Antibiotic resistance and molecular typing of Photobacterium damselae subsp. damselae, isolated from seafood

Authors


Correspondence

Ming-Lun Chen, Department of Food Science, National Penghu University of Science and Technology, No.300, Liuhe Rd., Magong City, Penghu County 880, Taiwan. E-mail: vanco@gms.npu.edu.tw

Abstract

Aim

The objectives of our study is to determinate the antibiotic susceptibility of this organism to different antibiotics to determine the discriminatory power of the molecular typing methods.

Methods and Results

In this study, 50 Photobacterium damselae subsp. damselae isolates from Scomber australasicus and Rachycentron canadum were collected in Taiwan and their resistance to 15 different antimicrobial agents was determined. In addition, random amplification of polymorphic DNA (RAPD) and pulsed-field gel electrolysis (PFGE) were performed to study the epidemiology and clonal relationship of P. damselae subsp. damselae. The results showed that the 50 isolates generated 25 typeable profiles with multidrug resistance to 3–7 antimicrobials. The results also indicate that the RAPD and PFGE methods have high discriminatory power for molecular subtyping.

Conclusion

Photobacterium damselae subsp. damselae isolates from fish to examine for multidrug resistance to antimicrobials. RAPD and PFGE methods revealed the high discriminatory power for molecular subtyping and provided information that could be used for risk assessment of P. damselae subsp. damselae infections.

Significance and Impact of the Study

These results may help in epidemiological investigations of P. damselae subsp. damselae and may be useful in controlling or treating P. damselae subsp. damselae infections in aquaculture and clinical therapy.

Introduction

Photobacterium damselae includes two subspecies, subsp. damselae and subsp. piscicida (Gauthier et al. 1995). Both subspecies are Gram-negative bacterial fish pathogens and are widespread in marine environment. P. damselae subsp. piscicida is the causative agent of fish pasteurellosis. P. damselae subsp. damselae acts as a pathogen to not only wild and cultivated fish but also marine animals and humans. P. damselae subsp. damselae has been reported to cause wound infections and fatal necrotizing fasciitis in humans (Clarridge and Zighelboim-Daum 1985; Tang and Wong 1999; Barber and Swygert 2000; Yamane et al. 2004; Aigbivhalu and Maraqa 2009). The pathogenicity of this organism is due to the production of toxins. P. damselae subsp. damselae strains have been shown to possess similar virulence determinants for poikilothermic and homeothermic hosts, secreting a potent, lethal phospholipase toxin with hemolytic and cytotoxic activities (Kothary and Kreger 1985; Toranzo and Barja 1993). Most of the reported infections in humans have their origin in wounds inflicted during the handling of fish, exposure to seawater and marine animals, and ingestion of raw seafood (Vaseeharan et al. 2007; Pedersen et al. 2009; Rivas et al. 2011). Moreover, the histamine-producing ability of P. damselae subsp. damselae has been reported in some studies, and some organisms are such strong producers of histamine that they can produce histamine in toxicologically sufficient amounts to cause histamine food poisoning (Kanki et al. 2007; Bjornsdottir et al. 2010). Thus, P. damselae subsp. damselae is an important clinical and foodborne pathogen.

Antibiotics are used in treatment for human infections caused by bacteria and widely used in aquaculture for prevention and treatment for fish diseases. At present, chemotherapy is the only option against pasteurellosis in Taiwan (Liao et al. 2007). For successful and effective administration of antibiotic treatment, knowledge about the antibiotic susceptibility of the target bacteria is important. Because pasteurellosis is a major cause of large economic losses in aquaculture, the antibiotic susceptibility of the P. damselae subsp. piscicida has received more attention than that of Pdamselae subsp. damselae. P. damselae subsp. piscicida strains isolated from different countries were found resistant to some antibiotics, such as ampicillin, carbenicillin, cefalothin, cefazolin, erythromycin, kanamycin and tetracycline (Thyssen and Ollevier 2001; Kawanishi et al. 2006; Martínez-Manzanares et al. 2008; Laganà et al. 2011). In addition, P. damselae subsp. piscicida has been found to harbour conjugative transferable multidrug resistance plasmids (Kim et al. 2008). Because P. damselae subsp. damselae often exist with P. damselae subsp. piscicida in the marine environment or in aquaculture farms, acquiring drug resistance by horizontal gene transfer between these two subspecies is possible. Furthermore, P. damselae subsp. damselae has been recognized as an emerging pathogen for marine animals and humans (Labella et al. 2011). But only a few P. damselae subsp. damselae isolates have been examined for antibiotic susceptibility (Fouz et al. 1992; Stephens et al. 2006; Pedersen et al. 2009; Jun et al. 2010). Therefore, more surveys on antibiotic resistance and susceptibility of P. damselae subsp. damselae are important for effective antibiotic use in aquaculture and clinical therapy.

