Identification and Characterization of pvuA, a Gene Encoding the Ferric Vibrioferrin Receptor Protein in Vibrio parahaemolyticus

Abstract

We previously reported that Vibrio parahaemolyticus expresses two outer membrane proteins of 78 and 83 kDa concomitant with production of siderophore vibrioferrin in response to iron starvation stress and that these proteins are the ferric vibrioferrin receptor and heme receptor, respectively (S. Yamamoto, T. Akiyama, N. Okujo, S. Matsuura, and S. Shinoda, Microbiol. Immunol. 39:759-766, 1995; S. Yamamoto, Y. Hara, K. Tomochika, and S. Shinoda, FEMS Microbiol. Lett. 128:195-200, 1995). In this study, the Fur titration assay (FURTA) system was applied to isolate DNA fragments containing a potential Fur box from a genomic DNA library of V. parahaemolyticus WP1. Sequencing a 3.2-kb DNA insert in one FURTA-positive clone revealed that an amino acid sequence deduced from a partial gene, which was preceded by a full-length gene (psuA) encoding a receptor for a siderophore of unknown origin, was consistent with the N-terminal amino acid sequence of the 78-kDa ferric vibrioferrin receptor. Then, the full-length gene (pvuA) encoding the ferric vibrioferrin receptor was cloned and characterized. The deduced protein encoded by pvuA displayed the highest similarity (31% identity; 48% similarity) to RumA, a ferric rhizoferrin receptor of Morganella morganii. Primer extension and Northern blot analyses indicated that psuA and pvuA constitute an operon which is transcribed from a Fur-repressed promoter upstream of psuA. The product of the pvuA gene and its function were confirmed by generating a pvuA-disrupted mutant, coupled with genetic complementation studies. A mutant with disruption in the upstream psuA gene also displayed a phenotype impaired in the utilization of ferric vibrioferrin.

Iron is an essential element for almost all living organisms by virtue of its two valences, which act as cofactors in various oxidative-reductive enzymatic reactions. However, in an aerobic environment at neutral pH, iron exists as insoluble iron complexes that are largely unavailable to bacteria. In the mammalian host most of the iron is intracellular in the form of heme and the small amount of extracellular iron is sequestered in high-affinity iron-binding proteins such as transferrin in serum and lymph and lactoferrin in mucosal secretions. To overcome such iron-restricted conditions, most potential pathogens can express high-affinity transport systems to efficiently obtain iron from one or more of these host iron sources. One mechanism is the direct assimilation of heme and iron bound to transferrin or lactoferrin through specific bacterial receptors (6, 8, 18, 49). Another mechanism involves the production and excretion of soluble siderophores that can chelate free ferric ion or iron bound to transferrin and lactoferrin. Siderophores complexed with ferric ions are subsequently conveyed into the bacterial cell by ligand-specific cell surface receptors and ABC transporters (2, 6, 8, 9, 18, 44, 49). In addition, to exploit the availability of diverse iron sources that may be present in their surroundings, some bacteria express additional iron transport systems including the receptors that enable them to utilize exogenous or heterologous siderophores produced by other microbial species (6, 8, 11, 33). Thus, the ability to utilize the host iron sources has been frequently associated with bacterial pathogenesis (8, 18, 49). Moreover, there is increasing evidence that the restricted availability of iron in the host constitutes a major signal which coordinately regulates the expression of a number of virulence and metabolic genes (37, 39).

In Escherichia coli, the iron transport systems characterized to date are all regulated at the transcriptional level by iron binding protein Fur (ferric uptake regulation), which requires ferrous ion as a cofactor and which acts as a repressor when environmental iron levels are high (2, 20). Fur homologs with similar functions have been identified in many other bacterial species (15).

The genus Vibrio includes 46 species, and 11 of these are human pathogens or have been isolated from human clinical specimens (38). Vibrio parahaemolyticus, one of the pathogenic vibrios, is a gram-negative, halophilic bacterium that naturally inhabits marine and estuarine environments and that causes three major syndromes of clinical illness: gastroenteritis (the most common syndrome), wound infections, and septicemia (10, 41). Past epidemiological studies revealed that thermostable direct hemolysin (TDH) is closely related to the pathogenesis of the bacterium (26, 45, 62). More recently, several workers (23, 59) demonstrated that TDH-negative clinical strains produce TDH-related hemolysin (TRH), suggesting that TRH is another important virulence factor of V. parahaemolyticus. In addition, heat-labile hemolysin(s) and lethal toxin(s) (57), vascular permeability factor(s) (24), and other enteropathogenic factors (22) have been proposed as virulence factors but have not been well characterized. Moreover, a virulence-enhancing effect of iron loading (27), enhanced production of TDH in response to iron limitation (67), and production of TDH under control of a Vibrio cholerae ToxR-like regulator (35) have been demonstrated in this species. Besides these pathogenic factors, adherence to and proliferation within the host intestine are the prerequisites for pathogenesis of V. parahaemolyticus. It seems likely that the ability to acquire iron for proliferation in the host is another important virulence factor in this species.

Under iron-limited conditions, V. parahaemolyticus produces native siderophore vibrioferrin to facilitate iron acquisition (72) and also utilizes heme as a sole source of iron (70). V. parahaemolyticus expresses two iron-repressible outer membrane proteins of 78 and 83 kDa, which were identified as the receptors for ferric vibrioferrin (68) and heme and hemoglobin (70), respectively. In addition, the fur gene of this bacterium was cloned (69) and shown to mediate iron regulation both in the production of vibrioferrin and in the expression of the outer membrane proteins (16). However, little is known about iron assimilation systems in this bacterium at the gene level.

