elmer/Applied Biosystems 377-18 system.

Salmonella serovar Typhi harboring plasmid pFMTrc12 (strain IMSS-II) (12) or plasmid pFMTrcleuO-50 (strain IMSS-III) (5) and Salmonella serovar Typhi ΔleuO harboring plasmid pFMTrc12 (strain IMSS-32) were grown in MA medium supplemented with ampicillin and isopropyl-β-d-thiogalactopyranoside (IPTG) (50 μM) to an optical density at 595 nm (OD₅₉₅) of 0.6. Salmonella cultures (100 ml) were pelleted and washed with 1× phosphate-buffered saline. Cellular proteins were obtained by sonication at 24 kHz for 1 min in the on position and 1 min in the off position for five cycles at 4°C using a Vibra Cell (Sonics, United States) in the presence of a protease inhibitor (Complete tablets; Roche Diagnostics GmbH, Mannheim, Germany). To further limit proteolysis, protein isolation was performed using phenol extraction (17). To solubilize proteins and to obtain completely denatured and reduced proteins, pellets were dried and resuspended as previously reported (6). Prior to electrophoresis samples were mixed with 7 M urea, 2 M thiourea, 4% 3-[(3-choloamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS) (Roche Diagnostics GmbH, Germany), 2 mM tributhyl phosphine, 2% ampholytes, and 60 mM dithiothreitol.

The methods used for sample preparation, analytical two-dimensional gel electrophoresis (2-DGE), image analysis, and preparative 2-DGE have been described previously (7), and pH gradients were determined using a two-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis standard (Sigma, United States). For isoelectric focusing 500 μg of total protein was loaded. All gel experiments were performed more than three times.

Selected spots from Coomassie blue-stained preparative two-dimensional gels were excised manually and frozen at −70°C until they were used. Samples were prepared for mass spectrum analysis by using a slight modification of a previously described procedure (6, 7, 49). Protein spots were destained, reduced, alkylated, and digested with trypsin (Promega, Madison, WI). Before the mass spectra of the peptide mixtures were obtained, the mixtures were desalted using a C₁₈ Zip Tip (Millipore, Bedford, MA) according to the manufacturer's recommendations.

Mass spectra were obtained using a Bruker Daltonics Autoflex (Bruker Daltonics, Billerica, MA) operated in the delayed extraction and reflectron mode. Spectra were externally calibrated using a peptide calibration standard (Bruker Daltonics 206095). Peptide mixtures were analyzed using a saturated solution of alpha-cyano-4-hydroxycinnamic acid in 50% acetonitrile-0.1% trifluoroacetic acid. Peak lists of the tryptic peptide masses were generated and searched against the NCBI nr databases using the Mascot search program (Matrix Science, Ltd., London, United Kingdom; http://www.matrixscience.com ). Proteins not identified by matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) analysis were excised from preparative gels and analyzed by liquid chromatography-MS/MS using a nanoflow chromatograph (Agilent 1100 LC NanoPump; Agilent, Waldbronn, Germany) coupled to a hybrid triple quadrupole linear ion trap (3200 Q TRAP system; Applied Biosystems/MDS Sciex, Ontario, Canada) equipped with a Nanospray II source and using information-dependent acquisition. Digested proteins were separated using a 60-min acetonitrile-0.1% formic acid elution gradient (Mallinckrodt Baker, United States) and a flow rate of 0.4 μl/min with a Zorbax C₁₈ column (75 μm by 150 mm; 3.5 μm). Precursor ions were determined using an enhanced MS scan over a mass range from m/z 400 to 1,500 at 4,000 atomic mass units/s (with no trapping in Qo and a LIT fill time of 20.00 ms) with an ion spray voltage of 2,400 V applied to a Picotip (FS360-75-15; New Objective, Woburn, MA) with ion spray gas (nitrogen). Precursors ions were collided in Q2 using rolling collision energy (maximum allowed collision energy, CE-80). Enhanced product ion scans (MS/MS) were performed over a mass range from 50 to 1,700 m/z at 4,000 atomic mass units/s, and collision voltages were determined dynamically. The entire precursor ion mass/charge ratio was confirmed with an Enhance Resolution scan. Proteins were identified by using the Mascot algorithm (available at http://www.matrixscience.com ) (59).

