Regulation of Acid Resistance by Connectors of Two-Component Signal Transduction Systems in Escherichia coli

Abstract

Two-component signal transduction systems (TCSs), utilized extensively by bacteria and archaea, are involved in the rapid adaptation of the organisms to fluctuating environments. A typical TCS transduces the signal by a phosphorelay between the sensor histidine kinase and its cognate response regulator. Recently, small-sized proteins that link TCSs have been reported and are called “connectors.” Their physiological roles, however, have remained elusive. SafA (sensor associating factor A) (formerly B1500), a small (65-amino-acid [65-aa]) membrane protein, is among such connectors and links Escherichia coli TCSs EvgS/EvgA and PhoQ/PhoP. Since the activation of the EvgS/EvgA system induces acid resistance, we examined whether the SafA-activated PhoQ/PhoP system is also involved in the acid resistance induced by EvgS/EvgA. Using a constitutively active evgS1 mutant for the activation of EvgS/EvgA, we found that SafA, PhoQ, and PhoP all contributed to the acid resistance phenotype. Moreover, EvgS/EvgA activation resulted in the accumulation of cellular RpoS in the exponential-phase cells in a SafA-, PhoQ-, and PhoP-dependent manner. This RpoS accumulation was caused by another connector, IraM, expression of which was induced by the activation of the PhoQ/PhoP system, thus preventing RpoS degradation by trapping response regulator RssB. Acid resistance assays demonstrated that IraM also participated in the EvgS/EvgA-induced acid resistance. Therefore, we propose a model of a signal transduction cascade proceeding from EvgS/EvgA to PhoQ/PhoP and then to RssB (connected by SafA and IraM) and discuss its contribution to the acid resistance phenotype.

Microorganisms respond to the environmental stresses they encounter by regulating gene expression with factors such as two-component signal transduction systems (TCSs), sigma factors, and transcription factors. These factors are usually coordinated to form cascades and networks of regulation to adjust themselves appropriately to fluctuating environments. One example of such regulation is the intensively studied Escherichia coli acid response network (10). In this microorganism, at least 14 gene products are implicated as direct participants in the biochemistry of acid resistance (AR) and at least 15 regulators govern the expression of one or more of the 14 AR genes (10). Among the regulators, GadE, a LuxR family regulator, serves as the central activator protein of AR genes such as gadA and gadBC (14). Regulation of gadE expression involves such regulators as EvgA, YdeO, GadE, PhoP, TrmE, GadX, and GadW (12, 21, 24, 30, 33), which makes GadE the center of regulation for the induction of acid resistance.

Among TCSs, EvgS/EvgA is the major system for conferring acid resistance to exponential-phase cells (15, 21, 23, 24). Activation of this system initiates a transcriptional cascade of AR genes encoding regulators EvgA, YdeO, and GadE (see Fig. 6). Indeed, transcriptome analysis of an EvgS-activated strain resulted in the induction of many AR genes. In addition to those AR genes, members of a group of PhoP-regulated genes were also induced, indicating activation of another TCS, the PhoQ/PhoP system (8). A further genetic screening assay revealed a 65-amino-acid small membrane protein that was directly induced by EvgA and directly interacted with the sensor PhoQ at the inner membrane (7). We have annotated this small protein as SafA (Sensor-associating-factor A) (15) and termed it a “connector” (26), because it connects the EvgS/EvgA and PhoQ/PhoP systems.

Several other small-sized proteins that connect TCSs, such as PmrD, connecting the PhoQ/PhoP and PmrB/PmrA systems in Salmonella bacteria (16, 19), MzrA in E. coli, connecting the CpxA/CpxR and EnvZ/OmpR systems (11), and IraM in E. coli, connecting PhoQ/PhoP and response regulator RssB (3), have been previously reported. However, the biological roles of these connectors remained unclear. Since the PhoQ/PhoP system induces expression of AR genes gadW and hdeA (33), we focused on whether connector SafA played a role in the acid resistance conferred by the EvgS/EvgA system. In the present study, we constructed safA, phoP, and phoQ mutants of an EvgS-activated strain and examined how the deletion affected the acid resistance phenotype in exponential-phase cells. The SafA-activated PhoQ/PhoP system induced another connector, IraM, which bound to RssB, thus raising the cellular RpoS level in exponential-phase cells. IraM also participated in the acid resistance phenotype conferred to exponential-phase cells by EvgS/EvgA activation. We show that two connectors, SafA and IraM, contributed to the acid resistance phenotype in E. coli.