To gain insight into the Fur-regulated genes in V. parahaemolyticus, we used the Fur titration assay (FURTA) system, originally established for E. coli (60), to isolate Fur target genes from V. parahaemolyticus. As a result, we identified an operon consisting of psuA (V. parahaemolyticus siderophore utilization), encoding a new TonB-dependent receptor for an unidentified ligand, probably an exogenous siderophore, and pvuA (V. parahaemolyticus vibrioferrin utilization), encoding the ferric vibrioferrin receptor. The function of the PvuA protein in vibrioferrin-mediated iron transport was confirmed by insertion mutation studies coupled with genetic complementation studies.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

The strains and plasmids used in this study are listed in Table 1. Unless otherwise noted, the V. parahaemolyticus strains were cultured with shaking in Luria-Bertani (LB) broth (1% tryptone [Difco], 0.5% yeast extract [Difco], 3% NaCl, pH 7.5), which was determined to contain ferric ion at a sufficient concentration (approximately 8 μM). The E. coli DH5α and JM109 strains (74), used for propagation of various plasmids, and the E. coli λ pir strains (40), used for construction of the V. parahaemolyticus AQ3354 mutants by homologous recombination, were grown in LB broth or on LB agar containing 0.5% NaCl. When required, appropriate antibiotics were added to the media as follows: ampicillin at 50 μg/ml, chloramphenicol at 10 μg/ml, and tetracycline at 10 μg/ml. All strains were grown at 37°C. Iron-free water was obtained by passage of distilled water through a Milli-Q water filtration unit (Millipore Corp.). All reagent solutions were made with iron-free water, and all glassware was washed with 6 M HCl and rinsed several times with iron-free water.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant featuresa	Reference or source
V. parahaemolyticus
WP1	Clinical isolate	70
AQ3354	Clinical isolate	70
MY-1	Spontaneous vibrioferrin-deficient AQ3354 mutant	71
VPTF1	AQ3354, psuA disrupted	This study
VPTF2	AQ3354, pvuA disrupted	This study
E. coli
H1717	araD139 rpsL150 Δ(argF-lac) relA1 U169 flbB5301 deoC1 ptsF25 rbsR aroB fhuF::λ p lacMu; host strain for FURTA	60
DH5α	endA1 hsdR17 (rk− mk+) supE44 thi-1 recA1 gyrA96 relA1 Δ(argF-lacZYA)U169 deoR[(φ80d lac Δ(lacZ)M15)]; general cloning host	Promega
JM109	recA1 endA1 gylA96 thi hsdR17 supE44 relA1 Δ(lac-proAB)/F" (traD36 proAB+laclq λαχZΔM15); general cloning host	74
SY327 λ pir	Δ(lac pro) argE(Am) recA56 gyrA rpoB λ pir; host for π-requiring plasmids	40
SM10 λ pir	thi thr leu tonA lacY supE recA::RP4-2-Tc::Mu, λ pir; Kmr; host for π-requiring plasmids; conjugal donor	40
Plasmids
pUC19	High-copy-number cloning vector; Apr	65
pBluescript II KS(+)	High-copy-number cloning vector; Apr	Stratagene
pMW118	Low-copy-number cloning vector; Apr	Nippon Gene
pKTN701	R6K-ori suicide vector for gene replacement; Cmr	46
pRK415	Broad-host-range plasmid (∼10.5 kb); Tcr	28
pVP3151	Initially isolated FURTA-positive clone; pUC19 containing chromosomal 3,151-bp PstI fragment from AQ3354 in the orientation opposite to the lac promoter; Apr	This study
pVPD1	pUC19 containing the EcoRI (in pUC19)-EcoRI fragment from pVP3151; Apr	This study
pVPD2	pUC19 containing the KpnI-EcoRI fragment internal to psuA from pVP3151; Apr	This study
pVPV2995	pBluescript II KS(+) containing the chromosomal 2,995-bp EcoRI-SacI fragment from AQ3354 in the same orientation relative to the lac promoter; Apr	This study
pVPV2995-1	pMW118 containing 2,995-bp EcoRI-SacI fragment from pVPV2995; Apr	This study
pVPSV4780	pMW118 containing 39-bp EcoRI-PstI (from pUC19) and 4,780-bp PstI-SacI fragments; ligated 1.8-kb EcoRI (in pUC19)-EcoRI fragment from pVPD1 into the EcoRI site of pVPV2995-1 in the same orientation relative to the lac promoter; Apr	This study
pVPSV4780R	pMW118 containing 4,813-bp SacI-SacI fragment of pVPSV4780 in the opposite orientation relative to the lac promoter; Apr	This study
pTF1	pKTN701 containing 1.4-kb KpnI-EcoRI fragment from pVP3151; Cmr	This study
pTF2	pKTN701 containing 877-bp EcoRI-KpnI fragment PCR amplified with primers 5 and 6 using pVPV2995 as a template; Cmr	This study
pRKVP4.8	pRK415 containing 4,813-bp SacI-SacI fragment of pVPSV4780 in the same orientation relative to the lac promoter; Tcr	This study
pRKVP3	pRK415 containing 2,995-bp EcoRI-SacI fragment of pVPV2995 in the opposite orientation relative to the lac promoter; Tcr	This study

^aAp^r, ampicillin resistant; Cm^r, chloramphenicol resistant; Km^r, kanamycin resistant; Tc^r, tetracycline resistant.

General DNA manipulations.

Chromosomal DNAs of V. parahaemolyticus WP1 and other Vibrio species were extracted from overnight cultures with a Wizard genomic DNA purification kit (Promega) according to the manufacturer's protocol. Cloning, restriction endonuclease digestion, and DNA ligation were carried out according to standard protocols (56). Electroporation was performed in a Gene Pulser apparatus (Bio-Rad) at a capacitance of 25 μF, a potential difference of 2.5 kV, a resistance of 200 Ω, and an electrode distance of 2 mm. Restriction enzymes and a DNA ligation kit (version 2) were purchased from Takara Biomedicals (Kyoto, Japan).

Nucleotide sequence determination.

Nucleotide sequencing was carried out by a Hitachi (Tokyo, Japan) DNA sequencer (SQ5500E) with the Thermo Sequenase premixed or core cycle sequencing kit and appropriate primers, which were labeled with a 5"-oligonucleotide Texas red labeling kit (Amersham Pharmacia Biotech). Sequence analysis was conducted with the Genetyx-Mac, version 9.0, software package (GENETYX Software Development Co., Tokyo, Japan). The BLASTP and FASTA programs (1, 50) of the Institute for Chemical Research, Kyoto University, were used to determine homologies of the deduced amino acid sequences to other proteins.

Oligonucleotide primers.

Primers 1 and 2 (nucleotide positions 2140 to 2159 and 2807 to 2826) and primers 3 and 4 (nucleotide positions 3634 to 3653 and 4259 to 4278) were used for preparation of digoxigenin (DIG)-labeled hybridization probes A and B, respectively, under the PCR conditions recommended in the PCR DIG probe synthesis kit (Roche Diagnostics). The EcoRI-KpnI fragment internal to pvuA, which was ligated into pKTN701 to construct pvuA disruptant VPTF2, was prepared with primers 5 (AACTGAATTCAGTAACCGCC; nucleotide positions 3046 to 3065) and 6 (ATACGGTACCCACTGATCGG; nucleotide positions 3912 to 3931) (the nucleotides changed to generate the respective restriction enzyme sites are underlined) under the following PCR conditions. KOD-plus DNA polymerase (Toyobo, Osaka, Japan) was used, and after initial denaturation of 94°C for 2 min, a cycle of 94°C for 15 s, 55°C for 30 s, and 68°C for 1 min was repeated 30 times.

