Isolation and molecular characterization of streptococcal species recovered from clinical infections in farmed Nile tilapia (Oreochromis niloticus) in the Philippines

Streptococcosis outbreaks occur in a wide range of farmed fish species, globally (Mishra et al., 2018). Outbreaks in farmed freshwater tilapia continue to threaten global production, contributing to severe economic losses encountered worldwide (Amal & Zamri-Saad, 2011; Li et al., 2015; Liu et al., 2018; Mishra et al., 2018). Numerous incidences of streptococcosis have been reported in intensive tilapia culture systems in South-East Asia and have caused tremendous financial damage since tilapia farming is the most important, if not globally, in the region (Kayansamruaj, Areechon, & Unajak, 2020). While a range of bacterial aetiological agents have been identified from streptococcosis infections in fish, by far the greatest causes of these diseases are the Gram-positive, Streptococcus agalactiae and S. iniae (Agnew & Barnes, 2007; Mishra et al., 2018; Zhou et al., 2008). Understanding the pathogenesis in aquatic outbreaks is complex, as Received: 8 June 2020 | Revised: 25 July 2020 | Accepted: 27 July 2020 DOI: 10.1111/jfd.13247

both of these bacterial species possess and express an array of virulence genes and factors related to their pathogenicity. Identification of virulence gene patterns can be used to determine the genetic diversity of fish isolates belonging to streptococcal species and provide understanding of the evolutionary relationship between pathogen virulence and host adaptation (Godoy et al., 2013). Investigation of the genetic and epidemiological relationship of piscine streptococci is important in finding the appropriate health strategies, which may be unique depending on the geographical area. To support the tilapia farming sector, the "holy grail" is to produce highly efficacious prevention and control measures against fish loses due to streptococcal infections. Disease outbreaks from both of these bacterial pathogens often present with similar clinical signs and can be confused with other bacterial infections. In addition, concurrent infections with both S. agalactiae and S. iniae (Conroy, 2009) can be seen on single farms. These factors have contributed to a slower development of effective prevention and control strategies within tilapia farming systems globally and have resulted in more fragmented approaches.
To reduce disease outbreaks, reliable data to discriminate between streptococcal species, their serotypes and genetic profiles are required. Uptake of molecular typing methods for S. agalactiae includes molecular typing by capsular polysaccharide (cps) gene, multi-locus typing (MLST) and genotyping of virulence genes present (Delannoy et al., 2013;Godoy et al., 2013;Kannika et al., 2017). Data from such studies have contributed towards improved knowledge on the aetiological identification, geographical distribution and molecular profile during disease outbreaks (LaFrentz et al., 2018). Currently, identification including molecular typing status of S. agalactiae and S. iniae affecting farmed tilapia in the Philippines is unknown. The objectives of the present study were to identify, characterize and profile streptococcal species recovered from clinically sick tilapia farmed in the Philippines, with a view to improving targeted disease prevention and control measures.

| Fish sampling
Fish were sampled from different tilapia farms that were currently suffering from a clinical disease outbreak. This was identified as morbidity/mortalities with external clinical signs as described in Table 3. Farmers were unable to provide exact data on mortalities, but the morbidities occurred all year. The sampling was performed in collaboration with the Fisheries Biotechnology Centre and Bureau of Fisheries and Aquatic Resources. A total of 106 fish were sampled during this study from 10 farms located in 3 geographical areas (Luzon, Visayas and Mindanao) of the Philippines ( Figure 1). Of these farms, nine were grow-out and one was a hatchery where only broodstock fish were sampled. These farms were representative of the varied tilapia farming systems practised in the Philippines. At each cage or pond of a farm, a total of 7 fish were sampled for bacterial recovery and histopathology including a minimum of 2 apparently healthy fish. All tilapia sampled in this study were from natural clinical outbreaks with a body weight range of 30-500 g and exhibited various clinical signs of disease. These included lethargy and erratic swimming, petechiae, discoloration, exophthalmia (uni-and bilateral) ( Figure 2) and corneal opacity. Internally, the moribund fish presented with enlarged spleen and kidney, ascites, gill pallor, brain haemorrhages and congested heart. The apparently healthy fish were judged as those that did not display any observable behavioural or clinical abnormalities. Samples of spleen, kidney and brain tissues from all fish were aseptically inoculated onto tryptone soya agar (TSA; Oxoid) and onto the selective Edward's medium (Oxoid) supplemented with colistin sulphate (5 mg/L). This medium is selective for the isolation of S. agalactiae. A laboratory-based pilot study was performed to confirm the specificity of Edward's medium for growth of Gram-positive S. agalactiae prior to use in the field study (data not presented). With the addition of liver and heart F I G U R E 1 Sampling sites of the different tilapia farms in the Philippines (number indicates the number of farm/s sampled in the area) [Colour figure can be viewed at wileyonlinelibrary.com] samples, sections of the same tissues were also preserved in 10% (v/v) neutral buffered formalin for histopathology processing and screening. The inoculated agar plates were incubated at 28°C for 48 hr. Single colony subcultures were performed onto TSA to obtain pure isolates prior to identification. These were then stored using Protect Beads (Thermo Scientific, UK) at −70°C, and working stocks of 1 ml aliquots in 15% (v/v) glycerol were stored at −20°C.

