Logo of jbacterPermissionsJournals.ASM.orgJournalJB ArticleJournal InfoAuthorsReviewers
J Bacteriol. 2008 Mar; 190(5): 1638–1648.
Published online 2007 Dec 21. doi:  10.1128/JB.01484-07
PMCID: PMC2258666

A Common Virulence Plasmid in Biotype 2 Vibrio vulnificus and Its Dissemination Aided by a Conjugal Plasmid


Strains of Vibrio vulnificus, a marine bacterial species pathogenic for humans and eels, are divided into three biotypes, and those virulent for eels are classified as biotype 2. All biotype 2 strains possess one or more plasmids, which have been shown to harbor the biotype 2-specific DNA sequences. In this study we determined the DNA sequences of three biotype 2 plasmids: pR99 (68.4 kbp) in strain CECT4999 and pC4602-1 (56.6 kb) and pC4602-2 (66.9 kb) in strain CECT4602. Plasmid pC4602-2 showed 92% sequence identity with pR99. Curing of pR99 from strain CECT4999 resulted in loss of resistance to eel serum and virulence for eels but had no effect on the virulence for mice, an animal model, and resistance to human serum. Plasmids pC4602-2 and pR99 could be transferred to the plasmid-cured strain by conjugation in the presence of pC4602-1, which was self-transmissible, and acquisition of pC4602-2 restored the virulence of the cured strain for eels. Therefore, both pR99 and pC4602-2 were virulence plasmids for eels but not mice. A gene in pR99, which encoded a novel protein and had an equivalent in pC4602-2, was further shown to be essential, but not sufficient, for the resistance to eel serum and virulence for eels. There was evidence showing that pC4602-2 may form a cointegrate with pC4602-1. An investigation of six other biotype 2 strains for the presence of various plasmid markers revealed that they all harbored the virulence plasmid and four of them possessed the conjugal plasmid in addition.

Vibrio vulnificus is a gram-negative bacterial species that is ubiquitous in marine environments. This organism causes infections that are characterized by severe wound infection and primary septicemia with high mortality in humans, particularly those with underlying diseases such as liver cirrhosis and hemochromatosis (33, 40). It also causes systemic infections, called vibriosis, with high mortality in brackish-water-cultured eels (4, 16). This organism is considered an emerging pathogen, since cases in fish and human were first reported in 1976 and 1979, respectively (8, 34).

Strains of V. vulnificus are divided into three biotypes, biotype 1 (BT1), BT2, and BT3, based on their differences in phenotypic traits, serological type, and host range (6, 7, 42). All three biotypes can cause sporadic cases of human diseases and have been shown to be virulent for the mouse, an animal model; however, only the BT2 strains produce epizootic or outbreaks of vibriosis, mainly in eels (4, 34). The lesions on the body of a V. vulnificus-infected eel, similar to those caused by Vibrio anguillarum, include external skin ulcer, hemorrhagic fins, protrusion of the rectum, and hemorrhages and inflammation of the internal organs (4). The BT2 strains are further subdivided into several serovars (6, 16, 22), among which serovar E is not only the most virulent and prevalent but also the only one isolated from sporadic human infections associated with handling of contaminated fish (1).

The virulence mechanism of BT2 V. vulnificus strains in eels remains unclear, although some virulence factors have been proposed. The extracellular products of BT2 strains exhibit hydrolytic/toxic activities and lethality for the eel similar to those produced by the BT1 strains (3), suggesting that they may not be the sole virulence determinants. Nevertheless, the degree of virulence for eels is directly correlated with the ability to resist and grow in fresh eel serum (2).

To find out the genetic basis of eel virulence in BT2 V. vulnificus strains, the whole genome of one serovar E BT2 strain (tester strain) was compared with those of three BT1 clinical strains (driver strains) by suppression subtractive hybridization (SSH) (28). The distribution of the tester-specific DNA sequences thus identified was surveyed in a collection of more than 100 strains of the three biotypes by PCR. Three of these DNA sequences were present in all strains of BT2 but none of other biotypes. Further, these DNA sequences were all located in the plasmids, suggesting that the virulence for eels could be conferred by a plasmid. In fact, previous studies had shown that all BT2 strains harbored one or more plasmids of high molecular weight (6, 22, 30).

The purpose of this study was to investigate the role of plasmids in the virulence of BT2 V. vulnificus for eels. Toward this end, we determined and analyzed the nucleotide sequences of three plasmids in two serovar E BT2 strains, CECT4602 (containing two plasmids) and CECT4999 (containing one plasmid). We then isolated a strain that was cured of the plasmid from CECT4999 and examined its virulence for eels and mice. We found that these two strains shared a highly homologous plasmid that was essential for both the resistance to bactericidal activity of eel serum and virulence for eels but not for growth in human serum and virulence for mice. A gene in the virulence plasmid encoding a novel protein was further shown to be essential for virulence in the eel. We also demonstrated that the virulence plasmids of CECT4602 and CECT4999 could be transferred to the plasmid-cured strain by conjugation in the presence of the other plasmid of CECT4602, which was shown to be self-transmissible.


Bacterial strains, plasmids, and oligonucleotides.

The various V. vulnificus strains, plasmids, and primers used in this study are listed in Table Table11 .

V. vulnificus strains, plasmids, and primers used in this study

Determination and analyses of plasmid nucleotide sequences.

The plasmid DNA was extracted from bacterial cells harvested from a 100-ml overnight culture in Luria-Bertani (LB) medium (BD Biosciences) with 6 ml of TENS solution (0.09 N NaOH and 0.45% sodium dodecyl sulfate in Tris-EDTA buffer) followed by 3 ml of sodium acetate (3 M, pH 5.2). This mixture was centrifuged, and the plasmid DNA in the supernatant was extracted with equal volume of phenol-chloroform (1:1) several times, precipitated with ethanol, dissolved in 10 ml of Tris-EDTA buffer, and purified by CsCl gradient centrifugation (37).

The nucleotide sequences of the plasmids were determined as described previously (12). Briefly, the DNA sequences of clones in a shotgun library of the plasmid with over 10-fold coverage were determined with an autosequencer (Applied Biosystems Prism 3700) and then jointly assembled using Phred/Phrap/Consed software provided by University of Washington, Seattle (20, 21, 25). The sequence gaps were closed by editing the end sequences of each contig and primer walking on the linking clones. The assembled plasmid DNA sequence was confirmed by restriction mapping with a variety of enzymes.