To evaluate the LeuO concentration used in this work, Salmonella serovar Typhi IMSS-III was grown in MA medium (supplemented with 50 μM IPTG) to an OD₅₉₅ of 0.6; a total-extract sample was collected, boiled, and loaded on a 12.5% sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel. Proteins were transferred to nitrocellulose membranes using the Trans-Blot SD system (Bio-Rad). Membranes were probed using rabbit anti-LeuO His-tagged polyclonal antibody. Protein concentrations were determined with the Bradford protein assay reagent using bovine serum albumin as a standard (Bio-Rad). To calculate the LeuO concentration, a comparative analysis of the LeuO blot signals from the total extract and from various amounts of the LeuO purified protein was performed.

Oligonucleotides (see Table S2 in the supplemental material) were designed to PCR amplify the complete 5′ intergenic regions of the genes regulated by LeuO. PCR fragments were double digested with BamHI-KpnI or BamHI-HindIII and ligated into pKK232-8 or pKK232-9 (see Table S1 in the supplemental material), which contained the promoterless cat gene. Fusions were sequenced in order to verify the DNA sequences of the PCR fragments.

The chloramphenicol acetyltransferase (CAT) assays were performed as follows (31). Salmonella serovar Typhi strains were grown in MA medium supplemented with ampicillin or kanamycin and with IPTG (50 μM) to OD₅₉₅ of 0.4, 0.6, 0.8, and 1 and for 12 h. Salmonella serovar Typhi STYhns99 was also grown in MA medium. Portions (1.5 ml) of bacterial cultures were collected by centrifugation and washed with 0.8 ml of TDTT buffer (50 mM Tris-HCl [pH 7.8], 30 μM dl-dithiothreitol). Bacterial cells were resuspended in 0.6 ml of TDTT buffer and sonicated on ice using 10-s bursts with 10-s rest periods until the extract was clear. The homogenate was centrifuged, and the supernatant was used to measure activity. For CAT assays 5 μl of each extract was added in duplicate to a 96-well enzyme-linked immunosorbent assay plate, which was followed by addition of 0.2 ml of a reaction mixture that contained 1 mM 5,5′-dithiobis(2-nitrobenzoic acid), 0.1 mM acetyl coenzyme A, and 0.1 mM chloramphenicol in 0.1 M Tris-HCl (pH 7.8). The absorbance at 412 nm was measured every 5 s for 5 min using a scanning autoreader and Ceres 900 microplate workstation. The protein concentrations of the cell extracts used in the CAT assays were determined using the bicinchoninic acid protein assay reagent (Pierce). The protein values and the mean rate of product formation by cat were used to determine the CAT specific activity, which was expressed in micromoles per minute per milligram of protein. The results presented below are the means of three independent experiments.

Salmonella serovar Typhi strains were grown at 37°C in MA medium for 12 h. Bacterial cells (5 ml) were collected, and total RNA was isolated using a commercial kit (RNeasy; Qiagen). The concentration of RNA was determined by measuring the absorbance at 260 nm. The integrity of RNA was determined by using a 1.5% agarose gel. Five to 20 μg of total RNA was denatured at 95°C for 3 min and then slowly cooled to 45°C. The RNA was annealed with [γ-³²P]ATP-labeled primers. Primers were extended with reverse transcriptase at 42°C for 1 h, and then the extended products were purified by ethanol precipitation and analyzed by electrophoresis in 8% polyacrylamide-8 M urea gels alongside sequencing ladders (39). Sequencing ladders were generated from plasmids that contained the regulatory regions of the LeuO-controlled genes.

