H-NS Represses Salmonella enterica Serovar Typhimurium dsbA Expression during Exponential Growth

Abstract

Disulfide bond formation catalyzed by disulfide oxidoreductases occurs in the periplasm and plays a major role in the proper folding and integrity of many proteins. In this study, we were interested in elucidating factors that influence the regulation of dsbA, a gene coding for the primary disulfide oxidoreductase found in Salmonella enterica serovar Typhimurium. Strains with mutations created by transposon mutagenesis were screened for strains with altered expression of dsbA. A mutant (NLM2173) was found where maximal expression of a dsbA::lacZ transcriptional fusion occurred in the exponential growth phase in contrast to that observed in the wild type where maximal expression occurs in stationary phase. Sequence analysis of NLM2173 demonstrated that the transposon had inserted upstream of the gene encoding H-NS. Western immunoblot analysis using H-NS and StpA antibodies showed decreased amounts of H-NS protein in NLM2173, and this reduction in H-NS correlated with an increase of StpA protein. Northern blot analysis with a dsbA-specific probe showed an increase in dsbA transcript during exponential phase of growth. Direct binding of H-NS to the dsbA promoter region was verified using purified H-NS in electrophoretic mobility shift assays. Thus, a reduction in H-NS protein is correlated with a derepression of dsbA in NLM2173, suggesting that H-NS normally plays a role in suppressing the expression of dsbA during exponential phase growth.

Salmonella enterica serovar Typhimurium is a major cause of gastroenteritis or food poisoning in humans (34). This gram-negative, facultative, intracellular pathogen has evolved a number of distinct strategies to survive and propagate in a wide variety of cell types in the host. Many of these strategies involve proteins that are exported from the cytoplasm to either the periplasm or outer membrane or secreted out of the cell (15). Some proteins are transported or assembled by means of specialized secretory systems, but many of these proteins pass through the periplasm, where they undergo some degree of folding into their native conformation. Disulfide bonds usually contribute to the stabilization of a folded protein conformation (2, 35). In gram-negative bacteria, disulfide bond formation is mediated by the foldase DsbA, which is part of a disulfide oxidoreductase system that includes other Dsb proteins, such as DsbB, DsbC, and DsbD (2, 24, 35). DsbA, a soluble periplasmic disulfide oxidoreductase, was first discovered in Escherichia coli (4) and has also been characterized from a number of gram-negative bacteria, including S. enterica serovar Typhimurium (49). Disulfide bonding is an essential step for the proper folding and hence, function, of a number of disulfide bond-containing proteins that are bacterial virulence factors, such as exotoxins, fimbriae, and adhesins (52). Although DsbA is not essential for growth under laboratory conditions, lack of disulfide oxidoreductase activity in serovar Typhimurium renders cells nonmotile and slows growth in defined minimal medium (49). Interestingly, in contrast to the observations made in E. coli (6), DsbA is growth phase regulated in S. enterica serovar Typhimurium, with expression levels increasing during late exponential phase of growth and remaining elevated for at least 72 h in liquid culture (16). This stationary-phase regulation is not dependent upon RpoS (16), a common stationary-phase sigma factor (27, 36) or SlyA, a serovar Typhimurium stationary-phase transcriptional regulator (8).

This study details the investigation of a new facet of DsbA regulation involving the global regulator H-NS. By characterizing mutants that were derepressed for expression of dsbA from a plasmid-encoded dsbA::lacZ construct in S. enterica serovar Typhimurium during exponential phase growth, it was determined that H-NS was involved in the growth phase-dependent regulation of dsbA. H-NS is a major protein of the bacterial nucleoid and is involved in the regulation of both housekeeping and virulence genes in E. coli (10, 21). H-NS is a small, abundant protein that has affinity for all types of nucleic acids but binds preferentially to curved DNA substrates (37, 47). A number of hns mutant alleles have been shown to cause slow growth, reduce motility, and confer mucoid appearance on the mutant strain (5, 19). H-NS has been shown to negatively or positively regulate more than 200 genes in E. coli (21). Many of the target genes that are affected by H-NS are also regulated by other global transcription factors, such as LRP, VirF, CfaD, RpoS, and the DNA-binding protein FIS (1, 41). Hence, the effect of H-NS on many target genes is not straightforward. In this study, we demonstrate that H-NS binds to the dsbA promoter region and that a reduction in the amount of H-NS protein derepresses dsbA expression early in the growth cycle, suggesting that H-NS normally represses dsbA until late log or early stationary phase.

MATERIALS AND METHODS

Bacterial strains, media, and culture conditions.

