OUP user menu

Occurrence and potential pathogenesis of Vibrio cholerae, Vibrio parahaemolyticus and Vibrio vulnificus on the South Coast of Sweden

Betty Collin, Ann-Sofi Rehnstam-Holm
DOI: http://dx.doi.org/10.1111/j.1574-6941.2011.01157.x 306-313 First published online: 1 November 2011


During the summer of 2006, several wound infections – of which three were fatal – caused by Vibrio cholerae were reported from patients who had been exposed to water from the Baltic Sea. Before these reports, we initiated a sampling project investigating the occurrence of potential human pathogenic V. cholerae, Vibrio vulnificus and Vibrio parahaemolyticus in The Sound between Sweden and Denmark. The Blue mussel (Mytilus edulis) was used as an indicator to follow the occurrence of vibrios over time. Molecular analyses showed high frequencies of the most potent human pathogenic Vibrio spp.; 53% of mussel samples were positive for V. cholerae (although none were positive for the cholera toxin gene), 63% for V. vulnificus and 79% for V. parahaemolyticus (of which 47% were tdh+ and/or trh+). Viable vibrios were also isolated from the mussel meat and screened for virulence by PCR. The mortality of eukaryotic cells when exposed to bacteria was tested in vivo, with results showing that the Vibrio strains, independent of species and origin, were harmful to the cells. Despite severe infections and several deaths, no report on potential human pathogenic vibrios in this area had been published before this study.

  • Vibrio cholerae
  • Vibrio vulnificus
  • Vibrio parahaemolyticus
  • ecology
  • microbiology
  • wound infections


Despite the vast majority being harmless to humans, some field strains of the widely spread aquatic bacteria Vibrio spp. are potential human pathogens. The three species in focus in the present study are Vibrio cholerae, Vibrio vulnificus and Vibrio parahaemolyticus. Vibrio cholerae serogroups O1 (classical and El Tor biotypes) and O139 are associated with epidemic and pandemic cholera and other serogroups are known to cause nonepidemic cholera (Banwell et al., 1970; Holmgren, 1981). Vibrio vulnificus may cause fatal wound infections, sepsis (Torres et al., 2002; Oliver, 2005a) and food-related infections (Hlady et al., 1993; Mead et al., 1999). Vibrio parahaemolyticus is one of the most common causes of seafood-borne bacterial gastroenteritis (Joseph et al., 1982; Janda et al., 1988; Honda & Iida, 1993) and can also rarely cause wound infections (Ellingsen et al., 2008; Tena et al., 2010). Little attention, however, has been paid to the ability of V. cholerae to cause severe wound infections (Johnston et al., 1983; Morris, 1990; Oliver, 2005a; Tsai et al., 2009). During the summer of 2006, such infections – including three fatal cases – were recorded among Swedish patients who had been exposed to Baltic Sea water (Shönning et al., 2008), with another fatal case earlier reported from Finland (Lukinmaa et al., 2006). Several cases of wound infections caused by V. vulnificus have also been described (Melhus et al., 1995; Dalsgaard et al., 1996; Ruppert et al., 2004), but as yet, no clinical case of V. parahaemolyticus has been reported from the Baltic Sea.

The few previous environmental studies of Vibrio spp. from northern temperate waters have focused on the total counts of each Vibrio spp., without discriminating between the potential pathogenic and nonpathogenic strains of which they are composed. Bauer (2006) analyzed the Blue mussel Mytilus edulis for human pathogenic Vibrio spp. at sites along the Norwegian coast using cultivation methods, i.e. enrichment and subsequent plate growth. The results from the identification of isolates showed only low occurrences of potentially virulent strains. Eiler (2006) used 16S rDNA analyses to investigate Vibrio populations in Swedish coastal waters and reported a high occurrence, but with a clear difference along the salinity gradient. However, the occurrence of virulence genes was not analyzed. Thus, despite several fatal infections caused by Vibrio spp. from the Baltic Sea, the present study is the first to focus on the actual occurrence of potentially virulent V. cholerae, V. vulnificus and V. parahaemolyticus in the transition waters between the Baltic Sea and the Kattegat (The Sound, in Swedish; Öresund, between Sweden and Denmark; Fig. 1). Mytilus edulis are known to efficiently ingest and accumulate bacteria from the surrounding water (Haamer, 1996; Hernroth et al., 2000) and were therefore used in this study to identify the occurrence of the Vibrio species. Strains were identified through biochemical analysis and molecular determination of species-specific genes, and by investigating the cytopathic effect on Chinese Hamster Ovary cells (CHO cells). Sequences of virulence genes of V. cholerae isolated from the mussel tissue were compared with those isolated from infected patients.


