Isolation and characterization of two virulent Aeromonads associated with haemorrhagic septicaemia and tail-rot disease in farmed climbing perch Anabas testudineus

Diseased Anabas testudineus exhibiting signs of tail-rot and ulcerations on body were collected from a fish farm in Assam, India during the winter season (November 2018 to January 2019). Swabs from the infected body parts were streaked on sterilized nutrient agar. Two dominant bacterial colonies were obtained, which were then isolated and labelled as AM-31 and AM-05. Standard biochemical characterisation and 16S rRNA and rpoB gene sequencing identified AM-31 isolate as Aeromonas hydrophila and AM-05 as Aeromonas jandaei. Symptoms similar to that of natural infection were observed on re-infecting both bacteria to disease-free A. testudineus, which confirmed their virulence. LC50 was determined at 1.3 × 104 (A. hydrophila) and 2.5 × 104 (A. jandaei) CFU per fish in intraperitoneal injection. Further, PCR amplification of specific genes responsible for virulence (aerolysin and enterotoxin) confirmed pathogenicity of both bacteria. Histopathology of kidney and liver in the experimentally-infected fishes revealed haemorrhage, tubular degeneration and vacuolation. Antibiotic profiles were also assessed for both bacteria. To the best of our knowledge, the present work is a first report on the mortality of farmed climbing perch naturally-infected by A. hydrophila as well as A. jandaei, with no records of pathogenicity of the latter in this fish.


Results
Morphological and biochemical characterization. Dominant colonies were obtained of the two bacterial isolates, AM-31 and AM-05, which displayed circular morphology with average diameter of 2-3 mm and appeared yellowish on nutrient agar. Both isolates were motile, Gram-negative, oxidase positive, catalase positive, O/129 resistant and produced gas on glucose fermentation. All tested biochemical characteristics of both isolates are presented in Table 1.  Table 1. Selected biochemical characteristics of two bacterial isolates (AM-31 and AM-05) retrieved from swabs of infected body parts and kidney of naturally-infected Anabas testudineus collected from a fish farm of Assam, India (ND not determined, + positive, − negative). ). An almost-complete 16S rRNA and rpoB sequence containing < 1% undetermined positions was obtained for both the isolates. The gene sequences determined for the two representative isolates in this study were deposited in GenBank (Accession numbers: MN097841 and MN204041 for 16S rRNA gene, and MN977195 and MN977194 for rpoB gene). The phylogenetic tree (Fig. 1) clustered the isolate AM-31 with Aeromonas hydrophila and isolate AM-05 with A. jandaei, respectively. These clusters were also strongly supported by their high bootstrap values.

Molecular identification based on
Clinical signs and virulence. Intraperitoneal injection of AM-31 and AM-05 isolates in disease-free Anabas testudineus individuals resulted in ulcerations and fin-rot, similar to as observed in natural infections (Fig. 2). The fishes developed symptoms of excessive mucus secretion, followed by gradual development of grayish-white lesions towards the posterior half of body that later extended to the caudal fin. The anal region exhibited scale loss, which then developed into ulcers, and the fins appeared reddish that later formed conspicuous fin rot. All clinical signs commenced within 48 h of artificial infection.
Mortality was recorded within 15 days, and the calculated lethal concentrations (LC 50 ) for AM-31 and AM-05 were 1.3 × 10 4 and 2.5 × 10 4 CFU/fish, respectively ( Table 2). The infected fish showed severe haemorrhage and blood congestions notably in the liver (Fig. 3). However, no mortality was recorded in the control groups (i.e. the PBS-injected and without injection).
Hemolysin assay. Positive haemolytic activity (yellow colour) was observed for both isolates with halo diameters of 0.9 mm (AM-31) and 1.4 mm (AM-05), respectively.   Histological changes in kidney and liver. Histopathological examination of the experimentally-challenged A. testudineus showed haemorrhages and vacuolar degeneration of renal tubules of kidney and severe vacuolation in liver (Fig. 4).

