Alternative Sigma Factor RpoE Is Important for Vibrio parahaemolyticus Cell Envelope Stress Response and Intestinal Colonization

Abstract

Vibrio parahaemolyticus is a halophile that inhabits brackish waters and a wide range of hosts, including crustaceans, fish, mollusks, and humans. In humans, it is the leading cause of bacterial seafood-borne gastroenteritis. The focus of this work was to determine the role of alternative sigma factors in the stress response of V. parahaemolyticus RIMD2210633, an O3:K6 pandemic isolate. Bioinformatics identified five putative extracytoplasmic function (ECF) family of alternative sigma factors: VP0055, VP2210, VP2358, VP2578, and VPA1690. ECF factors typically respond to cell wall/cell envelope stress, iron levels, and the oxidation state of the cell. We have demonstrated here that one such sigma factor, VP2578, a homologue of RpoE from Escherichia coli, is important for survival under a number of cell envelope stress conditions and in gastrointestinal colonization of a streptomycin-treated adult mouse. In this study, we determined that an rpoE deletion mutant strain BHM2578 compared to the wild type (WT) was significantly more sensitive to polymyxin B, ethanol, and high-temperature stresses. We demonstrated that in in vivo competition assays between the rpoE mutant and the WT marked with the β-galactosidase gene lacZ (WBWlacZ), the mutant strain was defective in colonization compared to the WT. In contrast, deletion of the rpoS stress response regulator did not affect in vivo survival. In addition, we examined the role of the outer membrane protein, OmpU, which in V. cholerae is proposed to be the sole activator of RpoE. We found that an ompU deletion mutant was sensitive to bile salt stress but resistant to polymyxin B stress, indicating OmpU is not essential for the cell envelope stress responses or RpoE function. Overall, these data demonstrate that RpoE is a key cell envelope stress response regulator and, similar to E. coli, RpoE may have several factors that stimulate its function.

INTRODUCTION

Vibrio parahaemolyticus is a Gram-negative organism, commonly found within brackish waters, such as coastal marine and estuarine waters, worldwide (1–3). V. parahaemolyticus is frequently associated with bacterial seafood-borne gastroenteritis following the consumption of raw or undercooked fish and shellfish, resulting from this organism's ability to colonize shellfish in high numbers (3–5). Disease caused by V. parahaemolyticus is marked by diarrhea and abdominal pain. In addition, septicemia and mortality have been documented in immunocompromised individuals and in cases following exposure of open wounds to the organism (6). In the United States the number of incidences of illness associated with Vibrio, including V. parahaemolyticus, has increased. According to the Centers for Disease Control and Prevention, the rate of infections caused by Vibrio species in 2012 had increased 43% since 2006-2008, whereas the rates of infections caused by other enteric species, such as Escherichia coli O157 and Salmonella enterica, showed no change during that same time period (7).

V. parahaemolyticus must be able to alternate between natural and host environments and possess the ability to respond to rapid changes in the extracellular environment in order to survive and cause disease. Previously, we developed a streptomycin-treated adult murine model of colonization to study the bacterial factors required for host colonization (8, 9). This animal model was used to demonstrate the importance of the Vibrio specific regulatory system ToxRS in colonization of the mouse gastrointestinal tract (8). ToxRS was shown to be important for survival under acid (organic and inorganic) stresses, sodium dodecyl sulfate (SDS), and bile salt stresses in V. parahaemolyticus through its positive regulation of the outer membrane protein (OMP) OmpU (8, 10). A toxRS mutant was defective in mouse intestinal colonization, indicating its importance in host-pathogen interactions (8).

Another method by which bacteria regulate gene expression in response to changing extracellular conditions is through the use of alternative sigma factors. The stress response sigma factor, RpoS, has been studied in V. cholerae, V. vulnificus, V. anguillarum, V. harveyi, and V. alginolyticus and has been shown to be involved in a number of stress responses in these species, including starvation, osmolarity, ethanol, hydrogen peroxide, and acid stress responses (11–17). Previously, RpoS was also studied in V. parahaemolyticus and was found to play a limited role in the stress response in this organism (10). In addition, it has been demonstrated that preadaptation of V. parahaemolyticus to high-salinity results in enhanced survival under lethal acid stress, and this phenotype is independent of RpoS (18). RpoS was also shown not to be required for oyster colonization (19). More recently, we examined RpoN, a sigma factor that has been shown to regulate over 500 genes in V. cholerae (20–26). The V. parahaemolyticus rpoN deletion mutant was nonmotile and defective in biofilm formation (9). In in vivo intestinal colonization assays, the rpoN mutant was shown to be a hypercolonizer compared to the wild type (WT) in the streptomycin-pretreated mouse model (9). To determine whether loss of motility was the cause of the increased fitness in vivo, analysis of deletion mutants in the polar flagellum sigma FliAP and a double mutant in FliAP and the lateral flagella FliAL sigma factor were examined (9). It was found that these mutants were slightly better colonizers than the WT but not to the same extent as the rpoN mutant, suggesting that motility was not the cause of the phenotype. It was shown that the rpoN mutant had a metabolic advantage over WT since it grew at a higher rate than WT in intestinal mucus, suggesting that carbon utilization is an important colonization factor (9).