Until now, only few studies have reported on the antibiotic susceptibility of P. damselae subsp. damselae, especially in Taiwan. The first objective of our study is to determinate the antibiotic susceptibility of this organism to different antibiotics. In addition, in our previous report, we showed that P. damselae subsp. damselae isolates collected from Taiwan displayed a universal ability to produce high levels of histamine leading to food poisoning (Chen et al. 2012). Hence, research in genetic characterization for P. damselae subsp. damselae is important for epidemiologic studies. Bacterial subtyping methods have been used successfully not only in detection and tracking of foodborne outbreaks but also in the analysis of genetic diversity and clonal relationship (Swaminathan et al. 2001; Foley et al. 2009). Therefore, the second objective of our study is to determine the discriminatory power of the molecular typing methods used [random amplification of polymorphic DNA (RAPD) and pulsed-field gel electrolysis (PFGE)] for characterization of P. damselae subsp. damselae-isolated strains.

Materials and methods

Bacterial strains

Scomber australasicus and Rachycentron canadum were collected from a local fish market during May–August 2006. The gills and internal organs of fish were taken aseptically into sterile plastic bags and homogenized with sterile saline. The homogenate was then diluted serially 10-fold. An aliquot (0·1 ml) of dilution was plated on to thiosulfate citrate bile salt sucrose agar (TCBS) supplemented with 1·5% sodium chloride and incubated at 25°C for 1–2 days. After incubation, the colonies with green centres were selected for Gram staining and oxidase test. The isolates of Gram-negative rod and oxidase-positive were then identified as P. damselae subsp. damselae by the API 20E test kit (bioMérieux Inc., Durham, NC, USA). In the present study, fifty P. damselae subsp. damselae isolates were obtained from S. australasicus (S1–S25) and R. canadum (C1–C25). The following reference strains were used for bacterial analyses: P. damselae subsp. damselae BCRC 12906 and 13854. The reference strains were obtained from Bioresource Collection and Research Center (BCRC), Hsin Chu, Taiwan. All strains were grown on tryptic soy broth with 0·5% sodium chloride (TSBN) at 25°C and were stored frozen at −80°C for future transfer.

Susceptibility testing

All 50 isolates and the two reference strains were subcultured onto Mueller-Hinton agar (Becton Dickinson, Cockeysville, MD, USA) for disk diffusion analysis. The antibiotic susceptibility of the isolates was analysed according to the CLSI standard (CLSI 2008). Resistance to 15 antimicrobial agents was examined using disks supplied by Oxoid (Oxoid Ltd, Hampshire, UK). These disks contained amoxicillin (AML, 25 μg per disk), ampicillin (AMP, 10 μg per disk), chloramphenicol (C, 30 μg per disk), enrofloxacin (ENR, 5 μg per disk), erythromycin (E, 15 μg per disk), florfenicol (FFC, 30 μg per disk), flumequine (UB, 30 μg per disk), gentamicin (CN, 10 μg per disk), kanamycin (K, 30 μg per disk), oxolinic acid (OA, 2 μg per disk), oxytetracycline (OT, 30 μg per disk), penicillin G (P, 10 units per disk), streptomycin (S, 10 μg per disk), sulphamethoxazole/trimethoprim (SXT, 19 : 1 25 μg per disk) and tetracycline (TE, 30 μg per disk). The quality control strain used was Escherichia coli ATCC 25922. Breakpoints for susceptibility testing were categorized according to Thyssen and Ollevier (2001).

Extraction of genomic DNA

For genomic DNA extraction, the following procedures were used. A portion (0·5 ml) of the overnight culture was centrifuged (14 000 g, 1 min). After removing the supernatant, the cell pellet was resuspended in 100 μl of lysozyme (20 mmol l−1 Tris-HCl, 2 mmol l−1 EDTA, 1% Triton X-100, 20 mg ml−1 lysozyme, pH 8·0); it was then incubated for 60 min at 37°C. Next, 350 μl of extraction buffer with 4 μl of proteinase K solution (20 mg ml−1) was added, and the mixture was incubated at 56°C for 1 h. For DNA isolation, the Gene-Spin™ Genomic DNA isolation kit (Protech Technology Enterprise Co., Ltd, Taipei, Taiwan) was used. According to the manufacturer's protocol, the nucleic acid was absorbed into the special silica membrane, and genomic DNA was then eluted using preheated water.