The open reading frames (ORF) were predicted with two programs, Glimmer (36) and GeneMark (10). The predicted ORFs were then submitted to AutoFACT (27), a tool that combines the BLAST reports from UniRef90, UniRef100, NCBI's nrdb, COG, KEGG, PFAM, and SMART, for gene annotation. The graphic view of plasmid DNA sequence comparison was generated with the Artemis comparison tool (11) according to the results obtained from bl2seq (41). The percentage of DNA sequence identity between the plasmids was determined by ClustalW (13) via sequence alignment. For a more accurate result, the DNA sequences were rearranged with traA and rtxE as the initial sequences for the pC4602-1/pYJ016 and pC4602-2/pR99 alignments, respectively.

Isolation of deletion mutants and a plasmid-cured strain from CECT4999.

A mutant, CT188, with a 225-bp in-frame deletion between the two HincII sites in ORFvep07 of pR99 and another mutant, CT095, with a chloramphenicol resistance (Cmr) cassette (44) inserted between these two HincII sites in ORFvep07 were isolated from CECT4999 by allelic exchange as described previously (38).

To facilitate the isolation of a cured strain from CECT4999, a 255-bp in-frame deletion was introduced into mazF of CT095 by allelic exchange. To cure the plasmid, this CECT4999 derivative was cultured in LB broth containing 3 μg/ml of acridine orange (Sigma-Aldrich) at 39°C with agitation overnight and then plated on the LB plates. Among the 4,900 colonies tested, only one was Cms. This strain (designated CT218) was further confirmed to be plasmid free by plasmid extraction, restriction pattern analysis, and Southern hybridization (data not shown).

Cloning of ORFvep07.

The entire ORFvep07 together with its putative promoter was amplified by PCR with primers C51F and C51R and then cloned into pIT009, a derivative of the broad-host-range plasmid pJRD215 (17), at the XbaI site. The resultant plasmid, pCT166, was transformed into Escherichia coli S17-1λpir (38), from which it was transferred into the various V. vulnificus strains by conjugation.

Serum resistance assay and growth in whole blood.

Eel blood and serum were prepared from European elvers (Anguilla anguilla) of 10 to 20 g as described previously (2). Human blood and serum were purchased from Sigma-Aldrich. To each sample, an equal volume of a bacterial suspension of 103 to 4 CFU/ml in phosphate-buffered saline was added, and the mixture was incubated at 28°C (for eel samples) or 37°C (for human samples) for 4 h. The viable counts were then determined on 1% NaCl-containing tryptic soy agar plates at 0 and 4 h after incubation. The significance of differences was analyzed by analysis of variance with the GraphPad Prism program.

Virulence assays.

Bacterial virulence in eels was assayed as described previously (5). Briefly, nonimmunized European elvers with an average weight of 10 g were injected intraperitoneally with 0.1 ml of 10-fold serially diluted bacterial suspension in phosphate-buffered saline containing 1% NaCl, pH 7.4. Three doses were used for each strain, and six eels were injected per dose. The eels were maintained in aquaria (1% salinity) with aeration at 25°C. The mortality was recorded daily for 7 days after challenge. The bacteria recovered from the internal organs of dead eels were confirmed serologically.

Bacterial virulence in mice was determined in C3H/HeN mice (6 to 8 weeks old, purchased from the Animal Center of College of Medicine, National Cheng-Kung University) infected by intraperitoneal injection (26). Four 10-fold serially diluted doses were used for each strain, and five mice were challenged per dose.

The 50% of lethal dose (LD50) of each strain was calculated from the number of confirmed deaths by a method described previously (35).


Two hundred microliters of an overnight culture of the donor strain and 20 μl of an overnight culture of the recipient strain were mixed in 5 ml of conjugation buffer (1% NaCl and 10 mM MgSO4) and filtered through a membrane with 0.45-μm pores (Millipore). The membrane was left on the LB plate for 4 h at 30°C. The bacteria were then collected from the membrane with 5 ml of LB broth, incubated at 37°C with agitation for 1 h, and plated on an LB plate containing 2% 5-bromo-4-chloro-3-indolyl-β-galactopyranoside (X-Gal) (MDBio) or appropriate antibiotics such as Cm (10 μg/ml) and streptomycin (Sm) (100 μg/ml).

PCR, RT-PCR, Southern hybridization, and colony hybridization.

PCR was performed as described previously (28). rtxA transcripts were detected by reverse transcription-PCR (RT-PCR). Briefly, total RNA was extracted from a 4-h bacterial culture with Tri reagent (Molecular Research Center), treated with RQI DNase (Promega), and reverse transcribed with Moloney murine leukemia virus reverse transcriptase (Promega) to prepare cDNA. PCR was then performed on the cDNA with primers P6 and P7 derived from rtxA.

In Southern hybridization, the uncut and restricted plasmid DNAs were separated on a 0.8% agarose gel, transferred onto a Hybond-N+ nylon membrane (Amersham Biosciences), and then hybridized with fluorescein-labeled PCR products prepared with the Gene Image random prime labeling kit (GE Healthcare Life Sciences). Primer pairs VF51/VR51, AD1-R/D1, Maz6/Maz7, JB02F/L1, and K1/JB02R (Table (Table1)1) were used to prepare the probes for seq51, traH, mazF, ID1, and ID2, respectively. The signals were visualized with the Gene Image CDP-Star detection module (GE Healthcare Life Sciences). Colony hybridization (37) was used to detect clones containing BT2 plasmids with the fluorescein-labeled probes prepared as described above.


Features of plasmid DNA sequences.

Plasmids pC4602-1 and pC4602-2 of strain CECT4602 and plasmid pR99 of strain CECT4999 were 56,628 bp, 66,946 bp, and 68,446 bp in size, respectively, and their G+C contents were 45.2%, 43.3%, and 43.8%, respectively. The numbers of predicted ORFs for pC4602-1, pC4602-2, and pR99 were 69, 67, and 71. The main features of the plasmids are shown in Fig. Fig.11 and and2.2. According to the annotation, the ORFs in these plasmids could be grouped into five categories: those related to conjugal transfer, those related to RTX toxin, those related to transposition, those with other functions, and those with unknown functions (Fig. (Fig.11 and and2;2; see Tables S1 to S3 in the supplemental material).

