Isolation and Characterization of rpoS from a Pathogenic Bacterium, Vibrio vulnificus: Role of {sigma}S in Survival of Exponential-Phase Cells under Oxidative Stress

A gene homologous to rpoS was cloned from a fatal human pathogen,Vibrio vulnificus. The functional role of rpoS in V. vulnificuswas accessed by using an rpoS knockout mutant strain. This mutantwas impaired in terms of the ability to survive under oxidativestress, nutrient starvation, UV irradiation, or acidic conditions.The increased susceptibility of the V. vulnificus mutant inthe exponential phase to H₂O₂ was attributed to the reducedactivity of hydroperoxidase I (HPI). Although ${sigma}$ ^S synthesis wasinduced and HPI activity reached the maximal level in the stationaryphase, the mutant in the stationary phase showed the same susceptibilityto H₂O₂ as the wild-type strain in the stationary phase. Inaddition, HPII activity, which is known to be controlled by ${sigma}$ ^S in Escherichia coli, was not detectable in V. vulnificus strainsunder the conditions tested. The mutant in the exponential phasecomplemented with multiple copies of either the rpoS or katGgene of V. vulnificus recovered both resistance to H₂O₂ andHPI activity compared with the control strain. Expression ofthe katG gene encoding HPI in V. vulnificus was monitored byusing a katG::luxAB transcriptional fusion. The expression ofthis gene was significantly reduced by deletion of ${sigma}$ ^S in boththe early exponential and late stationary phases. Thus, ${sigma}$ ^S isnecessary for increased synthesis and activity of HPI, and ${sigma}$ ^Sis required for exponentially growing V. vulnificus to developthe ability to survive in the presence of H₂O₂.

INTRODUCTION

Top

Introduction

The life cycles of pathogenic bacteria involve periods in whichthey exist in a nongrowing state in stressed environments. Onlyif they survive such conditions are they able to proliferatewith high metabolic activity in the proper host environments(7, 36). Thus, these organism have evolved several mechanismsthat allow them to survive under stressful conditions, suchas starvation, temperature fluctuation, oxidative stress, andosmotic shock, and that enable them to resume growth once thestress is removed (27). The cellular responses to environmentalstimuli have been extensively studied in many bacterial species,most notably Escherichia coli.

To respond properly to diverse stresses, E. coli requires therpoS gene product, which is a second principal sigma factor( ${sigma}$ ^S); this product induces expression of many genes and allowsthe organism to mediate changes in cellular physiology and structureand to adapt, resist, and survive under stress conditions (9,16, 19). ${sigma}$ ^S is also required for eliciting phenotypes relatedto virulence in many pathogenic bacteria belonging to the ${gamma}$ subdivisionof Proteobacteria (21, 32, 39, 45, 50).

It is generally believed that most microorganisms that communicatewith, associate with, or colonize host animals are relativelywell equipped with defense mechanisms to deal with oxidativestress (6, 15, 43). E. coli produces at least two enzymes toovercome the presence of hydrogen peroxide and to maintain arelatively constant concentration of intracellular H₂O₂ (8);these enzymes are KatG (hydroperoxidase I [HPI]), which hasboth catalase and peroxidase activities, and monofunctionalKatE (HPII), which has catalase activity (25). KatG, one ofthe members of the OxyRS regulon, is induced by direct exposureto H₂O₂ (37). In contrast, KatE is known to be regulated by ${sigma}$ ^S, and consequently cellular expression of this enzyme increasesat the onset of the stationary phase (25, 30). Open readingframes homologous to both katG and katE are present in the genomesof Vibrio cholerae and Vibrio parahaemolyticus. The presenceand role of monofunctional catalases have been studied in Vibriofischeri and Vibrio rumoiensis, but the regulation of theseenzymes has not been described (11, 44, 51).

The causative agent of septicemia, Vibrio vulnificus, has beenconsidered an important pathogen in humans due to its rapidpathogenic progress and its high mortality rate (10, 38). Anumber of studies have been performed on virulence factors ofthis organism, including metalloprotease (13), hemolysin (48),and siderophores (35). Several regulators, including cyclicAMP receptor protein (CRP)/LuxR (12, 33), ToxRS/CRP (1, 17),and Fur (18), have been reported to control expression of thesevirulence factors. While survival of this bacterial specieshas been studied under diverse conditions (26), the molecularmechanisms underlying its survival strategies have not beenstudied well.

