Identification and virulence of Aeromonas dhakensis, Pseudomonas mosselii and Microbacterium paraoxydans isolated from Nile tilapia, Oreochromis niloticus, cultivated in Mexico

Authors


Correspondence

Sonia A. Soto-Rodriguez, CIAD, A.C. Mazatlan Unit for Aquaculture and Environmental Management, A.P. 711 Mazatlan, Sinaloa 82010, Mexico. E-mail: ssoto@ciad.mx

Abstract

Aims

To identify bacterial pathogens of diseased NiIe tilapia Oreochromis niloticus and determine their virulence.

Methods and Results

Sixteen bacterial isolates were recovered from diseased Nile tilapias (O. niloticus) reared in floating cages in Adolfo Lopez Mateos (ALM), Sanalona and Dique IV dams in Sinaloa, Mexico, from February to May 2009. The bacterial isolates were identified by phenotypic and molecular (rep-PCR and 16S rRNA sequencing) methods and were mostly isolated from the kidneys and the brain of tilapias. Bacterial cells and extracellular products (ECPs) of strains were characterized and used in experimental infections with sole Solea vulgaris and Mozambican tilapia Oreochromis mossambicus. The fish challenged with Aeromonas dhakensis sp. nov. comb nov, Pseudomonas mosselii and Microbacterium paraoxydans (3·1 × 106 CFU g1) exhibited mortality between 40 and 100% starting at 6 h postinoculation. The ECPs displayed gelatinase, haemolytic and cytotoxic activity, causing the total destruction of the HeLa cell lines.

Conclusions

Aeromonas dhakensis and Ps. mosselii were virulent to O. mossambicus, whereas Mic. paraoxydans displayed virulence to Svulgaris.

Significance and Impact of the Study

This the first time that Aeromonas dhakensis and Ps. mosselii are reported as pathogens to tilapia and Mic. paraoxydans was isolated from fish; then, these fish pathogens could be a threat to farmed Nile tilapia in Mexico.

Introduction

Several species of tilapia are cultured commercially, but Nile tilapia (Oreochromis niloticus) is the predominant cultured species worldwide (FAO 2012). Strains of Aeromonas hydrophila, Edwarsiella tarda, Pseudomonas fluorescens and Streptococcus sp. have been reported to be virulent to farmed tilapia around the world (Al-Harbi and Uddin 2005; El-Sayed 2006). The heterogeneity of Aeromonas populations has also been observed in wild and farmed freshwater fish (Burr et al. 2012). Aeromonas species, although they have been misidentified many times, are well-known agents of fish disease, and two major groups are recognized. Aeromonas salmonicida sensu stricto causes fish furunculosis, particularly in salmonids, and mesophilic species (Aer. hydrophila and Aer. veronii) cause a similar assortment of diseases in fish, including motile aeromonas septicaemia (MAS), red-sore disease and ulcerative infections in carp, tilapia, perch, catfish, salmon, cod and goby (Joseph and Carnahan 1994).

Moreover, the genus Pseudomonas includes metabolically versatile organisms utilizing a wide range of organic compounds. The bacteria belonging to the genus Pseudomonas are present in most natural waters and infect a variety of fish. Some Pseudomonas species can cause bacterial haemorrhagic ascites, pseudomonad septicaemia, red spot disease, fin/tail rot and other fish diseases (Miyashita 1984; Nishimori et al. 2000; López-Romalde et al. 2003). Currently, Ps. anguilliseptica is considered the most significant pathogen of cultured fish (López-Romalde et al. 2003; Magi et al. 2009).

At present, the genus Microbacterium comprises 81 species (www.bacterio.cict.fr/m/microbacterium.html), which are common in dairy products and the mammalian intestinal tract (Buczolits et al. 2008). Microbacterium species are a rare cause of bacteraemia, although cases of peritonitis due to an uncommon Microbacterium sp. have been reported (Adames et al. 2010). Despite the importance of fish pathogens, to date, there has been no health surveillance programme initiated in Mexico to detect potential bacterial pathogens in farmed tilapia. In this study, bacterial isolates were recovered from the internal organs of diseased Nile tilapia cultured in floating cages. Strains of Aeromonas dhakensis sp. nov. comb nov., Microbacterium paraoxydans and Pseudomonas mosselii were identified by molecular methods. Therefore, given the importance of the Nile tilapia in Mexico, the present work was conducted to isolate and identify bacteria recovered from diseased Nile tilapia and also to evaluate their pathogenicity to flat fish (Solea vulgaris) and Mozambican tilapia (O. mossambicus).