FURTA.

The FURTA was essentially performed as described by Stojiljkovic et al. (60). V. parahaemolyticus WP1 chromosomal DNA fragments (2 to 5 kb) completely digested with PstI were cloned into the PstI site of pUC19. The resulting recombinant plasmids were introduced into E. coli H1717 carrying the chromosomal Fur-repressible fhuF::lacZ fusion, and ampicillin-resistant transformants were screened for the Lac⁺ phenotype on MacConkey lactose agar plates (Difco) supplemented with 20 μM ferrous ammonium sulfate after 15 h of growth at 37°C. Following several rounds of screening, we isolated more than 20 FURTA-positive clones with inserts of different sizes (data not shown).

N-terminal amino acid sequence determination.

Sarkosyl (Sigma)-insoluble outer membrane proteins were prepared from V. parahaemolyticus WP1 cells grown in LB broth supplemented (iron-limited) or not (iron-sufficient) with ethylenediamine-di(o-hydroxyphenylacetic acid) (EDDA; Sigma) at 25 μM, and the proteins were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) with a 5% stacking gel and a 10% separating gel (34). The separated proteins were electroblotted to a prewetted polyvinylidene difluoride (PVDF) membrane (ProBlott; Applied Biosystems) using a Trans-Blot semidry electrophoretic transfer cell (Bio-Rad) essentially as described by Towbin et al. (63) and stained with Coomassie brilliant blue. The membrane was rinsed several times with distilled water and air dried, and the iron-repressible protein bands of 78 and 83 kDa were excised from the membrane with a razor blade. The N-terminal amino acid sequence was determined by automated Edman degradation with a model 491 protein sequencer (Applied Biosystems) equipped with an online model 120A PTH-amino acid analyzer. The amino acid sequences were compared with those deduced from the nucleotide sequences of the FURTA-positive clones.

Southern blot analysis and colony hybridization.

These procedures were performed according to the DIG system user's guide for filter hybridization (Roche Diagnostics). Digested chromosomal DNA was resolved in 1% agarose gel, transferred onto a positively charged nylon membrane (Roche Diagnostics) with a model 785 vacuum blotter (Bio-Rad), and fixed to the membrane by baking it for 30 min at 120°C. Colonies on a nylon membrane for colony and plaque hybridization (Roche Diagnostics) were denatured and neutralized, and the transferred DNA was fixed to the membrane by baking it for 30 min at 80°C. Hybridization with an appropriate DIG-labeled probe was carried out overnight at 68°C, and, after treatment of the membrane with alkaline phosphatase-labeled anti-DIG Fab fragments, the hybridized DNA was detected by a CSPD reagent for Southern blot analysis and by colorimetric detection reagents nitroblue tetrazolium and BCIP (5-bromo-4-chloro-3-indolylphosphate) for colony hybridization, according to the DIG system user's guide.

Cloning of the pvuA gene.

V. parahaemolyticus WP1 chromosomal DNA was first digested by the combination of EcoRI with various restriction enzymes, and the DNA fragments were examined by Southern blotting with DIG-labeled probe A (see Fig. 1C). Then, approximately 3-kb band fragments in the EcoRI-SacI digest which hybridized with probe A as a single band were ligated into the same restriction sites of pBluescript II KS(+). Colonies on LB agar plates were screened by colony blot hybridization with the same probe. The nucleotide sequence of the insert of positive plasmid pVPV2995 was determined.

FIG. 1.

Restriction map of the 4.8-kb psuA-pvuA region from V. parahaemolyticus WP1 (A), the relevant plasmids (B), and the probes used for hybridization experiments (C). ||||, pUC19 DNA sequence that provided convenient sites for subcloning. Open arrows, ORFs (arrows point in the direction of transcription). pVP3151 was initially isolated as a clone conferring a Lac⁺ phenotype in the FURTA system. Small solid arrows, orientation of the vector lac promoter. E. coli H1717 carrying the indicated plasmid was evaluated by FURTA for the positive (+) and negative (−) phenotypes on MacConkey agar plates. Open inverted triangle, Fur box.

Western blot analysis.

Expression of the PvuA protein was examined by Western blot analysis using rabbit antiserum against the 78-kDa ferric vibrioferrin receptor of V. parahaemolyticus WP1 (68). The Sarkosyl-insoluble outer membrane fractions were prepared from washed stationary-phase cells broken by sonication, as previously described (68). Protein concentrations were determined by the Lowry method. After SDS-PAGE, protein bands were transferred onto a PVDF membrane (63) and the membrane was treated with blocking solution (8% skim milk, 20 mM Tris-HCl [pH 7.5], 10 mM NaCl, 0.02% NaN₃) for 1 h. The membrane was then incubated with antiserum diluted 1:500 with gentle agitation and was reacted for 1 h with alkaline phosphatase-conjugated goat anti-rabbit immunoglobulin G antibodies (Roche Diagnostics) diluted 1:2,000. After thorough washing, the membrane was dipped in a BCIP-nitroblue tetrazolium substrate solution (Bio-Rad).

RNA isolation and analysis.

V. parahaemolyticus WP1 was grown in LB broth in the presence (iron-limited cells) and absence (iron-deficient cells) of 25 μM EDDA to an A₆₆₀ of 0.5. Total RNA from each cell sample was prepared using ISOGEN (Nippon Gene, Tokyo, Japan) according to the manufacturer's instructions, and the amount of total RNA was quantified by measuring A₂₆₀.

(i) Primer extension.

Oligonucleotide primers 7 (nucleotide positions 414 to 433) and 8 (nucleotide positions 2610 to 2629), complementary to the 5"-end regions of psuA and pvuA, respectively, were 5" labeled with Texas red as described above. Each of the Texas red-labeled primers was mixed with total RNA (approximately 10 μg), and the primer was extended at 50°C for 60 min using avian myeloblastosis virus reverse transcriptase XL (Takara Biomedicals) according to the manufacturer's instructions. The extension products were sized on a 6% denaturing polyacrylamide gel by a Hitachi DNA sequencer (SQ5500E) alongside the DNA sequence ladder of each control region synthesized with the same labeled primer.

