Occurrence and antibiotic resistance of Vibrio parahaemolyticus isolated from the Tunisian coastal seawater

Vibrio parahaemolyticus is a natural halophilic bacterium found in estuarine and in shore marine water throughout the world (Letchumanan et al. 2015; Lopez-Joven et al. 2015). Since its isolation about 72 years ago (Fujino et al. 1953), infection by this bacteria has frequently been reported and has been closely associated with the consumption of raw or insufficiently cooked seafood or through wounds (Oliver 2005; Lopez-Joven et al. 2015). V. parahaemolyticus is recognized as a cosmopolitan bacterium frequently isolated from the USA, Canada (Newton et al. 2012; Taylor et al. 2018), Asia (Letchumanan Chan & Lee 2014; Yang et al. 2017), European countries (Cantet et al. 2013; Passalacqua et al. 2016), South America (Martinez-Urtaza et al. 2013; Raszl et al. 2016) and Africa (Malainine et al. 2013).

Some V. parahaemolyticus strains are pathogenic to humans and were responsible for 25% of seafood-borne diseases (Martinez-Urtaza et al. 2013; Tran et al. 2018). This virulence was associated especially with the production of thermostable direct haemolysin (tdh) and/or thermostable-related haemolysin (trh) (Wang et al. 2018), also known as the Kanagawa phenomenon (KP) (Leoni et al. 2016). Nevertheless, virulence was reported in some negative tdh and trh strains (Ottaviani et al. 2012; Chung et al. 2016). Some specific clones of V. parahaemolyticus were recognized as responsible for pandemic episodes such as O4:K68, O3:K6, and O3:K69 serovars (Haendiges et al. 2015; Han et al. 2017). Apart from its pathogenesis in human beings, this bacterium was previously reported in aquaculture industry outbreaks worldwide (Khouadja et al. 2013b; Soto-Rodriguez et al. 2015).

Several factors influence the distribution and abundance of V. parahaemolyticus in the sea or in estuarine ecosystems including water temperature (De Paola et al. 2003; Konrad et al. 2017), salinity (Raszl et al. 2016), oceanic circulation, zooplankton density, harmful algal (Vezzulli et al. 2016; Han et al. 2017) and faecal pollution (Nongogo & Okoh 2014; Okeyo et al. 2018). Overall, most studies agree that warmer water along with low salinity constitutes suitable conditions for V. parahaemolyticus presence and abundance (Martinez-Urtaza et al. 2008; Urquhart et al. 2016).

The large use of antibiotics in public health or on farm animals is partly responsible for the emergence of multidrug-resistant strains in common environmental bacteria (Letchumanan et al. 2015; Park et al. 2018). Generally, the development of antibiotic resistance in bacteria is controlled by a variety of mechanisms, most commonly gene transfer, mutation (Dalsgaard et al. 2000; Bengtsson-Palm et al. 2018) and cell membrane modification.

While numerous foodborne infections are associated with V. parahaemolyticus worldwide (Wang et al. 2017), only a limited effort has been made to investigate this bacterium in the Tunisian seawater; however, most research has focused on the analysis and characterization of the Vibrio alginolyticus (Ben Kahla-Nakbi et al. 2009; Lajnef et al. 2012). To date, important data gaps are identified regarding the prevalence and the distribution of V. parahaemolyticus in the Tunisian coastal seawater. The few studies carried out on this bacterium reported a number of fish farm outbreaks caused by V. parahaemolyticus (Khouadja et al. 2013b).

In the present study, we provide information on the occurrence, the antimicrobial susceptibility and the potential virulence genes (tdh and trh) of V. parahaemolyticus isolated from 15 stations situated along the Tunisian coastal seawater.

During 2018, 66 surface seawater samples from 15 stations located along the Tunisian coast were collected and analysed specifically for the presence of V. parahaemolyticus while reporting the presence of V. alginolyticus, Vibrio vulnificus and Vibrio cholerae (Figure 1). Simultaneously, parameters including geographical coordinates, period of sampling, temperature (°C) of the water samples and the vocation of the sampling area were determined (Table 1). The sampling frequency varies from a minimum of twice a year to once a month depending on the logistics and the distance of each station.

Sampling stations were located in the coastal area delimited among 40–500 m from the coastline (Figure 1). The geographical coordinates of each station were obtained by a Garmin 60 GPS. Water samples were collected between 9 and 10 pm at 0.3 m depth in a 500 ml sterile glass bottle and then conserved in the darkness at 10 °C approximately, which is considered a non-stressful condition for Vibrio species (Shen et al. 2009; Wang et al. 2018). According to the proximity of the sampling stations to the laboratory, samples were processed within 7 min to 20 h after sampling.