The bacterial strains and plasmids used in this study are listed in Table 1. In general, bacteria were grown overnight at 30°C in Luria-Bertani (LB) medium (39) with the appropriate antibiotic selection. When required, antibiotics were used at the following concentrations: chloramphenicol (30 μg ml⁻¹), tetracycline (10 μg ml⁻¹), and ampicillin (100 μg ml⁻¹). When screening for blue or white colonies, 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) was used at a concentration of 40 μg ml⁻¹.

TABLE 1.

Bacterial strains and plasmids used in this study

Bacterial strain or plasmid	Relevant characteristic	Reference or source
Bacterial strains
CH1794	S. enterica serovar Typhimurium LT2 hns-106::Tn10Δ16Δ17	Hinton et al. (19)
SL1344	S. enterica serovar Typhimurium wild-type strain his rpsL	Wray and Sojka (51a)
NLM331	SL1344 dsbA::kan containing pMEG2	This study
NLM2160	SL1344 containing pMEG2	Goecke et al. (16)
NLM2167	SL1344 dsbA::kan	Turcot et al. (49)
NLM2173a	NLM2160 hns-112::Tn10d(T-POPII) containing pMEG2	This study
NLM2174	NLM2173 suppressor mutant containing pMEG2	This study
NLM2190	NLM331 hns-112::Tn10d(T-POPII) containing pMEG2	This study
SA4105	LT2 osmZ::Tn10	Salmonella Genetic Stock Centre
NLM2275	NLM2160 osmZ::Tn10 containing pMEG2	This study
Plasmids
pMP190	15-kb transcriptional fusion vector containing a promoterless lacZ gene	Spaink et al. (46a)
pMEG2	pMP190 with a 258-bp XhoI/BglII fragment from immediately upstream of the dsbA translational start site cloned into the multiple cloning site	Goecke et al. (16)
pBluescript	2.9-kb pBR322 vector derivative, Ampr	Stratagene

^aThe hns-112::Tn10d(T-POPII) mutation was originally designated zde-5A1t::Tn10d(T-POPII).

P22 transduction.

P22 transduction was performed by the method of Maloy (32). An aliquot of the lysate containing a pool of random Tn10d(T-POPII) insertions in S. enterica serovar Typhimurium LT2 was kindly provided by John Roth (38). Tn10d(T-POPII) is a Tn10 derivative that has been modified such that its insertion between any gene or operon and its promoter causes the expression of the gene or downstream gene to become tetracycline dependent. Strain NLM2275 was created by transducing a hns null allele (osmZ) from strain S4105 into a SL1344 background.

Motility assays.

Salmonella strains were grown overnight at 30°C with appropriate antibiotics. The following morning, all the strains were standardized to an A₆₀₀ of 0.04. A flat-ended sterile toothpick was dipped into standardized bacterial culture and stabbed into semisolid 0.3% agar LB plates (19). The plates were incubated for up to 16 h at 30°C, and the swarming behavior of each strain was measured as an indicator of motility.

β-Galactosidase assays.

Transcription of plasmid-borne dsbA-lacZ fusions were monitored by β-galactosidase assays of cells cultured to mid-exponential phase and stationary phase by the method of Miller (33). Assays were performed in triplicate, and experiments were done at least three times.

RNA protocols.

Total RNA was extracted from 3-ml samples of cells at the appropriate growth phase by the Trizol method (Gibco BRL). Diethyl pyrocarbonate-treated water or formamide was used to resuspend the RNA pellets. The concentrations and purity of the RNA samples were determined spectrophotometrically and by visual inspection of formaldehyde-agarose gels (see below).

For Northern blotting, samples of RNA (30 μg) were denatured at 65°C, loaded onto 1.5% formaldehyde-agarose gels, electrophoresed within buffer containing 20 mM MOPS [3(N-morpholino) propanesulfonic acid], 5 mM sodium acetate, and 1 mM EDTA at 80 V for 2 to 3 h, and transferred to Hybond-N nylon membranes (Amersham). The membranes were washed, denatured, neutralized, air dried, and cross-linked following established protocols (40). The templates used for the DNA probe were a PCR fragment amplified from the sequence of the dsbA gene using NLM88 (5′-CTGGCGAACCCCAGGTACTG-3′) and NLM78 (5′CGCATCAACGAACACTTTACGG-3′) or an amplicon specific for hns using primers NM93 (5′-ATAAGCTCTTTTTTGTGCGGTG-3′) and NM94 (5′-TATTTTTTTCGCGGCCTAAATG-3′). The DNA fragment was labeled with digoxigenin, and prehybridization and hybridization were performed as recommended by the manufacturer (Amersham). Chemiluminescence detection as described by the Genius guide was used for probe detection.