(ii) Northern blot analysis.

Total RNA (approximately 10 μg) from each cell sample was separated electrophoretically on a 1% agarose-2.2 M formaldehyde gel. The gel was rinsed in 20× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) buffer, and the RNA was transferred to a positively charged nylon membrane by a model 785 vacuum blotter. The SalI-EcoRI and PstI-ClaI fragments from pVP3151 and pVPV2995, respectively, were labeled with alkaline phosphatase according to the manufacturer's protocol (AlkPhos Direct; Amersham Pharmacia Biotech) to yield AlkPhos-labeled probes A and B, respectively (see Fig. 1C). Alkaline phosphatase activity was visualized fluorescently by using CDP-Star chemiluminescence reagent and by exposure to Fuji RX-U film. Perfect RNA markers (Novagen) were used as a standard.

Construction of mutant strains.

To investigate their specific functions, we attempted to inactivate the psuA and pvuA genes of V. parahaemolyticus by homologous recombination as described below. Unfortunately, no mutants with disruption in these genes were obtained from strain WP1. Since a similar event has been reported for the same strain by Nishibuchi et al. (46), we decided to use another strain, AQ3354, which appeared to have an arrangement of psuA and pvuA identical to that of strain WP1. The KpnI-EcoRI fragment derived from pVP3151 was subcloned into R6K-ori suicide vector pKTN701 (46) digested with the same restriction enzymes to yield pTF1. For the construction of pTF2, PCR was done with oligonucleotide primers 5 and 6 using pVPV2995 as the template to introduce EcoRI and KpnI linkers, and then an EcoRI-KpnI-digested fragment of this PCR product was inserted into pKTN701. The obtained plasmids were transformed into E. coli SM10 λ pir as a donor and transferred to V. parahaemolyticus AQ3354 by membrane-filter mating conjugation followed by incubation for 3 h at 37°C of a nitrocellulose membrane laid over LB agar plates with 1.5% NaCl. Transconjugants were selected by overnight incubation at 37°C on thiosulfate-citrate-bile-sucrose agar (Difco) (to select against the E. coli donor) containing chloramphenicol at 10 μg/ml. Some of the chloramphenicol-resistant colonies were isolated, and their single-crossover mutations with respect to psuA and pvuA were confirmed by Southern blot analysis with DIG-labeled probe A (data not shown). The mutants with disruption in psuA and pvuA thus obtained were designated VPTF1 and VPTF2, respectively.

For complementation studies, the SacI-SacI fragment of pVPSV4780 containing an intact copy of the V. parahaemolyticus WP1 psuA-pvuA locus was subcloned into the SacI site of broad-host-range plasmid pRK415 (28), and resulting replicative plasmid pRKVP4.8 was conjugated into pvuA mutant VPTF2, with E. coli SM10 λ pir used as a donor. One of the chloramphenicol- and tetracycline-resistant colonies was selected, and correct transfer of the plasmid was confirmed by restriction enzyme analysis. Plasmid pRKVP3, pRK415 bearing the intact copy of the pvuA gene, was prepared and introduced into VPTF1 and -2 in the same manner.

Growth assay.

Overnight cultures of V. parahaemolyticus AQ3354 and its mutant strains in LB medium were subcultured into the same medium containing 25 μM EDDA at an initial cell density corresponding to an A₆₆₀ of 0.05. Vibrioferrin (72) or ferrichrome (Sigma) was added to the medium at a final concentration of 20 μM. Cultures were shaken (125 rpm) at 37°C, and growth was evaluated by determining the A₆₆₀ of the culture. Three independent experiments were conducted on each strain.

Nucleotide sequence accession number.

The nucleotide sequence data have been deposited in EMBL, GenBank, and DDBJ databases under accession no. AB048250.

RESULTS

Identification of the V. parahaemolyticus ferric vibrioferrin receptor gene.

Application of the FURTA system to V. parahaemolyticus WP1 allowed us to isolate many different positive clones which contained potential Fur-regulated gene fragments. This was consistent with our hypothesis that the FURTA system might be applicable to V. parahaemolyticus because the Fur protein of this bacterium showed the high degree of homology (81% identity) with that of E. coli (16). Nucleotide sequences of the 5" and 3" regions in the insert of plasmid pVP3151 suggested the presence of full-length and partial open reading frames (ORFs) in the same direction (Fig. 1). Interestingly, as described below, the amino acid sequences deduced from these ORFs were each compatible with one of the extracted amino acid sequences in the 83-kDa band and with the N-terminal amino acid sequence of the 78-kDa ferric vibrioferrin receptor. The putative Fur box which overlaps the −10 region of the predicted promoter of the upstream gene and which shares 11 of 19 nucleotides (GTAAATAATATTTGTTCTT; the matched nucleotides are underlined) with the E. coli consensus Fur box (12) was found. In agreement with this, subclone pVPD1, containing the EcoRI (in pUC19)-EcoRI fragment of pVP3151 conferred a Lac⁺ phenotype to H1717 in the FURTA, whereas subclone pVPD2, with a deletion of the PstI-KpnI portion containing the putative Fur box sequence from pVP3151, did not (Fig. 1).

N-terminal amino acid sequences of iron-repressible outer membrane proteins from V. parahaemolyticus.

Separating the Sarkosyl-insoluble outer membrane proteins from V. parahaemolyticus WP1 grown in iron-limited medium revealed two major protein bands of 78 and 83 kDa, which were normally suppressed by growth in iron-replete medium (Fig. 2A). These iron-repressible proteins were blotted from a preparative gel onto a PVDF membrane, and the area of the membrane corresponding to each band was cut out. The amino acid sequences determined by Edman degradation are shown in Fig. 2B. The N-terminal amino acid sequence of the 78-kDa protein, which had been identified as the ferric vibrioferrin receptor protein (68), matched the amino acid sequence deduced from pvuA nucleotide positions 2588 to 2617 (Fig. 3B). Unexpectedly, sequence determination of the 83-kDa protein band, in which the heme and hemoglobin receptor had been identified (70), revealed the presence of three kinds of proteins, three amino acid residues being detected in most of the 10 cycles (Fig. 2B). However, the amino acid sequence (SEETNSTPSA) extracted from the amino acid residues detected in every cycle was correlated with that deduced from psuA nucleotide positions 393 to 422 (Fig. 3A). In addition, the other extracted amino acid sequence (AEQAQQLASQ) corresponded with the deduced N-terminal amino acid sequence encoded by an incomplete ORF detected in another V. parahaemolyticus FURTA-positive clone, whose deduced amino acid sequence had 27% identity (in a 73-amino-acid overlap) with the outer membrane receptor IutA in the E. coli ferric aerobactin transport system (T. Funahashi and S. Yamamoto, unpublished data). Therefore, the remaining amino acid sequence (EQHSTFNEVV) is suggestive of the 83-kDa heme and hemoglobin receptor (70) because it is similar to the N-terminal amino acid sequences of V. cholerae HutA (DDYASFDEVV) (21) and V. vulnificus HupA (QDAGLFDEVV) (36), identified as heme receptors; the amino acid residues identical to those of the TonB box sequences proposed for HutA and HupA are underlined.