FIG. 1.
Main features of the plasmids pC4602-1 (A) and pC4602-2 (B) in V. vulnificus strain CECT4602. Some of the predicted ORFs are indicated with arrows. The ORFs associated with production of an RTX toxin, conjugative transfer of plasmids, and other functions ...
FIG. 2.
Main features of the plasmid, pR99, in V. vulnificus strain CECT4999. Some of the predicted ORFs are indicated with arrows. The ORFs related to the production of an RTX toxin and those with other functions are labeled in gray and white, respectively. ...

In pC4602-1 about half of the ORFs (32/69) had no known function, among which 20 were annotated as conserved hypothetical. Among the genes with known functions, a cluster of 15 genes (traB, traC, traD, traE, traF, traG, traH, traI, traK, traL, traN, traU, traW, trbB, and trbC) involved in conjugal transfer was found in a 27.2-kb region (nucleotides [nt] 20993 to 48163). This cluster shared high homology with those of F plasmid (23) and pYJ016, a conjugal plasmid of 48.5 kb in the BT1 V. vulnificus strain YJ016 (12), in DNA sequence and gene organization (see Fig. S1A in the supplemental material). The overall identity between pC4602-1 and pYJ016 was 81%, and the regions with a DNA sequence identity of over 84% contained mainly the transfer genes (see Fig. S1B in the supplemental material). An operon homologous to mazEF, a member of the toxin-and-antitoxin (TA) modules involved in maintenance of plasmids by postsegregational killing of the host (32), was also found in this plasmid. seq10, a BT2-specific DNA sequence identified previously by SSH (28), was found spanning two ORFs (ORFvpa21 and ORFvpa22), both of which were annotated as transposase (Fig. (Fig.1A;1A; see Table S1in the supplemental material).

Half of the ORFs (34/67) in pC4602-2 were annotated as hypothetical, and 21 of them were conserved. A notable feature of this plasmid was a cluster of RTX (repeats-in-toxin) genes homologous to those in the chromosomes of Vibrio cholerae and V. vulnificus, including those encoding RTX toxin (rtxA), toxin transporter (rtxB, rtxD, and rtxE), and an enzyme for toxin activation (rtxC) (9, 12, 45). A comparison of the rtx gene cluster DNA sequence of pC4602-2 with those of V. cholerae and BT1 V. vulnificus revealed dissimilar regions that were all located in rtxA. As shown in Fig. Fig.3,3, the rtxA gene of V. cholerae (rtxAVc) contains an actin-cross-linking domain (ACD) (39) that is not present in BT1 V. vulnificus rtxA (rtxAVvbt1). On the other hand, rtxAVvbt1 contains two regions, U1 and U2, that are absent in rtxAVc, with U1 being at the same location as the ACD in rtxAVc. Surprisingly, the rtxA gene of pC4602-2 (rtxAVvbt2) showed a mosaic structure composed of DNA fragments from rtxAVc and rtxAVvbt2. In rtxAVvbt2 the region corresponding to U1 of rtxVvbt1 was replaced by the entire ACD of rtxAVc, and only the 5′ portion (1,078 bp) of U2 was retained in rtxAVvbt2. Moreover, a 1,872-bp DNA sequence immediately downstream of the ACD was replaced by another copy (identity, 95%) of this 5′ portion of U2.

FIG. 3.
Comparison of the rtxA nucleotide sequences in V. cholerae strain N16961 (rtxAVc), V. vulnificus BT1 strain YJ016 (rtxAVvbt1), and pC4602-2 (rtxAVvbt2). The comparison was done with MUSCLE, a tool for multiple-sequence alignment (19). DNA sequences shared ...

seq25 and seq51, two BT2-specific sequences previously identified by SSH in our laboratory (28), were found to be close to and overlapping with, respectively, ORFvpb22, which was annotated as RTX toxin and related Ca2+-binding protein (Fig. (Fig.1B;1B; see Table S2 in the supplemental material).

Forty-five percent (32/71) of the ORFs in pR99 were annotated as hypothetical, among which 21 were conserved. This plasmid showed 92% identity with pC4602-2. Like pC4602-2, pR99 carried the rtx gene cluster and the ORF (ORFvep07) that contained part of seq51 (Fig. (Fig.2),2), both of which were 100% identical to those of pC4602-2. However, the mazEF operon, which was not present in pC4602-2, was found in pR99.

We further detected a copy of the rtx gene cluster exhibiting a restriction pattern indistinguishable from that of pR99 (plasmid-safe DNase-treated total DNA of CECT4999), but not the BT1 strain YJ016, in the plasmid-free derivative of CECT4999, CT218 (Fig. (Fig.4A).4A). By using RT-PCR, we showed that both CECT4999 and CT218 expressed rtxA when they were cultured in rich medium (Fig. (Fig.4B).4B). PCR products showing restriction patterns identical to those of products amplified from CECT4999 or CT218 were also obtained from strain CECT4602 with primer pairs P1/P2 and P3/P4 (Fig. (Fig.4A).4A). However, an additional restriction fragment of 1.7 kb, which could be derived from the chromosomal rtxA gene, was observed in the PCR product amplified from CECT4602 with primers P5 and P6 (Fig. (Fig.4A4A).

FIG. 4.
Analyses of the rtx gene clusters in the chromosomes of strains CECT4999 and CECT4602. (A) Restriction patterns of PCR products amplified from the rtx gene clusters. The DNA fragments amplified by PCR with the indicated primer pairs were digested with ...

Virulence of a CECT4999 derivative that had been cured of plasmid.

To determine whether the plasmid was involved in virulence for eels, we isolated a plasmid-free derivative, CT218, from strain CECT4999. The growth of CT218 in LB medium was similar to that of the parent strain, CECT4999 (data not shown). The LD50 of this cured strain in eels was at least 4 logs higher than that of CECT4999 and close to that of a BT1 strain, YJ016 (Table (Table2).2). However, the virulence of CT218 in mice was comparable to that of CECT4999 or YJ016 (Table (Table2).2). The resistance of CT218 to eel serum and growth in whole blood were also greatly impaired compared to those of the parent strain (Table (Table22).

