Prevalence, antibiotic resistance, virulence and plasmid profiles of Vibrio parahaemolyticus from a tropical estuary and adjoining traditional prawn farm along the southwest coast of India

Abstract

Vibrio parahaemolyticus, autochthonous to estuarine, marine, and coastal environments throughout the world, is a leading cause of foodborne gastroenteritis in Asia as well as in other countries. Considering their role in disease outbreaks, water and sediment samples from two different environments along the southwest coast of India, such as the Cochin estuary, which is a tourist hot spot and a traditional prawn farm, were examined for the detection of V. parahaemolyticus. A total of 120 presumptive vibrios were further screened using a chromogenic medium. The isolates producing the specific colour for V. parahaemolyticus on the chromogenic medium were confirmed by molecular methods. Using a PCR assay targeting the species-specific tlh gene, 75 isolates were confirmed as V. parahaemolyticus. The incidence of V. parahaemolyticus was found to be 71.6 % in the Cochin estuary and 53.3 % in the shrimp farm. The virulence genes tdh and trh were detected in two isolates from the Cochin estuary by multiplex PCR. Using 16S rRNA sequence analysis, our isolate exhibited 100 % similarity to the V. parahaemolyticus O3:K6 pandemic clone. The isolates were evaluated for their susceptibility towards 16 different antibiotics. All the isolates exhibited multiple antibiotic resistance (MAR). The MAR index of the isolates from the Cochin estuary ranged from 0.312 to 0.75 and that from the shrimp farm ranged from 0.1875 to 0.5. The plasmid profiles of the isolates were also checked. About 50.6 % of the isolates harboured plasmids and ten different profiles were observed. No correlation was observed between the antibiotic resistance patterns and plasmid profiles.

Introduction

Vibrios are Gram-negative halophilic bacteria, found naturally in shallow coastal waters to the deepest parts of the ocean, and are highly abundant in aquatic environments, including estuaries, marine coastal waters and sediments, and aquaculture settings worldwide. Vibrios comprise more than 74 species (Thompson et al. 2004), and most of the Vibrio sp. are known to be pathogenic to both humans and animals. V. parahaemolyticus is regarded as the primary source of increase in vibriosis incidence (Newton et al. 2012), and highly pathogenic serotypes of the species are emerging on a global scale. The pandemic serovar O3:K6 that emerged in India in 1996 has since been found to account for many cases of V. parahaemolyticus outbreaks worldwide. V. parahaemolyticus is a leading cause of bacterial gastroenteritis in Asia as well as in other countries (Chiou et al. 2000; Daniels et al. 2000; Wong et al. 2000; McLaughlin et al. 2005; Shigeaki et al. 2012), and is attracting increasing interest worldwide where raw or undercooked seafood is often consumed (Chao et al. 2009). It causes wound infections in those exposed to contaminated water (Miyoshi et al. 2008). It also causes septicemia, particularly in immunocompromised people (Daniels et al. 2000). Hence, the prevalence and distribution of Vibrios in aquatic environments is of great public health importance.

During the past few decades, there has been a rapid emergence of antimicrobial resistance in many bacterial genera, due to the excessive use of antibiotics in humans, agriculture, and aquaculture systems (Cabello 2006). Release of sewage also results in the entry of a large number of drug resistant bacteria from various sources into the environment. Resistant genes are further transferred from non-pathogens to pathogens through horizontal gene transfer via conjugation, transduction and transformation. This could lead to the transfer of drug resistance features to extremely autochthonous microflora such as vibrios. Thus, the search for genetic elements like plasmids, transposons, and integrons associated with antibiotic resistance in microorganisms also becomes important. Multiple drug resistance among Vibrio sp. in estuarine-marine environments may have future implications for those who consume seafood contaminated with these pathogenic vibrios, and also for the recreational and commercial users of these environments (Shaw et al. 2014).