DNA fragments encompassing the promoter regions of the LeuO-controlled genes were PCR amplified using labeled primers. LeuO was purified as previously described (12). Binding of the recombinant LeuO protein to the labeled promoter was performed at room temperature for 30 min in a reaction buffer containing 20 mM HEPES (pH 7.9), 100 mM KCl, 2 mM MgCl₂, 0.1 mM EDTA, 0.1 mM dithiothreitol, 20% glycerol, and 1.0 μg of poly(dI-dC) (Boehringer)/ml. The reaction mixture was treated without DNase I and with 0.4 U of DNase I. After precipitation with ethanol, the digested DNA fragments were separated by electrophoresis in 8% polyacrylamide-8 M urea gels alongside sequencing ladders. Sequencing ladders were generated, with the labeled primer used for PCR amplification, from plasmids containing the regulatory regions of the LeuO-controlled genes.

To determine whether LeuO recognizes a specific DNA sequence, two bioinformatics approaches were used: pattern matching and pattern discovery. In both approaches, we used 400 bp upstream and 50 bp downstream of the putative ATG translational start codon. In the first method, the experimental LeuO regions protected from DNase I were used as a training set to run the consensus program (16); thus, we obtained alignments and weight matrices of binding sites that were 8 to 12 bp long. To assign match scores, we used an “alphabet” based on the frequency of occurrence of each base in the regulatory regions analyzed. The search was conducted with a single strand. The weight matrix generated (highest information content) was used to search binding sites. Using patser, we searched for the highest-scoring sequence in each region of the training set. To evaluate the quality of the matrix, we evaluated different sets, including upstream regions of LeuO-regulated genes, 1,000 groups of six randomly selected regions, and non-LeuO-regulated genes. The mid-value of the scores was used as a threshold because negative scores were obtained with the training set. The second method, pattern discovery, assumed that LeuO binding sites are unknown; thus, the regulated LeuO sequences together with the Oligo and dyad analysis programs were used to identify a motif in the LeuO interacting sites (57). To evaluate this approach, we counted the number of times that we could locate a known binding site in a collection of upstream regions. The dyad and Oligo analysis programs were designed to find overrepresented “small words” (57). Overrepresented “words” would be expected to occur at the binding sites, and thus we determined whether the resulting Oligo or dyad analysis was in agreement with the LeuO sites detected experimentally. The parameters used to determine the LeuO consensus sequence are available at http://www.ccg.unam.mx/Computational_Genomics/tools/LeuO/ .

In order to define the LeuO-regulated proteins, Salmonella serovar Typhi harboring plasmid pFMTrc12 (strain IMSS-II) (12) or pFMTrcleuO-50 (strain IMSS-III) (5) and Salmonella serovar Typhi ΔleuO containing plasmid pFMTrc12 (strain IMSS-32) were grown to an OD₅₉₅ of 0.6 in MA medium supplemented with IPTG (50 μM). These conditions were the conditions determined previously for simultaneous expression of the two minor porins OmpS1 and OmpS2, as the activation of the corresponding genes is dependent on the level of induction of LeuO (5). Whole-cell proteins were obtained and separated by 2-DGE; about 600 proteins with pI values ranging from 4 to 8 and with molecular masses ranging from 20 to 90 kDa were resolved in each gel. The protein patterns of IMSS-II and IMSS-32 were similar. However, a comparison of the proteomic profiles of IMSS-III and IMSS-32 showed that overexpression of LeuO quantitatively up-regulated and down-regulated 10 and 3 protein spots, respectively (Fig. 1). To identify the LeuO-regulated proteins, spots of interest were excised, trypsin digested, and analyzed by MALDI-TOF MS. These analysis showed that LeuO activated the expression of OmpS1 (five isoforms), OmpS2, AssT (two isoforms), and STY3068, the third gene of the STY3070-STY3064 operon. In contrast, LeuO repressed the expression of OmpX, STY1978, and Tpx. As expected, recombinant LeuO was also detected in the 2-DGE gel (Table 1).