MATERIALS AND METHODS

Strains, plasmids, and growth media.

The E. coli strains and plasmids used in this study are listed in Table 1. The strains were aerobically grown at 37°C in LB medium (1% [wt/vol] Bacto tryptone [BD Diagnostics, Sparks, MD], 0.5% [wt/vol] Bacto yeast extract [BD], 1% [wt/vol] NaCl, pH 7.5) or minimal A medium (25) (pH 5.5 and 7.5, respectively) with the addition of 0.2% glucose, 0.0005% vitamin B₁, and 0.004% Casamino Acids. When necessary, selective antibiotic (100 μg/ml ampicillin, 12.5 μg/ml tetracycline, or 25 μg/ml kanamycin) was added to the medium. Mutants were obtained by a series of P1 transductions (Table 1). To prevent polar effects, kanamycin-resistant cassettes were eliminated from the safA and phoP mutants by transforming pCP20 as described by Datsenko and Wanner (5). Fragments of safA and iraM were amplified by PCR using specific primers and template MG1655 genomic DNA and were cloned into pBAD18, an arabinose promoter vector, to obtain pBADsafA and pBAD iraM. Arabinose (at a final concentration of 0.01%) was added to the culture for cells transformed with arabinose promoter vectors.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant genotype	Reference, source, or description
Strain
MG1655	Wild type
MG1655 evgS1	MG1655 evgS1 fadL71::Tn10	Eguchi et al. (7)
MG1655 b1500 (MG1655 safA)	MG1655 b1500	Eguchi et al. (7)
MG1655 evgS1 b1500 (MG1655 evgS1 safA)	MG1655 evgS1 fadL71::Tn10 b1500	Eguchi et al. (7)
MG1655 phoQ	MG1655 phoQ::kan	MG1655 × P1(JW1115)a
MG1655 evgS1 phoQ	MG1655 evgS1 fadL71::Tn10phoQ::kan	MG1655 evgS1 × P1(JW1115)
MG1655 phoP	MG1655 phoP	MG1655 × P1(JW1116)
MG1655 evgS1 phoP	MG1655 evgS1 fadL71::Tn10phoP	MG1655 evgS1 × P1(JW1116)
MG1655 evgS1 safA phoP	MG1655 evgS1 fadL71::Tn10 safA phoP	MG1655 evgS1 safA × P1(JW1116)
MG1655 rpoS	MG1655 rpoS::kan	MG1655 × P1(ZK1000)
MG1655 evgS1 rpoS	MG1655 evgS1 fadL71::Tn10rpoS::kan	MG1655 evgS1 × P1(ZK1000)
MG1655 iraM	MG1655 iraM::kan	MG1655 × P1(JW1147)
MG1655 evgS1 iraM	MG1655 evgS1 fadL71::Tn10iraM::kan	MG1655 evgS1 × P1(JW1147)
JW1115	BW25113 phoQ::kan	Keio collection (2)
JW1116	BW25113 phoP::kan	Baba et al. (2)
JW1147	BW25113 iraM::kan	Baba et al. (2)
ZK1000	W3110 ΔlacU169tna2 rpoS::kan	Gowrishankar et al. (13)
M15(pREP4)	Expression host	Qiagen
Plasmid
pCP20	FLP expression plasmid; amp, temperature-sensitive replication, and FLP synthesis	Datsenko and Wanner (5)
pBAD safA	Region encoding safA cloned into pBAD18	This study
pHO119 (pMW phoPQ)	Region encoding phoPQ cloned into a low-copy-number vector, pMW119	Kato et al. (18)
pBAD iraM	Region encoding IraM cloned into pBAD18	This study
pPhoP	Region encoding PhoP cloned into pQE30	Yamamoto et al. (31)

^aRecipient strain × P1 (donor strain) for P1 transduction.