| Histopathology
The tissues were fixed in 10% neutral buffered formalin for 24 hr before processing as described by Del-Pozo, Crumlish, Turnbull, and Ferguson (2010). Once embedded in paraffin, duplicate 5-µm-thick sections were cut and stained with haematoxylin and eosin and tissue Gram stain with crystal violet and counterstained with 1% neutral red, following routine laboratory methods. The sections were examined under light microscopy at up to 100× magnification and images captured using a digital slide scanner (ZEISS Axio Scan.Z1, ZEISS Germany).

| Bacterial identification
All pure cultures recovered were identified using traditional bacteriology methods including Gram stain, oxidase, catalase and motility (Barrow & Feltham, 2003). Additional tests included measuring haemolysis on 5% sheep blood agar (SBA), starch hydrolysis using the method of Cowan and Steel (Barrow & Feltham, 2003) and capsule formation following the protocol of Anthony in Cowan and Steel (Barrow & Feltham, 2003). Presumptive identification of the bacteria based on the above tests was accomplished following Buller (2004).

| Bacterial DNA extraction and identification using species-specific duplex PCR and 16S rRNA sequencing
DNA was extracted from single purified colonies following a crude boiling method with modification from Seward, Ehrenstein, Grundmann, and Towner (1997). The concentration of the DNA extracts was measured by NanoDrop (Thermo) spectrophotometer.
The DNA samples were stored in sterile tubes in 20 µl aliquots at −20°C until required.   To complement the duplex-PCR results, all S. iniae isolates identified from the duplex PCR as S. iniae (n = 7) and S. agalactiae (n = 2)

F I G U R E 2
were randomly selected and processed for 16S rRNA PCR sequence analysis. The two S. agalactiae selected were S. agalactiae NFFTC and S. agalactiae BSMF1. The 16S rRNA gene was PCR-amplified using universal primers 20F (5′-AGAGTTTGATCATGGCTCAG-3′) and 1500R (5′-CGGTTACCTTGTTACGACTT-3′) (Weisburg, Barns, & Lane, 1991) which amplifies approximately 1501-bp region of the gene. Sequences were aligned with ClustalW algorithm against phylogenetically related organisms available in GenBank in the National Centre for Biotechnology Information website (http://www.ncbi. nlm.nih.gov).

| Molecular serotyping of Streptococcus agalactiae isolates
Molecular serotype based on the cps gene was determined using a multiplex PCR described by Imperi et al. (2010) with minor modifications. Instead of 19 primers, seven primers (Table 1) were used to amplify serotypes Ia, Ib, II and III as these were found to be the main serotypes infecting fishes (Delannoy et al., 2016). The S. agalactiae

| Antibiotic susceptibility assay
All the isolates were tested for antibiotic sensitivity to amoxicillin

| Detection of virulence genes
Isolates were screened for a total of 6 and 7 virulence genes for the S. iniae and S. agalactiae, respectively (Table 2)

| Fish sampling
A total of 106 tilapia were sampled during this study which represented 68 moribund fish presenting with a range of clinical signs of disease and 38 were observed as apparently normal as judged by naked eye. A total of 25 Streptococcal isolates were recovered from 24 (35%) moribund tilapia which presented a wide range of clinical signs, typical of streptococcosis in fish (Table 3). No streptococcal species were isolated from apparently healthy tilapia from any of the farms visited. The streptococcal species were recovered from 7 grow-out floating cages, 1 hatchery and 1 earthen pond, all freshwater, and included in 3 sites in Taal Lake, 1 site in Calauan, Laguna, and 1 site in Nueva Ecija, Philippines (Table 3). Only S. agalactiae were recovered from the brain samples of the fish (Table 3), and only in 1 site, both bacterial species were recovered (Table 3). A total of 7 S. iniae and 18 S. agalactiae were isolated from the moribund tilapia within this study (Table 3). From a single moribund tilapia in Taal Lake Site 2 cage 2, we recovered S. agalactiae serotypes Ia and Ib from the spleen and brain tissue, respectively (Table 3). From the remaining 53 fish sampled in this study, either no bacteria were recovered (apparently normal) or only Gram-negative or bacteria were isolated.
These were not included further in this study.