Discussion
Based on the combination of biochemical and molecular characterization, the present study confirms the occurrence of Aeromonas hydrophila (isolate AM-31) and A. jandaei (isolate AM-05) in naturally-infected farmed Anabas testudineus with visible signs of tail-rot and ulceration on body.
Starch hydrolysis is one of the important tests to identify species of Aeromonas 18 . As A. hydrophila and A. jandaei both tested positive to starch hydrolysis (undetermined so far), this biochemical property will help in improved characterization and identification of the two species in future. Sorbitol fermentation and (lactose + urea) assimilation were found positive for both species, which is also an important phenotypic  www.nature.com/scientificreports/ characteristic to distinguish mesophilic A. hydrophila from its complex members [19][20][21] . A. hydrophila differs from A. jandaei in characteristic of its sucrose and esculin hydrolysis positivity. Besides sucrose negative, A. jandaei is mannitol-positive; which further supports its earlier description as a genospecies DNA group 9 A. sobria, later re-named as A. jandaei 14 . Molecular markers like 16S rRNA and rpoB have wide uses in identification, characterization and measurement of microbes. Sequencing of these two housekeeping genes has proven to be a useful tool in species delineation of the genus Aeromonas, and previous studies have successfully identified Aeromonads using the same [22][23][24][25] .  Virulence-associated factors in bacteria may include secreted enzymes 9 , cytotoxic enterotoxins 26,27 , hemolysins 28 , siderophores 29 and aerolysins 30 , respectively. The positive response to haemolysin activity by A. hydrophila as well as A. jandaei in the present study confirms the toxicity of these isolates. Earlier studies on haemolytic activity of members of Aeromonas, too, reported toxicity to be strongly associated with enterotoxin production 31 , and that, haemolytic and cytotoxic activities could frequently relate a predominant bacteria to A. hydrophila 32 .
The pathogenicity tests were performed to analyse the capacity of A. hydrophila and A. jandaei to induce disease in healthy A. testudineus. The tests indicated their virulence to Anabas and the clinical signs of artificially-infected fish were similar to those observed in the naturally-infected individuals. Re-isolation of the same bacteria from experimentally-infected fish also fulfilled Koch's postulates 33 . But, the nature of symptoms and the percentage mortality caused by the bacterial isolates at varied doses of intraperitoneal injection suggest that low concentrations of bacterial load (or with minimal stress factors) in the surviving environment leads to mild pathogenicity. Similar findings were reported in Nile tilapia re-challenged with A. veronii and A. jandaei 13 . Pathological lesions and mortality rate were milder in natural infection than in fishes infected artificially. Confinement of the fishes in aquaria during the experimental infection with increasing bacterial load on the host fish may have played a vital role to exhibit such symptoms.
Screening for virulence-associated genes is one of the rational approaches in determining whether Aeromonas strains have the potential to be virulent 50 . Present study showed that A. hydrophila (isolate AM-31) and A. jandaei (isolate AM-05) were positive for the enterotoxin gene, which corroborates with the findings of Pablos et al. 35 . Further, this result is also concurrent with the haemolytic activity of these bacteria discussed earlier and, thus, both proving to be potent pathogens of A. testudineus. Contrary to the positive detection of aerolysin gene (aerA) in A. jandaei, A. hydrophila was, however, aerA negative. Although aerA gene is closely related to haemolytic, cytotoxic and enterotoxic activities of enterotoxin gene as reported in most Aeromonas strains 32 , the possible contradiction in the present study may be due to variations in prevalence of aerA in Aeromonas with respect to different geographical locations 36 .
Several studies reported that chronic infections of A. hydrophila cause dermal ulcerative lesions with focal hemorrhages and inflammation. Haemorrhage and vacuolar degeneration in renal tubules of kidney in experimentally-challenged specimens corroborates with similar findings made on Oreochromis niloticus experimentallychallenged with A. hydrophila 7 . Marked cytoplasmic vacuolations in liver tissues also correlate with previous studies on vacuolar degeneration in the parenchymatous organs and necrosis and haemorrhages caused by non-proliferative bacteria in teleosts 37 . These alterations may be toxin-induced owing to the presence of the enterotoxin gene in both bacterial isolates as discussed above.
AM-31 and AM-05 both, showed variable degrees of antibiogram resistance. Antibiotic resistance to penicillin and ampicillin as revealed for AM-31 isolate from the antibiogram study correlates with previous reports on multiple drug resistance of A. hydrophila 38 . However, A. jandaei was susceptible to all the antibiotics tested. Both bacteria exhibited high sensitivity for gentamycin and norfloxacin among all other antibiotics tested. Higher sensitivity for specific antibiotics (or the lack of any resistance towards them) may be due to the lack of awareness among local fish farmers within the Assam state (or northeast India) on the possible use of antibiotics towards disease prevention or treatment. Due to poor economic background, the farmers are more dependent on the traditional use of lime (calcium carbonate) and/or potassium permanganate as water disinfectant or disease treatment. Thus, antibiotic resistance may not have reached an alarming stage yet in these areas as compared to other regions of the world. Variation in sensitivity and resistance pattern may be partly due to different isolation sources and environmental conditions as well 39 . However, several antibiotics used to characterize the bacteria in our study, except florfenicol, erythromycin and sulphonamide group, are prohibited for use in aquaculture 40 .
In conclusion, the isolation of Aeromonas hydrophila (AM-31) and A. jandaei (AM-05) from diseased A. testudineus, their role in pathogenicity by re-infection and the presence of virulent genes corroborates the active contribution of motile Aeromonads towards pathogenesis in cultured fish species. We confirm that A. hydrophila and A. jandaei can infect farmed A. testudineus, evident as visible ulcers and haemorrhage on the body and fins as well as tissues within. To the best of our knowledge, this is a first record on the mortality of farmed Anabas naturally-infected by A. hydrophila as well as A. jandaei, with no available records on pathogenicity of the latter bacterium in this fish. As culture of A. testudineus in India is gaining momentum given its high market demand, the present study will act as baseline data on the pathogenic role of A. hydrophila and A. jandaei in this fish, opening up future scopes for investigation on their possible transmission (or opportunistic role) onto other Bacterial isolation. Live specimens (N = 50) of naturally-infected Anabas testudineus, exhibiting haemorrhagic ulcerations and tail-rot (see Supplementary Fig. S1 online), were collected from a local fish farm (26°27ʹ21″ N; 91°22ʹ53″ E) in Assam (India) between November 2018 and January 2019. The moribund Anabas individuals were euthanized using an overdose of clove oil prior-to-use for bacterial isolation and histopathological studies. After sedation, diseased fishes were aseptically washed with 70% ethyl alcohol and the ulcerated areas treated with a heated scalpel in order to reduce contamination. After sterilization, swabs were taken with a sterilized loop from the ulcerated body parts as well as kidneys and streaked on sterilized nutrient agar (NA) (Himedia) plates. The NA plates were then incubated at 29 (± 2) °C for 48 h and the growth of bacterial colonies (if any) was observed. Two dominant growing bacterial colonies were selected and purified through subculture by multiple streaking on NA medium. These were then marked and labelled as two isolates: AM-31 and AM-05. The isolates were further processed for biochemical and molecular identification. The bacterial isolates were also cultured in LB broth (Himedia) containing 10% glycerol and then kept in − 80 °C for long time storage.