Simultaneously to sampling, temperature (°C), salinity psu (practical salinity unit) and pH were recorded in situ using a thermo conductive probe (Thermo-salinometer WTW 320) and a pH-meter (WTW pH 3310). The turbidity expressed in NTU (Nephelometric Turbidity Unit) was measured in the laboratory using a turbidity meter Hach model Ratio XR 43900.

For Vibrio detection, 100 ml of seawater for each sample were filtered under sterile conditions through a 0.45 μm pore size cellulose nitrate filter (Whatman 7184-004). The filter was transferred in 200 ml of double concentrated alkaline saline peptone water (APW-salt 3% (w/v) NaCl, pH 8.4). After 18 h of incubation at 37 °C under shaking at 25 RPM (revolutions per minute), a loopful of each enrichment was streaked on Vibrio plates (CHROMagar™ Vibrio Microbiology, France), used for the selective isolation of some Vibrio species. The presence of presumptive colonies of V. parahaemolyticus, V. alginolyticus, V. cholerae and V. vulnificus was examined in the CHROMagar plate after 18 h of incubation at 37 °C based on their colours.

Only typical purple colonies of V. parahaemolyticus, in CHROMagar, were selected and plated onto thiosulphate citrate bile salt sucrose (TCBS, SCHARLAU, Spain) for further control. A typical colony with a purple colour in CHROMagar and a green one in TCBS is considered as a presumptive V. parahaemolyticus. Only eight samples from 66 showed the presence of purple colonies. In total, nine presumptive colonies were selected and purified on trypticase soy agar (TSA) (BIORAD) supplemented with 2% NaCl, then stored in the CRYO-BEADS (AES Laboratoire) at −80 °C for further analysis.

Catalase and cytochrome oxidase activities were tested, respectively, with hydrogen peroxide (30% volume) and oxidase discs (HIMedia product), while the Gram staining and motility were examined microscopically. The sensitivity to vibriostatic compound was tested by using O₁₂₉ (150 μg/disk, Oxoid) in TSA supplemented with 2% NaCl (BIORAD). Furthermore, the identification of each presumptive V. parahaemolyticus isolate was performed by commercially available miniaturized identification systems API 20E (API–bioMerieux, Marcy l'Etoile, France) following the manufacturer's instructions. The inocula were prepared in sterile distilled water supplemented with 1% NaCl. Positive results were recorded according to colour change and identified by using the APIWEB identification software (APIWEB™ identification software bioMérieux). The results of identification were presented with the aforementioned acceptability provided by the software. All biochemical tests were made in duplicate.

The antibiotic susceptibility analysis of V. parahaemolyticus was carried out using the disc diffusion method (Bonnet et al. 2018) on Muller Hinton agar plates (Pronadisa Laboratories CONDA, Spain) supplemented with 2% NaCl. The standard commercial antibiotic disks used (Bio-Rad Marnes-la-Coquette France) were as follows: colistin (COL) (50 μg), amikacin (AKN) (30 μg), nalidixic acid (NAL) (30 μg), nitrofurantoin (NFE) (300 μg), cefotaxime (30 μg) (CTX), chloramphenicol (CHL) (30 μg), gentamicin (GMI) (15 μg), tetracycline (TET) (30 μg), trimethoprim/sulphamethoxazole (SXT) (1.25/23.75 μg) and Penicillin (P) (6 μg).

After incubation at 37 °C for 18–24 h, the diameter of the inhibition zone was measured using a calliper, and the obtained values were interpreted according to the standard of Clinical and Laboratory Standards Institute (CLSI 2018). Antibiotics' susceptibilities were carried out in triplicate, and the results are expressed as the mean of the three experiments with standard deviation.

The preparation of template DNA for the polymerase chain reaction (PCR) was performed by a boiling technique. Briefly, 1 ml of overnight pure culture in tryptic soy broth supplemented with 2% NaCl of each presumptive isolate was centrifuged at 7,000 RPM, at 4 °C for 8 min. The pellet was suspended in 300 μL of Tris–EDTA (TE) buffer (10 mM Tris, 1 mM EDTA, pH 8.2) and boiled at 100 °C for 8 min. Cell suspension was centrifuged for 8 min at 9,000 RPM at 4 °C, and the obtained supernatant containing the DNA was carefully transferred to a new tube and stored at −20 °C for PCR analysis (Alexopoulou et al. 2006).