Materials and methods

Study site

The State of Sinaloa is located in north-western Mexico. It has 12 dams, which were built primarily for irrigation and for domestic water supply. Nile Tilapia has been traditionally fished in some dams. However, in recent years, floating cages made of metal frames and 1-in. mesh nets have been installed in water reservoirs and stocked with sexually reversed fish fry. Three dams were selected to collect Nile tilapia; the dams were selected according to the date of seeding, the number of farms, the number and size of floating cages in each farm, the information about health status and sanitary management, access to roads and distance to the laboratory. The dams sampled were Adolfo Lopez Mateos (ALM), with 70 cages of size 11·5 and 80·0 m3, Sanalona, with 26 cages of size 8·64 m3, and Dique IV, with 50 cages of size 75 m3. The temperature, pH and dissolved oxygen of the water inside the cages were recorded from February to May 2009. For the ALM and Sanalona dams, the in situ temperature, pH and dissolved oxygen of the water were measured every 6 h for 24 h (diurnal variation) twice a month. For the Dique IV dam, daily measures of the water were performed.

Nile tilapia sampling

Ten diseased Nile tilapias were collected using a casting net from three floating cages every 15 days for each farm. Nile tilapias were considered to be diseased when they presented clinical signs such as lethargy, erratic swimming, skin discoloration, scale loss, lateral blindness, red or opaque eyes and/or exophthalmia. Live fish were placed in plastic bags filled with local water in ice boxes, and aeration was provided by pumping oxygen into the bags. The fish were transported to the laboratory as soon as possible. After 4 h, the ice cubes were put around the bags to lower the temperature. Mortality events occurring during the sampling period were recorded.

Culturing for bacteria

The Nile tilapias were first anesthetized with 2-phenoxyethanol (Sigma-Aldrich©, St. Louis, MO, USA) at 0·2 ml l−1 of water during 5 min, and then another 0·2 ml l−1 was added two times, until fish were killed. Each Nile tilapia was weighed and measured, and any external abnormalities were recorded. The ventral approach to internal organs was selected. In a laminar flow hood, the flank was first swabbed with antiseptic (70% ethanol) and then with sterile distilled water, avoiding the anal area and any skin lesions. The area was then dried with a sterile gauze pad. The body wall was cut with a sterile scissors, avoiding the anus and intestine. A sterile loop was used to obtain an inoculum from the kidney, liver, brain and spleen. External lesions were obtained in the surface of the fish near the lesion was disinfected with 70% ethanol, and a sterile loop was inserted under the scales or eyes. Inoculum was streaked on general and selective bacteriological media selected to ensure the recovery of a wide variety of bacteria (Whitman 2004). Trypticase soy (TSA; Bioxon, Mexico City, Mexico) supplemented with 0·5% glucose, brain–heart infusion (BHI, BBL, US) supplemented with 5% sheep blood (selective media to isolate of Aer. hydrophila, Staphylococcus sp. and Streptococcus sp.), MacConkey (MAC), glutamate–starch–phenol-red (GSP; Merck, Whitehouse Station, NJ, USA) and thiosulfate–citrate–bile–saccharose (TCBS, Bioxon) agars were used to isolate bacteria. All plates were incubated at 30 ± 1°C for 24–48 h. After incubation, only plates with dense culture growth were selected for examination for colony morphology, pigmentation, shape and haemolysis. TSA plating media were the most beneficial to isolate bacteria from lesions and internal organs of tilapia, followed by MacConkey and BHI media. Five to ten representative colonies from TSA and MacConkey media were first Gram-stained and examined microscopically (Carl Zeiss, Kirchdorf, Germany), and then three colonies were selected and further purified in TSA (30 ± 1°C for 24 h). The isolates were temporarily preserved in 1·5-ml plastic tubes filled with TSA and covered with sterile mineral oil. The tubes were stored in the dark for no more than 2 months. Selected strains were then cryopreserved at −80°C in tryptic soy broth (TSB; Bioxon) with 15% (v/v) glycerol.