| Histopathology
A range of histopathology changes were observed from the mori- agalactiae Ia presented high number of bacteria n the sampled internal organs most notably in the atrium of the heart.

| Phenotypic observations and identification of streptococci
Bacterial recovery on TSA plates showed colonies were small, whitish-grey and smooth-edged, while in Edward's medium the colonies were small, blue to blueish colour with a white smooth edge. This was typical appearance on the agar types. Three isolates on TSA media produced very small colony variants (SCVs) that did not in-  (Table 4).

| Molecular identification and virulence profile of Streptococcal species
The species-specific duplex PCR discriminated S. iniae from S. agalactiae, where a single band was observed at 870 bp for S. iniae and at 270 bp for S. agalactiae (Figure 4, Table 5). Good correlation was found between the 16S rRNA sequence results and the duplex-PCR results for S. iniae and S. agalactiae strains ( Table 4).
All six virulence genes (simA, scpI, pgm, cpsD, pdi and sagA) were absent for gene cylE. This led to 2 virulence profiles for the S.
agalactiae strains: profile I (all virulence genes present) and profile II (one virulence gene absent).

| Molecular serotyping of Streptococcus agalactiae isolates
From the 18 isolates identified as S. agalactiae, 15 (83%) isolates were identified as serotype Ia, while the remaining 3 (17%) were identified as serotype Ib. Serotype Ia was indicated by the presence of cpsL (688bp) and cpsG (272 bp), while serotype Ib by cpsL (688bp), cpsJ (621 bp) and cpsG (272 bp) ( Figure 6, Table 6). Serotype Ia and Ib shares the presence of cpsL gene. No other serotype was detected in this study.

| Antibiotic susceptibility
The antibiotic susceptibility showed that 100% (n = 7) of S. iniae isolates were resistant only to oxolinic acid only and susceptible to all other antibiotics tested. For the S. agalactiae isolates, 100% (n = 18) were resistant to oxolinic acid and 17% (n = 18) were resistant to sulphamethoxazole trimethoprim. They were susceptible to all other antibiotics tested.