FIG. 2.

(A) SDS-PAGE of the outer membrane proteins from V. parahaemolyticus WP1. Lane 1, molecular mass markers (kilodaltons) indicated on the left; lanes 2 and 3, 30 μg of outer membrane protein preparations from V. parahaemolyticus WP1 grown in iron-sufficient and iron-limited conditions, respectively. (B) N-terminal amino acid sequences determined for the iron-repressible 78- and 83-kDa proteins. Designation of the sequences determined for the 83-kDa protein band as IutA and heme receptor are putative. For details for PsuA and PvuA, see the text.

FIG. 3.

Portions of the nucleotide sequences and predicted protein sequences of the psuA and pvuA genes. Nucleotide sequences of the promoter and the N-terminal coding regions of psuA (A), the intervening region between psuA and pvuA and the N-terminal coding region of pvuA (B), and the C-terminal coding region of pvuA including a putative transcription terminator (C) are shown. The Shine-Dalgarno (SD) sequences and the putative TonB box amino acid sequences are indicated. The putative Fur box is double overlined. Vertical arrow, transcription initiation site determined by primer extension analysis. The deduced amino acid sequences compatible with the N-terminal sequences of the mature PsuA and PvuA are double underlined. Converging arrows, inverted repeats located downstream of psuA and pvuA. Two pairs of direct repeat sequences (A1 and A2; B1 and B2) in the region between psuA and pvuA are overlined. Asterisks, stop codons.

Cloning of the full-length pvuA gene.

V. parahaemolyticus WP1 genomic DNA samples digested with various combinations of restriction enzymes were separated by agarose gel electrophoresis followed by hybridization with DIG-labeled probe A (Fig. 1). The gel-extracted 3-kb EcoRI-SacI fragments that would be sufficient in length to encode pvuA were ligated into the same sites of pBluescript II KS(+). Positive clones were screened by colony blot hybridization with the same probe, and recombinant plasmid pVPV2995, which produced a Lac⁻ phenotype when introduced into E. coli H1717, was selected (Fig. 1).

Nucleotide sequences of the psuA and pvuA genes and homology search of the deduced amino acid sequences.

The combined nucleotide sequences of pVP3151 and pVPV2995 yielded two complete ORFs in the same orientation and 122 bp apart. The G+C content of the sequenced DNA was 49%, which is similar to the V. parahaemolyticus overall G+C content of 46 to 47% (3). The nucleotide sequences, including the promoter region of psuA, the intergenic region between psuA and pvuA, and the end of pvuA including the putative transcriptional termination signal, are presented in Fig. 3. Possible Shine-Dalgarno sequences (GGAA for psuA and AAGGAG for pvuA) are located just upstream of the starting methionine codons. In contrast to what was found for psuA, the obvious promoter sequences similar to both components of the E. coli ςsgr;⁷⁰-like promoter consensus were not detectable in the upstream region of pvuA. An inverted repeat suggestive of a potential rho-independent transcriptional terminator begins 28 bp beyond the termination codon of pvuA. Interestingly, an inverted repeat and two pairs of direct repeats are also present in the intergenic region between psuA and pvuA, but their roles are unclear because, as described below, primer extension and Northern blot analyses indicated that pvuA and psuA were cotranscribed under iron-limited conditions.

Comparison of the N-terminal amino acid sequences determined for PsuA and PvuA with the amino acid sequences deduced from the nucleotide sequences of psuA and pvuA disclosed additional amino acid residues at the amino termini, indicating that the psuA and pvuA genes encode 25- and 37-amino-acid signal peptides, respectively, which are cleaved during translocation across the membrane. Indeed, each of the putative signal peptides has a typical signal sequence and a potential peptidase cleavage site (66). The entire protein encoded by psuA consists of 678 residues, and the mature protein has a calculated molecular mass of 72,364 Da. This is much less than the 83-kDa molecular mass of PsuA estimated from the electrophoretic mobility by SDS-PAGE. The difference may be due to aberrant migration of PsuA on SDS-PAGE gel as frequently reported for outer membrane proteins, but the possibility that this protein is posttranslationally modified prior to incorporation into the outer membrane cannot be ruled out. The second gene, pvuA, encodes a protein of 712 residues, and the molecular mass of the mature protein is 75,080 Da, which is almost compatible with the 78-kDa estimate based on SDS-PAGE. Furthermore, the predicted isoelectric points of the mature PsuA and PvuA, 4.50 and 4.71, respectively, are similar to the acidic isoelectric points of TonB-dependent outer membrane proteins of E. coli (43).

The BLASTP or FASTA algorithm for protein homology was used to compare PsuA and PvuA with other proteins. The deduced PsuA protein sequence had homology with those of a variety of the known TonB-dependent siderophore receptors from many bacteria and was the most similar to that of the putative ferrichrome iron receptor of Synechocystis species (PIR accession no. S74457; 27% identity, 41% similarity). However, PsuA exhibited no homology to the recently characterized V. cholerae FhuA, needed for ferrichrome iron utilization (55). On the other hand, PvuA had homology with RumA (31% identity, 48% similarity), the ferric rhizoferrin receptor of Morganella morganii (33), and FecA (31% identity, 47% similarity), the receptor of the E. coli ferric citrate transport system (53), but no PvuA homolog was found in the genomic sequence of V. cholerae. This is reminiscent of some structural analogy because rhizoferrin consists of two citrate moieties linked to putrescine (13). However, the mature PvuA lacked the homologous counterpart of the N-terminal extension of FecA, which is required to mediate transcription induction by the cognate ferric citrate (30). The predicted TonB box amino acid sequences which may be involved in direct interaction with the TonB protein are detected near the N-termini of PsuA (ETIQV) and PvuA (ETVVV) (Fig. 3A and B); the amino acid residues identical to those highly conserved in many TonB-dependent ferric siderophore receptors are underlined (4, 29, 33, 43). In addition, various outer membrane proteins possess the highly conserved C-terminal sequences which were proposed to form an amphipathic β-sheet important for the correct assembly of the protein into the outer membrane (4, 29, 61). This peculiar sequence motif also exists in PsuA and PvuA and in TonB-dependent outer membranes of other Vibrio species listed in Fig. 4.