Phenotypic characterization of the isolates

Phenotypic characterization of isolates was performed following the criteria described in Bergey's Manual of Determinative Bacteriology (Holt et al. 1994) for identification to the genus or species level. An API 20 NE bacterial identification kit was utilized according to manufacturer's instructions (bioMérieux®, Craponne, France).

Molecular identification by sequencing the 16S rRNA gene

DNA was extracted from isolates with the Wizard® DNA purification kit (Promega, Madison, WI, USA), according to the manufacturer's instructions. The DNA obtained was spectrophotometrically adjusted to 50 ng μl−1. DNA fingerprinting of strains was performed with repetitive extragenic palindromic elements (rep) PCR (rep-PCR) using the (GTG)5 primer (Gomez-Gil et al. 2004), and the PCR amplification was performed as described earlier (Wong and Lin 2001) with Taq® polymerase (Promega 5U) using a MyCycler™Thermal cycler (Bio-Rad©, Hercules, CA, USA). PCR products were electrophoresed in 2·25% 20 × 20 cm agarose gels (Promega) in TAE buffer 0·5 × for 260 min at 70 V and 4–8°C. The gels were stained with ethidium bromide for 30 min at room temperature and visualized after integration in a gel documentation system (UVP, Upland, CA, USA). The TIFF files obtained were analysed using Gel-Compar II software (version 4.5; Applied Maths, Austin, TX, USA). The similarity matrix was calculated with the Jaccard coefficient as recommended by Kosman and Leonard (2005), and the dendrogram was constructed using Ward (optimization 0·35, position tolerance of 1·82%). The 16S rRNA gene was amplified and sequenced as previously described (Gomez-Gil et al. 2010). In addition, 16S rRNA sequences were used for identification by BLAST with the EzTaxon-e server (http://eztaxon-e.ezbiocloud.net/; Kim et al. 2012).

Characterization of the virulence factors of the isolates

Bacterial ECPs were obtained following the cellophane plate technique described by Zhang and Austin (2000). The total protein concentration and the proteolytic and enzymatic activities of the ECPs were determined using the methods described by Soto-Rodriguez et al. (2012). The production of haemolysins was assayed on Müeller–Hinton agar (Oxoid, Basingstoke, England) supplemented with 5% (v/v) sheep blood; the plates were incubated for 24–48 h at 37°C. A commercial deoxyribonuclease test medium (Biolife, Milan, Italy) was used for the DNase test. Esculin hydrolysis was performed on TSA plates supplemented with 0·1% (w/v) esculin and 0·05% ferric ammonium citrate as an indicator. Cytotoxicity assays with ECPs obtained from strains were conducted as described by Toranzo et al. (1983) using HeLa cell lines (human cervical carcinoma) that were inoculated with 100 μl of ECPs. PBS was used as a negative control. The development of cytotoxic effects was observed at 6, 12, 24 and 48 h postinoculation (pi).

Virulence (bacterial cells and ECPs) of the strains

Filtered ECPs from the strains were subjected to a sterility test conducted via inoculation with 20 μl on TSA agar (Bioxon) and incubation overnight at 30°C. All ECP samples were stored at −20°C until used. Bacterial cells were recovered from the cryovials, cultured on TSB (Bioxon) and incubated at 30°C for 24 h. Bacterial suspensions were plated on TSA (Bioxon) and incubated overnight at 30°C. The resulting colonies were suspended in sterile PBS 0·85% NaCl at pH 7·0 and centrifuged at 5724 g for 10 min at 15°C. The bacterial suspension was adjusted to an optical density of 1·0 at 610 nm, with approximately 108 colonies forming units (CFU) ml−1 (Soto-Rodriguez et al. 2003), and serially diluted to achieve densities from 103 to 107 CFU ml−1. These suspensions were plated onto TSA after serial dilution to determine the real density of the isolates used in the challenges.

Three challenges were performed to evaluate the pathogenicity of the strains. The first trial was performed to screen the pathogenicity of the strains. It was conducted by intraperitoneal (IP) inoculation of sole (Solea vulgaris, 5–10 g in weight). Groups of five fish were inoculated with 100 μl per fish of ECPs from the following strains: C4 1 m, C4 3 m, C11 1 m, C1b 4 m, C5 10 m, C4 12 m, C7b 4 m and C2 6 m. The fish were held in 60-l aquaria at 22°C for 5 days. Groups of control fish were inoculated with 100 μl of sterile PBS (pH 7·2). All aquaria had individual aeration and free-flowing water. The mortality rate of each group was monitored over a 5-day period.