The focus of the present study was to elucidate the role of the extracytoplasmic function (ECF) sigma factor, RpoE, in stress response and colonization of V. parahaemolyticus. RpoE was first identified and studied in E. coli and has been shown to be an essential protein, necessary for cell envelope integrity (27–30). Under optimum growth conditions in E. coli, RpoE activity is low, and it is bound to the inner membrane by RseA, an antisigma factor (28, 31). Under stress conditions that result in the presence of misfolded proteins in the periplasmic space, RseA is degraded by DegS with release of RpoE (32). RpoE has been characterized in a number of Vibrio species (33–38). For example, in V. vulnificus an rpoE mutant was found to be sensitive to ethanol, SDS, and hydrogen peroxide but was not attenuated for virulence in mice (36). In V. harveyi, rpoE appeared to be essential since an rpoE mutant could not be constructed (38). It was shown that overexpression of rpoE in this species resulted in a reduction of hemolytic activity and attenuation for colonization in shrimp (38). In a classical biotype V. cholerae strain an rpoE mutant was important for intestinal survival and virulence in an infant mouse model (34). The rpoE mutant strain was sensitive to 3% ethanol stress but resistant to heat stress, bile salts, hydrogen peroxide, the antimicrobial peptide polymyxin B, osmolarity, and pH stresses (34). In contrast, in an El Tor biotype V. cholerae strain, an rpoE mutant was determined to be sensitive to a bioactive peptide P2 and polymyxin B (35). This group previously showed that the outer membrane protein OmpU was important for resistance to the bioactive peptide P2 and polymyxin B (39). It was shown that in V. cholerae an rpoE mutant could only be made in the presence of suppressor mutations, 75% of which occurred in the promoter region of ompU, and that OmpU is a key requirement for RpoE function (37).

The role of alternative sigma factors in V. parahaemolyticus has not been extensively studied. The aim of the present study was to determine the function of the rpoE homologue VP2578 in V. parahaemolyticus. To accomplish this, an in-frame deletion of VP2578 was constructed and examined under a number of stress conditions. The effect of the deletion of this global regulator, RpoE, on the ability of V. parahaemolyticus to colonize the murine gastrointestinal tract was also examined.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

All bacterial strains and plasmids used in the present study are listed in Table 1. A streptomycin-resistant V. parahaemolyticus RIMD2210633 O3:K6 clinical isolate and a streptomycin-resistant β-galactosidase-positive RIMD2210633 isolate named WBWlacZ were used (8–10). Genetic manipulations to construct the ΔrpoE mutant strain used E. coli strains DH5α λpir and β2155 λpir. Unless otherwise noted, all strains were grown at 37°C in Luria-Bertani (LB) broth (Thermo Fisher Scientific, Waltham, MA) with aeration (250 rpm). The final NaCl concentration was adjusted to either 1% for E. coli strains or 3% for V. parahaemolyticus strains. The E. coli β2155, a diaminopimelic acid (DAP) auxotroph, was grown on medium supplemented with 0.3 mM DAP (Sigma-Aldrich, St. Louis, MO). When needed, antibiotics were added to the medium at the following concentrations: streptomycin, 200 μg/ml; chloramphenicol, 25 μg/ml; and ampicillin, 100 μg/ml.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristicsa	Source or reference(s)
Strains
Vibrio parahaemolyticus
RIMD2210633	O3:K6 clinical isolate; Strr	44
WBWlacZ	RIMD2210633::lacZ; Strr	8, 9
BHM2578 (ΔrpoE)	RIMD2210633 ΔrpoE (VP2578)	This study
LMN2553(ΔrpoS)	RIMD2210633 ΔrpoS (VP2553)	10
BHM2467(ΔompU)	RIMD2210633 ΔompU (VP2467)	This study
BHM2578C	ΔrpoE pBA rpoE	This study
WBW0819-820(ΔtoxRS)	RIMD2210633 ΔtoxRS (VP VP0819-20)	8
Escherichia coli
DH5αλpir	Δlac pir
B2155 DAP	ΔdapA::erm pir for bacterial conjugation
DH5pΔrpoE	DH5α λpir containing pJΔrpoE	This study
B2155pΔrpoE	B2155 DAP containing pDSΔrpoE	This study
DH5pΔompU	DH5α λpir containing pJΔompU	This study
B2155 DAP-ΔompU	B2155 DAP containing pDSΔompU	This study
B2155 DAP-prpoE	B2155 DAP containing pBrpoE	This study
Plasmids
pJET1.2	Cloning vector
pJEΔrpoE	pJET1.2 harboring truncated rpoE	This study
pJEΔompU	pJET1.2 harboring truncated ompU	This study
pDS132	Suicide plasmid; Cmr; SacB	52
pDSΔrpoE	pDS132 harboring truncated rpoE	This study
pDSΔompU	pDS132 harboring truncated ompU	This study
pBAD33	Expression vector, araC promoter; Cmr	53
pBArpoE	pBAD33 harboring rpoE gene	This study

^aCm^r, chloramphenicol resistance; Str^r, streptomycin resistance.

Phylogenetic analysis.