In an effort to isolate global regulators involved in survivalof V. vulnificus, we cloned the rpoS gene and defined its physiologicalrole in survival of V. vulnificus in the presence of variousstresses. These analyses showed that V. vulnificus in the exponentialphase requires ${sigma}$ ^S for survival in the presence of low concentrationsof hydrogen peroxide. In the present study we also observedregulation of the expression and activity of a catalase involvedin this response, and the results were quite different fromthose obtained with E. coli.

MATERIALS AND METHODS

Top

Materials and Methods

Isolation of the rpoS gene from V. vulnificus. The genomic DNA of V. vulnificus ATCC 29307 was prepared bya standard technique (29) and then partially digested with Sau3AIand size fractionated by agarose gel electrophoresis. The DNAfragments, which ranged from 2 to 6 kb long, were pooled andligated with the pUC19 vector which had been digested with BamHIand subsequently treated with bacterial alkaline phosphatase.The V. vulnificus library obtained was introduced into E. coliZK918 having a deletion in its rpoS gene and a ${sigma}$ ^S-dependent bolA::lacZfusion in its chromosome (2). After transformation with thelibrary, colonies were screened on Luria-Bertani (LB) mediumsupplemented with ampicillin (100 µg/ml) and X-Gal (5-bromo-4-chloro-3-indolyl-ß-D-galactopyranoside)(10 µg/ml), which was blue as a result of induced expressionof bolA::lacZ after addition of the RpoS homolog of V. vulnificus.Sequencing of the double-stranded DNA of the plasmid selected,pINE32, containing a 2,693-bp insert, was performed with anApplied Biosystems 373A automated DNA sequencer. The remaininggaps in the sequence were filled in by sequence analysis byusing specifically designed internal primers that annealed tothe insert region. Sequence analysis and database searches wereperformed by using the National Center for Biotechnology InformationBLAST server.

Construction of rpoS knockout mutant KPR101. A 762-bp NruI fragment containing two-thirds of the rpoS codingsequence was deleted from pINE32. The resultant plasmid, pKP11,was digested with SmaI and XbaI, which resulted in a DNA fragmentcontaining a region adjacent to the rpoS gene but not the rpoSgene. This DNA was cloned into suicide vector pDM4 (23), whichwas digested with ApaI and XbaI, yielding pKP13. pKP13 in E.coli SM10 ${lambda}$ pir was mobilized into strain AR, a rifampin-resistantderivative of the wild-type strain V. vulnificus ATCC 29307.Conjugal transfer was performed by mixing aliquots of the strainsthat contained about 10⁸ donor cells and about 10⁸ recipientcells and then incubating the preparation overnight at 37°Cin close contact on a membrane filter. The cell mixture wasthen resuspended in LBS (LB medium containing NaCl at a finalconcentration of 2.5%) broth and plated onto selective plates(LBS agar plates supplemented with 4 µg of chloramphenicolper ml and 50 µg of rifampin per ml). A colony showingindications of a double homologous recombination event (resistanceto 5% sucrose and sensitivity to chloramphenicol) was isolated,and deletion of its rpoS region was confirmed by PCR by usingprimers F2 and R2 (Table 1).

View this table:
[in this window]
[in a new window]
TABLE 1. Strains, plasmids, and oligonucleotide primers used in this study

Determination of survival of V. vulnificus under various stress conditions. Exponential-phase cultures (optical density at 600 nm [OD₆₀₀]in LBS broth, 0.15 to 0.3) or stationary-phase cultures (OD₆₀₀in LBS broth, about 3 to 4; usually 8 to 10 h after an overnightculture was added to fresh LBS broth) were collected by centrifugation,washed with artificial seawater (ASW) (0.6 M NaCl, 0.1 M MgSO₄,0.02 M CaCl₂, 0.02 M KCl, 50 mM Tris-HCl [pH 8.3] [28]), andresuspended in the appropriate medium at a density of approximately10⁶ to 10⁷ cells/ml. During incubation under stress conditionsat 30°C, aliquots were removed and spread on LBS agar platesto monitor the viability by determining the number of CFU. Forsurvival studies under stress conditions, we used hyperosmosis(LBS medium containing 5 M NaCl), starvation (nutrient-freeASW), acidity (LBS medium titrated to pH 4.0 with HCl), UV irradiation(wavelength, 254 nm; energy, 120,000 µJ/cm²), and hydrogenperoxide (0.088 to 10 mM H₂O₂ in ASW).