The sequence coverage of the LeuO-dependent proteins ranged from 7.2 to 54%, and the Mascot database search algorithm revealed that the proteins identified are identical to the corresponding proteins of Salmonella serovar Typhi CT18 (41) (Table 1). OmpS1 and AssT were detected in five and two isoforms, respectively. The OmpS1 pI values were 4.44, 4.25, 5.09, 5.06, and 6.12, and the AssT pI values were 6.71 and 6.51. The variation in pI values was not accompanied by a substantial change in molecular weight (Table 1), suggesting that the isoforms may involve modifications such as phosphorylation, methylation, or deamidation rather than introduction of high-molecular-weight groups. Previous studies with prokaryotes showed that posttranslational modifications, such as methylation, are relevant in bacterial pathogenesis (42, 54). Multiple protein isoforms have been found previously in Salmonella, including isoforms involved in carbon metabolism in S. enterica serovar Typhimurium and S. enterica serovar Pullorum (8). Previously, other loci regulated by LeuO, such as bgl, rovA, dsrA, cadC, and leuO, have been reported (9, 27, 45, 50, 55); however, they were not detected in our proteomic analysis. The bgl and rovA genes are not present in Salmonella serovar Typhi, and dsrA codes for a small RNA. CadC and LeuO were not found, perhaps due to mRNA instability or low expression, because they overlapped with other proteins on the 2-DGE, or because the LeuO concentration was not the optimal concentration for induction of these genes.

The cellular LeuO concentration in IMSS-III harboring pFMTrcleuO-50 with 50 μM IPTG was calculated by Western blot analysis of total proteins and comparison to purified LeuO (data not shown). The data obtained indicated that LeuO was present at a level of 6 × 10⁵ molecules per cell (or 0.4 μg/100 μg of total cellular protein).

The proteins identified in this study thus corresponded to outer membrane proteins, such as OmpS1, OmpS2, and OmpX (11, 12, 33); to proteins involved in detoxification, such as Tpx and AssT (20, 22, 24); and to the hypothetical STY1978 and STY3068 proteins (41). Previously, we described isolation, characterization, and regulation analysis of the quiescent porins OmpS1 and OmpS2; furthermore, we showed that the genes encoding these proteins are LeuO targets, and thus we considered these proteins relevant controls for proteomic analysis.

To define the role of LeuO in the regulation of the identified genes, transcriptional fusions were constructed with each of the LeuO-regulated genes. PCR fragments containing the putative promoter sequence of these genes were obtained and cloned into pKK232-9, which contained the cat reporter gene. The plasmids harboring the fusions (see Table S1 in the supplemental material) were transformed independently into Salmonella serovar Typhi IMSS-III and Salmonella serovar Typhi IMSS-II, which contained pFMTrcleuO-50 and pFMTrc12, respectively. To measure CAT specific activity, Salmonella cells were grown to different optical densities in MA medium supplemented with 50 μM IPTG when necessary.

The effect of LeuO on the transcription of the STY3068 gene, the third gene of a putative large operon consisting of seven open reading frames, was analyzed by fusing the STY3068 upstream region to the cat reporter. Interestingly, we were unable to detect expression (data not shown), indicating that a LeuO-dependent promoter is not present in the closest 5′ region (602 bp) of STY3068. Thus, a gene reporter fusion was constructed with the complete 5′ intergenic region of the first gene (STY3070) of the operon, showing that this gene was positively regulated by LeuO, since we found CAT activity at all ODs tested (Fig. 2a). Low transcriptional activity of STY3070 was also detected at 12 h without IPTG, suggesting that the basal accumulation of LeuO was enough for induction. This indicates that LeuO positively regulates the putative STY3070-STY3064 operon.

In the case of assT and ompS1, no transcriptional activity was detected when LeuO was absent, but upon induction with IPTG, high transcriptional activity of the assT and ompS1 fusions was detected (Fig. 2b and c). For ompS2, we found that in the absence of IPTG considerable transcriptional activity was detected only at 12 h, showing that, as observed for STY3070, the basal expression from the recombinant LeuO plasmid was enough for induction of the quiescent porin and the putative operon. Expression from the ompS2 fusion was also detected in the presence of IPTG at different ODs (Fig. 2d); however, the activities obtained at different ODs with IPTG were three times lower than the activity detected without IPTG at 12 h, in accordance with the previous observation that the amount of LeuO determines the expression level of ompS2 (5). It is relevant that in the case of STY3070, assT, ompS1, and ompS2, transcriptional activity was not detected with pFMTrc12 (Fig. 2a to d), indicating that these genes are strictly LeuO dependent.