Random amplification of polymorphic DNA analysis for Photobacterium damselae subsp. damselae isolates

Random amplification of polymorphic DNA analyses of the isolates were conducted using a primer P6 (5′-CCC GTC AGC A-3′) (Amersham Pharmacia Biotech, Uppsala, Sweden). The PCR was performed in a 25 μl of solution containing 2.5 mol l−1 of each of the dNTPs, 25 mmol l−1 primer, 0·04 units of ProZyme (Biotech, Taipei, Taiwan), and 10 ng of template DNA. The amplification was carried out for 35 cycles, with each cycle consisting of 94°C for 30 s, 35°C for 30 s, 72°C for 30 s and finally, an additional 72°C for 7 min. The amplicons were analysed by electrophoresis at 100 V in 2·0% agarose gel and visualized under UV light.

Pulsed-field gel electrolysis analysis for Photobacterium damselae subsp. damselae isolates

A total of 50 P. damselae subsp. damselae isolates and two reference strains were analysed by PFGE according to the methods of Chiu, et al. (2007) with modifications. Each isolate was grown in 3 ml TSBN at 25°C for 4 h. The culture was centrifuged at 8000 g for 10 min at 25°C. Pelleted cells were washed once with suspension solution and resuspended in the suspension solution. The cell suspension was heated at 55°C for 3 min in a dry block heater and then mixed with pulsed-field certified agarose (Bio-Rad Laboratories, Hercules, CA, USA) solution prepared using the suspension solution. The mixture was poured into 1·5-mm-thick moulds and allowed to solidify at 4°C for 15 min.

The agarose plugs were placed in lysis buffer [2·0 mg ml−1 lysozyme, 10 mmol l−1 Tris-HCl (pH 7·2), 50 mmol l−1 NaCl, 100 mmol l−1 EDTA, 0·2% sodium deoxycholate and 0·5% N-laurylsarcosine] at 37°C for 24 h with gentle shaking to allow cell lysis followed by proteolysis in 3 ml of proteolysis solution (100 mmol l−1 EDTA, 1% N-laurylsarcosine, 0·2% sodium deoxycholate, and 0·15 mg ml−1 proteinase K) at 55°C for 24 h. The agarose plugs were washed several times in washing buffer [20 mmol l−1 Tris-HCl (pH 8·0) and 50 mmol l−1 EDTA] with gentle shaking for 30 min at 25°C and stored at 4°C in storage solution [2 mmol l−1 Tris-HCl (pH 8·0) and 5 mmol l−1 EDTA] until use. For enzyme digestion, a slice of gel plug (1 mm in thickness) was equilibrated with NEBuffer 2 (New England Biolabs, Beverly, MA, USA) and digested with 20 units of SfiI (New England Biolabs) at 50°C for 4 h. The enzyme-digested plugs were placed in slots of a 1·2% agarose gel (pulsed-field certified agarose; Bio-Rad). Electrophoresis was conducted using a CHEF DRIII system (Bio-Rad) with circulating 0·5× Tris-Borate-EDTA (TBE) buffer (14°C) using 6·0 V cm−1 with pulsed times ranging from 5 to 50 s for 24 h. Lambda Ladder PFG Marker (New England Biolabs) was used as the molecular weight marker. The gels were stained with ethidium bromide and visualized under UV light.

Similarities among patterns

Pulsed-field gel electrolysis and RAPD patterns obtained from P. damselae subsp. damselae isolates were analysed for genetic similarities using the BioNumerics software (Bio-Rad). Furthermore, the index of discrimination (DI) value was used to compare typing methods according to Hunter and Gaston (1988). This probability can be calculated by Simpson's index of diversity (Simpson 1949), which was developed for the description of species diversity within an ecological habitat.

Results

Susceptibility testing

The results of susceptibility testing of the 50 P. damselae subsp. damselae isolates are shown in Table 1. All isolates generated 25 typeable profiles showing multidrug resistance to 3–7 antimicrobials. The profile type showing major antimicrobial drug resistance, that is, resistant to 7 drugs, was the A11 type. Two strains of the A11 type were isolated from the S. australasicus samples and 6 strains from the R. canadum samples. The profile types showing minor antimicrobial drug resistance were A10, A9 and A20. The rate of resistance of the P. damselae subsp. damselae isolates to the 15 antimicrobial agents used in the study can be divided into three groups: high rates of resistance were for chloramphenicol (49/50, 98%), gentamicin (48/50, 96%), amoxicillin (47/50, 94%) and penicillin G (46/50, 92%); intermediate rates of resistance were for flumequine (27/50, 54%), enrofloxacin (25/50, 50%) and sulphamethoxazole/trimethoprim (22/50, 44%); low rates of resistance were for ampicillin (5/50, 10%), oxolinic acid (2/50, 4%), florfenicol (2/50, 4%), tetracycline (2/50, 4%), oxytetracycline (2/50, 4%), kanamycin (1/50, 2%) and erythromycin (1/50, 2%). All isolates were sensitive to streptomycin (Fig 1).