Vembanad Lake is connected to the Arabian Sea through the Cochin estuary. This is the largest brackish, tropical wetland ecosystem along the southwest coast of India, and is of extraordinary importance for its hydrological function, biodiversity and rich fishery resources. It is also one of the famous hot spots for tourism. The backwaters of Kerala support as much biological productivity and diversity as can be found tropical rain forests. These waters are responsible for the rich fishery potential of Kerala. The backwaters also act as nursery grounds for commercially important prawns and fishes. The fields around the backwaters are suitable for aquaculture. These areas support traditional, seasonal and perennial prawn fishery. The fishery products from these areas are exported worldwide.

Considering all these factors, the present study aimed to investigate the incidence of V. parahaemolyticus at various stations of the Cochin estuary (influenced by urban, industrial, human and hospital waste water) and the adjoining traditional shrimp farm at Edavanakkadu. The study area is one of the most famous tourist hot spots, and shrimps grown here are exported to various countries. The study also checked for multiple drug resistance and plasmid profiling among the isolates.

Materials and methodology

Sample collection, processing and enrichment

Considering the size of the estuaries, water and sediment samples were collected from ten stations in the Cochin estuary (9°40′- 10°12′ N, 76°10′- 76°30′ E). Four traditional shrimp farms and the feeder canal to these ponds located along Cochin backwaters at Edavanakkadu were also selected for the study. The farms in this area are well connected to the Cochin estuary through the feeder canal system, which brings water to this area from the Cochin backwaters. The water samples were collected using a Niskin water sampler, and sediment samples were collected using the Van-Veen grab. All samples were processed within 2 h of collection. Aseptic procedures were strictly followed during collection, transportation and analysis of the samples. Five hundred millilitres of water samples from each station were filtered using 0.45 μm bacteriological filters, and each filter was transferred to 100 ml alkaline peptone water for pre-enrichment and incubated at 37 °C for 18–24 h. Sediment samples were analysed after making tenfold dilutions in isotonic saline. One millilitre of the diluted soil sample was transferred to 99 ml of alkaline peptone water and enriched by incubation at 37 °C for 18–24 h.

Isolation and identification of Vibrio sp.

A loopful of enrichment broth was aseptically streaked onto a sterile surface of dried thiosulphate citrate bile salt sucrose (TCBS; Himedia, India) agar plates, and incubated at 37 °C for 24 h. About three to four typical sucrose non-fermenting colonies having a green or bluish-green colour were picked from each TCBS plate. Preliminary screening of isolates was done on the basis of Gram staining, oxidase test, catalase test and reactions on triple sugar iron agar. The presumptive isolates were then streaked onto a chromogenic medium for further confirmation (Hichrome Vibrio agar, Himedia). The isolates producing bluish-green colonies on Hichrome Vibrio agar were picked and maintained on nutrient agar slants supplemented with 3 % NaCl for further molecular identification.

Extraction of genomic DNA

Genomic DNA of these isolates was extracted by the boiling method (Devi et al. 2009). The colonies were inoculated in Luria Bertani (LB) broth supplemented with NaCl (2 % w/v), and incubated at 37 °C with shaking (120 rpm) for 16–18 h. The broth cultures were centrifuged (10,000 rpm, 4 °C, 1 min) to obtain a pellet, which was then washed with normal saline (0.8 % NaCl w/v) and re-suspended with DNA-free sterile distilled water (0.5 ml). The resulting suspension was heated at 98 ± 2 °C for 15–20 min in a water bath to lyse the cells, and release the DNA. The lysate was centrifuged to remove the cell debris (10,000 rpm, 4 °C, 5 min), and the supernatant was stored (−20 °C) until further use.

Detection of species specific tlh gene by PCR

The detection of the tlh gene was performed using the primers TLHF (5′-AAAGCGGATTATGCAGAAGCACTG-3′) and TLHR (5′-GCTACTTTCTAGCATTTTCTCTGC-3′). PCR amplification was optimized in a total reaction volume of 25 μl consisting of sterile Milli Q water (15.5 μl), 10× PCR buffer (2 μl), primer (1 μl each), dNTP mix (1 μl, 200 mM), template (4 μl), and Taq DNA polymerase (0.5 μl). The PCR conditions included an initial denaturation of 94 °C for 3 min, followed by 30 cycles of denaturation (94 °C for 1 min), primer annealing (58 °C for 1 min), and primer extension (72 °C for 1 min), followed by a final extension (72 °C for 5 min).