Acid resistance assay.

Acid resistance was determined by the method of Masuda and Church (23). A single colony of an E. coli strain was inoculated in 5 ml of LB medium (containing appropriate antibiotics) and grown overnight with aeration at 37°C. LB or minimal A medium (10 ml [with ampicillin for transformants]) was inoculated with 0.1 ml of the overnight culture and grown at 37°C with aeration. When the culture reached a cell optical density at 600 nm (OD₆₀₀) of 0.75 to 0.8 for LB and 0.15 to 0.2 for minimal A medium, 50 μl was transferred to 1.95 ml of prewarmed LB medium (adjusted to pH 2.5 with HCl) and incubated at 37°C for 1 h. Viable cells in the acidified culture were counted by plating serial dilutions onto LB plates. The percentage of acid survival was calculated as the number of viable cells remaining after acid treatment divided by the viable counts before acid treatment. Each experiment was repeated at least three times, and the means and standard errors (SEs) were determined.

Total RNA preparation and S1 nuclease assay.

Total RNA was extracted with hot phenol from a mid-exponential-phase culture (OD₆₀₀ = 0.75 to 0.8) at 60°C (1). An S1 nuclease assay was carried out as described by Kato et al. (17). In brief, ³²P-end-labeled probe was prepared by PCR using specific primers, MG1655 genomic DNA as the template, and ExTaq DNA polymerase (Takara Bio, Otsu, Japan). A mixture of ³²P-end-labeled probe and 100 μg of total RNA was incubated for 10 min at 75°C, gradually cooled to 37°C, and incubated overnight for hybridization followed by S1 nuclease (Takara Bio) digestion for 10 min. Undigested RNA-probe DNA was extracted with phenol, precipitated with ethanol, and subjected to electrophoresis on a 6% (wt/vol) polyacrylamide sequencing gel containing 8 M urea. The radioactivity of the transcripts was measured by the use of an FLA-7000 laser scanner (Fuji Film, Tokyo, Japan) optimized for quantitative phosphorimaging and Multi Gauge version 3.0 analytic software (Fuji Film).

RpoS detection.

The cells were grown in LB medium with aeration, as described for the acid resistance assay. At an OD₆₀₀ of 0.6, 500 μl of the culture was sampled and added to 500 μl of ice-cold 10% trichloroacetic acid (TCA), vortexed, and put on ice for 30 min. For the RpoS stability assay, chloramphenicol was added to the culture at an OD₆₀₀ of 0.6 and a final concentration of 100 μg/ml, and the culture was sampled after 0, 1, 5, 15, and 20 min. The TCA-denatured samples were centrifuged for 20 min at 17,800 × g and 4°C. The pellets were washed with 500 μl of acetone and centrifuged for 5 min at 17,800 × g and 4°C. The resultant pellets were dissolved in 50 μl of 1× sample buffer for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and boiled for 5 min, and 20 μl of each sample was subjected to SDS-PAGE. The electrophoresed proteins were then transferred to a polyvinylidene difluoride membrane (Immobilon-P transfer membranes; Millipore, Billerica, MA) and probed with anti-RpoS antiserum. Detection was carried out with an ECL peroxidase-labeled anti-rabbit antibody and ECL Plus solution (ECL Plus Western blotting detection system; GE Healthcare Bio-Science, Piscataway, NJ), and the signals were acquired by the use of an ImageQuant Imager 400 system (GE Healthcare Bio-Science).

PhoP purification.

For a gel shift assay, His-tagged PhoP was subjected to affinity purification using a lysate of M15(pREP4) transformed with pPhoP, as described previously (31). Protein concentrations were determined by the method of Bradford (Bio-Rad protein assay; Bio-Rad Laboratories, Hercules, CA) using bovine serum albumin (BSA) as a standard.

Gel shift assay.