Determination of catalase activities. Cellular extracts of E. coli and various V. vulnificus strains(Table 1) were prepared in cold 50 mM potassium phosphate buffer(pH 7.0) by sonication (Vibracell; Sonics & Materials, Inc.)in ice. The amount of protein in a cell lysate was determinedby the Bradford assay by using bovine serum albumin as the standard.After separation on an 8% nondenaturing polyacrylamide gel,the locations of HPI and HPII were visualized by staining thegel with a solution containing 1% K₃Fe(CN)₆ and 1% FeCl₃ (47)and compared with the locations of the corresponding enzymesof E. coli.

For the catalase assay, cell samples were obtained at variousgrowth phases (OD₆₀₀, 0.1 to 4.2) and were resuspended in chilledcatalase buffer (5 mM potassium phosphate buffer [pH 7.0], 5mM EDTA, 10% glycerol, 25 µM phenylmethylsulfonyl fluoride)(42) and sonicated in ice. Each cellular extract was then mixedwith 25 mM potassium phosphate buffer (pH 7.0) containing 5.9mM H₂O₂, and the amount of remaining H₂O₂ in the reaction mixturewas estimated by monitoring the absorbance at 240 nm at 30-sintervals for 10 min. One unit of specific activity was definedas 1 µmol of H₂O₂ degraded per min per mg of protein (42).

Construction of katG::luxAB transcriptional fusion and measurement of expression of the katG gene. The 504-bp DNA fragment which included nucleotide positions–318 to 115 relative to the initiation codon of katG (geneVV12755 in GenBank accession no. NC_004459) was amplified byPCR by using two primers, katG-F-KpnI and katG-R-XbaI (Table1). The PCR product was digested with appropriate restrictionenzymes (KpnI and XbaI) and inserted into luxAB-based plasmidpHK0011 (12) digested with the same enzymes. The resultant transcriptionalfusion, pHL-03, was transferred into V. vulnificus cells viaconjugative transfer. Overnight (16- to 18-h) cultures of thecells containing pHL-03, which exhibited the basal level ofbioluminescence, were inoculated into fresh LBS medium containing3 µg of tetracycline per ml, and the expression from thekatG promoter was measured by monitoring light production inthe presence of 0.006% n-decyl aldehyde by using luminometers(TD-20/20; Turners Designs). The specific bioluminescence wascalculated by normalizing the relative light units with cellmass (OD₆₀₀).

Complementation of KPR101 with a broad-host-range vector containing V. vulnificus rpoS (pKP14) or V. vulnificus katG (pKP51). An EcoRI-HindIII DNA fragment that included the intact rpoSgene was obtained from pINE32 and cloned into the broad-host-rangevector pLAFR5 (14) to obtain plasmid pKP14. The 2,569-bp DNAfragment that included the intact katG gene was amplified byPCR by using two primers, katG-F-KpnI and katG-R-HindIII. ThePCR product was digested with appropriate restriction enzymes(KpnI and HindIII) and cloned into the broad-host-range vectorpRK415 (14) to obtain plasmid pKP51. Each plasmid was introducedinto the rpoS mutant by conjugative transfer, as described above.

Western analysis of ${sigma}$ ^S. Two oligonucleotides, rpoS_vv-F and rpoS_vv-R (Table 1), wereused to amplify a 1,032-bp DNA fragment containing the completesequence of the rpoS gene from the genomic DNA of V. vulnificus.BamHI and HindIII sites located at both ends of the resultantrpoS DNA were used to clone this DNA into the pQE30 expressionplasmid (Qiagen) to generate plasmid pQErpoS. Recombinant ${sigma}$ ^Swas overexpressed in E. coli JM109 by adding isopropylthio-ß-D-galactoside(Sigma) at a concentration of 1.0 mM and was purified by usingan Ni⁺-nitrilotriacetic acid affinity column as directed bythe manufacturer (Qiagen). Polyclonal antibodies against V.vulnificus ${sigma}$ ^S were produced in a rabbit by intravenous immunizationwith 200 µg of the recombinant V. vulnificus ${sigma}$ ^S, and thisinitial immunization was followed by additional immunizationsat 1 and 4 weeks. Ten days after the last injection, the bloodof the immunized rabbit was collected, and its serum was usedfor Western blot analysis. Cell extracts of the V. vulnificuswild type and the rpoS mutant containing either pLAFR5 or pKP14were prepared by sonication in TNT buffer (10 mM Tris-HCl [pH8.0], 150 mM NaCl, 0.05% Tween 20) (29), and 40-µg aliquotsof the extracts were fractionated by sodium dodecyl sulfate-polyacrylamidegel electrophoresis. After transfer to a Hybond P membrane (Amersham),Western blot analysis was performed by serially incubating thefilter with V. vulnificus ${sigma}$ ^S polyclonal antibodies (1:1,000 dilution)and alkaline phosphatase-conjugated rabbit anti-rat immunoglobulinG (1:1,000 dilution; Sigma). The V. vulnificus ${sigma}$ ^S band was visualizedby using the nitroblue tetrazolium-BCIP (5-bromo-4-chloro-3-indolylphosphate)system (Promega).