FIG. 4.

Comparison of the last 11 C-terminal amino acid residues of PsuA and PvuA with those of TonB-dependent siderophore and hemin receptor proteins in Vibrio species. All amino acid sequences revealed the presence of a potential amphipathic β-sheet with hydrophobic residues at positions 1 (preferentially Phe), 3, 5, 7, and 9 and Arg at position 11 from the C terminus. The amino acid residues of PsuA at position 11 and of FhuA at position 1 are exceptionally alanine and tryptophan, respectively, which are boxed. GenBank accession numbers: V. anguillarum FatA, J03529; V. cholerae FhuA, AF203702; V. cholerae IrgA, M63192; V. cholerae HutA, L27149; V. cholerae ViuA, M90461; V. vulnificus HupA, AF047484; V. vulnificus VvuA, AF156494; Vibrio orientalis IutA, AB010890.

Localization of the transcriptional start site by primer extension.

To identify the transcriptional start sites of psuA and pvuA and to clarify whether transcription of these genes is regulated by iron, primer extension analysis with total RNA from strain WP1 cells was performed with oligonucleotide primers 7 and 8, complementary to the DNA sequences located 97 and 134 bp downstream of the methionine start codons of psuA and pvuA, respectively. As shown in Fig. 5A, primer extension analysis with primer 7 for total RNA from the iron-limited cells identified a potential transcriptional start site 34 bp upstream of the translational start site (Fig. 3A). However, the same analysis with primer 8 was unable to identify any transcriptional start site within the junction between the two genes, even when total RNAs from both iron-limited and iron-sufficient cells were used (data not shown). Therefore, we assumed that the pvuA gene might be cotranscribed with the psuA gene under iron-limited conditions.

FIG. 5.

Primer extension (PE) (A) and Northern blot analysis (B) of the total RNAs. (A) Total RNAs were prepared from V. parahaemolyticus WP1 cells grown in LB broth to an A₆₆₀ of 0.5 in the absence (+) or presence (−) of EDDA at a final concentration of 25 μM. Arrow, position (C284) corresponding to the 5" terminus of psuA in the DNA sequence ladder of the noncoding strand. (B) The same total RNAs (plus and minus) as in panel A were used. Blots were probed with AlkPhos-labeled probe A, internal to psuA (a), and AlkPhos-labeled probe B, internal to pvuA (b). Positions of RNA standards (in kilobases) are shown on the right.

Identification of the transcript by Northern blot analysis.

To clarify the aforementioned assumption, Northern blot analysis was performed for the same total RNA preparations as in primer extension analysis. The blots were probed with AlkPhos-labeled probes A and B, internal to psuA and pvuA, respectively (Fig. 1C). A 4.5-kb single band that hybridized with either probe A or probe B was detected in total RNA from the iron-limited cells; no corresponding band was seen in total RNA from cells grown under iron-sufficient conditions (Fig. 5B). These results indicate that both genes are cotranscribed as a single mRNA species under iron-limited conditions.

Expression of pvuA in an E. coli background.

Contrary to our finding that the pvuA gene has no promoter, Western blotting of the outer membrane preparation from E. coli JM109 carrying pVPV2995 [pBluescript II KS(+) containing the chromosomal EcoRI-SacI fragment in the same orientation relative to the lac promoter] revealed the production of a small amount of PvuA irrespective of iron status in the growth medium (data not shown). However, E. coli JM109 carrying pBluescript II SK(+) containing the EcoRI-SacI fragment of pVPV2995 in the opposite orientation relative to the lac promoter produced no PvuA, indicating that pvuA has no promoter functional in an E. coli background and therefore that the production of PvuA by E. coli JM109 carrying pVPV2995, the derivative of pBluescript II KS(+), may be due to leaky activity of the lac promoter. Taking these results into consideration, we attempted to reconstruct the entire psuA and pvuA genes including the promoter region in a single high-copy-number plasmid to examine the iron-regulated production of the PvuA protein by Western blotting. However, such a plasmid was not obtained, suggesting that some overproduction of PsuA and/or PvuA may be toxic to E. coli. Then, pVPSV4780R was constructed by means of low-copy-number plasmid pMW118, into which the SacI-SacI fragment from pVPSV4780 was inserted in the orientation opposite to that of the lac promoter to transcribe psuA and pvuA from the promoter upstream of psuA. E. coli JM109 carrying either pMW118 itself or pVPSV4780R was then grown under iron-limited conditions, and the outer membrane preparations were analyzed by SDS-PAGE and Western blotting. Although the outer membrane proteins expressed by the host strain itself hindered observing any apparent difference in the SDS-PAGE band profile between these two outer membrane preparations (Fig. 6A, lanes 2 and 3), a band reactive with the antiserum against the V. parahaemolyticus WP1 78-kDa outer membrane protein was seen in E. coli JM109 carrying pVPSV4780R (Fig. 6B, lane 3). In contrast, E. coli JM109 carrying pMW118 displayed no corresponding band (Fig. 6B, lane 2). These results proved that in an E. coli background also the potential Fur box located upstream of psuA can mediate iron regulation, leading to the transcription of these two genes as a bicistronic message.

FIG. 6.