Morphological and biochemical characterization of bacterial isolates.
Pure cultures of the AM-31 and AM-05 isolates were inoculated on NA medium overnight at 29 °C and the colony morphology was observed. The isolates were subjected to Gram staining followed by microscopic observation. Sub-culturing of both isolates in LB at 29 °C was done prior to biochemical characterization as per Martin-Carnahan et al. 42 .
Probable species level identification of the bacterial isolates followed Popoff and Véron 43 and Abbott et al. 44 .

Molecular identification based on 16S rRNA and rpoB gene amplification. The genomic DNA
of both isolates AM-31 and AM-05, cultured in LB at 29 °C overnight in an orbital shaking incubator, was extracted using QIAamp DNA Mini Kit (Qiagen, Germany) following the manufacturer's protocol and stored at − 20 °C until use. The 16S rRNA gene was amplified using universal bacterial primers according to the standard protocol of Martin and Collen 45 and the rpoB (RNA polymerase beta subunit) gene was amplified following the protocol of Liu et al. 46 , respectively. PCR reactions for each gene comprised of 25 µl ready-to-use master mix (R2523-100RXN, Sigma), 2.5 µl each of the forward and reverse primers and 5 µl DNA templates, and the final volume was made to 50 µl using nuclease-free water. A negative control (without DNA template) was taken as well. The qualities of the PCR-amplified products for both genes were checked by 1% (w/v) agarose gel (containing ethidium bromide) electrophoresis in TBE buffer. The PCR products were then purified using QIA quick Gel Extraction Kit (Qiagen, Germany) following the manufacturer's protocol, and outsourced to AgriGenome Labs Private Limited, Cochin (India) for Sanger sequencing. Sequences obtained for the gene datasets were edited using Clustal W (inbuilt in MEGA 7) 47 and a BLAST search was performed in the National Centre for Biotechnology Information (NCBI) database to identify their nearest neighbour(s). Phylogenetic trees were constructed from evolutionary distances using the unweighted pair group method with arithmetic mean (UPGMA) with Kimura 2-parameter (K2P) evolutionary sequence models. The tree branches were authenticated by bootstrap analyses of 1000 replicates. Pseudomonas sharmana (NR 043470.1) and P. pectinilytica (NR 156860.1) were taken as out-groups for rooting trees of the 16S rRNA dataset.