The proteomic data mentioned above also revealed that LeuO represses the expression of tpx, ompX, and STY1978. The results obtained with the fusions showed that in the absence of LeuO these three loci are constitutively expressed, but when LeuO was induced with 50 μM IPTG, strong repression was observed for tpx (Fig. 2e). In the case of ompX and STY1978, down-regulation by LeuO was also observed at different ODs (Fig. 2f and g).

To demonstrate the specificity of the LeuO-mediated regulation, a transcriptional fusion containing the complete 5′ intergenic region of the ompC gene was evaluated. As we expected, no regulation of ompC by LeuO was observed since no differences in CAT activity were detected in the presence or absence of LeuO (data not shown). The results obtained with the transcriptional fusions support the proteomic data in demonstrating the positive and negative regulatory roles of LeuO in Salmonella serovar Typhi.

To determine the promoters associated with each of the LeuO-dependent genes, primer extension experiments were performed. Salmonella serovar Typhi IMSS-II and Salmonella serovar Typhi IMSS-III, transformed with pKK232-9 derivatives harboring the 5′ regions of the LeuO-dependent genes (see Table S1 in the supplemental material), were grown in MA medium to stationary phase. Using total RNA isolated from MA medium-grown cells, the transcription start sites of STY3070, assT, ompX, tpx, and STY1978 were determined (Fig. 3). The transcription initiation sites of ompS1 and ompS2 were reported previously (12, 39). The primer extension results showed that STY3070 contains two transcription start sites (Fig. 3a, panel A), whereas assT, ompX, tpx, and STY1978 each had a single transcription initiation site (Fig. 3b, panel A; Fig. 3d, panel A; Fig. 3e, panel A; and Fig. 3f, panel A, respectively). Our results also show that in the case of STY3070 (Fig. 3a, panel A) and assT (Fig. 3b, panel A), the initiation sites were detected only in IMSS-III (containing pFMTrcleuO-50), while no signals were detected in the absence of LeuO (pFMTrc12). In contrast, for ompX, tpx, and STY1978 the initiation sites were detected in the presence of pFMTrc12, but diminished signals were observed with pFMTrcleuO-50 (Fig. 3d, panel A; Fig. 3e, panel A; and Fig. 3f, panel A, respectively). This observation supports the hypothesis that there is strict LeuO dependence of STY3070 and assT, as well as the hypothesis that down-regulation of ompX, tpx, and STY1978 is exerted by the LeuO regulator. In Fig. 3, the nucleotide sequences corresponding to the region containing the transcription start sites of STY3070, assT, ompX, tpx, and STY1978 are shown.

To further support the experimentally determined transcription start sites, we used neural network promoter prediction (38). The in silico promoter prediction results were similar to the experimental data, since for STY3070 the experimental data showed that there were two initiation sites at 85 and 90 bp upstream of the ATG codon, whereas the in silico analysis predicted that there was a single promoter at 85 bp upstream of the ATG codon. For ompX, tpx, assT, and STY1978 the experimental results showed that the initiation sites were 240, 33, 165, and 11 bp upstream of the ATG codon, respectively, and the in silico results predicted that the transcription initiation sites were located 240, 34, 135, and 29 bp upstream of the ATG codon, respectively. Hence, the locations of two LeuO-dependent promoters (STY3070 and ompX) were in agreement as determined by the experimental and in silico approaches.