Map of the sampling sites in The Sound. *56°06′58″N 12°36′12″E **55°59′31″N 12°44′30″E.

Because temperature is known to regulate the growth and persistence of vibrios (Kaneko & Colwell, 1973; Motes et al., 1998; Bauer et al., 2006; Hernroth et al., 2010), sampling took place from June to September to incorporate the most relevant period, which also coincides with the time when people are most frequently exposed to seawater.

Materials and methods

Area description

The Sound (Fig. 1) between Sweden and Denmark is one of three transitional areas where marine water from the Kategatt (Northern Sea) mixes with brackish Baltic Sea water. During the sampling period of 2006, the highest temperatures were recorded in July (24.1 °C) and the lowest in September (16.6 °C). Salinity ranged between 9.1‰ and 17.4‰, largely due to frequent strong currents affecting the area.

Collection and preparation of samples

Mytilus edulis were collected on 19 separate occasions during the period from June 13 until September 25, from two sites: Domsten (56°06′58″N, 12°36′12″E), 11 km north of Helsingborg, and Råå (55°59′31″N, 12°44′30″E), 7 km south of Helsingborg, both in the northern part of The Sound (Fig. 1). Mussels were collected from near-shore rocks, often close to marine sediment and still minimally affected by the tide. On each visit, a minimum of 12 mussels were collected and kept cold until preparation, with the water temperature also measured (HI-98130 HANNA instruments Inc. 584 Park East Drive Woonsocket, RI 02895). After being measured in terms of length and weight, the mussels were scrubbed and rinsed in distilled water, before being opened with a sterilized shucking knife. The mussel meat, including the mussel liquid, was then collected in a sterilized blender and homogenized for 2 min at maximum speed.

PCR and strain isolation by cultivation

DNA was extracted from the homogenized mussel meat using the QIAamp DNA Stool Mini Kit (Qiagen, Valencia, CA) and the extracted DNA was subsequently tested for species-specific and virulence genes for the three Vibrio spp. by PCR (Table 1). The amplification of genes was performed according to the respective protocols (Table 1); 1 μL of each primer (2 μL of primer viuB) and 10 μL of extracted DNA were added to 25 μL Master mix (HotStarTaq Master Mix, QIAamp, Qiagen). H2O was then added to obtain a total volume of 50 μL. The PCR products were examined by agarose gel electrophoresis with ethidium bromide (80 V, 45 min), using DNA ladder GeneRuler 100 bp as a size standard, before being visualized under UV-light illumination. The PCR was subsequently repeated in order to detect any potentially low concentration of genes, but instead of adding extracted DNA as a template, 1 μL of the PCR product was used. The same procedure was performed with the negative control. This method was favored compared with enrichment cultivation in order to avoid the risk of alternating the ratio of different Vibrio spp. in the sample. To verify the accuracy of the chosen primers, selected PCR amplicons of each positive gene fragment were sequenced (Genomic Core Facility, The Sahlgrenska Academy at University of Gothenburg) and the results were compared using blast (http://www.ncbi.nlm.nih.gov). To isolate Vibrio spp. from the homogenate, 25 g of mussel homogenate was added to 225 mL of pH 8.5 alkaline peptone water and incubated at 37 °C for 18±2 h. Ten microliters of the pre-enrichment was then inoculated onto thiosulfate citrate bile sucrose (TCBS) agar plates (Merck, Darmstadt, Germany) and incubated at 37 °C for 24±2 h. Assumed Vibrio spp. colonies were picked from the plates and identified by API 20NE (bioMérieux Inc., Hazelwood, MO) and subsequently by PCR (as shown above). The isolated Vibrio alginolyticus strains were tested for the same primers as V. parahaemolyticus.