Detection of virulence genes by multiplex PCR

The detection of the genes tdh (thermostable direct hemolysin) and trh (thermostable direct hemolysin-related hemolysin) was done using the primer pairs TDHL (5′-GTAAAGGTCTCTGACTTTTGGAC-3′) and TDHR (5′- TGGAATAGAACCTTCATCTTCACC-3′) for tdh and TRHL (5′-TTGGCTTCGATATTTTCAGTATCT-3′) and TRHR (5′-CATAACAAACATATGCCCATTTCCG-3′) for trh. PCR amplification was performed in a total reaction volume of 25 μl consisting of sterile Milli Q water (13.5 μl), 10× PCR buffer (2 μl), primer (1 μl each), dNTP mix (1 μl, 200 mM), template (4 μl), and Taq DNA polymerase (0.5 μl). The PCR conditions were the same as for the tlh gene.

Gel documentation and image analysis

The PCR amplified products were separated by electrophoresis on an agarose (1.5 % w/v) gel in 1× TBE Buffer (Himedia, India) containing 0.5 μg/ml of ethidium bromide. The amplicon sizes were compared with a 100 bp DNA ladder. The gels were then visualized under a UV transilluminator and recorded as tiff files using the Gel Documentation System, (GelDoc EZ imager, Bio-Rad).

16S rRNA amplification, sequencing and accession number

To confirm the identification of V. parahaemolyticus, the isolate PM1S2 was subjected to 16S rRNA gene amplification using the universal primer set (27F: 5′-AGA GTT TGA TCC TGG CTC AG-3′, 149R: 5′-GGT TAC CTT GTT ACG ACT T-3′). The reaction conditions included an initial denaturation at 95° for 2 min, followed by 30 cycles of denaturation at 95° for 2 min, annealing at 58 °C for 1 min, and extension at 72 °C for 2 min. A final extension at 72 °C for 10 min was also included.

The PCR product was sent for sequencing at Scigenom, Kochi. The sequence obtained was submitted to the basic local alignment search tool (BLAST; http://www.ncbi.nlm.nih.gov/BLAST) at NCBI (National Center for Biotechnology Information) to determine the percentage similarity with already identified 16S rRNA sequences in the GenBank database. The sequence was deposited in the Genbank and an accession number was allotted to it.

Antibiotic sensitivity test

The antibiotic sensitivity of the 75 V. parahaemolyticus isolates, i.e., 43 isolates from the Cochin estuary and 32 isolates from the shrimp farm, were analysed and compared using the disc diffusion method (Bauer et al. 1966). Antibiotic impregnated discs (Himedia, India) of 8-mm diameter were used for the test. Discs containing the following antibacterial agents were placed on the plates swabbed with enriched bacterial culture and incubated overnight: amoxycillin (Amx-10 μg), ampicillin (Amp-10 μg), amikacin (Ak-10 μg), carbenicillin (Cb-100 μg), sulphamethoxazole (Sm-300 μg), oxytetracycline (O-30 μg), chloramphenicol (C-30 μg), nalidixic acid (Na-30 μg), norfloxacin (Nx-10 μg), nitrofurantoin (Nit- 100 μg ), enrofloxacin (Ex-5 μg), erythromycin (E-15 μg), streptomycin (S-10 μg), netillin (Net-10 μg), trimethoprim (Tr-5 μg), and furazolidone (Fr-50 μg). The antibiotics belonged to nine different classes according to their chemical structure: aminoglycosides (amikacin, streptomycin, gentamicin, netilmicin), macrolides (erythromycin), beta-lactams (amoxycillin, ampicillin, carbenicillin), quinolones (nalidixic acid), fluoroquinolones (norfloxacin, enrofloxacin), tetracyclines (tetracycline, oxytetracycline), nitrofurans (nitrofurantoin, furazolidone), sulphadrugs (sulphamethoxazole) and folate pathway inhibitors (trimethoprim).