The proteomic data, the gene fusions, and the primer extension experiments described above indicated that LeuO is a genetic component involved in the regulation of STY3070, assT, ompS1, ompS2, tpx, ompX, and STY1978. In order to elucidate whether LeuO directly interacted with the 5′ upstream regulatory region of these LeuO-dependent genes, footprinting experiments were performed. A previous report showed that LeuO bound to the promoter region of ompS2 at −93 to −109 and −139 to −152 bp with respect to the transcription start site (12). In this work the same footprinting profile was obtained (Fig. 3c). For ompS1, two LeuO binding sites were recently detected (5). Interestingly, for the novel LeuO-dependent genes, two regions protected from DNase I by LeuO were also detected. In the 5′ intergenic region of STY3070, the sites were identified at nucleotides −65 to −105 and −110 to −140 (Fig. 3a, panel B). In the case of the regulatory region of assT, the first protected sequence was located at nucleotides −112 to −122 and the second was located at nucleotides −132 to −165 (Fig. 3b, panel B). For the 5′ region of ompX, the LeuO DNase I-protected sites were detected at nucleotides 1 to 9 and 28 to 60 (Fig. 3d, panel B). The promoter region of tpx had two LeuO binding regions located between nucleotides −22 to −46 and between nucleotides −51 to −59 (Fig. 3e, panel B). In the footprinting experiments (Fig. 3), a LeuO concentration of 150 nM was high enough for protection of ompS2 (12), whereas a higher concentration was required for STY3070, assT, ompX, and tpx.

Interestingly, in the 5′ intergenic region (41 bp) of STY1978, no LeuO protected sites were detected, suggesting that LeuO indirectly regulates the expression of this gene. Furthermore, we evaluated 131 bp upstream and 38 bp downstream of the 5′ intergenic region of STY1978 in order to see whether LeuO bound at a greater distance, but protection from DNase I was not detected. Alternatively, binding of LeuO to the 5′ regulatory region of STY1978 might require an accessory molecule, as reported for other LysR family members (48), or even could occur outside the boundaries of the region studied.

Hence, the footprinting results showed that LeuO directly interacts with the regulatory regions of STY3070, assT, ompX, and tpx and with ompS2, as previously shown (12). Likewise, LeuO interacts directly with the ompS1 regulatory region, as has been recently illustrated (5). In addition, LeuO probably indirectly regulates STY1978. Figure 3 shows the DNA sequences recognized by LeuO in each member of the LeuO regulon.

To determine a consensus sequence for LeuO interaction, the LeuO sites determined in this study were analyzed by using two bioinformatics approaches. The results showed that the matrix generated with the pattern-matching method identified binding sites in the non-LeuO-regulated genes and in the random sequences and detected only one-half of the experimental DNA binding sites reported in this work. In this set of genes false DNA binding sites were also detected (see http://www.ccg.unam.mx/Computational_Genomics/tools/LeuO/ ), indicating that the interacting sites are not sequence specific. In addition, the lack of either significant oligonucleotides or dyads in the upstream regulatory region of the LeuO-dependent genes supported the notion that the LeuO binding sites are not sequence specific and may not be easily represented with a consensus. The results of the complete analysis are available at http://www.ccg.unam.mx/Computational_Genomics/tools/LeuO/ .

Since the proposed role of LeuO has been as an antagonist of H-NS action (2, 5, 55), the functional effect of the global regulatory protein H-NS on the genetic expression of the LeuO-regulated genes was evaluated. To this end, transcriptional fusions of each member of the LeuO regulon were transformed independently into Salmonella serovar Typhi STYhns99 and CAT specific activity was measured in MA medium at different ODs. The transcriptional profile obtained was consistent with the hypothesis that H-NS represses the expression of ompS1 (5) (Fig. 2c), assT (Fig. 2b), and STY307O (Fig. 2a), suggesting that the STY3070-STY3064 operon also belongs to the H-NS regulon. For ompS2 and ompX expression, no change in the transcriptional pattern was observed in the hns mutant (Fig. 2d and f). In the case of tpx and STY1978, the CAT activity decreased 60% in each case in the absence of H-NS, suggesting that H-NS enhances positive regulation (Fig. 2e and g). These data are in agreement with the notion that LeuO and H-NS regulate the same genetic loci (2, 5, 27, 45, 50, 55).