SDS-PAGE (A) and Western blotting (B) of the outer membrane proteins prepared from the relevant strains. The outer membrane protein fractions of strains V. parahaemolyticus AQ3354, E. coli JM109 carrying pMW118 or pVPSV4780R, and V. parahaemolyticus AQ3354(pRKV4.8) were prepared from cells grown to stationary phase in LB broth in the presence of 25 μM EDDA. For preparation of the outer membrane proteins from the psuA and pvuA disruptants of V. parahaemolyticus AQ3354 and V. parahaemolyticus AQ3354(pRK415), the culture previously grown in LB broth to an A₆₆₀ of 0.3 was supplemented with EDDA at a final concentration of 100 μM and then the cultures were further incubated to an A₆₆₀ of 0.5 to induce iron-regulated outer membrane proteins. Outer membrane proteins (25 μg) were electrophoresed in an SDS-containing 10% polyacrylamide gel in duplicate; one was strained with Coomassie blue, and the other was electroblotted and immunostained with the antiserum against the PvuA protein. Lanes: 1, V. parahaemolyticus AQ3354 (wild type, positive control); 2, E. coli JM109(pMW118); 3, E. coli JM109(pVP4780R); 4, VPTF1; 5, VPTF2; 6, VPTF2(pRKVP4.8); 7, VPTF2(pRK415). Arrows, PvuA protein (78 kDa) and 83-kDa protein.

Characterization of the psuA and pvuA mutants.

Gene-disrupted mutants VPTF1 and VPTF2 were created with suicide plasmids pTF1 and pTF2, respectively. Table 2 shows the abilities of the wild-type and the mutant strains to utilize vibrioferrin or ferrichrome for growth. Wild-type strain AQ3354 grew well even without supplementation of vibrioferrin under the iron-restricted conditions imposed by EDDA, whereas mutant strains VPTF1 and VPTF2 as well as MY-1, a spontaneous vibrioferrin-deficient mutant of wild-type AQ3354 (71), failed to grow under the same conditions. The result for MY-1 indicates that vibrioferrin can capture iron chelated to EDDA and deliver it to the cytoplasm. Failure of VPTF1 to grow under the same conditions as those for VPTF2 suggested a polar effect of the psuA disruption on the transcription of downstream gene pvuA. However, psuA mutant VPTF1 was still able to utilize ferrichrome iron, at least suggesting that the psuA gene is not responsible for the utilization of ferrichrome iron.

TABLE 2.

Utilization of vibrioferrin and ferrichrome by V. parahaemolyticus AQ3354 and its mutant strains

LB broth supplement	Growth of indicator straina:
	AQ3354	MY-1b	VPTF1	VPTF2	VPTF2(pKVP4.8)	VPTF2(pRK415)
EDDA (25 μM)	++	−	−	−	+	−
EDDA (25 μM) + 20 μM vibrioferrin	++	++	−	−	+	−
EDDA (25 μM) + 20 μM ferrichrome	++	++	++	++	ND	ND

^aA₆₆₀ of the culture after incubation for 25 h at 37°C was measured except for the AQ3354 culture, whose A₆₆₀ was measured after incubation for 17 h. ++, A₆₆₀ = 0.77 to 0.94; +, A₆₆₀ = 0.55 to 0.59; −, no growth; ND, not determined. The data from three independent experiments are shown.

^bMY-1 is a spontaneous vibrioferrin-deficient mutant derived from wild-type AQ3354 (71).

To confirm by Western blotting that disruptants VPTF1 and VPTF2 indeed lack the 78-kDa protein, their outer membrane proteins were prepared from a culture treated as follows. The culture was grown in LB broth to an A₆₆₀ of 0.3 and was split into aliquots; one aliquot was left untreated, and the other was supplemented with EDDA at a final concentration of 100 μM to achieve iron depletion. Then, both aliquots were further incubated until an A₆₆₀ of 0.5 was reached. SDS-PAGE analysis of the outer membrane proteins from these disruptants revealed loss of the 78-kDa protein band, in contrast to apparent induction of the 83-kDa protein in response to suddenly imposed iron restriction (Fig. 6A, lanes 4 and 5). Consistent with this result, no band corresponding to the 78-kDa protein was detected by Western blotting (Fig. 6B, lanes 4 and 5).

For complementation analysis, pvuA mutant VFTF2 was transformed with either pRK415 or pRKVP4.8, a pRK415 derivative bearing an intact copy of the psuA and pvuA genes. Growth of transformant VPTF2(pRKVP4.8) regardless of the supplementation with vibrioferrin was restored to about 60% of the level of wild-type AQ3354, whereas that of the mock transformant was not restored at all (Table 2). Incomplete restoration observed for VPTF2(pRKVP4.8) may be due to the maintenance of a relatively large plasmid (ca. 15 kb) with concomitant expression of the two antibiotic resistance genes. Consistent with the results of growth complementation assay, both SDS-PAGE and Western blotting of the outer membrane preparation from VPTF2(pRKVP4.8) cells grown under iron-limited conditions disclosed the expression of the 78-kDa protein (Fig. 6, lane 6). The mock transformant induced only the 83-kDa protein (Fig. 6, lane 7). Moreover, growth of VPTF1 and VPTF2 was not complemented with pRKVP3, a pRK415 derivative bearing an intact copy of the pvuA gene (data not shown), indicating that the pvuA gene has no promoter. This constitutes better proof that the insertion mutation in the psuA gene was polar to pvuA expression. These results clearly demonstrated that the pvuA gene in fact encodes the 78-kDa receptor protein necessary for ferric vibrioferrin utilization.

Distribution of pvuA in other Vibrio species.

Southern blot analysis with DIG-labeled probe B (Fig. 1C) prepared with primers 3 and 4 internal to pvuA was performed to detect homologous pvuA genes in other pathogenic Vibrio species. Genomic DNAs from Vibrio species were completely digested with SacI, and hybridization was carried out under a stringent conditions (at 68°C). All of the seven other V. parahaemolyticus clinical and environmental isolates and the four Vibrio alginolyticus environmental isolates tested exhibited a single hybridization band of ca. 5.2 kb, similar to the WP1 strain. However, the American Type Culture Collection type strains of V. cholerae, Vibrio mimicus, Vibrio vulnificus, Vibrio furnissii, Vibrio fluvialis, and Vibrio hollisae gave no signal (data not shown), demonstrating that the pvuA gene occurs exclusively in V. parahaemolyticus and V. alginolyticus. The detection of pvuA in V. alginolyticus is reinforced by the fact that the strains of this species produced vibrioferrin and expressed the iron-repressible outer membrane protein, which strongly cross-reacted with the antiserum against the 78-kDa V. parahaemolyticus outer membrane protein (68).