After incubation, the diameter of the zone of inhibition was measured and the results were interpreted based on recommendations of Clinical Laboratory Standards Institute (CLSI 2011). Based on resistance to more than three antibiotics, the isolates were grouped as multiple antibiotic resistant isolates. Multiple antibiotic resistance (MAR) indexing of the isolates was determined by calculating the ratio between the number of antibiotics to which an isolate is resistant and the total number of antibiotics to which the isolate was exposed (Krumperman 1983).

Plasmid profiling

The multiple drug resistant strains were screened for the presence of plasmids. Bacterial strains were grown in 10 ml of Luria Bertani broth (Hi media, India) containing 1 % sodium chloride and ampicillin, and were incubated overnight at 37 °C in a shaker incubator (200 rpm) (Scigenics Biotech, India) for 16–18 h. About 1.5 ml of this culture was used for plasmid extraction by the alkaline lysis method (BirnBoim and Doly 1979). Briefly, 1.5 ml of the culture was transferred into a microfuge tube and centrifuged at 6,000 rpm for 5 min. Supernatant was removed and the pellet was resuspended in 100 μl distilled water followed by 100 μl lysis buffer (10 % SDS, 0.5 M EDTA, 10 N NAOH). The tubes were kept in a boiling water bath for 10 min, then 50 μl of 1 mM MgCl₂ was added to them, in the hot condition itself. Tubes were kept on ice for 2 min, and then centrifuged at 12,000 rpm for 2 min. Then 3 mM potassium acetate was added and kept on ice for 2 min and then centrifuged at 12,000 rpm for 2 min. The supernatant was transferred to a new tube and 600 μl of isopropanol was added and the tube was kept on ice for 10 min. Tubes were then centrifuged at 10,000 rpm for 10 min. The supernatant was discarded and the pellet was rinsed in ice-cold 70 % ethanol at 5,000 rpm for 5 min and air-dried for about 10 min to allow the ethanol to evaporate. The pellet was resuspended in 30 μl distilled water and kept at 4 °C overnight for dissolving.

Electrophoresis was carried out on a 0.8 % agarose gel (Agarose, Himedia, Mumbai, India) 1 % (w/v) in 1× TBE Buffer (Himedia, India) containing 0.5 μg/ml of ethidium bromide. Electrophoretic separation was carried out at 75 V for 1 h and a molecular weight marker (Supercoiled DNA ladder, Himedia, India) was included. The gels were then visualized under a UV transilluminator and recorded as tiff files using the Gel Documentation System, (GelDoc EZ imager, Bio-Rad).

Results

Characterisation of the bacterial isolates and detection of virulence genes

A total of 120 V. parahaemolyticus-presumptive, bluish-green colonies (60 each from the Cochin estuary and shrimp farm) were isolated on the TCBS agar. Of the 120, 75 isolates, i.e., 43 isolates from the Cochin estuary and 32 isolates from the shrimp farm, produced bluish-green–coloured colonies (specific for V. parahaemolyticus) on Hichrome Vibrio agar. All 75 isolates were found to be V. parahaemolyticus by PCR assay targeting the species-specific tlh (450 bp) gene (Fig. 1). The incidence of V. parahaemolyticus was found to be 71.6 % in the Cochin estuary and 53.3 % in the shrimp farms. Multiplex PCR assay revealed two V. parahaemolyticus isolates from the Cochin estuary that possessed the virulent genes tdh and trh, with amplicon sizes of 270 and 500 bp, respectively (data not given).

Fig. 1

PCR amplified V. parahaemolyticus-specific tlh gene. Lane M: 1 kb molecular weight ladder; lanes 1–7: V. parahaemolyticus isolates

Confirmation by 16S rRNA sequencing analysis

When the sequence similarity search was done using the BLAST tool, our sequence was found to be 100 % identical with the 16S rRNA sequences of the pandemic clone V. parahaemolyticus O3:K6 in GenBank (RIMD 2210633). The sequence was submitted to GenBank and was assigned the accession number KM406325.