The phylogenetic tree was constructed from the alignment of sigma-70 (σ⁷⁰) family sigma factors using the amino acid sequences of the highly conserved domains 2 and 4. The software, Molecular Evolutionary Genetic Analysis version 5 (MEGA5), was used to construct a neighbor-joining tree using the Poisson model, complete deletion, and a bootstrap value of 1,000 (40–42). The locus tags used in the construction of the phylogenetic tree were as follows: VP2578, b2573, VMC_13050, VIBHAR_03542, VC2467, VVMO6_00468, VS_2625, VAA_03781, VF_2093, VSAL_I2531,vfu_A00830; VP2210, VIBHAR_03122, VVM_01689, vfu_A02554, VAA_02363,VC1045, VS0863, VPA1690, VMC_06000, VVM_01689, VS_0863, VIBHAR_03122, VAA_02363, VC1045, VF0972; VP2358, VMC_17260, VIBHAR_03284, VAA_03623, vfu_A02820, VC2302, VS_II1448, VF_A0766, VF_A0820; VP0055, VMC_10870, VIBHAR_00504, VVM_00142, VF2498, vfu_A00240, VSAL_I0117; and b4293.

Construction of V. parahaemolyticus ΔrpoE and ΔompU mutants.

Splicing by overlap extension (SOE) PCR and homologous recombination (double-crossover event) were used to make the in-frame deletions of genes of interest (43). Using the V. parahaemolyticus RIMD2210633 genome sequence as the template, primers were designed and purchased from Integrated DNA Technologies (Coralville, IA) to perform SOE PCR and construct an in-frame deletion mutation of the rpoE gene VP2578 (44). A 177-bp truncated version of rpoE was constructed. Briefly, the rpoE truncated PCR fragment was constructed by amplifying two products by using the primer pairs SOEArpoE (CCGTATTGCTGCACACCTAA)/SOEBrpoE (ATTGCTGAAGAGATGGATTG) and SOECrpoE (CAATCCATCACTTCAGCAATCAGCAGGTTGAATGCTTGCT)/SOEDrpoE (TGCGTGACATCCGTCACTAAG), which were ligated and cloned into the pJET1.2 vector and transformed into the E. coli strain DH5α λpir. Plasmid DNA was extracted from DH5α λpir harboring pJEΔrpoE and ligated into the suicide vector pDS132 which was designated pDSΔrpoE. pDSΔrpoE was subsequently transformed into the E. coli strain β2155 and then conjugated with V. parahaemolyticus RIMD2210633 via cross-streaking onto LB plates containing 0.3 mM DAP. Growth from these plates was streaked onto LB medium containing 3% NaCl (LB–3% NaCl) plus streptomycin and chloramphenicol to select only for V. parahaemolyticus containing pDSΔrpoE. Exconjugate colonies (positive for single-crossover event) were cultured overnight in the absence of antibiotics to promote recombination and serial dilutions were plated on LB–3% NaCl plus 10% sucrose to select for cells which had lost pDSΔrpoE. Double-crossover deletion mutants were then screened and confirmed by PCR using the SOEFLrpoEF (ATTCTTACTCGCCTCGCTCA) and SOEFLrpoER (GACACGTAAAGCCAACGACA) primers. The same protocol was followed to construct the ompU (VP2467) deletion mutant strain BHM2467 using the primer pairs SOEAompU (CAGCATAACGAACCGAATCA)/SOEBompU (AGAAGTGCCGTCTTGGTTGT) and SOECompU (ACAACCAAGACGGCACTTCTGGTGGCAACACTACAGCAT)/SOEDompU (GTTGGACGGATACCATCGAG). A double-crossover deletion mutant was confirmed by PCR using the SOEFLompUL (CCACGTAGGGTCATTGGAAC) and SOEFLompUR (CGCAGGTGGAAATAGTTGGT) primer pair.

Construction of rpoE complement.

The ΔrpoE strain was complemented with the rpoE gene creating strain BHM2578C. PCR primers rpoEF (AGAAGAGTAGGGGCATAACAAA) and rpoER (TGTTCTTTGTCAGCCATTGTTT) were designed to amplify a promoterless copy of VP2578 encoding rpoE from V. parahaemolyticus RIMD2210633, which was cloned into vector pJET1.2 and transformed into E. coli DH5α λpir. The fragment was then subcloned into the vector pBAD33, resulting in pBArpoE, and subsequently transformed into E. coli β2155 λpir, which was then cross-streaked with the rpoE mutant strain BHM2578 onto LB plates containing 0.3 mM DAP. The resulting bacterial growth was then streaked onto LB–3% NaCl plates containing chloramphenicol and streptomycin (but no DAP) to positively select for ΔrpoE cells harboring pBArpoE. To induce the expression of the complemented gene, BHM2578C harboring pBArpoE were grown in the presence of 0.10% arabinose.

Growth analysis.

Strains were grown overnight at 37°C with aeration in 5 ml of LB–3% NaCl. To set up growth assays, 5 μl from each overnight culture was used to inoculate 200 μl of LB–1% NaCl or LB–3% NaCl. Alternatively, late-log-phase cultures (4 h) were used in place of stationary-phase cultures for assays in LB–9% NaCl. Growth assays were carried out at 37°C for a 24 h period. The optical density at 595 nm (OD₅₉₅) was determined hourly by using a Sunrise plate reader (Tecan Group, Ltd.) and Magellan plate reader software. The data were plotted in Origin 8.5. The experiment was repeated as described above but grown at 42°C when appropriate.