The PCR was performed both on the DNA extract obtained from sample enrichment and from the pure culture of presumptive isolates. The species-specific toxR gene (368 bp) was conducted according to the method developed by Kim et al. (1999). DNA extracted from V. parahaemolyticus isolates was further examined to detect the presence of virulence genes tdh (251 bp) and trh (250 bp) (Tada et al. 1992; Kim et al. 1999). Primers used in this study are listed in Table 2.

DNA from V. parahaemolyticus AQ 4037 was used as a positive control for trh gene and V. parahaemolyticus ATCC 43996 was used as a positive control for toxR and tdh genes. In all PCRs, DNA extracted from Salmonella typhimurium ATCC 14028s was used as a negative control. Amplification was performed in a thermal cycler ‘TECHNE 3000’. PCR products were electrophoresed on a 1.5% agarose gel, stained with ethidium bromide, visualized and photographed under a UV transilluminator (Bio-Rad ChemiDoc XRS System).

All 66 seawater samples analysed by cultural methods using CHROMagar culture media showed the absence of typical blue-turquoise colonies relative to V. cholerae and green colonies relative to V. vulnificus as well. Our analysis showed the presence of yellow colonies in CHROMagar relative to V. alginolyticus in all seawater samples. These entire presumptive colonies specific to V. alginolyticus in CHROMagar indicated a positive amplification by PCR using collagenase-targeted gene amplification (Di Pinto et al. 2005) (data are not shown in this paper).

Only eight samples revealed the presence of purple colonies in CHROMagar (Table 1) characterizing a typical V. parahaemolyticus. In total, nine presumptive colonies were selected from CHROMagar, grown on TCBS agar for further differentiation and then isolated in pure culture on TSA agar supplemented with NaCl at 2%.

Hence, the analysis showed that all nine purified isolates are motile, Gram-negative, cytochrome oxidase-positive, catalase-negative and sensitive to vibriostatic compound O₁₂₉ (150 μg/disk). Furthermore, the API 20^E biochemical identification confirmed, with a level of acceptance up to 99%, that all presumptive colonies are V. parahaemolyticus (Table 3).

Biochemical characterization performed in API 20E revealed that all these isolates were β-galactosidase-negative and positive for glucose fermentation and indole. Some variations were registered on enzyme activity, such as tryptophan desaminase, melobiose and citrate metabolism. All the isolates (Isol) were urease-negative except those from station nos 11 and 12 (Isol 1 and Isol 2), indicating a capacity to metabolize urea.

All the colonies identified as V. parahaemolyticus by the API 20^E test were found to be positive for the toxR gene amplification (Figure 2). Simultaneously, perfect agreement has been found for the toxR amplification that was performed on DNA extracted from the enriched samples and presumptive V. parahaemolyticus isolates obtained by a cultural method. Only the two isolates, Isol 1 and Isol 2, obtained from stations 11 and 12, were found positive for trh gene (Figure 3). However, all isolates were negative for the tdh gene amplification (Figure 4).

The obtained results showed a limited presence of V. parahaemolyticus in the Tunisian coastal seawater, which was found only in seven stations (Table 1). All V. parahaemolyticus were isolated at temperatures ranging from 13.8 to 29.3 °C and at salinity between 37.8 and 41.1 psu (Table 1). The two positive trh V. parahaemolyticus were isolated from stations 11 and 12 during the month of February (16.7 °C temperature, 38 psu salinity and 0.722 NTU turbidity) (Figure 1; Table 1). V. parahaemolyticus was not found in stations 1, 3, 4, 9, 10, 13, 14 and 15 by both cultural and molecular methods.

Antibiotic susceptibility (Table 4) of different V. parahaemolyticus isolates shows that all isolates exhibited resistance to colistin, amikacin, penicillin and cefotaxime. However, all strains were sensitive to ceftazidime, tetracycline, ciprofloxacin, chloramphenicol, nitrofurantoin and nalidixic acid, and were registered. Different resistance patterns were recorded for tetracycline (66% of isolates), gentamicin (88% of isolates), nitrofurantoin (33% of isolates) and sulfamethoxazole-trimethoprim (11% of isolates). The multiple antibiotic resistance (MAR) index values show that eight of nine V. parahaemolyticus isolates have a value of 0.4, while only one has a 0.5 value (Isol 9).

To our knowledge, this study constitutes the first extended monitoring of the presence of V. parahaemolyticus in Tunisian marine waters. This reveals the low occurrence of V. parahaemolyticus in Tunisian coasts compared to investigation results found in other Mediterranean countries such as Italy, France and Spain (Lopez-Joven et al. 2015; Caburlotto et al. 2016). Our results are in agreement with those reported by Snoussi et al. (2008), Khouadja et al. (2013a), Zrelli et al. (2015) and Gdoura et al. (2016), who also mentioned the rare presence of V. parahaemolyticus in Tunisian seawaters.