Antibiotic sensitivity of Cochin estuary isolates

All 43 isolates from the Cochin estuary were found to be multiple drug resistant. The MAR index of the isolates ranged from 0.312 to 0.75. The MAR index and antibiotic resistance pattern are shown in the Table 1. The most repeated resistance pattern was Amp, Amx, Cb, Sm, Nit, Ex, Fr. The percentage resistance of the isolates is shown in Fig. 2. All of the isolates (100 %) were resistant to amoxycillin (beta lactam) and furazolidone (nitrofuran). A high percentage of resistance was also exhibited towards ampicillin and carbenicillin (beta lactams), nitrofurantoin (nitrofuran), sulphamethoxazole (sulphonamide) and enrofloxacin (fluoroquinolone). A medium-to-low level resistance was shown towards erythromycin (macrolide), streptomycin and amikacin (aminoglycoside), trimethoprim (folate pathway inhibitors), oxytetracycline (tetracycline), gentamicin (aminoglycoside) and nalidixic acid (quinolone). All of the strains were sensitive to netillin (aminoglycoside) and norfloxacin (fluoroquinolone).

Table 1 Antibiotic resistance patterns and plasmid profiles of V. parahaemolyticus isolates from Cochin estuary

Fig. 2

Antibiotic resistance of V. parahaemolyticus strains from Cochin estuary. Antibiotics tested are as follows: Amx—amoxycillin (10 μg), Amp—ampicillin (10 μg), Ak—amikacin (10 μg), Cb—carbenicillin (100 μg), Sm—sulphamethoxazole (100 μg), Nit—nitrofurantoin (100 μg), Ex—enrofloxacin (5 μg), E—erythromycin (15 μg), Fr—furazolidone (50 μg), Tr—trimethoprim (5 μg), O—oxytetracycline (30 μg), S—streptomycin (10 μg), Gen—gentamicin (10 μg), Na—nalidixic acid (30 μg), Nx—norfloxacin (10 μg), Net—netillin (10 μg)

Antibiotic sensitivity of shrimp farm isolates

All the isolates from the shrimp farm were also multiple drug resistant. The MAR index of the isolates ranged from 0.1875 to 0.5. The MAR index and antibiotic resistance patterns are shown in Table 2. The most repeated resistance pattern was Amp, Amx, Cb. The percentage resistance of the isolates is shown in Fig. 3. All ofthe isolates (100 %) exhibited resistance towards amoxycillin (beta lactam). A high percentage of resistance was also observed towards ampicillin (beta lactam) and sulphamethoxazole (sulphonamide) and erythromycin (macrolide). A low level of resistance was observed towards nitrofurantoin (nitrofuran), enrofloxacin (fluoroquinolone), furazolidone (nitrofuran), trimethoprim (folate pathway inhibitors), oxytetracycline (tetracycline), carbenicillin (beta lactam), nalidixic acid (quinolone) and norfloxacin (fluoroquinolone). All isolates from the shrimp farm demonstrated sensitivity towards streptomycin, gentamicin and netillin (aminoglycoside).

Table 2 Antibiotic resistance patterns and plasmid profiles of V. parahaemolyticus isolates from a traditional shrimp pond

Fig 3

Antibiotic resistance of V. parahaemolyticus strains from traditional shrimp farm. Antibiotics tested are as follows: Amx—amoxycillin (10 μg), Amp—ampicillin (10 μg), Ak—amikacin (10 μg), Cb—carbenicillin (100 μg), Sm—sulphamethoxazole (100 μg), Nit—nitrofurantoin (100 μg), Ex—enrofloxacin (5 μg), E—erythromycin (15 μg), Fr—furazolidone (50 μg), Tr—trimethoprim (5 μg), O—oxytetracycline (30 μg), S—streptomycin (10 μg), Gen—gentamicin (10 μg), Na—nalidixic acid (30 μg), Nx—norfloxacin (10 μg), Net—netillin (10 μg)