The ³²P-labeled probe for the iraM promoter used in the S1 nuclease assay was also used in the gel shift assay. The probe was incubated at 37°C for 10 min with purified His-tagged PhoP or BSA in a binding buffer (50 mM Tris-HCl [pH 7.8], 50 mM NaCl, 3 mM magnesium acetate, 0.1 mM EDTA, 0.1 mM dithiothreitol [DTT], 25 μg/ml BSA). After addition of the dye solution (40% [vol/vol] glycerol, 0.025% [wt/vol] bromophenol blue, 0.025% [wt/vol] xylene cyanol), the mixtures were directly subjected to 6% polyacrylamide gel electrophoresis in 1× Tris-acetate-EDTA (TAE) (pH 6), and the results were measured by FLA-7000.

RESULTS AND DISCUSSION

Contribution of SafA to acid resistance conferred by EvgS/EvgA activation.

When the EvgS/EvgA system was activated by a mutation in sensor EvgS (evgS1 mutation), 1% of the exponential-phase cells survived after acid treatment in LB medium (pH 2.5) at 37°C for 1 h, whereas only 0.004% of the wild-type cells survived under that set of conditions (Fig. 1A). This result is consistent with our previous findings (15). To determine whether the signal transduction cascade from EvgS/EvgA to PhoQ/PhoP contributed to the acid resistance phenotype conferred by EvgS/EvgA activation, we deleted safA, phoQ, or phoP from the evgS1 mutant by P1 transduction (Table 1). When these mutants were grown to mid-exponential phase and subjected to acid resistance assay, the evgS1 safA double mutant showed only 0.03% survival. Complementation of the evgS1 safA mutant with a safA-expressing plasmid rescued the safA phenotype and increased the survival rate to 2%. However, pBADsafA was not able to rescue the evgS1 phoP or evgS1 phoQ mutant cells. This clearly demonstrates that connector SafA contributed to the acid resistance phenotype via the PhoQ/PhoP system. For complementation of the evgS1 phoP and evgS1 phoQ mutants, we used a low-copy-number plasmid that included the phoPQ operon (pMWphoPQ). This approach was chosen because increasing the amount of either the response regulator or the sensor of a TCS may cause an imbalance between the two proteins, resulting to unexpected responses. As shown in Fig. 1B, complementation with pMWphoPQ partly restored the acid resistance phenotype of the evgS1 phoP and evgS1 phoQ mutants, resulting in survival rates of 1.5 and 0.8%, respectively. The reason why the survival rates did not recover to the level of evgS1 plus vector is not clear. Enhanced expression of phoP and phoQ from pMWphoPQ may have hindered the hyperactivation of the PhoQ/PhoP system by SafA. Next, a triple mutant, evgS1 safA phoP, was compared to the double mutants evgS1 safA and evgS1 phoP (Fig. 1C). The triple mutant showed a level of survival comparable to that of the two double mutants, confirming that SafA acted on the same pathway as PhoP.

FIG. 1.

Connector SafA contributes to acid resistance conferred by EvgS/EvgA activation via the PhoQ/PhoP system and IraM. Cells grown to mid-exponential phase in LB medium (ampicillin added for transformants) were treated in LB medium (pH 2.5) for 1 h, and their survival rates were determined. (A) Complementation of the evgS1 safA mutant with a safA-expressing plasmid. (B) Complementation of the evgS1 phoP and evgS1 phoQ mutants with a phoPQ-expressing p1asmid. (C) Comparison of the evg1 safA phoP mutant with the evgS1 safA and evgS1 phoP mutants. (D) Complementation of the evgS1 iraM mutant with an iraM-expressing plasmid. The vector used in the experiments represented by panels A and D was pBAD18 and in that represented by panel B was pMW119. Results are given as averages (values shown in parentheses) of the results of three independent culture experiments with standard errors.

Upregulation of RpoS-regulated genes by EvgS/EvgA activation.