DISCUSSION

Siderophores detected to date in pathogenic vibrios have displayed considerable structural diversity based on their main chelating groups, but Vibrio species generally produce a single cognate siderophore. V. cholerae (19), V. fluvialis (73), and V. vulnificus (47) produce catecholate-type siderophores, all of which characteristically contain a norspermidine (one of the unusual polyamines) moiety as a framework. We also found that V. furnissii produces the same catecholate as V. fluvialis. Anguibactin, a member of a unique structural class of plasmid-mediated siderophores isolated from Vibrio anguillarum, possesses both catecholate and hydroxamate functional groups (25). Interestingly, V. mimicus and V. hollisae produce hydroxamate-type siderophore aerobactin (48), which is mainly found in species in the family Enterobacteriaceae (49). In contrast, V. parahaemolyticus, as well as V. alginolyticus, produces structurally novel siderophore vibrioferrin (68, 72), which is classified as a citrate-based polyhydroxycarboxylate along with staphyloferrins A and B in staphylococcal species (14, 32) and rhizoferfin in fungus Rhizopus microsporus and other zygomycetes (13). Transport components for this type of siderophores in the producer strains have been poorly characterized at the gene level, although genes rumAB, essential for uptake of ferric rhizoferrin as an exogenous siderophore, have been characterized in M. morganii (33).

In this study, by a combination of the FURTA system to isolate the Fur box-containing gene fragments with information on the N-terminal amino acid sequences of the iron-repressible outer membrane proteins, we identified V. parahaemolyticus psuA and pvuA genes, which encode the TonB-dependent outer membrane receptors for a putative ferric siderophore and ferric vibrioferrin, respectively. The same genetic approach using the FURTA system led to the recent identification of Fur-repressed genes in bacteria other than E. coli, for example, in Bordetella, Salmonella, and Vibrio spp. (5, 7, 17, 31, 51, 52, 64). A sequence homologous to the E. coli Fur box consensus was also detected upstream of psuA. Primer extension analysis of mRNA from V. parahaemolyticus grown under iron-limited or iron-sufficient conditions defined a transcriptional start site adjacent to the proposed Fur box and demonstrated iron regulation of these genes at the transcriptional level. The iron regulation of these genes through the Fur box was consistent with the FURTA-positive phenotype of the isolated clone and with the constitutive expression of the iron-repressible outer membrane proteins, including the 78-kDa protein in the manganese-resistant mutants (fur mutants) (16). Examination of the size of the psuA-pvuA transcript by Northern blot analysis indicated that these genes are cotranscribed as a single unit from a Fur-repressed promoter and that the inverted repeat located just downstream of the second gene, pvuA, functions as a transcriptional termination signal. Such an operon comprising the two different siderophore receptor genes whose transcription is controlled by a common Fur box is unique, since many of the siderophore receptor genes characterized to date are always arranged in an operon with the relevant genes encoding the other iron transport components and the siderophore biosynthesis enzymes (6). However, genetic organization around the psuA and pvuA genes appeared to be conserved among the V. parahaemolyticus strains tested. In all of the seven other strains tested, the probe was hybridized with the SacI fragments of a similar size, suggesting that this operon may play an important role in iron assimilation of this species.

On the other hand, there is a stem-loop structure located 68 nucleotides downstream of the psuA translational stop codon, which seems to fit the criteria for a potential rho-independent transcriptional termination site (54). However, either this transcriptional termination site may be inefficient or there may be an antitermination mechanism operating only under iron-limited conditions, since Northern blot analysis showed that the psuA and pvuA genes are organized in an operon structure. Interestingly, two pairs of seven-nucleotide direct repeat sequences, 5"-TTTTGCC-3" and 5"-ATGTTTT-3", are identifiable in the same intergenic region. They seemed to be typical of sites of the binding of an antitermination factor to the mRNA under transcription, and binding may prevent the formation of the termination structure (42, 58). However, since E. coli JM109 carrying pVPSV4780R also expressed the PvuA protein under iron-limited conditions, it seems unlikely that the antitermination mechanism, if one exists, is specific to V. parahaemolyticus. Further studies are needed to clarify possible functions of these characteristic sequences.

Because of the similarity of PvuA to the putative ferrichrome iron receptor of Synechocystis species and because of the ability of V. parahaemolyticus to utilize ferrichrome as a exogenous siderophore, we hypothesized that psuA was the ferrichrome iron receptor gene. The disruption of this gene, however, did not affect ferrichrome utilization. Since the PsuA protein is in fact expressed by V. parahaemolyticus, it is not unexpected that this protein may be a receptor for a ferric siderophore of unknown origin. Therefore, the present study suggests that V. parahaemolyticus has two other siderophore-mediated iron acquisition systems, which may be associated with ferrichrome and aerobactin. Recent studies have demonstrated that, besides their cognate siderophores, many bacteria can utilize a wide variety of different exogenous siderophores (6, 8, 11, 33). This strategy may enhance the organism's ability to acquire iron under a range of environmental conditions and reflects the importance of iron competition in the natural habitat of the bacteria. An operon required for utilization of ferrichrome as an exogenous siderophore has recently been identified in V. cholerae (55). Studies in our laboratory to isolate and characterize the genes involved in the ferrichrome- and aerobactin-mediated iron acquisition systems in V. parahaemolyticus are under way.

Presumptive proof for the function of pvuA in ferric vibrioferrin utilization was obtained by phenotypic analysis of V. parahaemolyticus pvuA insertion mutants VPTF1 and VPTF2. These mutants were incapable of utilizing ferric vibrioferrin in the growth assay and of expressing the 78-kDa iron-repressible outer membrane protein. Moreover, the psuA insertion mutation resulted in a polar effect on pvuA expression, confirming that these two genes are organized in an operon structure. Introduction of pRKVP4.8, containing the wild-type psuA-pvuA locus, into mutant VPTF2 restored its abilities to utilize vibrioferrin and to express the 78-kDa protein for growth under iron-limiting conditions.

Genetic and biochemical evidence presented in this study corroborates our previous identification of the 78-kDa outer membrane protein as the receptor in the vibrioferrin-mediated iron uptake system of V. parahaemolyticus. At the same time, FURTA methodology originally established for E. coli has proven to be an effective approach to isolate the iron-regulated genes in V. parahaemolyticus. DNA probes internal to the isolated Fur box-containing gene fragments will be very useful in exploring the related neighboring genes by gene walking. Since siderophore biosynthesis genes as well as ferric siderophore transport genes are frequently organized in an operon structure or are closely linked, further investigation into the regions upstream and downstream of the psuA-pvuA operon may disclose the genes necessary for vibrioferrin biosynthesis and uptake of ferric vibrioferrin into the cytoplasm.