| Streptococcus agalactiae typing profile
Based on the phenotypic and molecular characterization, a typing scheme (  (Li et al., 2014;Su et al., 2019) where S.
agalactiae is the primary streptococci species associated with fish disease. This observed prevalence of S. agalactiae over S. iniae is similar to reports worldwide, especially in the major tilapia-producing regions in Asia and Latin America . Over the last 20 years, tilapia production in the Philippines has declined, and several contributory factors have been identified (Guerrero, 2019). The results from this study would confirm that infectious disease outbreaks are a contributing factor to the decline in tilapia production in the Philippines.
A wide range of identification methods have been applied to differentiate different streptococcal species and confirm serotypes in S. agalactiae populations occurring in farmed tilapia (Barkham et al., 2019;Mishra et al., 2018). In this study, we found excellent agreement between the duplex species-specific PCR and the 16S  Su et al., 2019;Syuhada et al., 2020). In this study, cps gene serotyping was applied which is used widely by others as a reliable way to serotype S. agalactiae (Kannika et al., 2017). While we recognize the limited sample size in this study, our findings were similar to surrounding SEA countries, but further characterization is needed to determine the sequence types, genetic and proteomic profiles, and exposure/transmission routes of the serotypes which will be vital for future vaccination programmes in the Philippines.
Histopathological changes in this study were similar to previous reports for streptococcal infections (Ferguson, 2006). In this study, it was not possible to differentiate histologically between S. iniae and S. agalactiae Ib. However, there were non-significant  in the atrium of the heart and in capillaries of the optic lobes of the brain; although a mild meningitis was consistently seen, there was no meningoencephalitis. Our data are consistent with previous histological studies on streptococcal infections in tilapia (Chen, Chao, & Bowser, 2007;Ferguson, 2006). The histopathology results obtained in this study support the highly invasive ability of S. agalactiae serotype Ia and its ability to cause acute infection compared with S. iniae which is more chronic (Chen et al., 2007;Ferguson, 2006). While a comparative or sequential pathology study was not performed, the histological results provided support the need for a more refined understanding of the complex pathogenesis of streptococcal infections in fish.
In this study, all three functional categories of virulence genes namely adhesins, invasins and immune evasins were present in all isolates revealing their pathogenic and invasive abilities. Identical virulence gene profiles were found for all the S. iniae isolates recovered from infected fish which may suggest that they have arisen from a single clone as these strains were all recovered from a single geographical location. In the S. agalactiae isolates, only serotype Ib lacked the cylE gene. This is the structural gene involved in β-haemolysis/cytolysis of the red blood cells (Pritzlaff et al., 2001). The lack of cylE gene was confirmed by the lack of haemolysis expressed in the 3 non-haemolytic S. agalactiae isolates tested when grown on sheep blood agar. The serotype Ia isolates from Thailand and Vietnam (Kayansamruaj et al., 2019) share the same patterns of virulence genes with the Philippine isolates. Studies have shown that β-haemolysin is a virulence factor that influences S. agalactiae survival in macrophages (Doran, Liu, & Nizet, 2003;Sagar et al., 2013) and promotes infection of the less or non-haemolytic strains by their ability to evade the host immune response and remain dormant inside macrophages until suitable conditions for their reactivation. The presence of cylE gene in S. agalactiae serotype Ia is believed to promote invasiveness (Chu et al., 2016) supporting rapid spread in the bloodstream and organs of the infected host, while its absence is considered a factor in the development of more chronic infection in fish (Li et al., 2014). For the non-haemolytic S. agalactiae serotype Ib isolates, the observed numerous bacteria-filled macrophage cells in the brain resulting in meningoencephalitis were indicative of a more chronic type infection.
In the Philippines, the use of antibiotics in tilapia aquaculture is not common because of the prevailing view that tilapia is resistant to diseases. However, recent episodes of disease outbreaks leading to mass mortality have led to some fish farmers starting to use antibiotics particularly amoxicillin which is administered orally mixed with feed. The antibiotic susceptibility profile of the Philippine strains was similar to previous reports in Thailand (Dangwetngam, Suanyuk, Kong, & Phromkunthong, 2016;Jantrakajorn et al., 2014;Kannika et al., 2017;Klingklib & Suanyuk, 2017).
While in general the results from this study are consistent with previous studies for SEA countries, this is the first report of the appearance of small colony variants (SCVs) from fish S. agalactiae strains. Small colony variants are a slow-growing subpopulation of bacteria with distinctive phenotypic and pathogenic traits that are involved in chronic and recurrent infections (Proctor et al., 2006).
In this study, after second subculture of S. agalactiae three isolates (BSMF1, BSMF4-2 and SMF1) exhibited very slow growth and pinpoint colonies that did not change in size, which were considered as SCV. These SCVs all belonged to serotype Ib, were non-haemolytic, cylE gene-deficient and SXT-resistant strains. Currently, there have been no reports of SCV appearance of piscine S. agalactiae, but in non-fish S. agalactiae strains reporting the appearance of SCV morphology, they are associated with subacute, recurrent and chronic infections, reduced antibiotic susceptibility and resistance to oxidative burst (Banno et al., 2014;Painter, Hall, Ha, & Edwards, 2017;Proctor et al., 2006). The piscine S. agalactiae SCV strains in this study are similar to human S. agalactiae SCV strains in that they both are non-haemolytic or have reduced haemolytic activity, small pinpoint colonies and resistance SXT (Banno et al., 2017). A change from wild-type phenotype to SCV could be a survival strategy for S. agalactiae serotype Ib fish strains similar to strategies used by SCV of human S. aureus and Salmonella sp. (Proctor et al., 2006).
It is clear that as with other SEA countries, streptococcal infections cause disease and fish losses within the tilapia production systems in the Philippines. Similar to previous reports from neighbouring countries, both S. iniae and S. agalactiae are present, able to cause disease with clinical signs similar to those previously reported, and S. agalactiae type Ia was the most prevalent pathogen. This study is the first to confirm that a range of streptococcal species is causing disease outbreaks in tilapia farms in the Philippines and that uptake of these data will better inform the disease prevention and control strategies for the Philippine sector and contribute towards a geographically distinct vaccine.

ACK N OWLED G EM ENTS
We thank the staff of Fisheries Biotechnology Centre-NFRDI (Jean

CO N FLI C T O F I NTE R E S T
The authors whose names are listed immediately below certify that they have NO affiliations with or involvement in any organization or entity with any financial interest (such as honoraria; educational grants; participation in speakers' bureaus; membership, employment, consultancies, stock ownership or other equity interest; and expert testimony or patent-licensing arrangements) or non-financial interest (such as personal or professional relationships, affiliations, knowledge or beliefs) in the subject matter or materials discussed in this manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
No supporting data have been shared in any public repositories.