Pathogenicity test.
To test the pathogenicity of isolates AM-31 and AM-05, disease-free, healthy individuals of A. testudineus (N = 160; weighing approximately 22-40 g) were collected from aquaculture farms of the Barpeta District of Assam, India, with a constant practice of carp, murrel, catfish and/or Anabas culture. After collection, the fishes were acclimatized in 80 L glass aquaria (@ 30 individuals in each aquaria) 14 days priorto-challenge. All disease-free individuals were anaesthetized in a 50 ppm clove oil solution following Shameena et al. 41 , with a recovery time of 4-5 min, prior to proceeding with the experimental infections. AM-31 and AM-05 isolates were cultured in LB at 30 °C overnight the day before the experimental challenge with continuous shaking. The isolates were then centrifuged and suspended in PBS at six different concentrations of each as follows-for AM-31, the concentrations ranged from 3.4 × 10 2 -3.4 × 10 7 CFU/fish; while for AM-05, the concentrations ranged from 4.2 × 10 2 -4.2 × 10 7 CFU/fish, respectively (CFU values were calculated from conventional plate counting) ( Table 2).
An experimental setup was designed comprising 12 groups. Each group was represented by 10 individuals of A. testudineus in a 30 L aquarium. Of these, six groups were separated for the bacterial isolate AM-31, and 0.1 ml of each of the six prepared AM-31 bacterial suspensions was injected intraperitoneally to the individuals of each group in order of their increasing concentrations. The same experiment was conducted for the remaining six groups with the prepared suspensions of the bacterial isolate AM-05.
Also, two groups were prepared as control against the bacterial isolate AM-31, both comprising of 10 individuals of A. testudineus in 30 L aquaria. Of these, one group was injected with 0.1 ml of PBS to each individual; while the other group was maintained as without any injection. Similarly, two additional groups (each with 10 individuals) were prepared as control against the isolate AM-05.
Clinical signs and mortality (if any) of fish in all the respective groups were then recorded everyday up to 15 days post challenge. Detection of virulence-associated genes. PCR amplification of the virulence-associated genes [aerolysin (aerA) and enterotoxin (act)] from the genomic DNA was performed using different set of primers, amplification conditions and the expected product sizes (Table 4).
Histopathological study. Kidney and liver tissues from Anabas post artificial infection with the bacterial isolates (@ dozes of 3.4 × 10 6 CFU/fish for AM-31 and 4.2 × 10 6 CFU/fish for AM-05; each with 100% morality) were fixed in 10% neutral buffer formalin (NBF) for histological studies. Tissues were then subjected to dehydration through a series of increasing alcohol grades, cleared with xylene and embedded in paraffin following standard method. The paraffin-embedded tissues were sectioned at 5 µm using a microtome and stained with haematoxylin and eosin 49 . Pathological changes manifested in the tissue sections were examined under a Leica DM 3000 microscope at 40 × magnification.

Antimicrobial susceptibility test.
Antibiotic sensitivity or resistivity was tested on Mueller-Hinton agar plates inoculated with 0.1 ml of overnight broth culture of both isolates (1 × 10 7 CFU/ml of AM-31 and 1 × 10 6 CFU/ml of AM-05) following Bauer's disc diffusion method 50 . Antibiotic-impregnated discs (Himedia, India), comprising of 20 antimicrobials, were placed on the solid medium and the plates were incubated at 30 °C. After 24 h, the antibiotic sensitivity was determined by measuring the diameter (in mm) of the zone of inhibition formed around the disc, which were expressed as mean ± standard deviation of five replicates. The radius of zone of inhibition was scaled from the centre of the antibiotic disc to the end of the clear zone where bacteria could be seen growing. Results were interpreted on the basis of zone diameter as sensitive and resistant following Clinical and Laboratory Standards Institute 51 .
Statistical analysis. The mortality data obtained were subjected to Probit Analysis following Finney 52

Data availability
All data generated during the current study are included in this article and its Supplementary Information files. The nucleotide sequences are freely accessible at the NCBI database (Accession numbers: MN097841, MN204041, MN977195 and MN977194).