Table 1. Antibiotic resistant patterns for Photobacterium damselae. damselae-isolated strains
TypeAntibiotics patternsNo. of drugsNo. of strains
  1. a

    AML, amoxicillin; AMP, ampicillin; C, chloramphenicol; ENR, enrofloxacin; E, erythromycin; FFC, florfenicol; UB, flumequine; CN, gentamicin; K, kanamycin; OA, oxolinic acid; OT, oxytetracycline; P, penicillin G; S, streptomycin; SXT, sulphamethoxazole/trimethoprim; TE, tetracycline.

A1Ca SXT FFC CN P AML UB71 (S1)
A2OA C P AML41 (S2)
A3C CN P AML42 (S3, S8)
A4C SXT AMP CN P AML61 (S4)
A5C TE AMP CN ENR P AML71 (S5)
A6C CN P AML K UB61 (S6)
A7C CN AML UB E51 (S7)
A8C SXT CN P41 (S9)
A9C SXT CN P AML55 (S10, S15, S17, S21–22)
A10C CN P AML UB56 (S11–12, S16, C2, C4, C25)
A11C SXT CN ENR P AML UB78 (S13, S23, C6, C9, C16–17, C20, C24)
A12C CN P31 (S14)
A13C SXT CN AML UB51 (S18)
A14C CN ENR P AML53 (S9, C1, C8)
A15C SXT CN P AML OT61 (S20)
A16C SXT CN P AML UB61 (S24)
A17C SXT CN ENR P AML63 (S25, C14, C21)
A18OA C CN ENR P AML UB71 (C3)
A19TE CN ENR OT UB51 (C5)
A20C CN ENR P AML UB65 (C7, C11–12, C15, C18)
A21C AMP CN ENR P AML UB71 (C10)
A22C CN ENR AML41 (C13)
A23C AMP CN ENR P AML61 (C19)
A24C P AML31 (C22)
A25C FFC AMP CN P AML61 (C23)
Figure 1.

Percentage of Photobacterium damselae subsp. damselae isolated strains resistant to 15 antimicrobial drugs.

Random amplification of polymorphic DNA and pulsed-field gel electrolysis analysis of Photobacterium damselae subsp. damselae

By using P6 primer, fragments with molecular sizes between 200 and 2000 bp were obtained. All the 50 isolates were divided into three major groups (Fig 2) and a total of 40 profiles (Table 2). The R26 pattern was the major type, which included C3, C7, C9, C15, C17, C18 and C24 strains. For this type, all six strains were isolated from R. canadum samples. The antimicrobial drug resistance profiles of the six strains were distinct. The C3 strain was of the A18 type; C7, C15 and C18 strains were of the A20 type; and C9, C17 and C24 strains were of the A11 type.

Table 2. Genotypes by random amplification of polymorphic DNA (RAPD) and pulsed-field gel electrolysis (PFGE) methods for Photobacterium damselae subsp. damselae-isolated strains
StrainRAPD typePFGE type
S1R3P3
S2R4P4
S3R5P5
S4R6P6
S5R7P7
S6R8P8
S7R9P9
S8R10P10
S9R11P11
S10R12P12
S11R13P13
S12R14P14
S13R15P15
S14R16P16
S15R17P17
S16R18P18
S17R17P17
S18R8P19
S19R19P20
S20R20P21
S21R21P22
S22R22P17
S23R23P23
S24R22P24
S25R13P25
C1R24P26
C2R25P27
C3R26P28
C4R27P29
C5R28P30
C6R29P31
C7R26P28
C8R30P32
C9R26P28
C10R31P33
C11R32P34
C12R31P33
C13R33P35
C14R27P29
C15R26P28
C16R34P36
C17R26P28
C18R26P28
C19R35P37
C20R36P38
C21R37P39
C22R38P40
C23R39P41
C24R26P28
C25R40P42
BCRC 12906R1P1
BCRC 13854R2P2
Figure 2.

Dendrogram showing the genetic similarity between the random amplification of polymorphic DNA patterns from Photobacterium damselae subsp. damselae strains. The dendrogram was generated by BioNumerics software. Similarity percentages are shown above the dendrogram.