In the second trial, healthy Mozambican tilapia (O. mossambicus) were obtained from a local tilapia supplier and acclimated for 1 week prior to challenge. Fish (5–10 g in weight) in groups of 10, with two replicates, were IP-injected with washed bacterial cells in PBS 3·1 × 106 CFU g−1. In this trial, the same strains used in the first trial were tested as well as the strains: C2 7 m, C41 1 m, C2b 4 m and C2 5 m. Escherichia coli CAIM 21 2·0 × 108 CFU ml−1 was included as an innocuous bacteria to rule out mortality due to bacterial toxic shock. The fish were held in 56-l fibreglass aquaria filled with filtered (10 μm), UV-sterilized and aerated seawater at 28–30°C with a 12 h/12 h photoperiod and observed daily for pathological signs.

In the third trial, 100 μl of the ECPs obtained from the strains previously tested in the second trial was IP-injected into tilapia (O. mossambicus). The fish (5–10 g in weight) were in groups of 10, with three replicates, and groups of control fish were inoculated with 100 μl of sterile PBS (pH 7·2). The fish were fed ad libitum twice a day with a commercial diet (protein 33% Purina™, Culiacan, Sinaloa, Mexico). The experimental conditions were similar in the second trial. The mortality rate of each group was registered over 7 days for the second and third trials, and mortality due to strain inoculated was considered the cause only if a pure culture of the tested strain was obtained from the internal organs of fish (brain, heart, liver, kidney and spleen). A piece of each internal organ was inoculated on TSA plates and incubated at 30°C. DNA fingerprinting of isolates was performed with rep-PCR using the (GTG)5 primer (Gomez-Gil et al. 2004) and the inoculated strain as control.

Results

The temperature, pH and dissolved oxygen of the water inside floating cages were within the acceptable to optimal levels to grow Nile tilapia (El-Shafai et al. 2004; El-Sayed 2006). The mean water temperature inside the floating cages registered during the period of study was 25·3, 24·9 and 27·1°C for the Sanalona, Adolfo Lopez Mateos and Dique 4 dams, respectively (Fig. 1). Strains were mostly recovered from the kidneys and the brain and mainly from Nile tilapia cultured from the Sanalona dam (Table 1). Five strains belonged to the Aeromonas genus, and the second most abundant bacteria were Plesiomonas shigelloides. Diseased Nile tilapia had symptoms of erratic swimming, loss of scales, skin darkening, blindness, exophthalmia and red or opaque eyes.

Table 1. Bacterial strains isolated from the Nile tilapia Oreochromis niloticus farmed in floating cages from Sinaloa
StrainMolecular identification% Sequence similaritySize bpGenBank Acc.SourceHealth statusOrigin
  1. ALM, Adolfo Lopez Mateos dam.

  2. a

    Collection of Aquatic Important Microorganisms (www.ciad.mx/caim).

C1b 4 m (=CAIMa 1876) Aeromonas ichthiosmia 100·0922 KC660987 KidneyScale loss, lateral blindnessSanalona
C5 10 m (=CAIM 1873) Aeromonas dhakensis 99·631133 HQ663904 Eye injuredScale loss, darkeningSanalona
C4 12 m (=CAIM 1879) Aeromonas popoffii 99·42931 HQ663920 LiverExophthalmiaDique IV
C4 11 m Aeromonas allosaccharophila 98·92793 KC660986 KidneyRed eyesALM
C1 11 m (=CAIM 1877) Aeromonas veronii 99·671,539 HQ663903 BrainScale lossALM
C3a 8 m  Delftia tsuruhatensis 99·88817 KC706668 Eye injuredScale lossALM
C7 5 m (=CAIM 1875) Edwardsiella tarda 97·74929 HQ663902 KidneyDarkeningALM
C2 6 m (=CAIM 1880) Microbacterium paraoxydans 99·15942 HQ663907 LiverOpaque eyesDique IV
C1b 6 m  Plesiomonas shigelloides 99·89925 HQ663900 BrainOpaque eyesDique IV
C1b 7 m  Plesiomonas shigelloides 99·54880 KC706669 Eye injuredScale lossSanalona
C2 5 m  Plesiomonas shigelloides 97·501561 KC706670 KidneyScale lossSanalona
C2 7 m Pseudomonas anguilliseptica 99·21918 KC660988 BrainScale lossSanalona
C4 1 m (=CAIM 1881) Pseudomonas mosselii 99·321503 HQ663894 KidneyNormal appearanceSanalona
C2b 4 mSoonwooa sp.95·09986 KC660989 LiverScale loss, lateral blindnessSanalona
C7b 4 m Soonwooa buanensis 97·76939 HQ663901 KidneyExophthalmiaSanalona
C2 10 m  Sphingobacterium hotanense  99·14939 HQ663897 BrainOpaque eyesSanalona
C10 4 m  Vogesella perlucida  99·501526 KC706672 KidneyExophthalmiaSanalona
Figure 1.