Association of BT2 plasmids with bacterial virulence in eels and mice and growth in eel and human sera and whole blood

Involvement of ORFvep07 in pR99 with virulence for eels.

ORFvep07, which contained a BT2-specific sequence, seq51, and shared low sequence homology with a protein annotated as RTX toxin and related Ca2+-binding protein, was tested for its involvement in virulence for eels. The Δvep07 deletion mutant, CT188, was isolated by allelic exchange and confirmed by Southern hybridization (see Fig. S2 in the supplemental material) and nucleotide sequence determination. The growth of this mutant in LB medium was not affected (data not shown). However, its survival in eel whole blood and serum as well as virulence in the eel was greatly reduced to a level comparable to that of plasmid-cured strain CT218 (Table (Table3).3). Reintroduction of ORFvep07 carried by a broad-host-range plasmid into this mutant restored these abilities (Table (Table3).3). However, introduction of ORFvep07 alone enhanced, but did not fully restore, the survival of CT218 in eel blood and its virulence for eels and had little effect on those of the BT1 strain YJ016 (Table (Table33).

Association of ORFvep07 with bacterial virulence in eels and growth in eel serum and whole blood

Formation of a cointegrate between pC4602-1 and pC4602-2.

Two DNA sequences, ID1 (nt 46873 to 53550) and ID2 (nt 56129 to 2511), separated by a region of about 2.6 kb in pC4602-1, were found to be identical to two adjacent DNA sequences (nt 59769 to 66446 and nt 66447 to 2511, respectively) in pC4602-2 (Fig. 1A and B). Thus, these two plasmids might form two types of cointegrate via homologous or site-specific recombination between the common regions, ID1 or ID2, in these plasmids (Fig. (Fig.5A).5A). However, no band with the predicted size of the cointegrate was observed either directly in agarose gels (Fig. (Fig.6A,6A, lane 3) or after Southern hybridization with probes derived from the plasmids (data not shown). Nevertheless, when we performed PCR with two pairs of primers, K1/99P and M2/179R, derived from the unique DNA sequences in each plasmid, we did obtain PCR products of the predicted sizes from strain CECT4602 (Fig. (Fig.5B,5B, lane 5). The DNA sequences of these PCR products were further shown to be as predicted by nucleotide sequence determination or restriction pattern analysis (data not shown). These results demonstrated that the cointegrates coexisted with pC4602-1 and pC4602-2, although in an amount too small to be detected unless PCR was used.

FIG. 5.
Formation of cointegrates of plasmids pC4602-1 and pC4602-2. (A) Model of cointegrates formation (solid arrows) and resolution (broken arrows) via recombination between the identical regions, ID1 (white bar) or ID2 (black bar), of pC4602-1 and pC4602-2. ...
FIG. 6.
Profiles of uncut (A) and BglII-restricted (B) plasmids of various transconjugants of strain CECT4999. Lanes 1 to 5, plasmids extracted from CT225, CT226, CECT4602, CT223, and CT227, respectively. Lane M1, uncut plasmids from Pantoea stewartii SW2 (15 ...

Transfer of pC4602-1 and pC4602-2 from strain CECT4602 to B1 and B2 strains by conjugation.

The colonies of CT218 and its transconjugants could be easily distinguished from those of strain CECT4602 by their colors on a plate containing X-Gal: CECT4602 was blue, while CT218 was white. After conjugation, the bacteria were plated on X-Gal-containing LB agar and the white colonies were tested for the presence of plasmids by colony hybridization with the probes derived from pC4602-1 and pC4602-2, respectively. It was found that both pC4602-1 and pC4602-2 could be transferred to CT218 at different frequencies by conjugation (Table (Table4).4). Among the 229 colonies tested, 64 hybridized only with the probe derived from pC4602-1, 3 hybridized only with that derived from pC4602-2, and 9 hybridized with both probes (Table (Table44).

Frequencies of transfer of pC4602-1 and pC4602-2 from strain CECT4602 to CT218 and biotype 1 strains by conjugation

All the transconjugants identified were tested for the activity of ornithine decarboxylase (ODC) to ensure that they were derived from CT218, an ODC-positive strain, but not CECT4602, an ODC-negative strain (data not shown). They were further subjected to plasmid profile analysis (some of the results are shown in Fig. Fig.6),6), and the presence of either or both plasmids was confirmed. Similar to strain CECT4602, the transconjugants containing both plasmids gave PCR products (Fig. (Fig.5B)5B) with the primer pairs K1/99P and M2/179R, which amplified DNA fragments only from the cointegrates as mentioned above.

Two of the nine transconjugants that contained two plasmids exhibited a plasmid profile different from those of the other seven transconjugants and strain CECT4602 (the results for one of them are shown in Fig. Fig.6).6). The plasmid sizes and two of the DNA fragments produced by BglII digestion had apparently been changed in these two transconjugants (Fig. 6A and B).

The transfer of plasmids from strain CECT4602 to two spontaneous Smr BT1 mutants, LF044 and CT162, was also investigated. The colonies grown on Sm-containing LB plates after mating were tested for the presence of plasmids by colony hybridization. None of the 1,000 randomly picked colonies hybridized with the probe derived from pC4602-1 or pC4602-2 for either LF044 or CT162 (Table (Table44).

Association of pC4602-1 and pC4602-2 with virulence for eels.

Four transconjugants of CT218 containing either or both pC4602-1 and pC4602-2 were further tested for their virulence in eels and ability to grow in eel serum. As shown in Table Table2,2, CT225 and CT226, containing both pC4602-1 and pC4602-2, as well as CT227, which contained pC4602-2 only, were virulent for eels, exhibiting an LD50 of around 1 × 104 CFU/fish. The virulence of bacteria recovered from the internal organs of killed eels was further shown to be enhanced to the wild-type level (LD50 of around 1 × 102 CFU/fish) for all of these strains (Table (Table2).2). In contrast, the transconjugant that contained pC4602-1 only was not virulent for eels (Table (Table22).

Transfer of pR99 in the presence of pC4602-1.