V. parahaemolyticus has received extensive attention in numerous countries for its recovery in seafood (Letchumanan et al. 2014; Tran et al. 2018). However, few resources have been allocated in Tunisia to V. parahaemolyticus due to the absence of epidemiological cases (annual report of Institut Pasteur 2017, 2018). A rare infection caused by V. parahaemolyticus is probably due to the absence of consumption of raw or undercooked seafood in Tunisian traditional food. The low presence of V. parahaemolyticus in the Tunisian coastal seawater compared to data in Asian and European countries (Zulkifli et al. 2009; Esteves et al. 2015; Odeyemi 2016) suggests that this species adapts poorly to southern Mediterranean conditions. In fact, the viability of V. parahaemolyticus in the south Mediterranean is likely influenced by the synergistic effect of high temperature, important salinities and sunlight intensity (Givens et al. 2014; Larsen et al. 2015).

However, the high sunning and good water transparency in the sampling sites (lighting superior to 25 Klux and turbidity ranging between 0.141 and 2.75 NTU) contribute to the bacterial die-off by the production of reactive oxygen species (Zaafrane et al. 2004; Schmitz-Valckenberg et al. 2016).

This phototoxicity action induces the rapid passage in viable, but non-culturable (VBNC) cell states in many bacterial species, which partly explains the low presence of V. parahaemolyticus and other pathogenic Vibrio in Tunisian seawater (Zaafrane et al. 2004; Chandran & Hatha 2005; Idil et al. 2013).

In addition to the direct effect of sunlight, the scarcity of rivers, discharging into the sea on the Tunisian context contributes to the elevated salinities, and the limitation of nutrient inputs in the coastal zone, which constitute a selective condition for several germs (Millot & Taupier-Letage 2005; Slimani et al. 2012).

However, it is important to note the limitation of some current methods in quantifying and identifying bacteria in the marine environment due to the presence of the VBNC state (Li et al. 2014; Ding et al. 2017).

Underestimated or undetected in environmental and clinical samples by conventional methods of analysis, this physiological state (VBNC) poses a risk to public health. It also requires the use and the development of efficient techniques to quantify and evaluate pathogenic bacteria in marine environments, especially the ones associated with epidemic and pandemic cases. In this context, the use of techniques such as a High-Performance Flow Cytometry and RT-PCR seems to be required for monitoring different states of pathogenic bacteria. However, these rigorous techniques are actually limited for rapid large-scale use (Zhong & Zhao 2018; Robben et al. 2019; Wagley et al. 2019).

In this study, all V. parahaemolyticus isolates were isolated at temperatures ranging from 15.8 to 22.5 °C and salinities ranging from 37.8 to 39 psu, constituting favourable environmental conditions for this bacterium (De Paola et al. 2003; Zulkifli et al. 2009). Other authors (Martinez-Urtaza et al. 2008; Caburlotto et al. 2016) showed that the presence of V. parahaemolyticus is correlated essentially with low salinity (30.9–36.2 psu).

The absence of V. parahaemolyticus during the summer season (Table 1) characterized by an important photoperiod also explains the stressing synergistic effect of sunlight, high temperature and salinity on the presence of this bacterium (De Paola et al. 2003; Martinez-Urtaza et al. 2008; Liu et al. 2015).

Considering the limited number of V. parahaemolyticus isolates, and the few salinity fluctuations, it would be difficult to correlate the presence of this bacterium with this parameter (Table 1). However, some monitoring performed in the Tunisian coastal seawater showed that the distribution of V. alginolyticus and other pathogenic bacteria is correlated essentially with the water temperature (Cherif et al. 2011).

The important presence of V. alginolyticus in all samples is due not only to its capacity to tolerate hostile environmental conditions, but also its ability to use a zooplankton chitin as a nutrient (Rao et al. 2013). Moreover, several authors demonstrate that V. alginolyticus exhibited a better survival and a larger distribution in seawater than V. parahaemolyticus and V. vulnificus (Munro et al. 1994; Eiler et al. 2007).

The amplification of the toxR gene by the PCR from suspected V. parahaemolyticus isolates is in perfect concordance with the results of PCR performed directly from the enriched samples.

The amplification of virulence genes shows only the presence of trh in two isolates (stations 11 and 12) (Figure 1). These virulent isolates are found at the same time and in the same area (situated near to the anchorage area of the commercial Sfax harbour) and seem to be derived from ballast water.