Dissolved oxygen (mg l−1), pH and temperature (°C) of the water inside the floating cages at the Sanalona (a), Adolfo Lopez Mateos (b) and Dique IV (c) dams in Sinaloa, Mexico. The samples were collected from February to May 2009. (image_n/jam12280-gra-0001.png) Oxygen; (image) pH; (image) Temperature.

Biochemical characterization and virulence factors of the strains

The strains were mostly motile Gram-negative bacteria that were oxidase and catalase positive with fermentation of glucose, and the isolates grew well on TSA, MacConkey and BHI agar (Table S1). None of the ECP samples were able to hydrolyse esculin except for the Sphingobacterium hotanense C2 10 m strain. Most of the ECPs displayed cytotoxic effects, with a total destruction of HeLa cell monolayers at 12 h postinoculation except for the Plesiomonas shigelloides C1b 6 m and Vogesella perlucida C10 4 m strains. Haemolysis, DNAse, gelatinase and lipase activities of the ECPs were observed from Aeromonas ichthiosmia, Aeromonas dhakensis, Aeromonas popoffii, Aeromonas veronii C1 11 m and Soonwooa buanensis.

Virulence of the strains

The Aer. dhakensis C5 10 m, Ps. mosselii C4 1 m and Mic. paraoxydans C2 6 m strains were the dominant colonies of the dense cultures from the TSA, MacConkey and BHI agar. ECPs from these strains were toxic to the tested fish. The mortality ranged from 40% for the sole challenged with Mic. paraoxydans C2 6 m to 100% for tilapia challenged with the Ps. mosselii C4 1 m strain and Aer. dhakensis C5 10 m strain (Table 2). Solea vulgaris presented erratic swimming and skin darkening after exposure to the ECPs of these strains, whereas O. mossambicus presented similar signs only after exposure to the ECPs from the Ps. mosselii C4 1 m. Erratic swimming, lethargy, skin darkening, swollen abdomens and bloody intraperitoneal fluid in moribund fish were observed in tilapia injected with the ECPs from A. dhakensis C5 10 m.

Table 2. Percentage of mortality/time postinoculation (pi) of Solea vulgaris (trial 1) and Oreochromis mossambicus (trials 2 and 3) injected with bacterial cells and extracellular products (ECPs)
StrainTrial 1Trial 2Trial 3
ECPs 100 μl per fish3·1 × 106 CFU g−1ECPs 100 μl per fish
  1. n.d., not determined.

Aeromonas ichthiosmia C1b 4 m (=CAIM 1876)000
Aeromonas dhakensis C5 10 m (=CAIM 1873)60/24 h pi100/12 h pi100/12 h pi
Aeromonas popoffii C4 12 m (=CAIM 1879)000
Aeromonas allosaccharophila C4 11 mn.d.00
Aeromonas veronii C1 11 m (=CAIM 1877)000
Microbacterium paraoxydans C2 6 m (=CAIM 1880)40/48 h pi00
Plesiomonas shigelloides C2 5 mn.d.00
Pseudomonas anguilliseptica C2 7 mn.d.00
Pseudomonas mosselii C4 1 m (=CAIM 1881)80/48 h pi0100/6 h pi
Soonwooa sp. C2b 4 mn.d.00
Soonwooa buanensis C7b 4 m000
Escherichia coli (2·0 × 108 CFU ml−1)n.d.00

In trials with bacterial cells only, Aer. dhakensis C5 10 m caused 100% tilapia mortality in the first 12 h. The clinical signs of moribund fish challenged with bacterial cells and ECPs were similar. No mortality or clinical signs were observed in fish injected with bacterial cells or ECPs from other strains. No mortality was registered in the fish control groups from the three trials, neither in the fish injected with bacterial cells or the ECPs from E. coli CAIM 21.