Nucleotide sequence accession number. The V. vulnificus rpoS nucleotide sequence has been depositedin the GenBank database under accession number AY187681.

RESULTS AND DISCUSSION

Top

Results and Discussion

Isolation of V. vulnificus rpoS by functional complementation. In E. coli, ${sigma}$ ^S is known to regulate the expression of severalgenes involved in cellular adaptation to diverse stresses. Sincethe ${sigma}$ ^S homologue of V. vulnificus, if there is one, may alsobe able to play an equivalent role in regulation of these genes,E. coli strain ZK918 containing the ${sigma}$ ^S-dependent bolA promoter::lacZfusion (2) was utilized. Upon transformation of ZK918 with thegenomic library of V. vulnificus, the bluish transformants onLB medium supplemented with X-Gal were chosen as candidatesfor complementing plasmids for the rpoS function. One of theplasmids, plasmid pINE32 carrying a 2,693-bp insert (Table 1),was used for further investigation.

Analysis of the pINE32 insert DNA. Sequence analysis of the insert in pINE32 revealed that it codedfor the proteins of V. vulnificus homologous to NlpD, RpoS,and MutS. Although the genetic organization of the open readingframes flanking rpoS is the same as that found in other bacteria,the lengths of the intergenic spaces between nlpD and rpoS andbetween pcm (the upstream gene of nlpD) and nlpD were quitedifferent from the lengths of the intergenic spaces in E. colior Pseudomonas. This finding suggests that the regulation ofrpoS expression at the transcriptional level in V. vulnificusmay be different from the regulation in other organisms whichhave been extensively studied (20, 30, 41). The amino acid sequencededuced from the gene homologous to rpoS, which codes for 343amino acid residues corresponding to ca. 39.6 kDa with a pIof 4.92, was aligned with other known ${sigma}$ ^S sequences and was foundto exhibit overall levels of identity of 83, 79, and 70% withthe sequences of V. parahaemolyticus, V. cholerae, and E. coli,respectively. There is complete homology in subregions 2.3 and2.4 of ${sigma}$ ^S, which are involved in promoter recognition (e.g.,RpoD box and 14-mer) (4), and there is significant conservationin subregion 2.1, which is involved in core binding. The helix-turn-helixmotif in subregion 3.1 and the –35 recognition regionin subregion 4.2 showed relatively weak similarity.

Generation of an rpoS-deficient strain, KPR101. A major portion of the V. vulnificus rpoS gene in pINE32 wasdeleted by digestion with NruI and subsequent ligation. ThenlpD-rpoS-mutS region with the rpoS gene deleted was transferredto a conjugative plasmid, pDM4, resulting in pKP13. Replacementof the wild-type rpoS gene located on the chromosome of V. vulnificusAR with this plasmid was accomplished through homologous recombination.Deletion of the rpoS gene in the V. vulnificus mutant was confirmedby PCR by using primers F2 and R2 (Table 1). A PCR analysisof a deletion of the internal region of the rpoS gene in themutant revealed the expected size for the DNA fragment (754bp); meanwhile, the intact rpoS gene in the wild-type produceda 1,516-bp DNA fragment (data not shown). The resultant mutantwas designated KPR101.

Survival characteristics of KPR101. The survival of KPR101 in the exponential phase was examinedwhen it was exposed to various stresses, including hyperosmoticconditions, starvation, an acid environment, UV irradiation,and oxidative conditions, and was compared to the survival ofthe wild type. In contrast to enteric bacteria (9, 19), rpoS-deficientV. vulnificus exhibited the same survival pattern in the presenceof a high salt concentration (5 M NaCl) that the wild type exhibited(data not shown). The survival of the mutant, however, was severelyimpaired in the presence of other stresses. For example, theabilities of the mutant to survive under starvation, acidic,and UV-irradiated conditions (6 days, 1 h, and 12 s, respectively)were estimated to be 25-, 1,300-, and 16-fold less than thoseof the corresponding controls, respectively (Fig. 1A $~$ C).