For reverse transcriptase PCR (RT-PCR), RNA isolated from S. enterica serovar Typhimurium strains at various phases of growth was subjected to DNase treatment (RQ1 RNase-free DNase; Promega) and subsequent purification using RNeasy columns (RNeasy Mini kit; Qiagen). Reverse transcription reactions (Retroscript kit; Ambion) using 100 pmol of primer NM112 hns primer (5′-GCAGTTTACGAGTGCGTTCTTCC-3′) were performed on approximately 3 μg of purified RNA (Trizol method). Reverse transcription negative-control reactions were performed simultaneously where water was added instead of the RT enzyme. PCR amplification was performed using forward NM111 (5′-TAGCGACAGACGGTGAGTATCC-3′) and reverse NM112 (5′-GCAGTTTACGAGTGCGTTCTTCC-3′) hns-specific primers. PCRs (50 μl) were performed using 2.5 U of Taq DNA polymerase (Gibco BRL), 5 μl of reverse transcription reaction mixture sample as the template, 50 pmol of each PCR primer pair, 1× PCR buffer, 0.2 mM (each) deoxynucleoside triphosphate, and 1.5 mM MgCl₂. Template cDNA was denatured for 2 min at 94°C before Taq DNA polymerase was added. Twenty cycles of PCR were performed, with 1 cycle consisting of denaturation (45 s at 4°C), annealing (30 s at 60°C), and extension (1 min at 72°C). The final extension step was 7 min at 72°C. To aid in qualitative analysis, we normalized the RT-PCR product to an established endogenous internal control (tsf encoding the elongation factor EF-Tsf) (20). After the PCR, 5 μl of PCR product was visualized by agarose gel electrophoresis.

Gel electrophoresis and Western blotting.

Proteins were separated by the method of Laemmli (26) using a sodium dodecyl sulfate-12% polyacrylamide gel. Cell cultures were centrifuged at a specific optical density, and the cell pellet was resuspended in loading buffer and boiled before loading. The amount of protein in each whole-cell lysate was determined, and equal amounts of protein (2 × 10⁷ cells) were loaded in each lane. The expression of StpA and H-NS proteins in S. enterica serovar Typhimurium was determined using anti-StpA polyclonal antibody (S100) that does not cross-react with H-NS and anti-H-NS monoclonal antibody (H113) that does not cross-react with StpA (45).

Electrophoretic mobility shift assay.

Purified H-NS protein from S. enterica serovar Typhimurium was kindly provided by John Ladbury (Department of Biochemistry and Molecular Biology, University College London). The band shift reaction contained various concentrations of H-NS protein in the picomolar range and 10 ng of radiolabeled probe DNA in binding buffer (10 mM Tris-HCl [pH 7.5], 15 mM KCl, 0.1 mM EDTA, 2 mM spermidine, 15% glycerol) and was performed as described previously (23). The template used for the DNA probe was a PCR fragment amplified from the promoter region of the dsbA gene using primer pair NLM22 (5′-ACAAGATCTATTAATACATTGGCGTT-3′) and NLM24 (5′-CCCCTCGAGAAGCTTATCAAGAAGTT-3′) and primer pair NM111 (5′-TAGCGACAGACGGTGAGTATCC-3′) and NM112 (5′-GCAGTTTACGAGTGCGTTCTTCC-3′) for the promoter region of hns. The reaction mixture was incubated at room temperature for 30 min, and the samples were loaded onto a 5% polyacrylamide gel in Tris-acetate-EDTA and electrophoresed at 35 mA for 2 h. After electrophoresis, the gel was dried, and radiolabeled DNA was detected by autoradiography.

RESULTS

Transposon mutagenesis strategy for the isolation of dsbA regulatory mutants.