In E. coli, stationary-phase cells are usually more acid resistant than exponential-phase cells (6). We wondered whether a change similar to that seen in the stationary phase occurs in the EvgS/EvgA-activated exponential-phase cells. In fact, in our former DNA microarray analysis comparing evgS1 mutants to wild-type E. coli (9), increases of expression in the evgS1 mutant were seen for as many as 25 genes (including katE, appA, dps, ficA, and osmY) among the 65 RpoS upregulated genes reported by Loewen et al. (20). Also, expression of two genes (mutS and sdhA) among the eight reported RpoS-downregulated genes (20) was repressed in the evgS1 mutant.

S1 nuclease assays were performed with three genes (katE, gadE, and ydeO) that were activated by the EvgS/EvgA system (Fig. 2A). The katE gene, encoding catalase, requires RpoS for transcription. The gadE and ydeO genes encode the transcriptional regulators of the AR genes, and GadE, especially, is the central regulator for AR genes. The evgS1 mutant (lanes 2) showed increased expression of these three genes compared with the wild type (lanes 1), confirming our microarray results. When rpoS was deleted from the evgS1 mutant, the expression of katE was abolished and that of gadE decreased (lanes 4). On the other hand, expression of ydeO did not change between the evgS1 mutant and the evgS1 rpoS double mutant. These results indicate that transcription of gadE was partly dependent on the presence of RpoS but that transcription of ydeO was totally independent of RpoS. Recently, the upstream region of gadE was intensively investigated by Sayed and Foster (29), who identified three promoters (P1 to P3). RpoS, which is directly involved in transcription from the P1 promoter, activates the P2 and P3 promoters via transcriptional regulators GadX and GadW. Thus, our finding that gadE transcription is partly dependent on RpoS is consistent with the previous results (29). As shown in Fig. 2, only the transcripts from the P1 and P2 promoters of gadE were detected, because the gadE probe used in Fig. 2 covered only the P1 and P2 promoters. Using a different gadE probe encompassing the P1 to P3 promoters, we confirmed that the P3 promoter was also induced by activation of EvgS/EvgA (data not shown).

FIG. 2.

Induction of gadE by EvgS/EvgA activation involves RpoS. S1 nuclease assays were performed with total RNA (100 μg/lane) from mid-exponential-phase cells. (A) Lanes 1, MG1655; lanes 2, MG1655 evgS1; lanes 3, MG1655 rpoS; lanes 4, MG1655 evgS1 rpoS. (B) Lanes 1, MG1655; lanes 2, MG1655 evgS1; lanes 3, MG1655 evgS1 safA; lanes 4, MG1655 evgS1 phoP; lanes 5, MG1655 evgS1 phoQ). ³²P-end-labeled probes of katE, gadE, and ydeO were used in the panels as indicated. Transcripts induced by EvgS/EvgA activation are indicated by arrows. Lanes A+G, Maxam-Gilbert sequencing ladder.

We next investigated how deletion of safA, phoP, and phoQ influences katE and gadE expression. As shown in Fig. 2B, levels of transcription from katE and gadE P1 decreased in the evgS1 safA, evgS1 phoP, and evgS1 phoQ double mutants. Levels of gadE P2 transcription also decreased, but the decrease was not as drastic as that seen for katE and gadE P1 transcription. These results indicated that the enhanced transcription from the RpoS-dependent katE and gadE promoters mediated by the activation of EvgS/EvgA required the SafA-PhoQ/PhoP pathway. Thus, it was presumed that the activation of EvgS/EvgA increased the cellular RpoS level in a SafA-PhoQ/PhoP-dependent manner.

Accumulation of RpoS induced by EvgS/EvgA activation at the exponential phase requires SafA.

To determine the cellular RpoS level, Western blotting using anti-RpoS antiserum was performed using proteins from the exponential-phase cells (Fig. 3). Compared to the wild type, the evgS1 mutant showed an increased RpoS level; however, that level decreased as a result of deletion of safA, phoP, or phoQ, indicating that the activation of the PhoQ/PhoP system by SafA increased the cellular RpoS level (Fig. 3A).

FIG. 3.