View larger version (17K):
[in this window]
[in a new window]
FIG. 1. Survival of exponential-phase wild-type and ${Delta}$ rpoS mutant KPR101 cultures under stress conditions. Wild-type and KPR101 cells in the exponential phase were challenged by starvation (ASW) (A), acidic conditions (LBS medium titrated to pH 4.0) (B), and irradiation with UV light (254 nm; 120,000 µJ/cm²) (C). At several times during exposure, aliquots of each culture were removed, and the numbers of CFU per milliliter were estimated as described in Materials and Methods. The resultant values were expressed as percentages of the initial cell density, which ranged from 10⁶ to 10⁷ CFU/ml.

The ${Delta}$ rpoS mutant cells in the exponential phase also showed significantlyincreased susceptibility to 880 µM H₂O₂; there was upto a 1,000-fold difference after 30 min of exposure (Fig. 2A).The KPR101 cells in the stationary phase showed more resistancethan the cells in the exponential phase, but the susceptibilitywas basically similar to the susceptibility of the wild typein the presence of various concentrations of H₂O₂ up to 10 mM(Fig. 2B). This ${sigma}$ ^S-independent increase in H₂O₂ resistance inthe stationary phase is an unusual observation, since synthesisof ${sigma}$ ^S is induced in the stationary phase in V. vulnificus (datanot shown), and one of catalases (HPII) is known to be inducedby ${sigma}$ ^S in E. coli (30, 37, 40). In addition, V. vulnificus cellswere generally more sensitive to H₂O₂ than other enteric bacteriawere, because exposure to the concentrations of H₂O₂ used forE. coli or V. cholerae (e.g., ca. 10 mM for 30 min) resultedin survival of only 0.1 to 0.01% of the cells present initially(Fig. 2B). Thus, it is possible that V. vulnificus may havemechanisms for oxidative stress response that are distinct fromthose found in E. coli, at least under the conditions whichwe used.

View larger version (15K):
[in this window]
[in a new window]
FIG. 2. Survival of the wild-type and ${Delta}$ rpoS mutant KPR101 cells in the presence of H₂O₂. (A) Wild-type and KPR101 cells in the exponential phase were challenged with an oxidative stress (ASW containing 880 µM H₂O₂). At several times during exposure, aliquots of each culture were removed, and the numbers of CFU per milliliter were estimated. The asterisks indicate values lower than the detection limit used in this analysis. (B) Wild-type and KPR101 cells in the stationary phase (OD₆₀₀ in LBS broth, approximately 3 to 4) were harvested and incubated in ASW containing concentrations of H₂O₂ ranging from 0.6 to 10 mM. After 30 min of exposure, the numbers of remaining culturable cells were estimated. The resultant values were expressed as percentages of the initial cell density, which was about 10⁷ CFU/ml.

Characterization of the catalase of V. vulnificus. The distinct susceptibility of KPR101 to H₂O₂, especially duringthe exponential phase, led us to examine the innate propertiesof hydroperoxidase(s) in wild-type V. vulnificus. Protein extractsof the type strain of V. vulnificus (ATCC 29307) were separatedon a nondenaturing polyacrylamide gel, and the bands representinghydroperoxidase activities were compared with those of a stationary-phaseE. coli extract. In contrast to E. coli, which has both KatG(HPI; a multimer composed of two 80.0-kDa subunits) and KatE(HPII; a multimer composed of four 84.2-kDa subunits), onlythe band corresponding to HPI was visualized in both exponential-and stationary-phase V. vulnificus extracts (Fig. 3A). The calculatedmolecular mass of the deduced KatG polypeptide of V. vulnificus(VV12755 in GenBank accession no. NC_004459) is estimated tobe 80.4 kDa. The absence of HPII activity seemed to be commonin this bacterial species, since various V. vulnificus strainsexhibited only HPI activity, at least under the conditions whichwe tested (Fig. 3B).

View larger version (81K):
[in this window]
[in a new window]
FIG. 3. Nondenaturing gels showing activity staining of catalases in bacterial crude extracts. (A) Crude extracts (120 µg) prepared from V. vulnificus ATCC 29307 in the exponential or stationary phase which had been treated with either 0 or 10 µM H₂O₂ for 10 min were loaded into an 8% polyacrylamide gel, and the HPI and HPII activities were examined. The typical HPI (dimer consisting of 80.0-kDa subunits) and HPII (tetramer consisting of 84.2-kDa subunits) bands found in E. coli AMS6 (60 µg) were visualized by staining the gel with 1% K₃Fe(CN)₆ and 1% FeCl₃, and the results were compared to the results obtained for V. vulnificus. (B) Crude extracts of various V. vulnificus strains (60 µg) were loaded into an 8% polyacrylamide gel. The strains used were ATCC 29307 (lane 1), MO6-24/0 (lane 2), clinical isolate CN7 (lane 3), clinical isolate CNUH94-4 (lane 4), environmental (seawater) isolate SC9649 (lane 5), environmental (sediment) isolate SC9720 (lane 6), and environmental (seafood) isolate SC97126 (lane 7).