We further tested the transfer of virulence plasmid pR99 in the presence or absence of pC4602-1 to confirm that the virulence plasmid by itself is not transmissible but that its transfer can be achieved with the aid of a conjugal plasmid. Plasmid pC4602-1 was transferred from CECT4602 to CT095, a CECT4999 derivative containing a Cmr cassette in pR99, by conjugation, and the plasmid compositions of a transconjugant, CT295, were confirmed by PCR with specific primer pairs. The frequencies of transfer of pR99::Cmr from CT095 and CT295 to CT237, a spontaneous Smr mutant of CT218, by conjugation were then determined. Colonies resistant to both Cm and Sm were obtained after conjugation at frequencies of 1.1 × 10−1 and < 10−5, respectively, for CT295 and CT095. The frequencies of spontaneous mutation that gave rise to colonies resistant to both Cm and Sm for CT095, CT295, and CT237 were determined in parallel as negative controls, and they were all found to be <10−5.

Distribution of seq51, traH, mazF, ID1 and ID2 in the plasmids of various B2 strains.

To investigate the prevalence of virulence and conjugal plasmids of strains CECT4602 and CECT4999 in BT2 V. vulnificus, six other BT2 strains from various sources were examined by Southern hybridization with a number of probes derived from these plasmids. Among them, two contained one plasmid and the other four contained two plasmids (Fig. (Fig.7A).7A). seq51, a virulence plasmid marker, and mazF, the toxin module associated with plasmid stability, were detected in all of the tested strains; seq51 was located in the large plasmids, while mazF was located in either the large or small plasmid in those with two plasmids (Fig. (Fig.7B).7B). The traH gene, a conjugal plasmid marker, was detected only in the small plasmid of strains with two plasmids (Fig. (Fig.7B).7B). ID1 and ID2, the common DNA sequences in pC4602-1, pC4602-2, and pR99, were detected in every plasmid of the tested BT2 strains (Fig. (Fig.7B).7B). The presence of a gap between ID1 and ID2 was further investigated by PCR with a pair of primers derived from ID1 and ID2, respectively. As shown in Fig. Fig.7C,7C, PCR products corresponding to those amplified from pC4602-1 and pC4602-2, respectively, were obtained from strains containing two plasmids, except for strain CECT4865. All the strains with only one plasmid and strain CECT4865 gave a PCR product corresponding to that amplified from pC4602-1.

FIG. 7.
Distribution of genes and regions involved in virulence for eels, plasmid transfer, and plasmid stability in the plasmids of various B2 strains. (A) Profiles of uncut plasmids. Lanes 1 to 7, plasmids from strains CECT4602, CECT4865, CECT4917, CECT897, ...


BT2 of V. vulnificus is unique among the three biotypes of this species in its ability to infect and cause septicemia in eels. Two previous findings, the presence of large plasmids in all BT2 strains (6, 16, 30) and the identification of BT2-specific plasmid DNA sequences found by SSH (28), suggested that the plasmids might be involved in virulence for eels. To confirm this, we selected two highly virulent serovar E strains, CECT4999 and CECT4602, that exhibited different plasmid profiles and determined the nucleotide sequences of their plasmids, followed by isolation of a plasmid-cured strain and testing of its virulence.

The completed plasmid nucleotide sequences revealed that both strains possessed a common plasmid (nucleotide sequence identity, 92%) of about 68 kb in size. Like many other virulence plasmids (14, 18, 24), the common plasmid of BT2 strains appeared to be a mosaic composed of regions from multiple sources that might be formed via horizontal gene transfer between the aquatic bacteria.

Based on the plasmid DNA sequences, we were able to identify and disrupt mazEF, a TA module, on the plasmid to facilitate the isolation of a plasmid-free derivative from strain CECT4999. Curing of pR99 resulted in loss of virulence for eels, and introduction of pC4602-2 into the cured strain restored the virulence (Table (Table2),2), indicating that both pR99 and pC4602-2 were virulence plasmids.

The virulence conferred by pR99 or pC4602-2 is different from that conferred by pJM1, the common virulence plasmid of V. anguillarum, another causative agent of vibriosis, which has been shown to encode the ability to take up iron for growth (18). The loss of virulence for eels of the plasmid-cured strain and CT188, a mutant with a deletion in ORFvep07 of pR99, was associated with diminished resistance to killing by eel serum and, consequently, an inability to grow in the eel blood. However, plasmid loss did not affect either the virulence in mice or growth in human blood, suggesting that different sets of virulence genes might be used in different hosts.

The virulence plasmid may be required for resistance to a bactericidal factor functioning in eel, but not human, serum. The identification of this bactericidal factor would be crucial for better understanding of the pathogenesis of vibriosis in eels. Previous studies on serovar E BT2 strains have suggested that different surface antigens may be involved in resistance to sera from different hosts: lipopolysaccharide for resistance to eel serum (2) and capsule for resistance to human serum (5, 46). We had checked the structure of lipopolysaccharide in CT218 and found that it was not altered (43), implying that the virulence plasmid might encode another factor(s) important for the resistance to eel serum.

The rtx genes encoding the production, export, and activation of an RTX family member (9, 45) were the only known virulence genes in the virulence plasmids of BT2 strains. An rtx gene cluster has also been found in the small chromosome of BT1 V. vulnificus (12) and has been shown to be associated with virulence in the mouse (29, 31). The rtxB, rtxC, rtxD, and rtxE genes and the promoter region of this gene cluster in pC4602-2 or pR99 were highly homologous to those of V. cholerae and BT1 V. vulnificus (80% and 94% identical, respectively). To our surprise, the rtxA gene in pC4602-2 or pR99 was a mosaic composed of DNA fragments from those of V. cholerae and BT1 V. vulnificus. A copy of the rtx gene cluster showing a restriction pattern identical to that of the virulence plasmid was also detected in the chromosome of CECT4999, suggesting a redundancy of this gene cluster in this strain. Such a redundancy was also observed in strain CECT4602, with only a minor difference in rtxA. Using RT-PCR, we demonstrated that the chromosomal rtxA gene of the plasmid-cured strain, CT218, was transcribed when the organism was cultured in rich medium (Fig. (Fig.4B).4B). Further, CT218, but not CT218 ΔrtxA, could lyse HEp-2 cells (our unpublished data), indicating that a functional RTX could be produced from the chromosome. While the role of RTX toxin in virulence for eels needs to be further defined, we feel that the rtx gene cluster in the virulence plasmid might not be indispensable, since there is another copy in the chromosome that is functional.