To assay growth in mouse intestinal mucus, the same experiment as outlined above was performed in M9 supplemented with 30 μg/ml large intestinal, small intestinal, or cecum streptomycin-treated mouse mucus as the sole carbon source. Intestinal mucus was extracted from the gastrointestinal tract of mice as follows. Mice were pretreated with streptomycin, and 24 h later mice were sacrificed, and their gastrointestinal (GI) tracts were dissected; mucus was collected by flushing the GI sections with phosphate-buffered saline (PBS) and then by gently scraping the walls of the intestine. Extracts from the small intestine, cecum, or large intestine (LI) were collected and stored at −80°C. Overnight cultures were set up as described for previous assays; however, cells were pelleted and washed in M9 (Thermo Fisher Scientific) medium and resuspended in M9 before the growth assay was performed.

Stress survival assays.

The cells were first grown overnight at 37°C with aeration; 100 μl of this culture was used to inoculate 5 ml of LB–3% NaCl, and the strains assayed (the WT; ΔrpoE, ΔrpoS, and ΔompU mutants; and a ΔrpoE mutant complemented with rpoE) were grown to early log phase (2 h) at 37°C with aeration. Cells were pelleted at 4,000 rpm and then resuspended in LB–3% NaCl containing either 200 μg of polymyxin B (Sigma-Aldrich), 15% bile salt (Sodium cholate) (Sigma-Aldrich), 0.5% SDS (Thermo Fisher Scientific), or 10% ethanol (EtOH). At 0, 30, and 60 min, the cells were serially diluted in PBS (Sigma-Aldrich) and plated on LB–3% NaCl agar plates (1.5% agar). Colony counts were used to determine CFU at the indicated time points, and the percent survival was determined by dividing the number of CFU at 30 or 60 min by the initial starting concentration at 0 min. For survival assays in LI mucus, the same procedure as outlined above was performed except that CFU were taken at the 6-, 12-, and 24-h time points. Each experiment was performed in duplicate with at least two biological replicates.

In vivo colonization.

This assay utilized a previously described streptomycin-treated mouse model of colonization and a LacZ-positive V. parahaemolyticus RIMD2210633 strain WBWlacZ (8). All experiments involving animals were approved by the University of Delaware Institutional Animal Care and Use Committee. Male C57BL/6 mice, aged 6 to 10 weeks were housed under specific-pathogen-free conditions in standard cages in groups (four to five per group) and provided standard mouse feed and water ad libitum. Treatment with streptomycin and inoculations were performed as previously described (8). Briefly, mice were treated orogastrically with streptomycin 24 h prior to infection. Food and water were returned upon antibiotic treatment. Prior to infection, the mice were fasted for 4 h and then inoculated with 100 μl of a bacterial suspension in PBS by gavage. Water was returned immediately upon infection, and food was returned at 2 h postinfection. The inoculum was prepared as follows. The WBWlacZ and rpoE mutant strains were grown in 5 ml of LB–3% NaCl overnight at 37°C, with shaking. The WBWlacZ strain is RIMD2210633 with a β-galactosidase gene knockin that enables a blue-white screen to differentiate between the ΔrpoE and WT strains (8, 9). Overnight cultures were diluted 1:40, allowed to grow to late log phase (4 h), and subsequently pelleted and washed at 4,000 rpm for 10 min. The cells were then resuspended in PBS to a concentration of ∼10¹⁰ CFU/ml. An aliquot of 1 ml was mixed in a 1:1 ratio of WT to WBWlacZ, ΔrpoE, or ΔrpoS strain in PBS, resulting in an inoculum of 10¹⁰ CFU/ml. Mice were inoculated orogastrically with 10⁹ CFU of that mixed culture. A 100-μl aliquot of the inoculum was added to 5 ml of LB–3% NaCl and incubated overnight with aeration at 37°C; serial dilutions were plated, and the CFU were counted to determine the in vitro competitive index (CI). Mice were sacrificed at 24 h postinoculation, and the gastrointestinal tract was harvested and homogenized in 8 ml of PBS. Samples were plated for CFU on LB–3% NaCl supplemented with streptomycin and X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) for a blue (WBWlacZ)-versus-white (WT or deletion mutant) screen after overnight incubation at 37°C. By dividing the ratio out by the ratio in, the CI was calculated: CI = ratio out_(mutant/WT)/ratio in_(mutant/WT). A CI of <1 would indicate a defect in vivo of the mutant strain, whereas a CI of >1 would indicate the mutant strain outcompeted WBWlacZ. An in vivo single localization colonization analysis was also performed to determine whether there were any differences in colonization localization between WT and ΔrpoE strains. The in vivo single infection localization analysis was performed: mice received 10⁹ CFU inoculum of either WT or ΔrpoE strain. Mice were sacrificed 24 h postinfection; the small intestine, cecum, and large intestine were harvested, and samples were placed in 8 ml of sterile PBS, mechanically homogenized, serially diluted in PBS, and plated on LB–3% NaCl plus streptomycin and X-Gal for CFU of each strain.

RESULTS

Identification of five putative ECF alternative sigmas in V. parahaemolyticus.