When the katG gene of V. vulnificus was disrupted, no hydroperoxidaseband appeared (data not shown), and the missing HPI band wasrestored when the V. vulnificus katG gene was supplied in amulticopy plasmid (J. H. Rhee, personal communication). ThekatE gene is present on chromosomal DNA of V. vulnificus (VV21473in GenBank accession no. NC_004460), and the molecular massof the deduced polypeptide is predicted to be 55.4 kDa. However,no apparent KatE (HPII) activity was observed under any of theconditions tested, even when we used large amounts of cellularextract (up to 200 µg of crude protein extract [data notshown]) and cellular extracts were exposed to H₂O₂ (Fig. 3A).In addition, no transcript for katE was detected in the totalRNA (30 µg) by Northern analysis with a 1,523-bp katEgene probe (data not shown). Exposure of the exponential-phasecells to H₂O₂ resulted in the induced level of HPI found inthe stationary-phase cells, which suggests that a redox-operatedregulator, possibly OxyR, is involved in activating katG (46;I. Kong, A. Huelsmann, and J. D. Oliver, 102nd Gen. Meet. Am.Soc. Microbiol., poster no. K-94, 2002).

Determination of HPI activity in the wild-type strain and KPR101. To identify the possible role of ${sigma}$ ^S in regulation of the activityof KatG, HPI activities were monitored along the growth curve(Table 2). The HPI catalase specific activity of the wild-typestrain was about 2 U during the early exponential phase (forthe first 2 h of growth) and gradually decreased during themid-exponential phase. Then it reached a maximal level, approximately6 to 8 U, after the cells entered the late stationary phase.The activity of HPI in V. vulnificus cells during the wholegrowth period, except for an initiation period during the stationaryphase (OD₆₀₀, 2.2 and 3.2 [Table 2]), was regulated by the presenceof ${sigma}$ ^S, since the rpoS mutant contained about twofold less HPIthan the wild-type strain contained. Thus, the induction ofHPI at the onset of the stationary phase may depend on otherregulators.

View this table:
[in this window]
[in a new window]
TABLE 2. Catalase activities of the wild type and ${Delta}$ rpoS mutant KPR101 in various growth phases

Similarly, the HPI catalase specific activity in KPR101 was2 to 3 U at the onset of the stationary phase, but it did notreach the maximal level observed in the wild type (Table 2).However, the twofold reduction in HPI activity in late-stationary-phaseKPR101 compared to the activity in the wild type did not resultin a difference in survival in the presence of H₂O₂ (Fig. 2B),because this amount of HPI activity (i.e., more than 2 to 3U) might provide as much resistance to the concentrations ofH₂O₂ used (up to 10 mM) as the amount in the wild type provides.

The medium used to grow KPR101 contained less HPI activity thanthe medium used to grow the wild type contained. No activitywas detected in a cell-free medium (glucose-based ASW) whichwas used to grow KPR101 at the exponential phase, whereas significantHPI activity (0.38 ± 0.05 U) was found in the spent mediumused for the wild type at the exponential phase. In the mediumused to culture KPR101 at the stationary phase, the extracellularHPI activity (0.97 ± 0.06 U) was about 70% of the activityof the wild type (1.32 ± 0.08 U). This suggests thatthe smaller amount of KatG activity in KPR101 than in the wildtype was not due to increased excretion of this enzyme.

Complementation of KPR101 with a broad-host-range vector containing the V. vulnificus rpoS gene, pKP14. After pKP14, a pLAFR5-based plasmid containing the V. vulnificusrpoS gene, was added to KPR101, the expression of V. vulnificus ${sigma}$ ^S in this strain was confirmed by Western blotting by usingV. vulnificus ${sigma}$ ^S-specific polyclonal antibody, which clearlyshowed the presence of an $~$ 39-kDa V. vulnificus ${sigma}$ ^S (Fig. 4A, inset).The exponential-phase KPR101 culture, containing pKP14 grownin glucose-based ASW supplemented with tetracycline, was exposedto 880 µM H₂O₂, and changes in the number of CFU weremonitored for 2 h. This strain was found to be resistant toH₂O₂, whereas KPR101 containing pLAFR5 was found to be sensitiveto the same concentration of H₂O₂ (Fig. 4A).