A pool of random Tn10d(T-POPII) (38) insertions from S. enterica serovar Typhimurium LT2 was obtained as a P22 phage lysate and used to transduce S. enterica serovar Typhimurium SL1344 carrying a dsbA::lacZ transcriptional fusion on a low-copy-number plasmid, pMEG2 (strain NLM2160). This dsbA::lacZ fusion is normally regulated by growth phase, with maximal induction of expression occurring upon entry into the stationary phase of growth (16). The pool of mutants resulting from T-POPII insertion mutations in strain NLM2160 were screened for those altered in dsbA regulation by monitoring β-galactosidase activity of isolated colonies using X-Gal. White colonies were chosen, as these colonies were considered potential regulatory mutants. Of several white colonies isolated, the phenotypes of three of these colonies was caused by a single transposon insertion, as confirmed by 100% linkage of the dsbA::lacZ phenotype with the tetracycline resistance marker upon transduction into a fresh strain of SL1344 (NLM2160). Two of these insertion mutants contained the T-POPII insertion in the same location, and this mutant was designated NLM2173, while the third remains to be characterized. The mutant phenotype of NLM2173 was somewhat unstable at temperatures higher than 30°C where secondary mutations led to renewed expression of β-galactosidase. These suppressor mutants were visible as blue sectors in the colonies on agar plates containing X-Gal. This suppressor phenotype, as exemplified by strain NLM2174, was able to grow in the presence of tetracycline, confirming the maintenance of the transposon. The suppressor phenotype was not, however, cotransduced with the tetracycline resistance marker, confirming that the secondary mutations were not closely linked to the transposon insertion site. The ratio of the appearance of suppressor mutants to the original mutant was observed to be much lower at 30°C (0.55%) than at 37°C (24%).

On solid LB agar plates, the mutant NLM2173 produced colonies that were generally smaller in size than strain SL1344 or the suppressor mutant, NLM2174. When grown in LB broth at 30°C, the growth rate of strain NLM2173 was lower than that of the parental strain (Fig. 1A). In addition to the difference in the growth rate of strain NLM2173, a decrease in motility was also observed. The parental strain NLM2160 was fully motile (diameter of the motility zone, 5.0 ± 0.4 cm), whereas strain NLM2173 had reduced motility (diameter of the motility zone, 2.8 ± 0.2 cm).

FIG.1.

Comparison of growth and the levels of transcription initiated from the dsbA promoter in the construct pMEG2 in the wild type (NLM2160), Tn10d(T-POPII) mutant (NLM2173), and suppressor strain (NLM2174). (A) Growth of NLM2173, NLM2174, and NLM2160 at 30°C. O.D.₆₀₀, optical density at 600 nm. (B) Expression of the dsbA::lacZ fusion in strains NLM2160, NLM2174, and NLM2173 (in the presence and absence of tetracycline [tet]). (C) Expression of the dsbA::lacZ fusion in a dsbA null S. enterica serovar Typhimurium SL1344 strain (NLM331) containing the transposon-interrupted locus (NLM2190) (in the presence and absence of tetracycline). The results are representative of three independent trials.

Expression of dsbA::lacZ transcriptional fusion in S. enterica serovar Typhimurium strains.

In order to measure the effects of the transposon insertion on dsbA promoter activity, the β-galactosidase activity from strain NLM2173 containing the dsbA::lacZ transcriptional fusion was measured throughout the growth cycle (Fig. 1B). As expected, maximal expression of the dsbA::lacZ fusion occurred at the onset of stationary phase in wild-type strain NLM2160. In contrast, strain NLM2173, in either the absence or presence of tetracycline, showed a shift in the activation of the dsbA promoter to earlier in the growth cycle, from stationary phase to log phase. If the Tn10d(T-POPII) transposon were to disrupt a promoter, then the expression of any downstream gene could become tetracycline dependent (38). This was not the case, however, for NLM2173. In addition, stationary-phase levels of dsbA promoter activity in NLM2173 are lower than in the wild type. The dsbA::lacZ expression was also assayed in the suppressor strain NLM2174, where no shift in the induction of dsbA::lacZ expression was observed and the stationary-phase β-galactosidase levels were the same as those of the wild type (Fig. 1B). When the Tn10d(T-POPII) interrupted locus was transduced into a dsbA null strain, the overall level of dsbA promoter induction was even higher than in the wild type but still occurred prior to the onset of stationary-phase growth (Fig. 1C). Previous studies have shown that the dsbA::lacZ activity from pMEG2 is higher in a dsbA null background than in a wild-type background (16). These observations led to the hypothesis that a feedback loop exists for the expression of DsbA in S. enterica serovar Typhimurium, whereby the absence of DsbA activity signals the cell to produce more DsbA (16), and this mechanism of autoregulation appears to be independent of the effect generated by the T-POPII insertion.

Localization and identification of the site of transposon insertion.

Genomic DNA from strain NLM2173 was digested with SacI, SalI, and HindIII, shotgun cloned into pBluescript, and selected by screening for clones that conferred tetracycline resistance. Using outward facing primers specific to the Tn10d(T-POPII) transposon, the regions flanking the transposon were sequenced and localized to a 3,629-bp contig (B-STM1107) from the S. enterica serovar Typhimurium sequencing project. The transposon insertion occurred 580 bp upstream of the hns coding region and 116 bp upstream of the tdk gene on the opposite strand encoding a thymidine kinase. Two other open reading frames could be recognized on this genomic DNA fragment; downstream of hns, a putative galU gene was detected, and upstream of hns, a putative adhE gene was also found (Fig. 2). Mutations at the hns locus are highly pleiotropic (5, 19, 53). In general, hns mutant strains grow more slowly than wild-type strains, show reduced motility, and are mucoid in nature (5). These phenotypic characteristics of an hns mutant were also shared by the mutant strain NLM2173, consistent with hns being the locus that affected dsbA regulation.