Alternative sigma factors are global regulators that enable bacteria to respond and adapt to changes in their environment. An analysis of the V. parahaemolyticus genome revealed 11 putative sigma factors in all: VP0055, VP0404, VP2210, VP2232, VP2358, VP2553, VP2578, VP2670, and VP2953 in chromosome I and VPA1555 and VPA1690 in chromosome II. A BLAST analysis of all sequenced Vibrio species (39 fully sequenced genomes available), using the sigma factors identified in V. parahaemolyticus as seeds, was conducted to determine the distribution of these sigma factors. VP0404 encoded RpoD (σ⁷⁰), the primary housekeeping sigma factor; VP2670 encoded the master flagellar regulator RpoN (σ⁵⁴); VP2553 encoded RpoS (σ³⁸), the stationary-phase stress sigma factor; and VP2953 encoded a homologue of RpoH (σ³²), the heat shock response sigma factor. All of these sigmas were present in all sequenced Vibrio species. The V. parahaemolyticus genome encoded two additional sigma factors involved in flagellar synthesis regulation: VP2232-encoded fliAP (σ²⁸), involved in polar flagellar synthesis, was present in all sequenced Vibrio species, and VPA1555-encoded fliAL (σ²⁸), required for lateral flagella synthesis, was present in only 13 Vibrio species. The V. parahaemolyticus genome encoded five putative ECF sigma factors: VP0055, VP2210, VP2358, VPA1690, and VP2578. VP0055 showed homology to an ECF in eight other Vibrio species. VP2210 encoded a putative ECF sigma and was present in nearly all species of sequenced Vibrio, with the exception of Vibrio fischeri and Vibrio salmonicida (also known as Allivibrio salmonicida). VP2358 encoded another putative ECF that was present in 30 Vibrio species. VPA1690 was a putative ECF sigma that was present in only three species of Vibrio. Lastly, VP2578 encoded a homologue of RpoE (σ²⁴) that was highly related (76% amino acid identity) to the E. coli RpoE and was present in all sequenced vibrios. VP2578 shared 91% amino acid identity with RpoE from V. cholerae, a species that has a total of eight sigma factors.

In order to examine the relationships among the five putative ECF alternative sigma factors further, a phylogenetic tree was constructed using the amino acid sequences from the following 10 representative species: E. coli, V. alginolyticus, V. anguillarum, V. cholerae, V. fischeri, V. furnissii, V. harveyi, V. parahaemolyticus, V. salmonicida, V. splendidus, and V. vulnificus. The sequences were aligned by CLUSTAL W and were used to construct a neighbor-joining tree. This phylogeny demonstrated that there were at least five major clades which encompass the identified putative ECF factors (Fig. 1). VP2578 clustered tightly with RpoE from all other Vibrio species examined and with the E. coli RpoE, indicating that this is the ancestral RpoE (Fig. 1). VP2210 also formed a tight clustering with ECF sigma factors from other Vibrio species, suggesting possible conservation in function. The other three V. parahaemolyticus ECF sigma factors each formed a distinct highly divergent branching pattern unlike the tight closely related clustering in both VP2578 and VP2210 lineages (Fig. 1).

FIG 1.

Phylogeny of ECF sigmas among Vibrio species. A phylogenetic tree was constructed from the alignment of alternative sigma factors by using the amino acid sequences of the highly conserved domains 2 and 4. The molecular evolutionary genetics analysis software, MEGA5, was used to construct a neighbor-joining tree using the Poisson model, complete deletion, and a bootstrap value of 1,000 (42). Abbreviations: Ec, E. coli; Vp, V. parahaemolyticus; Vc, V. cholerae; Vv; V. vulnificus; Vh, V. harveyi; Vf, V. fischeri; Va, V. alginolyticus; Vs, V. splendidus; Vaa, Vibrio anguillarum; Vfu, Vibrio furnissii; As, V. salmonicida (A. salmonicida). Phylogenetic analysis shows that there are five clusters of ECF-type sigma factors among Vibrio species and that VP2578 clusters closely with the conical E. coli RpoE, suggesting that this is the ancestral copy.

RpoE is required for V. parahaemolyticus cell envelope stress response.

In order to determine the role of the putative RpoE in the stress response of V. parahaemolyticus, we constructed an in-frame deletion mutation of the rpoE gene VP2578. This deletion mutant strain, BHM2578 (ΔrpoE), has 402 bp of the gene deleted. There was no detectable difference in growth between BHM2578 and WT in LB–3% NaCl, indicating that the deletion did not result in an overall growth defect (Fig. 2). We wanted to examine the role of RpoE on cell envelope stresses that would be encountered within the natural environment and within a host. In vivo bacteria encounter numerous antimicrobial peptides that are critical for innate antibacterial defense. We examined sensitivity to the peptide antibiotic polymyxin B of our rpoE mutant BHM2578 compared to WT by performing survival assays. In these assays, survival in 200 μg of polymyxin B showed the rpoE mutant is more sensitive in comparison to WT since the mutant strain had significantly reduced survival rate of 15% compared to 95% for WT after 30 min (Fig. 3A). The survival rate was restored to near WT levels when the mutant was complemented with rpoE (Fig. 3A). Next, we examined survival in LB–3% NaCl supplemented with 10% EtOH, a compound that targets the cell envelope. In these survival assays, the mutant was more sensitive, showing reduced survival rates in comparison to WT (Fig. 3B). Increased resistance to 10% EtOH stress was restored in the mutant similar to WT level via complementation (Fig. 3B). The survival of the rpoE mutant was also examined in the presence of 1 mM H₂O₂, and the mutant had reduced survival after 30 min in comparison to WT (data not shown). Growth analysis of the mutant and WT strains was examined at 42°C in LB–3% NaCl, and it was found that the mutant had a longer lag phase and reached a lower final optical density compared to WT, suggesting that it was more sensitive to high temperature. This temperature sensitivity in the mutant was alleviated via complementation with rpoE (Fig. 4A). V. parahaemolyticus is a moderate halophile, and whether or not RpoE played a role in the ability of the organism to adapt to changes in salinity was assayed. This was accomplished by comparing the growth of BHM2578 to that of the WT under low- and high-salt (NaCl) stress conditions (Fig. 2B and 4B). In high salinity (9% NaCl), there was no difference in the growth pattern between BHM2578 and WT, but a slight difference was noted in LB–1% NaCl (Fig. 2B and 4B). A previously described rpoS mutant, which encodes the stationary-phase sigma factor, was also examined under the same conditions described above and showed phenotypes similar to that of the WT (Fig. 2, 3, and 4).