Plasmid profiling of the isolates

Plasmid profiling of the 75 multiple drug resistant isolates was performed. Plasmids were found in 50.6 % (38/75) of the isolates and ten different plasmid profiles (P1–P10) were observed (Tables 1 and 2). Comparing between the two environments, 55.8 % of the isolates from the Cochin estuary revealed the presence of one to four plasmids of a size ranging from 0.5 to 33 kb (Table 1), and 43.75 % of the isolates from the shrimp farm revealed the presence of one to two plasmids of a size ranging from 0.75 to 33 kb (Table 2). Eleven isolates from the Cochin estuary harboured one 33-kb plasmid each, seven isolates had two plasmids, three isolates harboured three plasmids, and three isolates revealed the presence of four plasmids. Eleven isolates from the shrimp farm also harboured one plasmid each, and three isolates had two plasmids. The plasmid profile 1(P1: 33 kb) was the most common profile among the isolates from both of the environments. We had strains with the same resistance patterns, but some presented different plasmid profiles and some were devoid of plasmid. Hence, no correlation was observed between the presence of plasmids and antibiotic resistance among the isolates. Plasmid curing experiments have to be done in the future to confirm the correlation.

Discussion

Vibrio parahaemolyticus has been recognized as one of the most important foodborne pathogens and the leading causal agent of human acute gastroenteritis, following the consumption of raw, undercooked or mishandled seafood and marine products (Su and Liu 2007; Pal and Das 2010; Velazquez-Roman et al. 2012). The present study investigated the presence of V. parahaemolyticus in sediment and water samples from various stations of the Cochin estuary and the adjoining traditional shrimp farm. A high incidence of the species was found in both the environments. The Cochin estuary (71.6 %) showed a much higher incidence of V. parahaemolyticus when compared to the shrimp farm (53.3 %). Incidence of V. parahaemolyticus in India has almost doubled in the last 5 years (Chowdhury et al. 2000). By 16S r RNA sequencing, we observed that our isolate had 100 % similarity to the pandemic clone V. parahaemolyticus O3:K6. Strains belonging to pandemic O3:K6 have been previously isolated from environmental samples in several countries, including Bangladesh (Islam et al. 2004), Japan (Hara-Kudo et al. 2003), India (Deepanjali et al. 2005), and Italy (Caburlotto et al. 2010). They may prove endemic if favorable conditions occur in the environment. In our study, the virulence genes tdh and trh could be detected in two strains (2.8 %) that were isolated from the Cochin estuary. This is in agreement with previous reports stating that most of the environmental strains are not pathogenic to humans, and only 1–2 % of the environmental strains have been shown to be positive for these genes (Hervio-Heath et al. 2002; Robert-Pillot et al. 2004). On the contrary, a previous study revealed a much higher presence of tdh (8.4 %) and trh (25.3 %) genes in V. parahaemolyticus isolated from seafoods along the southwest coast of India (Devi et al. 2009).

Vibrio is generally considered to be highly susceptible to most clinically used antimicrobials (Oliver 2006). However, in the past few decades, antimicrobial resistance has emerged and evolved in many bacterial genera due to the excessive use of antimicrobials in human, agriculture, and aquaculture systems (Mazel and Davies 1999; Cabello 2006). This emerging issue has gained great concern due to an increased resistance of pathogenic V. parahaemolyticus towards clinically used antibiotics. Recently, a higher frequency of drug-resistant Vibrio has been reported (Okoh and Igbinosa 2010; Hua and Apun 2013). Our findings reveal that all the V. parahaemolyticus isolates from the Cochin estuary and shrimp farm have acquired multiple drug resistance. However, there was a difference in the pattern of antibiotic resistance from the two environments. This could be due the difference in the types and amounts of antibiotic residues the isolates were exposed to. As pointed out by Hsu et al. (1992), differences in the percentage of bacterial resistance to various antibiotics may reflect the history of antibiotic application, and hence there is a possibility of using bacterial drug resistance as an indicator of antibiotic application. A high percentage of β lactam resistance was exhibited by our isolates. Resistance towards β-lactam antibiotics has been previously reported in V. parahaemolyticus and other vibrios from different sources (Molina-Aja et al. 2002; Manjusha et al. 2005; Devi et al. 2009). A previous study reported sensitivity of V. parahaemolyticus from the southwest coast towards nitrofurantoin and trimethoprim (Devi et al. 2009). Similarly, in another study from Tunisia, more than 70 % of isolates of Vibrio showed susceptibility to trimethoprim-sulfamethoxazole (Lajnef et al. 2012). In contrast in the present study, Cochin estuary isolates demonstrated very high resistance towards the nitrofuran drugs (Nit—97.6 % and Fr—100 %) and sulphamethoxazole (88 %), and low level resistance towards trimethoprim (32.5 %). The shrimp farm isolates exhibited resistance towards the nitrofuran drugs (18 % and 33.3 %), trimethoprim (24.2 %), and sulphamethoxazole (51.5 %). Resistance towards enrofloxacin was shown by 95.3 % isolates from the Cochin estuary, while only 21 % of shrimp farm isolates had acquired resistance towards it. All isolates from the shrimp farm were sensitive towards gentamicin and streptomycin, whereas Cochin estuary isolates exhibited low level resistance towards both the drugs. Treatment recommendations for Vibrio infections include tetracyclines, flouroquinolones, aminoglycosides and folate pathway inhibitors (Daniels and Shafaie 2000; CDC 2013). In the present study, most of the V. parahaemolyticus isolates from the Cochin estuary have attained resistance towards life-saving drugs like enrofloxacin, nitrofurantoin, trimethoprim, sulphamethoxazole, streptomycin, amikacin and nalidixic acid. Shrimp farm isolates also exhibited resistance towards enrofloxacin, nitrofurantoin, trimethoprim, sulphamethoxazole, nalidixic acid. These strains act as potential reservoirs of drug resistant genes in the environment. Resistant genes are further transferred from non-pathogens to pathogens through horizontal gene transfer via conjugation, transduction and transformation. Based on our data, treatment of Vibrio infections may benefit from the use of aminoglycoside, netillin, which was the only aminoglycoside that was 100 % effective against vibrios recovered in this study.