FIG. 2.

Schematic diagram illustrating the position of the transposon insertion in S. enterica serovar Typhimurium SL1344. Sequencing analysis of strain NLM2173 demonstrated that the 2.5-kb Tn10d(T-POPII) transposon had inserted 580 bp upstream of the hns gene (encoding H-NS) and 116 bp upstream of the tdk gene (encoding a thymidine kinase). The positions of adhE (encoding an alcohol dehydrogenase) and galU (encoding glucose-1-phosphate uridyltransferase) are also shown. The position of the Tn10 insertion 377 bp upstream of the translational initiation codon of hns in strain CH1794 is shown by an asterisk.

It was not immediately apparent how the T-POPII insertion 580 bp upstream of the hns coding region, produced an hns phenotype. However, Hinton et al. (19) and Hulton et al. (22) have demonstrated that S. enterica serovar Typhimurium strain CH1794, which contains a Tn10 insertion 377 bp upstream of the translational initiation codon of hns, produced reduced levels of H-NS protein than other hns mutants and the wild type. Unlike mutations within the hns structural gene, this hns-106::Tn10d insertion caused only a low level of derepression of the proU locus and had no detected effect on DNA supercoiling (19). This hns mutant strain was also included in this study for comparison.

hns mutations differentially affect the levels of dsbA mRNA at mid-logarithmic growth.

To examine the effects of hns mutations on dsbA transcription, Northern blot analysis was performed on total RNA extracted from several strains using a probe complementary to the transcribed dsbA gene. Goecke et al. (16) previously showed that two transcripts were consistently detected for dsbA and that the amount of these two transcripts varied with growth conditions. In the present study, this transcription pattern was observed in the wild-type strain NLM2160 and the suppressor mutant NLM2174. However, there was a substantial increase in the amount of the shorter dsbA-specific transcripts compared to the larger transcript in strains NLM2173 and CH1794 relative to the wild type or the suppressor strain (Fig. 3). This increase in the amount of dsbA transcript during log phase growth correlated with the increase in expression of the dsbA::lacZ fusion, suggesting that dsbA promoter activity is elevated in strain NLM2173.

FIG. 3.

Northern blot analysis of the transcription of dsbA in several strains at different time points. Using a dsbA DNA probe, two transcripts of approximately 700 and 800 nucleotides (arrows) were detected in all the strains. The amount of the shorter dsbA transcript in strains NLM2173 and CH1794 at A₆₀₀s of 0.2 and 0.6 is higher than in strains NLM2160 (wild type) and NLM2174. Equal amounts of total RNA were loaded in each lane. Results are representative of three independent trials.

Comparison of expression from the dsbA promoter in hns mutant strains.

The effect on dsbA transcription of the Tn10 insertion 377 bp upstream of the translational initiation codon of hns (strain CH1794) was compared to that of the T-POPII insertion mutant NLM2173 under log phase growth conditions (Fig. 4). Both NLM2173 and CH1794 showed increased dsbA promoter activity at mid- and late log phase, suggesting that although the transposons are inserted 200 bases apart from each other, their effect on hns is similar. An hns null strain (NLM2275) was also tested and had even higher levels of β-galactosidase activity than NLM2173. NLM2174 was also included in this comparison, and although the dsbA promoter activity is closer to the wild type in this suppressor strain, it is not identical to that of the wild type.

FIG. 4.

Expression of the dsbA::lacZ fusion was measured in strains NLM2160, NLM2173, NLM2174, CH1794, and NLM2275 in mid- and late log growth phase. Cultures of the various strains were grown at 30°C, and samples were taken at A₆₀₀s of 0.3 and 0.6. β-Galactosidase activity was plotted for each strain. Data are means ± standard errors of the means (error bars) of three independent experiments, each with duplicate samples.

H-NS binds to the dsbA promoter region.

In order to determine whether H-NS could interact directly with the dsbA promoter region, band shift assays were performed. As H-NS had previously been shown to bind to its own promoter with high affinity (11, 51), the hns promoter region was used as a control for binding specificity. The results show that the dsbA promoter fragment begins to shift at an H-NS concentration equivalent to that required to demonstrate binding to the hns promoter (Fig. 5). This binding is specific, as demonstrated by the fact that a mobility shift does not occur in the 112-bp digested hns fragment but does occur in the 267-bp fragment previously shown to contain the H-NS binding domain (Fig. 5).