Discussion

In the present study, isolates were recovered primarily from the kidneys and the brain of diseased Nile tilapia and, based on the biochemical characteristics and 16S rRNA gene sequence results, most of the strains belonged to the Aeromonas genus. The ECPs obtained from Aer. ichthiosmia, Aer. popoffii, Aer. veronii and Aer. dhakensis displayed characteristics of potential virulence factors, although only the last one was toxic. To date, there is not a clear correlation between the crude ECPs obtained from pathogenic bacteria and their virulence to tropical fish. For example, in tilapia hybrids, virulence of Aer. hydrophila is due to the presence of bacterial cells and ECPs (Rey et al. 2009). However, Pridgeon and Klesius (2011) observed that the virulence of Aer. hydrophila isolates challenged in catfish was primarily due to bacterial cells and not their ECPs. We observed that ECPs from Aer. dhakensis C5 10 m, Ps. mosselii C4 1 m and Mic. paraoxydans C2 6 m were toxic to fish and had potential fish virulence factors, many of which may work in concert.

Recent developments in the field of Aeromonas taxonomy based on genetic identification have led to a reclassification of aeromonads and to the description of new species (Janda and Abbott 2010). Beaz-Hidalgo et al. in 2013 reported two subspecies of Aerhydrophila, that is, hydrophila and ranae and proposed a reclassification of Aer. dhakensis and Aer. hydrophila subsp. dhakensis as Aeromonas dhakensis sp. nov. comb nov. using a multilocus phylogenetic analysis (MLPA) constructed with the concatenated sequences of gyrB, rpoD, recA, dnaJ and gyrA. Species Aer. dhakensis are widely distributed throughout warm countries. Environmental strains have been recovered from wild European eel suffering from haemorrhagic septicaemia (Esteve and Alcaide 2009), isolated from aquarium water and ornamental fish (Martínez-Murcia et al. 2008). Strains have been also isolated from river water, faeces and fish (Aravena-Román et al. 2011; Esteve et al. 2012). Clinical strains have been associated with a variety of human diseases worldwide (Huys et al. 2002; Figueras et al. 2009; Wu et al. 2012). Aeromonas dhakensis was misidentified originally as Aer. caviae and then as Aer. hydrophila, which contradicts the longstanding notion that Aer. caviae, Aer. hydrophila and Aer. veronii bv. sobria represent the most frequently isolated aeromonads (Janda and Abbott 2010; Aravena-Román et al. 2011). Aeromonas hydrophila as causative agent of MAS has been detected in freshwater fish as tilapia (Abd-El-Rhman 2009), trout (Lui et al. 2010) and catfish (Pridgeon and Klesius 2011). The virulence factors of this bacteria include secreted enzymes (Lui et al. 2010), cytotoxic enterotoxins (Chopra et al. 1993; Santos et al. 1998), haemolysins (Ljungh et al. 1981), siderophores (Barghouthi et al. 1989) and aerolysins (Howard and Buckley 1985).

Aeromonas dhakensis C5 10 m recovered from an injured eye was a Gram-negative motile bacteria, grew well on MacConkey, TSA and BHI agar and utilized sodium pyruvate and citrate as sources of energy. Bacterial cells and the ECPs of the Aer. dhakensis C5 10 m strain were equally toxic to both fish species tested. This strain contains virulence factors that are lethal to fish, including gelatinase, DNAase, lipase, cytotoxicity and haemolytic activity (Toranzo and Barja 1993). Aeromonas dhakensis (formerly Aer. hydrophila subsp. dhakensis) was pathogenic to rainbow trout within 48 h and their ECPs displayed gelatinase, haemolytic and lipase activity (Orozova et al. 2009). Numerous virulence genes associated with diarrhoea, bacteraemia and wound infections were detected in Aer. dhakensis strains isolated from humans and the egg masses of midge Chironomous sp. (Figueras et al. 2009; 2011). Clinical species of Aer. dhakensis isolated in Malaysia have been demonstrated to be the most prevalent species containing an important subset of virulence genes lip/alt/ser/fla/aer, which may be responsible for a wide range of infections (Puthucheary et al. 2012). In silico genomic analysis of Aerdhakensis AAk1, isolated as the sole pathogen from the blood of a patient with septicaemia, found virulence genes, including those encoding for a cytotonic heat-labile enterotoxin, elastase, lipase, flagella, cytotoxic enterotoxins, the type III secretion system and the AexT toxin delivered by this system (Wu et al. 2012). The mechanism of pathogenesis of aeromonads is complex and unclear because virulence is considered to be multifactorial (Janda and Abbott 2010).