Analysis of the 50 isolated strains by PFGE using SfiI digestion generated 42 PFGE profiles (Table 2). All 50 isolates were divided into five major groups (Fig 3). The P28 pattern was the major pattern, which included six strains. The strains of major type identified by both RAPD and PFGE were identical.

Figure 3.

Dendrogram showing the genetic similarity between the pulsed-field gel electrolysis patterns from Photobacterium damselae subsp. damselae strains. The dendrogram was generated by BioNumerics software. Similarity percentages are shown above the dendrogram.

Analysis of the DI value, a single numerical index for discrimination power, for PFGE and RAPD methods showed that combined PFGE patterns had a DI value of 0·980, while the RAPD pattern had a DI value of 0·980. Both RAPD and PFGE methods used in this study had high DI values with strong discriminatory powers for molecular subtyping.

Discussion

Photobacterium damselae subsp. damselae strains from marine environment have been reported to exhibit multiple antibiotic resistance, which includes resistance to tetracycline, ciprofloxacin, florfenicol, ampicillin, sulfamethoxazole/trimethoprim, oxolinic acid, oxytetracycline, erythromycin and penicillin G (Fouz et al. 1992; Stephens et al. 2006; Jun et al. 2010). In Penghu, a wide variety of antimicrobial agents have been used for the treatment of pasteurellosis since 1999, including amoxicillin, chloramphenicol, flumequine, oxolinic acid, oxytetracycline and trimethoprim-sulfamethoxazole. Previously, Ku et al. (2009) have reported P. damselae ssp. piscida isolates from cobia that are resistant to trimethoprim-sulfamethoxazole, oxolinic acid and flumequine. Some studies have also indicated that P. damselae subsp. piscicida strains show high rates of multidrug resistance (Kim and Aoki 1993; Kawanishi et al. 2006). The transferable R plasmid, which carries several drug resistance genes, is present in both P. damselae subsp. piscicida strains and P. damselae subsp. damselae strains (Kim et al. 2008; Nonaka et al. 2012). In this study, the results show that the P. damselae subsp. damselae isolates are also resistant to the same antibiotic drugs as P. damselae subsp. piscicida strains. It is obvious that P. damselae subsp. damselae and P. damselae subsp. piscida grow in the same environment and carry similar antibiotic resistance. Hence, a serious problem is encountered in treatment of infections caused by these multidrug-resistant strains. No vaccine for treatment of pseudotuberculosis has been approved till now. Thus, periodical investigation of antimicrobial susceptibility of P. damselae subsp. damselae is important for the effective use of antimicrobials for the treatment of pseudotuberculosis.

Some studies have reported that using RAPD, amplified fragment length polymorphism (AFLP), and PFGE methods demonstrate high intraspecific variability among the strains of P. damselae subsp. damselae isolated from different sources (Botella et al. 2002; Takahashi et al. 2008; Labella et al. 2011). The results of this study indicate that the isolated strains of P. damselae subsp. damselae were different clones with a high level of genetic diversity among the P. damselae subsp. damselae isolates, as detected by antibiogram, RAPD, and PFGE methods, which in turn suggests that the RAPD and PFGE methods used in this study have a high discriminatory power for molecular subtyping. Labella et al. (2011) reported that no significant relationship could be established between molecular techniques and the origin, geographical location, host, or phenotypic characteristics. Similar results were confirmed in this study; different hosts have high diversity isolates. The subtyping data obtained from this study maybe more useful than the epidemiological data obtained from different countries and sources.

In addition to histamine poisoning, P. damselae subsp. damselae has also been reported to cause fatal infection in humans through water or food. Shin et al. (1996) reported that a 63-year-old man died of septicemia, suspected to have been caused by ingestion of raw eel contaminated with P. damselae subsp. damselae. Another case reported that a 46-year-old man died of hepatic dysfunction after oral ingestion of P. damselae subsp. damselae-contaminated raw fish (Kim et al. 2009). In this study, we have isolated P. damselae subsp. damselae strains from mackerel and cobia. Because a majority of aquacultured cobia is consumed as sashimi in Taiwan, more attention should be paid to the potential hazards caused by P. damselae subsp. damselae in cobia.

In summary, this study characterized 50 P. damselae subsp. damselae isolates from fish to examine for multidrug resistance to antimicrobials. Further analysis with RAPD and PFGE methods revealed the high discriminatory power of these two methods for molecular subtyping and provided information that could be used for risk assessment of P. damselae subsp. damselae infections. Furthermore, a long-term investigation of P. damselae subsp. damselae from seafood and environment should be conducted along with monitoring of dissemination by setting up a surveillance database in Taiwan.

Ancillary