Cellular RpoS level in exponential phase cells. Cells were grown to mid-exponential phase, precipitated by TCA, and subjected to SDS-PAGE (proteins from 200 μl of cell culture/lane). Western blotting was performed using anti-RpoS antiserum, and signals were detected by the use of ECL peroxidase-labeled anti-rabbit antibody and ECL Plus solution. (A) Accumulation of RpoS at the exponential phase by EvgS/EvgA activation requires SafA, PhoP, and PhoQ. (B) Degradation of RpoS at the exponential phase is blocked by EvgS/EvgA activation. Chloramphenicol was added to MG1655 and MG1655 evgS1 cells at the mid-exponential phase, and samples were taken after the indicated period. (C) Accumulation of RpoS at the exponential phase by EvgS/EvgA activation requires IraM.

It is common knowledge that RpoS turnover is determined at various stages (transcription, translation, and posttranslation) (20). Our transcriptome analysis of evgS1 mutant indicated that the transcriptional level of rpoS was not affected (9); this was also confirmed by the results of an S1 nuclease assay (data not shown). Thus, we investigated the RpoS level after adding chloramphenicol to the exponential-phase cells. As shown in Fig. 3B, aborting protein synthesis by adding chloramphenicol to the wild-type cells resulted in the loss of RpoS within 1 min. On the other hand, RpoS was stable in the evgS1 mutant for as long as 20 min after the addition of chloramphenicol. These results indicated that RpoS degradation was blocked by EvgS/EvgA activation.

Accumulation of RpoS by EvgS/EvgA activation at the exponential phase also requires IraM.

Response regulator RssB is one factor that controls the stability of RpoS (27). RssB binds to RpoS and delivers it to the ClpXP protease for degradation (32). Recently, antiadaptors (IraP, IraD, and IraM) of RssB have been reported (3). These small proteins bind to RssB and prevent it from binding to RpoS, increasing RpoS stability. One of the antiadaptors, a 107-amino-acid IraM, has been reported to be induced by the PhoQ/PhoP system under low Mg²⁺ conditions. Therefore, IraM can also be called a connector of TCSs because it connects the PhoQ/PhoP system to response regulator RssB (26). We examined whether IraM was involved in the accumulation of RpoS induced by EvgS/EvgA activation at the exponential phase. Deleting iraM from the evgS1 mutant decreased the RpoS level (Fig. 3C). Complementation of iraM with an iraM-expressing plasmid restored the RpoS accumulation. Therefore, RpoS accumulation by EvgS/EvgA required IraM.

Next, we checked whether iraM was actually induced by EvgS/EvgA activation. An S1 nuclease assay was performed with iraM (Fig. 4A). Expression of iraM was induced in the evgS1 mutant but not in the evgS1 safA, evgS1 phoP, or evgS1 phoQ double mutant. This indicated that iraM was indeed induced by the activation of PhoQ/PhoP via SafA. Since the S1 nuclease assay enabled us to determine the transcriptional start site of iraM (121 nucleotides [nt] upstream of the start codon), we further analyzed the promoter region of iraM and found a PhoP box near the −35 region (Fig. 4B) (3, 31). In fact, purified PhoP directly bound to the iraM probe used in the S1 nuclease assay (extending from +35 to −559 nt of the start codon) in a gel shift assay (Fig. 4C).

FIG. 4.

Induction of iraM by EvgS/EvgA activation requires SafA, PhoP, and PhoQ and is directly regulated by PhoP. (A) S1 nuclease assays against ³²P-end-labeled probes of iraM were performed as described for Fig. 2. (B) Promoter analysis of iraM. Thin and thick arrows indicate the transcription start site and direct repeat of the PhoP box, respectively. The −10 and −35 hexamers are underlined, and the start site and PhoP box are indicated in uppercase letters. (C) Gel shift assay. The iraM probes used for the S1 nuclease assay performed as described for panel A were incubated with 0, 0.05, 0.1, 0.2, 0.39, 0.75, 1.56, 3.13, and 6.25 pmol of purified PhoP and 6.25 pmol of BSA. Mixtures were subjected to 6% PAGE in 1× TAE (pH 6).