All the isolates from the Cochin estuary showed a MAR index higher than 0.2, and it ranged up to 0.75. About 72 % of isolates from the shrimp farm also had a MAR index that was greater than 0.2. MAR indices higher than 0.2 are often considered to have originated from higher-risk sources (Krumperman 1983) of contamination, such as those from hospital sewage, commercial poultry farm waste, etc., that somehow find their way to the open sea via illegal dumping of waste or transferral by infected humans. The study area is one of the most famous tourist hot spots, and shrimps grown in this area are exported to various countries, making the findings of our study all the more important. Therefore, continued monitoring of both the prevalence of and the antimicrobial susceptibility profile of V. parahaemolyticus is important to better ensure seafood and public health safety from our study area.

The plasmid is known to be one of the most important mediators facilitating the fast spread of antibiotic resistance among bacteria (Dale and Park 2004). In the present study, we could detect the presence of plasmids among 50.6 % of our isolates. Ten different plasmid profiles were obtained. The plasmid of size 33 kb (profile1) was the common profile observed among the isolates from both environments; this is similar to the finding by Zhang et al. (2006), stating the presence of > 30 kb plasmids in environmental Vibrio isolates. Bacterial antibiotic resistance patterns are known to sometimes be associated with the presence of large plasmids, as well as the abilities of plasmids in conjugation. Transferable R plasmids are usually as big as 30 kb, and the indispensable components of a conjugative plasmid make it big in size compared to other plasmids (Guiney and Landa 1989). In our study, when we compared the antibiotic resistance patterns and the plasmid profiles, we could not find any correlation. Even among the strains with same resistance pattern, the plasmid profiles were different and some strains even lacked plasmids, which was similar to findings by Lajnef et al. 2012. So, in some strains resistance may be plasmid coded, and in some it may be chromosomally borne. Further experiments have to be done to confirm the origin of antibiotic resistance among the isolates. Plasmid profiles have been previously studied in Vibrio species such as V. parahaemolyticus (Devi et al. 2009), V. ordalii (Tiainen et al. 1995), V. vulnificus (Radu et al. 1998) and V. salmonicida (Sorum et al. 1990), and most extensively in V. anguillarum (Pedersen et al. 1996; Pedersen 1999), where a high diversity of profiles have been observed. A high incidence of plasmids in Vibrio sp. of both polluted and pristine environments may be ecologically important to the survival of these bacteria in the environment (Zhang et al. 2006). In the present study, we were not able to find any correlation between antibiotic resistance and presence of plasmids in the isolates.