View this table:

Oligonucleotide primers and protocol used for PCR

Vibrio spp.GenePrimer sequence 5′–3′Product (bp)Annealing (°C s−1)Ref.
V. choleraetoxR (regulatory protein)CCT TCG ATC CCC TAA GCA ATA C77960/60Panicker (2004)
ctx (cholera toxin)CTCAGACGGGATTTGTTAGGCACG30255/120Brasher (1998)
V. parahaemolyticustlh (thermolabile hemolysin)AAA GCG GAT TAT GCA GAA GCA CTG45055/60Panicker (2004)
tdh (thermostable hemolysin)GTA AAG GTC TCT GAC TTT TGG AC26955/60Panicker (2004)
trh (TDH- related hemolysin)TTG GCT TCG ATA TTT TCA GTA TCT50055/60Panicker (2004)
V.vulnificusvvh (hemolysin)TTC CAA CTT CAA ACC GAA CTA TGA C20565/60Panicker (2004)
viuB (iron acquisition)GGT TGG GCA CTA AAG GCA GAT ATA50465/60Panicker (2004)

Cell toxicity test

To investigate the virulence of the isolated Vibrio spp. strains on eukaryotic cell cultures, CHO cells were exposed to the strains and the killing index (KI%) of the bacteria was calculated using the colorimetric method, based on enzymatic reduction of tetrazolium (MTS) and phenylmetasultazone (PMS) to formazan (Celltiter 96 aQuenous Non-Radioactive Cell Proliferation Assay G5421, Promega Corporation, WI). CHO cells were exposed to different concentrations of three V. parahaemolyticus strains (a, b and c), one V. vulnificus strain and four V. cholerae strains. Two of the V. cholerae strains originated from mussel meat (field a and b), with the other two being clinical non-O1/non-O139 isolates (clinical a and b). The bacteria were cultured overnight in a brain–heart infusion broth (Merck), pH 8.0–8.5 at 37 °C, with agitation at 150 r.p.m. The OD600 nm of the bacterial cultivation was measured (Eppendorf Biophotometer Plus) and adjusted to 1.0, corresponding to 109 mL−1, before the start of the experiment. The CHO cells were cultivated in GIBCO Dulbecco's modified Eagle medium with antibiotics [Antibiotic-Antimycotic (2184) 100X GIBCO] at 37 °C and 5% CO2. Cells were then trypsinated and the numbers were obtained by direct cell counts under a microscope using a hemocytometer. Using 96-well microtiter plates (Cellstar Tissue culture plate 96 W, Flat bottom, with a lid, sterile, Greiner bio-one), 104 CHO cells were exposed to three different bacterial concentrations: 104, 103 and 102. The plates were then incubated for 1 h at 37 °C and 5% CO2. After incubation, 15 μL of MTS/PMS dye was added to each well, followed by another 30 min of incubation under the same conditions. The formazan product was measured at 490 nm in an ELx 808IU plate reader from Bio-Tek Instruments Inc. The capacity of the bacteria to kill CHO cells, their KI%, was calculated in relation to cells incubated only in the medium and using the following formula: 100−(((AC+BAB) × 100)/AC), where A stands for absorbance, C for eukaryotic cells and B for bacteria. All experiments were performed three times, with a minimum of quadruplicates in each. Escherichia coli (K12) was used as a negative control.


The mean water temperature during the sampling period was 20.6 °C. The lowest temperatures were recorded on both the first and the last sampling dates: 17.0 °C on June 13 and 16.6 °C on September 25 and a maximum temperature of 24.1 °C recorded on July 29. Water salinity varied between 9.1‰ and 17.4‰ during the period, with the lowest value occurring in mid-June and the highest in mid-July. The mean weight of the mussel meat (excluding shell) was 1.3±0.9 g and the mean length (including shell) was 20.5±6.0 mm.

PCR and strain isolation by cultivation

DNA from the mussel meat was extracted and analyzed for the presence of the chosen genes (Table 1). Out of 19 samples, 53% were positive for V. cholerae specific toxR, while none were positive for the cholera virulence gene ctx. The specific V. vulnificus genes vvh and/or viuB were detected in 63% of the samples and 79% were positive for the V. parahaemolyticus tlh gene, of which 53% were positive for the virulence genes tdh and/or trh (Table 2; Fig. 2). On the first and the last sampling dates, when the water temperature was at or below 17 °C, all samples from both sampling sites were negative for all tested genes. The occurrence of vibrios at the two sampling sites was very similar.