FIG. 5.

Electrophoretic mobility shift assays showing an interaction between purified H-NS and the promoter regions of hns and dsbA. (A) All lanes contain a ³²P-labeled 379-bp amplicon covering the promoter region of hns that has been digested to give 112- and 267-bp fragments. The 112-bp fragment does not shift, while the 267-bp fragment, which contains the predicted H-NS binding region, does shift, in agreement with previously published results (11). (B) Lanes 1 to 7 lanes contain a ³²P-labeled 379-bp amplicon covering the promoter region of hns. Lanes 8 to 14 contain a ³²P-labeled 238-bp dsbA promoter amplicon and show that the dsbA fragment binds H-NS at concentrations as low as 20 pmol. Purified H-NS was added to each reaction mixture, as indicated at the bottom of the gel. These results are representative of several independent experiments.

H-NS expression is altered in strains NLM2173 and CH1794.

As the results indicated that H-NS was affecting the dsbA promoter activity, Western immunoblotting using a H-NS-specific monoclonal antibody was undertaken to monitor steady-state H-NS levels. As previous work had demonstrated that both H-NS and the homologous protein, StpA, are implicated in a global regulatory system and that the stpA gene is derepressed in hns mutants of E. coli (12, 44, 4), StpA protein levels were also monitored using an StpA polyclonal antibody. If a reduction in H-NS protein was occurring in NLM2173 and CH1794, increased expression of StpA would be expected. The amount of H-NS protein produced was lower in strain NLM2173 than in wild-type strain NLM2160, and the reverse pattern was observed when the same samples were probed with the StpA-specific antibody (Fig. 6). Thus, H-NS protein production was decreased in the transposon mutant, and this reduction in H-NS protein was associated with an increase in expression of StpA. A similar decrease in H-NS and increase in StpA levels were also observed in CH1794 (data not shown).

FIG. 6.

Western immunoblot analysis of H-NS and StpA levels in strains NLM2160 and NLM2173. Cells were grown at 30°C in LB and harvested sequentially at optical densities (O.D.) at 600 nm of 0.3, 0.6, and 1.2. Whole-cell lysates of samples, normalized to total cell number, of each strain were separated on a sodium dodecyl sulfate-15% polyacrylamide gel and probed with anti-H-NS (H113) antibody and anti-StpA (S100) antibody. Panel A is the Coomassie blue-stained gel. There is no visible difference in the banding pattern at the 17-kDa marker where both H-NS and StpA are located. Panel B shows the immunoblot from the same samples as in panel A using anti-H-NS antibodies. Under these growth conditions, H-NS protein levels are higher in the wild type than in strain NLM2173 during log phase growth, and the levels in both strains decrease later in the growth cycle. Panel C shows the immunoblot from the same samples as in panel A using anti-StpA antibodies. There is slightly more StpA protein in NLM2173 than in the wild type, and there is also an StpA-specific band running at the size of a StpA dimer.

hns transcription and growth phase.

Although the connection between decreased H-NS protein and increased dsbA transcription was established, it was not clear how transposons inserted either 377 or 580 bases upstream of the hns coding region could cause a decrease in H-NS protein levels. Transcriptional analysis of hns was undertaken to assess the effects of these transposons on hns transcript abundance. RT-PCR was used to assess the amount of hns transcript produced at different time points during the growth cycle (Fig. 7). For each strain, an hns-specific transcript was detected at all growth phases tested. In the wild-type strain (NLM2160), different amounts of the RT-PCR product were produced in the different growth phases, with more abundant transcript being detected earlier in the growth cycle, as expected. However, the relative intensities of the RT-PCR product in strains NLM2173 and CH1794 were higher in mid-log phase than those of the wild-type strain, NLM2160, and the suppressor strain, NLM2174. These RT-PCR results were surprising, as they did not correlate with the data showing a reduction in the steady-state H-NS protein levels, but these results were confirmed when Northern blotting or multiplex RT-PCR (17) were used to assess the hns transcript (data not shown).

FIG. 7.

RT-PCR was performed on RNA isolated from strains NLM2160, NLM2173, NLM2174, and CH1794 to detect hns-specific transcript at optical densities at 600 nm of 0.3 (mid-log phase) and 2 (stationary phase). The specific hns transcript is designated by an arrow and is approximately 378 bp in size. The 563-bp band corresponds to the amplified tsf gene that was used as the internal control (IC) in this experiment. These data are representative of three independent trials.