FIG 2.

Growth analysis of WT and mutant strains under optimum and high-salt conditions. Cells were incubated over a 24-h period. An absorbance reading (595 nm) was taking hourly, and the data were then plotted in Origin8.5. (A) In LB–3% NaCl the rpoE mutant BHM2578 grows similarly to the WT. (B) In LB–9% NaCl, BHM2578 (ΔrpoE), as well as LMN2553 (the rpoS mutant), grow similarly to the WT. All cultures were grown in triplicate, with at least two biological replicates. Error bars indicate the standard deviations.

FIG 3.

Cell envelope stress response. WT (■), ΔrpoE (BHM2578, ●), ΔrpoS (LMN2553, △), and ΔrpoE pBArpoE (BHM2578C, ▼) cells were grown in LB–3% NaCl to early log phase and then subjected for 60 min to either polymyxin B stress (A) or EtOH stress (B). Survivability was determined by dividing the surviving population at various time points by the initial population. Each experiment was performed in duplicate, with at least two biological replicates. Error bars indicate the standard deviations.

FIG 4.

Growth curves of V. parahaemolyticus under high temperature and NaCl stress conditions. (A) Heat stress was examined by growth analysis at 42°C for WT, ΔrpoE, and ΔrpoS strains. (B) Low-salt stress was examined by growth in LB–1% NaCl. All cultures were grown in triplicate, with at least two biological replicates. Error bars indicate standard deviations. An unpaired Student t test was used to determine the statistical difference between the final biomass of the WT cells compared to the rpoE mutant (*, P < 0.05; **, P < 0.005).

In addition, there was no difference in survival between the rpoE mutant and WT strains under acid stress conditions either in cells grown in LB–3% NaCl (pH 5.5) or in the presence of 4 mM acetic acid (data not shown). Similarly, mutant and WT strains behaved identical under anionic detergents stress conditions in 15% bile salt (see Fig. 8B) or 0.5% SDS, indicating that RpoE does not play a role in these stress responses under these conditions (data not shown).

FIG 8.

Phenotypes of ompU deletion mutant. (A) Strains were grown on TCBS agar to assess the sensitivity to bile salts. (B and C) Strains were grown to early log phase in LB–3% NaCl before being subjected to 15% bile (B) or 200 μg of polymyxin B (C) for 60 min. The percent survival was determined by dividing the number of viable cells at a given time point by the number of viable cells in the initial population. Each experiment was performed in triplicate and repeated at least twice. Error bars indicate the standard deviations.

RpoE is important for in vivo intestinal colonization in an adult mouse model.

Given the defect shown in survival under cell envelope stress conditions, we wanted to examine whether the rpoE mutant would have a defect in in vivo intestinal colonization. In order to accomplish this, the streptomycin-treated adult mouse model of colonization was utilized (8, 9). The β-galactosidase-positive RIMD2210633 strain WBWlacZ was used in order to allow for a blue-white screen differentiation with the ΔrpoE strain (8, 9). An in vivo competition assay in adult mice was carried out by pretreating mice with an orogastric dose of streptomycin (20 mg/mouse) 24 h prior to orogastric coinoculation with a mixture of WBWlacZ and either the WT strain (n = 10), the ΔrpoE mutant strain (n = 9) or the ΔrpoS mutant strain (n = 5). As previously shown, the WT and WBWlacZ strains do not outcompete each other, indicating that the lacZ knockin has no fitness effect under the conditions examined here (8) (Fig. 5A). Similarly, a WT-versus-WBWlacZ comparison in an in vitro assay in LB–3% NaCl also had a CI of 1. In contrast, in an in vivo assay between the rpoE mutant and WBWlacZ strains, there was a significant reduction (P < 0.001) in colonization, with the mutant having a CI of 0.07 (Fig. 5A). In an in vitro competition assay in LB–3% NaCl between the WT and the ΔrpoE mutant, a CI close to 0.4 was obtained after 24 h of incubation, showing an in vitro growth defect. This in vitro defect may reflect a requirement for RpoE in the stationary-phase response; therefore, we examined the CI of the mutant and the WT after 6-, 12-, and 24-h incubations. We found that there was no difference in CI at 6 h, but at 12 and 24 h the CIs between the strains were 0.45 and 0.55, indicating a defect in the rpoE mutant in stationary-phase cells (Fig. 5B). Overall, these data demonstrate that the rpoE mutant had a significant defect in vivo compared to the WT and that RpoE plays a significant role in in vivo survival. The importance of RpoE in intestinal colonization by V. parahaemolyticus is in contrast to the role of RpoS (VP2553). In an in vivo competition assay between WBWlacZ and the ΔrpoS mutant, it was found that the strains did not outcompete each other, having a CI close to 1 (Fig. 5A). This result indicates that in V. parahaemolyticus RpoS does not play a role in intestinal colonization.