Pseudomonas species have been recovered from diseased fish, including Ps. chlororaphis, Ps. fluorescens, Ps. putida, Ps. plecoglossicida and Ps. anguilliseptica, which is considered the most significant pathogen for cultured fish (Toranzo and Barja 1993; López-Romalde et al. 2003; Magi et al. 2009). Fluorescent pseudomonads, a group of Gram-negative, motile rods distributed in temperate as well as in tropical soils, predominate among the bacteria associated with plant rhizospheres (Sutra et al. 2001). Pseudomonas mosselii, a recognized species of the Ps. putida group, was first described by Dabboussi et al. (2002) as having rod-shaped cells producing a fluorescent pigment. The bacteria is an environmental species detected mainly in rhizospheric soil (Naik et al. 2008) and recently, was isolated from entomopathogenic nematodes (Tambong 2013). Overall, this bacterium is an unusual human opportunistic pathogen (Mclellan and Partridge 2009). Similar environmental species occasionally acting as opportunistic pathogens most likely play a role as shuttles for acquired metallo-β-lactamases genes from their as-yet-unknown natural reservoirs to the clinical setting (Giani et al. 2012). However, to date, Ps. mosselii has not been reported as fish pathogen. In this study, the Ps. mosselii C4 1 m strain recovered from the kidneys of O. niloticus was a Gram-negative motile bacteria, grew well on TSA, MacConkey and BHI agar and utilized sodium pyruvate and citrate as sources of energy. The strain's ECPs were lethal to sole and tilapia and contain potential virulence factors, such as gelatinase and cytotoxic and haemolytic compounds. Some fluorescent strains of Ps. mosselii possess antifungal activity, produce protease and pyoverdine siderophores (Meyer et al. 2002; Jha et al. 2009).

In addition, Microbacterium species belonging to the order Actinomycetales, known for their ability to produce antimicrobial compounds, were recently isolated from infective entomopathogenic nematodes (Tambong 2013). Such species are a rare cause of bacteraemia; however, Enoch et al. (2011) described an unusual case of central venous catheter-related infection due to Mic. paraoxydans. This species had been isolated from human clinical specimens (Gneiding et al. 2008) and from the liver of a laboratory mouse (Buczolits et al. 2008). ECPs from the Mic. paraoxydans C2 6 m strain recovered from the liver of the Nile tilapia in this study were pathogenic to sole. This strain was a Gram-positive motile bacteria, grew well on MacConkey agar and utilized sodium pyruvate as a source of energy. At present, this species has not been reported in environmental samples, and therefore, this is the first report of Mic. paraoxydans isolated from fish.

Here, we observed that the Aer. dhakensis C5 10 m strain and Ps. mosselii C4 1 m strain were fish pathogens, although 100 per cent mortality was registered only in tilapia. The virulence of Aer. dhakensis to tilapia is a novel observation and, as the species has been reported to be distributed throughout warm countries, therefore could be considered as a threat to aquaculture in Mexico. Mic. paraoxydans was pathogenic only to sole, and could be a potential fish pathogen. Apparently, the bacterial cells of Ps. mosselii and Mic. paraoxydans are not essential for pathogenesis in both fish. The virulence of these pathogenic strains might due to the presence of virulence genes encoding for the gelatinase, DNAase, lipase, haemolytic and cytotoxicity compounds. However, the virulence genes are not always expressed, and that expression can depend on the host–pathogen interaction. Therefore, research is needed to determine which virulence genes are present in those pathogenic strains and which genes are expressed during the infection process.

Acknowledgements

This study was in part supported by the Aquatic Animal Health Sinaloa State Committee (CESASIN in Spanish), Fundación Produce Sinaloa, and contributions from tilapia farmers. We thank Carmen Bolan and Francis Marrujo for technical assistance.

Ancillary