IraM contributes to the acid resistance conferred by EvgS/EvgA.

Finally, we examined whether IraM contributes to the acid resistance phenotype. Deletion of iraM from the evgS1 mutant decreased its survival rate to about one-third of the previous level (Fig. 1D). Complementation of iraM with an iraM-expressing plasmid restored the survival rate to the wild-type level or even higher. This may reflect the higher level of RpoS detected in MG1655 evgS1 iraM/pBADiraM compared to MG1655 evgS1/pBAD (Fig. 3C).

Instead of using the evgS1 mutation, the EvgS/EvgA systems of the wild-type strain and safA mutant were activated by growing the cells in minimal medium adjusted to pH 5.5 (4, 21) (Fig. 5). Compared with the pH 7.5 medium results, activation of the EvgS/EvgA system induced acid resistance. Deletion of safA reduced the survival rate to about one-half of the wild-type level, and complementation with a safA-expressing plasmid rescued the safA phenotype. Thus, SafA also contributed to the acid resistance induced by EvgS/EvgA activation in the wild-type strain but less than the contribution seen with the evgS1 mutant. The survival rates of the strains whose results are shown in Fig. 5 were much higher than those shown in Fig. 1A. It is presumed that cells were under more stress in minimal medium compared to LB medium and that the stressed state induced acid resistance via other factors.

FIG. 5.

Connector SafA contributes to acid resistance conferred by EvgS/EvgA activation by low pH in minimal medium. Cells grown to mid-exponential phase in minimal A medium (pH 7.5 or 5.5) were treated in LB medium (pH 2.5) for 1 h, and their survival rates were determined. pBAD18 was used as the vector. Results are given as averages (values shown in parentheses) of the results of three independent culture experiments with standard errors.

In conclusion, we propose a signal transduction cascade to regulate acid resistance genes. Activation of the EvgS/EvgA system directly induces connector SafA; SafA directly activates the PhoQ/PhoP system; the PhoQ/PhoP system directly induces connector IraM; and IraM directly binds to response regulator RssB and stabilizes RpoS (Fig. 6). An increase in the RpoS level enhances expression of gadE, as well as that of gadX (22); the enhancement of expression is caused by RpoS inducing a small RNA, gadY, that stabilizes the gadX mRNA (28). The transcriptional regulator GadX also induces gadE expression. In addition to the stabilization of RpoS, activation of the PhoQ/PhoP system also induces expression of acid resistance genes gadW and hdeA (33). This may explain why deletion of safA, phoP, or phoQ from an evgS1 mutant reduced the survival rate more than deletion of iraM (Fig. 1). Furthermore, response regulator PhoP has also been reported to repress safA expression, forming a negative-feedback loop (4). Thus, a complex network of AR gene regulation initiated by the EvgS/EvgA system is proposed (Fig. 6). EvgS/EvgA induces the EvgA-YdeO-GadE cascade of AR gene regulation, and a branched pathway operating via SafA activates the PhoQ/PhoP system, leading to an enhanced level of cellular RpoS. RpoS induces gadE, the central regulator of AR genes, enforcing the acid resistance phenotype. Since RpoS is a general stress response sigma factor, involved in responses to such stresses as starvation and high osmolarity, activation of the EvgS/EvgA system may induce responses to other stresses in addition to acid stress due to the enhanced level of cellular RpoS.

FIG. 6.

Model of regulation network of AR genes initiated by the EvgS/EvgA system. EvgS/EvgA induces the EvgA-YdeO-GadE cascade of AR gene regulation, and a branched pathway operating via SafA activates the PhoQ/PhoP system, leading to an enhanced level of cellular RpoS via IraM. RpoS induces gadE, the central regulator of AR genes, enforcing the acid resistance phenotype. The black bar in the safA promoter represents the inverted repeat for EvgA binding, gray bars in the iraM, hdeA, and gadW promoters represent PhoP boxes, and white bars in the gadA, hdeA, and gadB promoters represent GadE boxes.