FIG 5.

In vivo and in vitro competition assays in adult mouse intestinal colonization. (A) In vivo competition assay. Strains shown were grown to late log phase, mixed in a 1:1 ratio, and inoculated into mice treated with streptomycin. After 24 h, the mice were sacrificed, and each gastrointestinal tract was harvested and plated for CFU. The ratio of cells out was divided by the ratio of cells in to determine the competitive index (CI). The WT-versus-WBWlacZ assay was done concurrently with a previously published study from our group (9). The rpoE mutant, but not the rpoS mutant, shows a significant defect in colonization. (B) In vitro competition assay. Strains shown were grown to late log phase, mixed in a 1:1 ratio, and grown in LB–3% NaCl. The CFU were calculated at the indicated time points, and a CI was calculated for each time point. The rpoE mutant shows a defect in stationary-phase survival. P values were calculated by using Kruskal-Wallis one-way analysis of variance, followed by a Dunn multiple-comparison post test (***, P < 0.001).

An intestinal localization analysis was also performed in order to determine whether there was a difference in fitness of the WT strain and the rpoE deletion strain in different regions of the mouse intestinal tract. Mice were orogastrically inoculated with either the WT or the ΔrpoE strain alone. After 24 h, the small intestine, cecum, and large intestine were harvested separately and plated for CFU. The colonization levels between the two strains were highly similar in the small intestine (Fig. 6). In the cecum and large intestine there was a significantly (P < 0.05) greater amount of the WT present compared to the ΔrpoE strain, suggesting that the mutant was less fit in these environments (Fig. 6). Overall, our results demonstrate that the ΔrpoE mutant has a defect in colonization and that the cell envelope stress response is an important determinant for efficient in vivo survival.

FIG 6.

In vivo localization. Mice pretreated with streptomycin were inoculated with either the ΔrpoE or the WT strain. After 24 h, the mice were sacrificed, and the small intestines, ceca, and large intestines were harvested separately and plated for CFU. P values were calculated by using an unpaired Student t test with a 95% confidence interval. The rpoE mutant showed a defect in colonization of the cecum and large intestine (*, P < 0.05).

Role of RpoE on growth in mucus.

To complement the in vivo data and to elucidate whether there was a defect in growth on mucus that contributed to the defect in in vivo colonization, we analyzed the growth of the WT strain and the ΔrpoE strain in M9 media supplemented with mucus from the small intestine, cecum, or large intestine. The two strains grow similarly in the small intestine (Fig. 7A) and show a slight but not statistically significant defect in cecum mucus (Fig. 7B). In the large intestinal mucus, the ΔrpoE strain reaches a significantly lower optical density than did the WT (P < 0.05) (Fig. 7C). We speculate that the large intestinal mucus may contain more antimicrobial peptides than other regions of the intestine and that this may explain the defect, since the RpoE mutant is sensitive to antimicrobial peptides but not bile or acid stresses. To test this, a survival assay was performed in LI mucus, and the CFU were determined at the 6-, 12-, and 24-h time points (Fig. 7D). These data demonstrate that the mutant strain is more sensitive to LI mucus than is the WT.

FIG 7.

Growth analysis of V. parahaemolyticus on intestinal mucus as the sole carbon source. An rpoE mutant and WT strains were grown M9 supplemented with mucus from the small intestine (A), cecum (B), or large intestine (C) as the sole carbon source. (D) Survival assay in large intestinal (LI) mucus. All cultures were grown in triplicate with at least two biological replicates. Error bars indicate the standard deviations. An unpaired Student t test was used to determine the statistical difference between the final biomass (optical density) of the WT compared to the rpoE mutant (***, P < 0.001).

Deletion of OmpU does not affect the cell envelope stress response in V. parahaemolyticus.

It was proposed that in a V. cholerae El Tor strain in the presence of antimicrobial peptides, OmpU signals the release of RpoE from the membrane (35, 37). In order to test whether or not OmpU may be required for the RpoE stress response in V. parahaemolyticus, we constructed an ompU mutant strain. An unmarked nonpolar 849-bp deletion of the ompU gene (VP2467) was created. We compared growth of the mutant with that of the WT strain on LB–3% NaCl and found that both strains gave similar growth curves (data not shown). We reasoned that if there was a relationship between OmpU and RpoE function, there should be an overlap in phenotypes. For example, in V. cholerae both ompU and rpoE deletion mutant strains are sensitive to the antimicrobial peptide P2 and to polymyxin B (35, 37). Therefore, we compared growth of the WT, ΔompU, ΔtoxRS, and ΔrpoE strains on thiosulfate-citrate-bile salts-sucrose (TCBS) agar plates that contain bile salts as a selective agent (Fig. 8A). Both WT and ΔrpoE strains showed similar growth patterns, whereas both ΔompU and ΔtoxRS strains showed defects in growth (Fig. 8A). To examine this further, we performed survival assays in the presence of 15% bile salt. In this assay, the WT and rpoE mutant strains showed similar percent survivals, and the ΔompU and ΔtoxRS strains showed a significant defect (Fig. 8B). Next, a survival assay was performed on WT, ΔompU, ΔtoxRS, and ΔrpoE strains in the presence of polymyxin B. In this assay, ΔompU and ΔtoxRS strains survived similarly to the WT, and the ΔrpoE strain showed a significant defect (Fig. 8C). Overall, the results demonstrate that the ompU and rpoE mutant strains have no overlapping phenotypes, suggesting that under the conditions analyzed, OmpU is not essential in signaling the release of RpoE in V. parahaemolyticus.