DISCUSSION

The growth phase-regulated expression of dsbA in S. enterica serovar Typhimurium has been shown to be independent of the stationary-phase sigma factor, RpoS, and the transcriptional activator, SlyA (16). This study was thus initiated to determine the factors that may be involved in influencing the levels of dsbA expression in the cell. Screening a T-POPII mutant library using a dsbA::lacZ transcriptional fusion led to the isolation of strain NLM2173 that exhibited alterations in its ability to regulate dsbA. Cloning and sequencing of the transposon-containing DNA fragments from strain NLM2173 revealed that the zde-5A1t::Tn10d(T-POPII) locus is around 38.4 min on the chromosome of S. enterica serovar Typhimurium, 116 bp upstream of the thymidine kinase gene (tdk) and 580 bp upstream of the hns gene. Several phenotypic characteristics of the mutant, such as increased mucoidy, decreased growth rate, and decreased motility, suggested that modification of expression of the hns locus was being affected. The zde-5A1t::T-POPII mutation was later designated hns-112::Tn10d(T-POPII). NLM2173 was able to grow on plates containing thymidine kinase (25), confirming that the tdk locus was not affected by hns-112::Tn10d (data not shown). The motility of strain NLM2173 is about 56% that of the wild-type strain, suggesting that the transposon insertion does not completely eliminate the expression of H-NS but changes the level of H-NS expressed in the cell. Motility is reduced to different levels by several hns alleles, and it was previously shown in strain CH1794 that the insertion 377 bp upstream of the translation initiation codon of hns also resulted in a partial loss of motility (55% of the wild type), while a strain with an hns null allele lacks flagella (19). In E. coli, the presence of an hns mutation decreased the transcription of flhD and fliA genes required for the synthesis of flagella (7, 46), a rather rare example of H-NS acting as a positive regulator. DsbA is also involved in flagellar biosynthesis. Work done by Bardwell et al. (3) showed that DsbA is essential for flagellar assembly and function in E. coli, and Turcot et al. (49) demonstrated that an S. enterica serovar Typhimurium dsbA null strain was also immotile. Thus, the decrease in motility observed in the transposon mutant strains in the present study is unlikely to be related to reduced expression of the wild-type dsbA gene, since the expression of dsbA is derepressed in NLM2173.

The phenomenon of spontaneous second-site mutations arising in hns mutants has been observed previously (5, 18, 30). In addition, Barth et al. (5) found that hns suppressor strains had lost the increased mucoidy characteristic for hns mutants and grew faster, exhibiting shorter doubling times, than the parental strains, as was seen with the present study. Barth et al. (5) found that some of their suppressor mutants carried alterations at the rpoS locus, raising the possibility that the suppressor mutation in strain NLM2174 was in the rpoS gene. Qualitative assays of catalase activity showed an increase in NLM2173 relative to the wild type, but catalase activity in NLM2174 was similar to NLM2173 (data not shown). This increase in catalase activity in the hns mutant probably resulted from increased rpoS transcription (5) and, because activity was unchanged in NLM2174, suggests that the suppressor mutation in NLM2174 is not located in the rpoS gene.

Figure 1B clearly shows that expression of the dsbA::lacZ transcriptional fusion in strain NLM2173 occurs earlier in the growth cycle, with a twofold derepression of dsbA expression in mid- and late log phase (Fig. 4). This twofold derepression of dsbA was also seen for the expression of the dsbA::lacZ transcriptional fusion in another hns mutant strain (CH1794) that also contains a Tn10 insertion upstream of the hns coding region, while analysis of the hns null mutant NLM2275 showed even higher levels of dsbA promoter activity (Fig. 4). Taken together, these results imply that the level of H-NS in the cell influences the transcription of dsbA in S. enterica serovar Typhimurium. H-NS is known to act as a transcriptional repressor by binding to DNA in the promoter region (51) and shows a preference for binding to intrinsically curved DNA (9). Sequence analysis of the region upstream of the dsbA translation start site revealed the presence of a region predicted to bend (data not shown). Band shift assays with purified H-NS demonstrated high-affinity binding to the dsbA promoter region, suggesting that normally dsbA expression is directly repressed by H-NS (Fig. 5). By surveying the literature, Atlung and Ingmer (1) determined that H-NS has a larger effect on target gene expression in scenarios where expression is not also mediated by positive transcription factors and noted that any repression by H-NS is virtually eliminated when positive transcription factors are artificially induced. We see an increase in dsbA promoter activity in NLM2173 during exponential phase of only two- to threefold, the magnitude of which could be influenced by a positive regulator; it also could be due to the fact that H-NS is not completely abolished.