On the other hand, ORFvep07 of pR99 appeared to be essential for virulence in eels, because disruption of this gene in strain CECT4999 resulted in greatly reduced virulence in eels. ORFvep07 and its equivalent, ORFvpa22, in pC4602-2 shared low nucleotide sequence identity (21% in a 279-amino-acid region of this 388-amino-acid protein) with a gene annotated as the RTX toxin and related Ca2+-binding protein, whose function requires further experimental validation. These two genes may encode an outer membrane or extracellular protein, as was suggested from the result of computational analysis with CELLO, a subcellular localization predictor (47). However, ORFvep07 by itself was not sufficient for causing disease in eels, as the BT1 strain or plasmid-cured BT2 strain remained avirulent or became weakly virulent at most in the presence of this ORF alone. This suggested that other genes in the virulence plasmid might be required for the expression of full virulence in eels. The mechanism of resistance of BT2 strains to killing by eel blood and serum, in which ORFvep07 plays a role, is currently under investigation in our laboratories.

The third plasmid, pC4602-1, was shown to be an F-like conjugal plasmid that shared high sequence homology with pYJ016 of the BT1 strain YJ016. Interestingly, in a mating between strains CECT4602 and CT218, we obtained three groups of transconjugants, which contained pC4602-1 alone, both pC4602-1 and pC4602-2, and pC4602-2 alone, respectively. The frequency of transfer of pC4602-2, which did not carry any known transfer gene, was about one-sixth of that of transfer of pC4602-1. Transfer of pC4602-2 may occur in a form of cointegrate with pC4602-1 that could be generated by a recombination between the common regions, ID1 or ID2, in these plasmids. The existence of the cointegrate was suggested by the obtaining of predicted PCR products with two primer pairs that were both composed of primers hybridizing specifically to pC4602-1 and pC4602-2, respectively (Fig. (Fig.5B).5B). Once entering the recipient, the cointegrate may rapidly resolve into two plasmids, making detection of the cointegrate difficult. The fact that two of the nine transconjugants with both plasmids showed altered plasmid sizes and restriction profiles (Fig. (Fig.6)6) as well as PCR products amplified with primer pair M2/179R (Fig. (Fig.5B),5B), which could result from aberrant resolution, is consistent with this suspicion.

The three transconjugants with only the virulence plasmid may be derived from those containing both plasmids by losing the conjugal plasmid or, alternatively, may be a result of incomplete transfer of the cointegrate when the mating process was interrupted. The former could be readily ruled out because we found, by colony hybridization, that the conjugal plasmid was stably maintained in the transconjugants containing either the conjugal plasmid alone or both plasmids after overnight growth in rich medium (data not shown). Nevertheless, the possibility that the virulence plasmid may be mobilizable and that its transfer could be achieved once the mating pair is formed with the aid of a conjugal plasmid cannot be excluded. Although it demands experimental validation, a putative transfer origin (oriT), CGTGGTGTGTG, of F-like plasmids enabling transfer of the plasmid in the presence of a conjugal plasmid with the same oriT, has been found residing the ID1 regions of pC4602-1, pC4602-2, and pR99 between traI and pilT.

When we compared the nucleotide sequence of pR99 with those of pC4602-1 and pC4602-2, we found a common DNA sequence (100% identical), which contained ID1, ID2, the 2.6-kb sequence between ID1 and ID2, and four more ORFs, including mazEF, in pR99 and pC4602-1. The plasmids of two other single-plasmid-containing serovar E strains, CECT4917 and CECT4866, were also shown to contain ID1, ID2, and the DNA sequence between these two regions. These findings provided additional clues for the transmission of a virulence plasmid by conjugation with the aid of a conjugal plasmid. The single-plasmid-containing strains could have originated from a mating between a two-plasmid-containing donor and a plasmid-free recipient, as had been demonstrated in the conjugation between CECT4602 and CT218. Indeed, by introducing pC4602-1 into CECT4999, we were able to transfer pR99 into the plasmid-cured strain, even though pR99 per se was not transferable.

The frequency of transferring pC4602-1 from CECT4602 to CT218 by conjugation was close to 30%, but that of transferring it to two BT1 strains, LF044 and CT162, was less than 10−3, if any. This finding may partly explain why only a subset of V. vulnificus strains can cause disease in eels: the dissemination of the virulence plasmid is an uncommon process because the conjugation efficiency of the conjugal plasmid enabling its transfer is very low. The difficulty of transferring pC4602-1 to BT1 strains by conjugation may be due to defects in forming a mating pair between the two biotypes and/or replication of this plasmid in the BT1 strains or to rapid degradation of this plasmid by a restriction enzyme(s) in BT1 strains.

The mazEF operon, encoding a member of TA modules that contribute to the stability of plasmids (32), was found in pC4602-1, pR99, and the plasmids of all six other BT2 strains checked. This suggests that the BT2 strains might retain the plasmids via this system. In strain CECT4602, the mazEF operon was located in the conjugal plasmid but not the virulence plasmid. It is possible that mazEF of pC4602-1 stabilizes both plasmids in a form of cointegrates.

All of the tested BT2 strains, irrespective of their serotypes (E or A), phenotypic characteristics, sources (environments, fish, or humans), geographic origins (Europe, Asia, or Australia), or isolation dates (1979 to 1999), harbored a plasmid similar to the virulence plasmids characterized in this study. The majority harbored an additional plasmid that carried the transfer genes and might be needed for the transmission of virulence plasmid. Intriguingly, even though the size of the putative virulence plasmid differed among the strains (Fig. (Fig.7A),7A), the distances between ID1 and ID2 that were detected in all the plasmids appeared to be highly restricted: either none or 2.6 kb.

Collectively, our data demonstrated that the ability to colonize and develop disease in fish has emerged in V. vulnificus by acquisition of a virulence plasmid. The original virulence plasmid was probably a mosaic of genes from different aquatic bacterial species, including fish pathogens. This virulence plasmid might have been acquired by different ancestors (polyphyletic origin) to give rise to BT2 of the species, as phenotypically and serologically this biotype is composed of heterogeneous strains (1, 16, 28). Once it was acquired, the virulence plasmid could be stably maintained, via a mazEF TA module, in the BT2 strains and be disseminated with the aid of a conjugal plasmid. The virulence and conjugal plasmids can form a cointegrate, and the recombination events occurring during formation and resolution of the cointegrates could contribute to the diversification of the plasmids.