View larger version (25K):
[in this window]
[in a new window]
FIG. 4. Complementation of ${Delta}$ rpoS mutant KPR101 with intact V. vulnificus rpoS. (A) Complementation of KPR101 with the rpoS gene (in the broad-host-range pLAFR5 vector containing the V. vulnificus rpoS gene or pKP14) resulted in synthesis of ${sigma}$ ^S, as determined by Western blotting (inset). Exponential-phase KPR101 containing pKP14 or pLAFR5 was grown in ASW supplemented with tetracycline and exposed to 880 µM H₂O₂. At several times during exposure, aliquots of the culture were removed, and the numbers of CFU per milliliter were estimated as described in Materials and Methods. The resultant values were expressed as percentages of the initial cell density, which was approximately 10⁶ CFU/ml. (B and C) The HPI activity in a crude extract from an exponential-phase KPR101 culture containing pKP14 was compared to the HPI activity of KPR101 containing pLAFR5 by either the catalase staining method with a nondenaturing gel (B) or the H₂O₂ degradation assay with a spectrophotometer (C).

When the V. vulnificus rpoS gene was present in trans, apparentlyincreased HPI activity was observed by the gel staining assay(Fig. 4B), and the increase was estimated to be about fourfoldas determined by an H₂O₂ degradation kinetic study (Fig. 4C).These results show that in the exponential phase the rpoS mutantcomplemented with the V. vulnificus rpoS gene exhibited bothresistance to H₂O₂ and HPI activity.

Complementation of KPR101 with a broad-host-range vector containing the V. vulnificus katG gene, pKP51. If the effect of an rpoS mutation on increased sensitivity toH₂O₂ was mainly due to a lower level of HPI, KPR101 should becomeresistant to H₂O₂ if multiple copies of the katG gene are provided.Thus, KPR101 was complemented with pKP51, a pRK415-based plasmidcontaining the V. vulnificus katG gene. In the exponential phaseKPR101 containing pKP51 grown in LBS medium supplemented withtetracycline exhibited much greater HPI activity (Fig. 5A).This strain also exhibited greater resistance to 880 µMH₂O₂ than KPR101 containing pRK415 exhibited (Fig. 5B). However,it is possible that the other gene products regulated by ${sigma}$ ^S playan important role in survival in the presence of higher concentrationsof H₂O₂ than we tested. Currently, the proteins which are inducedin the wild type in the exponential phase by exposure to H₂O₂but are not induced in KPR101 in the exponential phase underthe same conditions are being investigated by proteome analysesby using two-dimensional matrix-assisted laser desorption ionization—timeof flight mass spectrometry.

View larger version (15K):
[in this window]
[in a new window]
FIG. 5. Complementation of the ${Delta}$ rpoS mutant KPR101 with V. vulnificus katG. (A) Complementation of KPR101 with the katG gene (in the broad-host-range vector pRK415 containing the V. vulnificus katG gene or pKP51) resulted in increased activity of HPI, as determined by the catalase staining method with a nondenaturing gel. The HPI activity in the crude extract of an exponential-phase KPR101 culture containing pKP51 was compared to the HPI activities of KPR101 and the wild type containing pRK415. (B) An exponential-phase KPR101 culture containing pKP51 or pRK415 in LBS medium was grown in LBS broth supplemented with tetracycline and exposed to 880 µM H₂O₂. At several times during exposure, aliquots of the culture were removed, and the numbers of CFU per milliliter were estimated as described in Materials and Methods. The resultant values were expressed as percentages of the initial cell density, which was approximately 10⁷ CFU/ml.

Determination of katG expression in the wild type and KPR101. Since the presence of ${sigma}$ ^S apparently increases HPI activity, itsrole in controlling HPI was analyzed further by monitoring expressionof the katG gene in the wild type and KPR101. To do this, aluxAB-based transcriptional fusion, pHL-03 containing the promoterregion of the katG gene, was constructed and transferred toboth strains. In the wild-type cells, induction of the katGfusion was apparently initiated during the early exponentialphase (during the first 1 to 2 h of incubation), and the leveldecreased to the basal level during the mid-exponential phase.A high level of expression during the initial incubation periodwas also observed in E. coli (8), and the HPI activity was alsohigh at this stage of growth (Table 2). Then induction occurredagain at the onset of the stationary phase, and the level ofexpression reached the maximal level (about 10-fold greaterthan the basal level) when the cells entered the late stationaryphase (Fig. 6). This expression profile essentially parallelsthe pattern of catalase activity measured directly (Table 2).