H-NS is autoregulated (11, 12, 50), positively regulated at the transcriptional level (41), and posttranscriptionally regulated by DsrA RNA (28). Free and Dorman (12) found that the hns transcript is virtually absent in stationary-phase cells but is present at high levels within 1 h of subculturing a stationary-phase culture. Additionally, Dorman et al. (10) showed that the ratio of H-NS synthesis to DNA synthesis was constant, which could explain why hns expression is reduced in stationary phase when DNA synthesis slows. In the present study, steady-state levels of H-NS in the wild-type strain were seen to be slightly higher in exponential phase growth than in stationary phase. The levels of H-NS were decreased in NLM2173 than in the wild type, correlating with the observed derepression of the dsbA promoter in exponential phase.

The steady-state levels of StpA were also examined. StpA is a paralogue of H-NS that shows 52% identity at the amino acid level and has a DNA-binding affinity that is comparable to that of H-NS (45). H-NS and StpA can act cooperatively to repress many H-NS-regulated genes (13, 14). Furthermore, it has been demonstrated that the expression of stpA is derepressed in an hns mutant strain (14, 44). This increase in StpA expression in an hns mutant strain appears to compensate for the lack of H-NS and allows repression of many H-NS-regulated genes (45), although not all H-NS-repressed loci can be regulated by StpA (10, 54). In this study, StpA levels were higher in strain NLM2173 than in the wild type (NLM2160), showing that the hns-112::Tn10d mutation caused a sufficient decrease in H-NS levels to exert a biologically relevant effect. However, the increase in StpA protein levels in NLM2173 did not allow StpA to substitute for H-NS in the repression of the dsbA promoter.

It was not clear how transposon insertions significantly upstream of the hns coding region caused a decrease in H-NS levels, especially since the transposons were inserted further from the promoter region than any previously described regulatory regions. Transcriptional autorepression occurs as a result of H-NS binding to extended regions of DNA 150 nucleotides upstream of its coding region (11, 48). In the present study, it was hypothesized that the T-POPII and Tn10 transposons somehow affect the hns promoter region, lowering the transcription of the hns locus. Using RT-PCR (Fig. 7), it was shown that the level of hns transcript is abundant early in the growth cycle and lower when the cells reach stationary phase in the wild-type strain. However, there was a marked increase in the amount of hns transcript in strains NLM2173 and CH1794 at mid-log phase compared to that of the wild type, suggesting that the transposon insertions enhanced hns transcription. Both Northern blotting and multiplex primer extension approaches to measuring the hns transcript abundance also showed increased transcription in NLM2173 and CH1794 than in the wild type (data not shown). The data clearly establish enhanced hns transcription, suggesting that the transposons have affected a previously uncharacterized regulatory element upstream of hns. The data also suggest that autoregulation of hns transcription has been disrupted, as the results are similar to that seen in an hns deletion strain where basal hns transcription levels are more than twofold higher than the level in the wild type (29).

With hns transcription increased, the observed decrease in H-NS protein levels still requires an explanation. It is hypothesized that DsrA, a small RNA, is involved in the repression of H-NS in these mutants. DsrA is an untranslated, regulatory RNA that is involved in the expression of RpoS (31, 43) and H-NS (29, 31, 42). It is thought that DsrA and hns mRNA interact and that this interaction enhances the turnover of hns mRNA, resulting in the production of less H-NS protein (28). In experiments performed by Lease et al. (29), DsrA expression decreased H-NS protein levels in a wild-type background and had no effect on the level of hns transcript. Lease et al. (29) also showed that StpA is still produced when DsrA is overexpressed, and it appears that the DsrA-mediated reduction in H-NS actually leads to an increase in StpA levels (29). A direct connection between increased hns transcription and increased DsrA activity remains to be established in future experiments.

In this study, we have demonstrated that a reduction of H-NS protein correlates with a derepression of dsbA expression in log phase. Since the regulation of DsbA is growth phase dependent, the involvement of H-NS, a protein abundant in log phase, fits with the expression profile of this disulfide oxidoreductase. There must also be as yet unidentified positively regulating factors involved in dsbA transcription to account for the increase in stationary-phase expression. DsbA appears to facilitate protein folding in stationary phase rather than exponential growth phase where it is expected that protein secretion would require foldases in order to be rapid and efficient. Given the involvement of H-NS in the expression of genes related to cell survival under stressful growth conditions, DsbA expression may reflect the need for foldase activity in the context of environmental factors causing stress to the bacterial cell.