DISCUSSION

In order to further our understanding of the factors important for V. parahaemolyticus stress response and colonization, we constructed an rpoE deletion strain; in other enteric species rpoE is required for the cell envelope stress response. We demonstrated that the loss of rpoE renders the mutant strain more sensitive to cell envelope stresses such as polymyxin B (a cationic antimicrobial peptide), ethanol, and high temperature.

We demonstrated that the loss of RpoE reduces fitness in vivo since the rpoE mutant was outcompeted by WBWlacZ and also showed significantly reduced colonization of the cecum and large intestine compared to the WT strain. This defect in colonization is predicted to be due to the increased sensitivity of the rpoE mutant to cationic antimicrobial peptides, since the rpoE mutant strain was not any more sensitive than the WT to bile salt, SDS, and acid stress conditions. The intestinal tract produces a consortium of cationic antimicrobial peptides such as α-defensins, β-defensins, and bactericidal/permeability-increasing protein (BPI), and the mechanism of action of many of these peptides is via the insertion and subsequent disruption of microbial membranes (45). BPI, for example, is specific to Gram-negative bacteria due to its high affinity for the lipid A region of lipopolysaccharide (LPS) and permeabilizes the outer membrane (45–47). Polymyxin B, used in the present study to mimic the antimicrobial peptides found throughout the intestinal tract, is also capable of destabilizing the LPS and insertion into microbial membranes (48–50). The sensitivity of our rpoE mutant to polymyxin B may indicate sensitivity to host cationic antimicrobial peptides in vivo and is predicted to be responsible for the colonization defect of the rpoE mutant. Of additional interest, secreted antimicrobial activity, such as that of α-defensins, has been shown to predominantly localize to the mucosal surface layer. By measuring the secreted antimicrobial activity of mouse intestinal extracts from the crypt/mucus/lumen compartments, it has been demonstrated that this antimicrobial activity is predominantly confined to the mucus layer (51). The idea that extracted mucus has the ability to retain antimicrobial activity from secreted host antimicrobials may explain the differences in biomass between the WT and rpoE mutant strains when grown in M9 supplemented with large intestinal mucus, supporting the notion that resistance to antimicrobial peptides is important for survival in the intestinal environment.

The role of RpoE in survival in the presence of antimicrobial peptides, as well intestinal colonization, has been studied in other Vibrio species. In V. cholerae, an rpoE mutant in a classical biotype strain was defective in colonization. Interestingly, this mutant was not more sensitive than WT to bile salts, hydrogen peroxide, or polymyxin B, suggesting that for this particular strain the importance of RpoE in colonization may be independent of antimicrobial peptide sensitivity (34). Conversely, an rpoE mutant of a V. cholerae El Tor biotype was shown to be more sensitive than WT to both the bioactive peptide P2 and polymyxin B, and it was demonstrated that in the presence of antimicrobial peptides misfolded OmpU in the periplasm signaled the release of RpoE from the membrane (35, 37). In V. harveyi, the disruption of two regulators of rpoE, rseB and rseC, resulted in an rseBC mutant strain that was outcompeted in vivo by the WT and which demonstrated reduced hemolytic activity compared to the WT, suggesting that changes in RpoE activity have an impact on virulence of V. harveyi (38).

Both ompU and toxRS mutant strains of V. parahaemolyticus are more sensitive to bile salt and SDS compared to the WT. In V. cholerae, it was proposed that in an El Tor biotype strain OmpU signals the release of RpoE in the presence of antimicrobial peptides (35). In V. parahaemolyticus, we created an ompU mutant and found no overlap in phenotype between the ompU and rpoE mutants, suggesting that OmpU may not signal release of RpoE. In V. parahaemolyticus, stimulation and release of RpoE may be regulated by several factors, as is the case in E. coli (29). In E. coli, RpoE activity and release from the membrane via proteolytic cleavage by DegS and YaeL is modulated through many different outer membrane proteins (OMPs) and OMP-like proteins (29). In V. parahaemolyticus, as in V. cholerae and E. coli, rpoE is found clustered in an operon with the putative anti sigma factor and the regulators, rseABC. RseA, RseB, and RseC in V. parahaemolyticus have 65, 72, and 57% and 43, 44, and 30% amino acid identities, respectively, to homologues found in V. cholerae and E. coli, respectively.

In conclusion, these results demonstrate that RpoE is a key cell envelope stress regulator and is important for intestinal survival. Here, we expand upon the use of our streptomycin-treated mouse model by showing the requirement for RpoE but not RpoS in intestinal survival and colonization for the first time in V. parahaemolyticus. Previously, it was shown that in V. parahaemolyticus RpoS plays a limited role in stress response (10). Our group has shown that deletion of RpoN in V. parahaemolyticus results in a hypercolonizer strain which outcompetes the WBWlacZ strain in vivo (9). This suggests that the different sigma factors and their regulons play discordant roles during the in vivo colonization and survival of V. parahaemolyticus. It will be interesting to examine further whether the additional ECFs are important for V. parahaemolyticus pathogenesis.