View this table:

Occurrence of genes during summer 2006

V. choleraeV. vulnificusV. parahaemolyticus
13 JuneRåå
13 JuneDomsten
04 JulyRåå++++
04 julyDomsten+++++
12 JulyRåå+++++++
14 JulyRåå++++++
19 JulyRåå++
19 JulyDomsten+++++++
21 JulyDomsten+++
25 JulyRåå++++++
25 JulyDomsten++
29 JulyRåå++++++++
14 AugustRåå++++++++
14 AugustDomsten+++++++
21 AugustRåå+++++
21 AugustDomsten+++++
25 AugustDomsten+++
25 SeptemberRåå
25 SeptemberDomsten
Number positive samples, n=19 (%)10 (53)0 (0)9 (47)9 (47)12 (63)15 (79)7 (37)7 (37)9 (47)

Occurrence of Vibrio spp. genes (%) in mussel samples (n=19) from The Sound during summer 2006, shown by PCR.

To establish the accuracy of the primers, selected PCR amplicons were sequenced and compared with reference strains in the NCBI database (http://www.ncbi.nlm.nih.gov/), with the results showing an almost perfect match with the corresponding Vibrio spp. genes in the database. Specifically, the sequencing results of tlh, tdh and trh displayed 98–100% similarity exclusively to V. parahaemolyticus, the toxR gene fragment was 98–99% similar to V. cholerae toxR and vvh and viuB were 98–99% similar to the corresponding genes in V. vulnificus. The toxR PCR amplicons from our field and clinical V. cholerae isolates were also aligned and compared, with the results showing 97–99% similarity (GenBank accession numbers: HM195237HM195246).

The enrichment of Vibrio spp. in mussel homogenate resulted in the successful isolation of all three Vibrio spp. Isolated strains were initially identified by API 20NE and then tested by PCR. All V. cholerae strains were positive for the toxR gene, while all were negative for ctx. The V. vulnificus isolate was vvh positive, but viuB negative. The V. parahaemolyticus strains were positive for the species-specific tlh gene, but negative for the virulence genes tdh and trh. Vibrio alginolyticus was also isolated from the mussel meat, and because V. alginolyticus and V. parahaemolyticus have molecular similarities, we tested this species for the tlh, tdh and trh genes used in the identification of V. parahaemolyticus. However, none of the V. alginolyticus strains were positive for the tested genes, thus confirming the V. parahaemolyticus specificity of the primers used.

Cell toxicity test

CHO cells were exposed to the isolated Vibrio spp. in order to test the bacterial cytotoxic abilities. The bacterial KI% represents the percentage of CHO cells that were killed during bacterial exposure. The mean KI% of the V. cholerae strains was 78% (SD 9%), of which the field strains had a KI% of 81% (SD 9%) and the clinical strains 76% (SD 11%). The KI% of V. vulnificus was 85% (SD 6%) and that of V. parahaemolyticus strains was 79% (SD 20%) (Fig. 3). Escherichia coli (K12) was used as a negative control, the KI% of this strain being 33% (SD 4%).


KI% of isolated Vibrio strains on CHO cells, with Escherichia coli cont (K12) as a nonvirulent control. Vc, Vibrio cholerae; Vv, Vibrio vulnificus; Vp, Vibrio parahaemolyticus. The cytotoxicity tests were performed three times, each in quadruplicate.