View larger version (18K):
[in this window]
[in a new window]
FIG. 6. Expression of the katG gene determined by using a katG::luxAB transcriptional fusion (pHL-03). Wild-type (•) and KPR101 ( ${circ}$ ) cells containing the katG::luxAB fusion were freshly grown in LBS broth supplemented with 3 µg tetracycline per ml by inoculating overnight cultures, whose levels of bioluminescence were the basal levels. Aliquots were removed, and the cell masses (OD₆₀₀) (A) and bioluminescence (relative light units [RLU]) (B) were estimated. The luciferase activities were expressed as normalized values obtained by dividing the number of RLU by the OD₆₀₀ of each sample. The activities of three independent experiments were averaged, and the error bars indicate the standard deviations.

The katG expression in KPR101 followed the same pattern as thekatG expression in the wild type, but the degrees of inductionin both the early exponential and late stationary phases weresignificantly reduced (Fig. 6). This pattern of V. vulnificuskatG expression is quite different from the katG expressionin E. coli. In E. coli, the maximal katG expression was onlyslightly reduced in an rpoS mutant (8), the maximal inductionof katG occurred during the late exponential phase (22), andthe duration of induced katG expression was not prolonged duringthe late stationary phase (8). Interestingly, an HPII-deficientE. coli mutant has been reported to exhibit higher HPI activitythan the parental strain exhibits (42).

While no difference in katG expression was observed during themid-exponential phase (Fig. 6), the estimated HPI activitiesin KPR101 were about 50% of those in the wild type at the samegrowth stage (Table 2). This difference in activity might resultfrom different expression during the early exponential phase.Otherwise, ${sigma}$ ^S might not directly regulate KatG at the transcriptionallevel, at least during the mid-exponential phase. Instead, somefactors regulated by ${sigma}$ ^S are involved in increased (or sustained)activity of HPI. It is necessary to identify these factors inorder to elucidate the regulatory pathways for catalase via ${sigma}$ ^S in exponential-phase V. vulnificus. The experiments in whichwe examined the effect of H₂O₂ on synthesis of ${sigma}$ ^S revealed aslightly increased amount of ${sigma}$ ^S in the exponential-phase cellsexposed to nonlethal concentrations of H₂O₂ (data not shown).However, additional studies are necessary to clarify whetherthe slightly increased amount of ${sigma}$ ^S upregulates the synthesisand/or activity of KatG during the exponential phase.

Several research groups have emphasized the importance of ${sigma}$ ^Sin bacterial survival in the stationary phase and in resistanceto H₂O₂ via regulation of HPII. Although it has been reportedthat in the stationary phase HPI is partially induced by ${sigma}$ ^S inE. coli (8, 24) and in the exponential phase some virulencegenes are regulated by ${sigma}$ ^S in Salmonella dublin (5), exponential-phaseinduction of HPI by ${sigma}$ ^S has not been examined previously. In thepresent study, however, we found that in V. vulnificus thisglobal regulator plays a role in the response to oxidative stressduring the exponential phase by increasing the amount of HPIwith no involvement of HPII. Investigation of the other rolesof ${sigma}$ ^S in hierarchical regulatory cascades and the expressionof rpoS in the presence of specific stresses is in progress.

ACKNOWLEDGMENTS

This research was supported by the 21C Frontier Microbial Genomicsand Applications Center Program, Ministry of Science and Technology(grant MG02-0201-004-2-1-0 to K.-H.L.), Republic of Korea.

We thank E.-K. Jeon and J. H. Lee for technical assistance,J. H. Rhee for providing a V. vulnificus ${Delta}$ katG mutant strain,H.-J. Myung for overexpression of V. vulnificus ${sigma}$ ^S, and J. K.Lee for helpful discussions.

FOOTNOTES

* Corresponding author. Mailing address: Department of Environmental Science, Hankuk University of Foreign Studies, Wangsan-Li, Mohyun-Myun, Yongin, Kyunggi-Do 449-791, Korea. Phone: 82-31-330-4039. Fax: 82-31-333-1696. E-mail: khlee{at}san.hufs.ac.kr.

Present address: Department of Microbiology, University of Idaho,Moscow, Idaho.

REFERENCES

Top