Supplementary Material

[HTML Page - index.htslp]


This work was supported by grants NSC 93-3112-B-006-007, NSC 94-2320-B-006-066, and NSC 95-3112-B-010-008 from the National Science Council of Taiwan and AGL2005-04688 from the Ministry of Science and Education of Spain and by the NCKU Project of Promoting Academic Excellence & Developing World Class Research Centers, Taiwan. We thank I. T. Lai for construction of pIT009 during a summer research program funded by the National Science Council of Taiwan (NSC-2815-C-006-058-B).


Published ahead of print on 21 December 2007.

Supplemental material for this article may be found at http://jb.asm.org/.


1. Amaro, C., and E. G. Biosca. 1996. Vibrio vulnificus biotype 2, pathogenic for eels, is also an opportunistic pathogen for humans. Appl. Environ. Microbiol. 621454-1457. [PMC free article] [PubMed]
2. Amaro, C., B. Fouz, E. G. Biosca, E. Marco-Noales, and R. Collado. 1997. The lipopolysaccharide O side chain of Vibrio vulnificus serogroup E is a virulence determinant for eels. Infect. Immun. 652475-2479. [PMC free article] [PubMed]
3. Biosca, E. G., and C. Amaro. 1996. Toxic and enzymatic activities of Vibrio vulnificus biotype 2 with respect to host specificity. Appl. Environ. Microbiol. 622331-2337. [PMC free article] [PubMed]
4. Biosca, E. G., C. Amaro, C. Esteve, E. Alcaide, and E. Garay. 1991. First record of Vibrio vulnificus biotype 2 from diseased European eel, Anguilla anguilla L. J. Fish Dis. 14103-109.
5. Biosca, E. G., H. Llorens, E. Garay, and C. Amaro. 1993. Presence of a capsule in Vibrio vulnificus biotype 2 and its relationship to virulence for eels. Infect. Immun. 611611-1618. [PMC free article] [PubMed]
6. Biosca, E. G., J. D. Oliver, and C. Amaro. 1996. Phenotypic characterization of Vibrio vulnificus biotype 2, a lipopolysaccharide-based homogeneous O serogroup within Vibrio vulnificus. Appl. Environ. Microbiol. 62918-927. [PMC free article] [PubMed]
7. Bisharat, N., V. Agmon, R. Finkelstein, R. Raz, G. Ben-Dror, L. Lerner, S. Soboh, R. Colodner, D. N. Cameron, D. L. Wykstra, D. L. Swerdlow, and J. J. Farmer III. 1999. Clinical, epidemiological, and microbiological features of Vibrio vulnificus biogroup 3 causing outbreaks of wound infection and bacteraemia in Israel. Lancet 3541421-1424. [PubMed]
8. Blake, P. A., M. H. Merson, R. E. Weaver, D. G. Hollis, and P. C. Heublein. 1979. Disease caused by a marine Vibrio. Clinical characteristics and epidemiology. N. Engl. J. Med. 3001-5. [PubMed]
9. Boardman, B. K., and K. J. Fullner Satchell. 2004. Vibrio cholerae strains with mutations in an atypical type I secretion system accumulate RTX toxin intracellularly. J. Bacteriol. 1868137-8143. [PMC free article] [PubMed]
10. Borodovsky, M., and J. McIninch. 1993. Recognition of genes in DNA sequence with ambiguities. Biosystems 30161-171. [PubMed]
11. Carver, T. J., K. M. Rutherford, M. Berriman, M.-A. Rajandream, B. G. Barrell, and J. Parkhill. 2005. ACT: the Artemis comparison tool. Bioinformatics 213422-3423. [PubMed]
12. Chen, C. Y., K. M. Wu, Y. C. Chang, C. H. Chang, H. C. Tsai, T. L. Liao, Y. M. Liu, H. J. Chen, A. B. Shen, J. C. Li, T. L. Su, C. P. Shao, C. T. Lee, L. I. Hor, and S. F. Tsai. 2003. Comparative genome analysis of Vibrio vulnificus, a marine pathogen. Genome Res. 132577-2587. [PMC free article] [PubMed]
13. Chenna, R., H. Sugawara, T. Koike, R. Lopez, T. J. Gibson, D. G. Higgins, and J. D. Thompson. 2003. Multiple sequence alignment with the Clustal series of programs. Nucleic Acids Res. 313497-3500. [PMC free article] [PubMed]
14. Chu, C., and C. H. Chiu. 2006. Evolution of the virulence plasmids of non-typhoid Salmonella and its association with antimicrobial resistance. Microbes Infect. 81931-1936. [PubMed]
15. Coplin, D. L., R. G. Rowan, D. A. Chisholm, and R. E. Whitmoyer. 1981. Characterization of plasmids in Erwinia stewartii. Appl. Environ. Microbiol. 42599-604. [PMC free article] [PubMed]
16. Dalsgaard, I., L. Hoi, R. J. Siebeling, and A. Dalsgaard. 1999. Indole-positive Vibrio vulnificus isolated from disease outbreaks on a Danish eel farm. Dis. Aquat. Organ. 35187-194. [PubMed]
17. Davison, J., M. Heusterspreute, N. Chevalier, V. Ha-Thi, and F. Brunel. 1987. Vectors with restriction site banks. V. pJRD215, a wide-host-range cosmid vector with multiple cloning sites. Gene 51275-280. [PubMed]
18. Di Lorenzo, M., M. Stork, M. E. Tolmasky, L. A. Actis, D. Farrell, T. J. Welch, L. M. Crosa, A. M. Wertheimer, Q. Chen, P. Salinas, L. Waldbeser, and J. H. Crosa. 2003. Complete sequence of virulence plasmid pJM1 from the marine fish pathogen Vibrio anguillarum strain 775. J. Bacteriol. 1855822-5830. [PMC free article] [PubMed]
19. Edgar, R. C. 2004. MUSCLE: a multiple sequence alignment method with reduced time and space complexity. BMC Bioinformatics 5113-131. [PMC free article] [PubMed]
20. Ewing, B., and P. Green. 1998. Base-calling of automated sequencer traces using phred. II. Error probabilities. Genome Res. 8186-194. [PubMed]
21. Ewing, B., L. Hillier, M. C. Wendl, and P. Green. 1998. Base-calling of automated sequencer traces using phred. I. Accuracy assessment. Genome Res. 8175-185. [PubMed]