This study discovered the presence of potentially human pathogenic V. cholerae, V. vulnificus and V. parahaemolyticus in M. edulis from The Sound during the summer months of 2006. The high similarities of Vibrio occurrence in mussels from the two different sampling sites (Fig. 1) enabled us to combine the results, leading to a more comprehensive picture of the area. Vibrio spp. were found to be present on every sampling occasion apart from the first and last dates (mid-June and late September), when none of the samples tested positive either by PCR or by cultivation. These two occasions coincided with the lowest water temperatures (≤17 °C) and subsequent attempts to isolate Vibrio spp. from water and sediment during colder water temperatures have been without success (B. Collin et al., unpublished data). Many studies have focused on how decreasing temperatures stress vibrios (Kelly, 1982; Singleton et al., 1982; Motes et al., 1998; Gonzalez-Acosta et al., 2006). A viable, but nonculturable (VBNC) state has been described for bacteria when they are alive, but do not grow on culture medium (Oliver, 2005b). Bacteria in the VBNC state can nevertheless be detected using the molecular method. However, in contrast to the hibernating theory presented by Kaneko & Colwell (1973), the negative results from PCR analyses during the low water-temperature periods here did not verify a VBNC state – at least not in concentrations high enough to be detected by PCR. Yet earlier studies in Scandinavian waters show different results. Bauer (2006) successfully isolated V. parahaemolyticus from rope-cultivated M. edulis when water temperatures were as low as 0.6 °C and Høi (1997) showed that it was possible to isolate V. vulnificus from a mussel-cultivating area in The Sound at 8 °C. This has also been shown from other parts of the world. DePaola (2003) isolated pathogenic V. parahaemolyticus (tdh+) from market oysters at 7 °C and V. vulnificus was isolated from oysters (Crassostrea virginica) at 10.8 °C by Motes (1998). Potentially human pathogenic Vibrio spp. might therefore be isolated even at very low water temperatures.

Human pathogenic V. cholerae is mostly known for its ability to produce the cholera toxin, encoded by the ctx gene, which is in turn regulated by ToxR (Sack et al., 2004). The samples (53%) in the present study were positive for the toxR gene, although none for the cholera-toxin gene. Even if extraintestinal infections caused by V. cholerae non-O1/non-O139 have been recorded in Scandinavia before, no molecular identification has been described as yet. Between 1994 and 1998, eight patients were diagnosed with extraintestinal V. cholerae non-O1/non-O139 infections in Denmark, with four of the patients declared to have been exposed to seawater. Six of the patients developed otitis, one ulcus cruris and one patient septicemia (Dalsgaard et al., 2000). The infections occurred when the water was warm, in July to September. In addition to the three deaths, several other wound and otitis media infections occurred in southern Sweden in 2006 (Shönning et al., 2008). The cytophatic test in the present study showed that clinical and environmental V. cholerae strains had similar cytotoxic effects, with a KI of 78% (SD 11%), indicating that the strains from the water might have a high pathogenic potential. The results may also confirm the strong assumption, based on the statement from the patient, that the clinical bacterial strain was acquired from the Baltic Sea. Bag (2008) studied the cytotoxic similarities of clinical and environmental V. cholerae strains, and showed that hemolytic activity may be due to genes other than those that have been previously identified as human virulence markers. For example, the hlyA gene in V. cholerae seems to be a good indicator of cytotoxicity on Vero cells (Ottaviani et al., 2009). In 2005, Dziejman presented a study on non-O1/O139 strains and showed that some environmental strains carry a type three secretion system (T3SS), which may be involved in virulence and environmental fitness.

The most common Vibrio spp. to cause severe wound infections is V. vulnificus. Several cases, some fatal, have been reported from the Baltic Sea (Bock et al., 1994; Dalsgaard et al., 1996; Ruppert et al., 2004; Frank et al., 2006). In the present study, genes specific for V. vulnificus were detected in 63% of the mussel samples, indicating that this Vibrio species is quite common in The Sound when water temperatures are above 17 °C. Both genes vvh and viuB were recorded, indicating virulence (Rho et al., 2002). Potential pathogenesis was confirmed by the cell toxicity test (KI 85%). Unlike Norwegian coastal waters (Bauer et al., 2006), the brackish water of the Baltic seems to favor the occurrence of V. vulnificus. During the period covered by the present study, two patients were diagnosed with wound infection caused by V. vulnificus, developed after bathing in the Baltic Sea on the German coast. Water from German Baltic beaches tested positive for V. vulnificus and V. parahaemolyticus, but not for V. cholerae (Frank et al., 2006).