The presence of positive trh strain is rarely observed in environmental isolates; however, some authors connect this presence with the season (Caburlotto et al. 2016; Di et al. 2019). Several studies demonstrate that the percentage of presence of tdh and trh strains in a marine environment remained between 2 and 9% for tdh and trh, respectively, and this proportion depends on environmental factors and geographical context (Caburlotto et al. 2016; Di et al. 2019). The two isolates possessing a trh gene also exhibit a positive urease activity in the API 20E test.

Normally, most urease-positive strains isolated from the environment are mainly positive for trh gene (Wang et al. 2017); yet, some research reported the presence of non-virulent strains (trh⁻ and tdh⁻) but having a capacity to hydrolyze urea (Ottaviani et al. 2012; Zrelli et al. 2015).

Although several authors criticize the use of API 20E systems for the identification of presumptive environmental isolate (Croci et al. 2007; Copin et al. 2012; Amalina & Ina-Salwany 2016), our results show a high degree of identification by phenotypic methods. In our investigation, cultural methods using CHROMagar and non-fermentation of the sucrose in TCBS seem to be an efficacious procedure for rapid control and differentiation of the three major pathogenic Vibrio species in environmental samples (Larsen et al. 2015; Nigro & Steward 2015).

Antibiotic susceptibility showed that most isolates are sensitive to sulphamide and chloramphenicol, yet resistant to amikacin, cefotaxime and penicillin. They presented intermediate susceptibility to gentamicin and tetracycline. These results are in agreement with those found by Letchumanan et al. (2015), Yang et al. (2017) and Elmahdi et al. (2018), which recorded an increase in multidrug resistance in environmental Vibrio and Aeromonas isolates. We note that all isolates are sensitive to the antibiotics used in the therapeutic treatment of V. parahaemolyticus such as amikacin sulphamide and nalidixic acid.

The tetracycline resistance observed in 66% of V. parahaemolyticus isolates is probably attributed to the large use of this antibiotic in aquaculture farms (Yang et al. 2017; Fri et al. 2018; Park et al. 2018). In fact, antibiotics and their residues have been reported frequently at sea and coastal areas (Smaldone et al. 2014; Patrolecco et al. 2018). In the same context, it is common that the isolation of V. parahaemolyticus strains is resistant to β lactams through the penicillinase production (Zago et al. 2020; Siddique et al. 2021). The recorded sensitivity to chloramphenicol in our case contrary to V. parahaemolyticus strains resistance observed in other countries (Han et al. 2017) is probably due to the limited use of this family of antibiotics in food product (Angulo et al. 2004; Chantziaras et al. 2014).

The MAR index ranges from 0.4 to 0.5 are greater than 0.2, which indicate that V. parahaemolyticus isolates have been exposed to antibiotics at least once in their cycles and probably have acquired a genetic resistance (Krumperman 1983; Letchumanan et al. 2015).

It has been largely demonstrated that antibiotic resistance acquired in bacteria is generally mediated by extrachromosomal plasmids and transmitted to the next generation or exchanged among the different bacterial population (von Wintersdorff et al. 2016). In the same framework, Letchumanan et al. (2015) demonstrated that the resistance to amikacin, ceftazidime, cefotaxime, kanamycin and levofloxacin depended on a plasmid carry in V. parahaemolyticus. Recent research has demonstrated the existence of two cellular signals inducted by bacteria for controlling antibiotic responses: a quorum-sensing signal responsible for inducing antibiotic resistance between bacteria and a quorum-quenching signal that altered this communication (Munita & Arias 2016). The transfer of mobile genetic elements of antibiotic resistance between marine bacteria (Partridge et al. 2018) and the large use of antibiotics is mostly responsible for the increase of multidrug resistance. This requires the development of environmental monitoring programmes in order to explore this emergence in marine bacteria.

This study conducted in 2018 showed the rare presence of V. parahaemolyticus in Tunisian coastal seawaters. This is probably due to poor adaptation of this bacterium for the south Mediterranean context (high sunning and salinity). However, the presence of virulent trh strains suggests the establishment of monitoring programmes for the detection and characterization of antibiotic resistance of V. parahaemolyticus and other pathogenic Vibrio species in Tunisian seawater and seafood product. Nevertheless, these results do not fail to record the expanding presence of V. parahaemolyticus and other pathogenic Vibrio species in the Tunisian context as a consequence of climate change as well as the disturbance in water quality.

The authors declare no conflicts of interest to disclose.

All relevant data are included in the paper or its Supplementary Information.