22. Fouz, B., F. J. Roig, and C. Amaro. 2007. Phenotypic and genotypic characterization of a new fish-virulent Vibrio vulnificus serovar that lacks potential to infect humans. Microbiology 1531926-1934. [PubMed]
23. Frost, L. S., K. Ippen-Ihler, and R. A. Skurray. 1994. Analysis of the sequence and gene products of the transfer region of the F sex factor. Microbiol. Rev. 58162-210. [PMC free article] [PubMed]
24. Golubov, A., H. Neubauer, C. Nolting, J. Heesemann, and A. Rakin. 2004. Structural organization of the pFra virulence-associated plasmid of rhamnose-positive Yersinia pestis. Infect. Immun. 725613-5621. [PMC free article] [PubMed]
25. Gordon, D., C. Abajian, and P. Green. 1998. Consed: a graphical tool for sequence finishing. Genome Res. 8195-202. [PubMed]
26. Hor, L. I., Y. K. Chang, C. C. Chang, H. Y. Lei, and J. T. Ou. 2000. Mechanism of high susceptibility of iron-overloaded mouse to Vibrio vulnificus infection. Microbiol. Immunol. 44871-878. [PubMed]
27. Koski, L., M. Gray, B. F. Lang, and G. Burger. 2005. AutoFACT: an automatic functional annotation and classification tool. BMC Bioinformatics 6151-161. [PMC free article] [PubMed]
28. Lee, C. T., C. Amaro, E. Sanjuan, and L. I. Hor. 2005. Identification of DNA sequences specific for Vibrio vulnificus biotype 2 strains by suppression subtractive hybridization. Appl. Environ. Microbiol. 715593-5597. [PMC free article] [PubMed]
29. Lee, J. H., M. W. Kim, B. S. Kim, S. M. Kim, B. C. Lee, T. S. Kim, and S. H. Choi. 2007. Identification and characterization of the Vibrio vulnificus rtxA essential for cytotoxicity in vitro and virulence in mice. J. Microbiol. 45146-152. [PubMed]
30. Lewin, A., B. Bert, A. Dalsgaard, B. Appel, and L. Hoi. 2000. A highly homologous 68 kbp plasmid found in Vibrio vulnificus strains virulent for eels. J. Basic Microbiol. 40377-384. [PubMed]
31. Liu, M., A. F. Alice, H. Naka, and J. H. Crosa. 2007. The HlyU protein is a positive regulator of rtxA1, a gene responsible for cytotoxicity and virulence in the human pathogen Vibrio vulnificus. Infect. Immun. 753282-3289. [PMC free article] [PubMed]
32. Moritz, E. M., and P. J. Hergenrother. 2007. Toxin-antitoxin systems are ubiquitous and plasmid-encoded in vancomycin-resistant enterococci. Proc. Natl. Acad. Sci. USA 104311-316. [PMC free article] [PubMed]
33. Muench, K. H. 1989. Hemochromatosis and infection: alcohol and iron, oysters and sepsis. Am. J. Med. 8740N-43N. [PubMed]
34. Muroga, K., Y. Jo, and M. Nishibuchi. 1976. Pathogenic Vibrio isolated from cultured eels. I. Characteristics and taxonomicstatus. Fish Pathol. 11141-145.
35. Reed, L. J., and H. Muench. 1938. A simple method of estimating fifty percent endpoints. Am. J. Hyg. 27493-497.
36. Salzberg, S. L., A. L. Delcher, S. Kasif, and O. White. 1998. Microbial gene identification using interpolated Markov models. Nucleic Acids Res. 26544-548. [PMC free article] [PubMed]
37. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
38. Shao, C. P., and L. I. Hor. 2000. Metalloprotease is not essential for Vibrio vulnificus virulence in mice. Infect. Immun. 683569-3573. [PMC free article] [PubMed]
39. Sheahan, K. L., C. L. Cordero, and K. J. Satchell. 2004. Identification of a domain within the multifunctional Vibrio cholerae RTX toxin that covalently cross-links actin. Proc. Natl. Acad. Sci. USA 1019798-9803. [PMC free article] [PubMed]
40. Tacket, C. O., F. Brenner, and P. A. Blake. 1984. Clinical features and an epidemiological study of Vibrio vulnificus infections. J. Infect. Dis. 149558-561. [PubMed]
41. Tatusova, T. A., and T. L. Madden. 1999. BLAST 2 S, a new tool for comparing protein and nucleotide sequences. FEMS Microbiol. Lett. 174247-250. [PubMed]
42. Tison, D. L., M. Nishibuchi, J. D. Greenwood, and R. J. Seidler. 1982. Vibrio vulnificus biogroup 2: new biogroup pathogenic for eels. Appl. Environ. Microbiol. 44640-646. [PMC free article] [PubMed]
43. Valiente, E., C. T. Lee, J. Lamas, L. I. Hor, and C. Amaro. 2007. Role of the virulence plasmid pR99 and the metalloprotease Vvp in resistance of Vibrio vulnificus serovar E to eel innate immunity. Fish Shellfish Immunol. 24134-141. [PubMed]
44. Wang, Y., and D. E. Taylor. 1990. Chloramphenicol resistance in Campylobacter coli: nucleotide sequence, expression, and cloning vector construction. Gene 9423-28. [PubMed]
45. Welch, R. A. 2001. RTX toxin structure and function: a story of numerous anomalies and few analogies in toxin biology. Curr. Top. Microbiol. Immunol. 25785-111. [PubMed]
46. Wright, A. C., L. M. Simpson, J. D. Oliver, and J. G. Morris, Jr. 1990. Phenotypic evaluation of acapsular transposon mutants of Vibrio vulnificus. Infect. Immun. 581769-1773. [PMC free article] [PubMed]
47. Yu, C. S., C. J. Lin, and J. K. Hwang. 2004. Predicting subcellular localization of proteins for Gram-negative bacteria by support vector machines based on n-peptide compositions. Protein Sci. 131402-1406. [PMC free article] [PubMed]

Articles from Journal of Bacteriology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...