The mussel samples (79%) in our study tested positive for V. parahaemolyticus genes, of which more than half were positive for the virulence genes tdh and/or trh. However, none of the V. parahaemolyticus strains isolated via cultivation were positive for these genes, a result corresponding with the findings of Bauer (2006). Previous studies claim that these virulence markers are present only in approximately 1–2% of environmental isolates (Nishibuchi & Kaper, 1995). We used PCR instead of cultivation, and were able to detect a high prevalence, but despite this, we cannot exclude the fact that our findings may differ from the others due to variation in location and habitat. The mussels used here were collected from near-shore rocks and concrete located close to marine sediment and were not rope-cultivated as in the Norwegian study, while the temperature and salinity also differed. Moreover, the results may indicate that enrichment and plating can alter the strain composition of Vibrio spp. found in mussel meat. In 1997, Pace investigated the possibility of V. parahaemolyticus being in a VBNC state in low-nutrient aquatic environments. Their results suggest that the addition of bile deoxycholic acid (at concentrations three times higher than bile in TCBS agar) to the culture media favors the viability of virulent over nonvirulent strains, potentially even inhibiting the growth of the latter. As the culture medium in this study did not have any extra bile added, this may explain why the isolated V. parahaemolyticus strains were tdh and trh, despite the finding of a high presence of virulence genes in mussel meat with the molecular technique. Although the cytophatic effect of V. parahaemolyticus was tested only using tdh and trh strains, the toxicity was comparable to that of the other virulent vibrios, indicating that not just the hemolysin (vvh) and iron acquisition (viuB) genes are responsible for its pathogenicity (Fig. 3). Vibrio parahaemolyticus causes gastroenteritis, most often acquired by the ingestion of undercooked or raw shellfish, as is generally the case for all Vibrio spp. E. coli (K12) was used as a nonvirulent control in the cytophatic test, with the results showing it had a much lower KI% than the vibrios. However, intestinal epithelial cells are the normal targets for gastroenteritis infection in humans, not CHO cells, meaning that additional studies of the strains' virulence are required in order to identify any clinically relevant cytotoxic effect. Previous studies of clinical and environmental V. parahaemolyticus strains show that the virulence markers tdh and/or trh are present in clinical fecal isolates and tdh and trh mutants lose their cytotoxic and enterotoxic ability (Nishibuchi et al., 1992; Xu et al., 1994). However, Vibrio species may also cause wound infections and the strains responsible are tdh and trh (Martinez-Urtaza et al., 2004; Ellingsen et al., 2008; Tena et al., 2010). Recent studies have shown that the T3SS may induce these infections. Zhou (2009) showed that wild-type V. parahaemolyticus caused the death of 95% HeLa cells while only 58% were killed by the T3SS mutant strain. Similar results have been presented earlier by Park (2004) and Ono (2006). Vibrio cholerae non-O1/O139 may also carry the T3SS genes, which can play a role in human virulence (Chatterjee et al., 2009), but this is not yet fully understood.

In order to investigate the similarities between the field and the clinical strains of V. cholerae, the 779-bp-long sequenced PCR amplicons of the toxR gene were aligned and compared. The results were highly similar to previously published sequences, indicating that potential pathogens may be common in the water. This is particularly reflected in the high similarity of 99% found between the amplicons of one field and one clinical strain. However, more genes must be sequenced in order to verify this result. Sequenced amplicons from V. vulnificus and V. parahaemolyticus also showed very high similarities to published sequences, indicative of the accuracy of the chosen primers.

In summary, the results from the present study show that potential human pathogenic Vibrio spp. are common in The Sound at water temperatures above 17 °C. Virulence genes specific to V. vulnificus and V. parahaemolyticus were present in about half of the samples, indicating a potential risk of catching infections for susceptible people exposed to the seawater. The results from the cytophatic tests further reinforce this assumption. The KI% of strains isolated from mussel meat was even higher than that of tested clinical strains. The similarity between strains of different origins was also confirmed on a molecular level, through alignments of PCR amplicons (99%). These amplicons also displayed a 99–97% correspondence to sequenced standards, confirming the accuracy of the chosen primers.

Despite several fatal cases caused by vibrios associated with the Baltic Sea being reported, screening for potential Vibrio pathogens has been neglected in this area. These results provide the first vital data for further studies, in which changes and variation of the occurrence will be of interest.


This study has been funded by research grant SWE-2005-397 from the Swedish International Development Cooperation Agency (SIDA). The two clinical V. cholerae strains were kindly provided by Johan Rydberg. We are also grateful to Annika Allard for the CHO cells, to Jim Collin, who assisted during the sampling period, and to Bodil Hernroth for discussions and support.


View Abstract