Molecular evidence for homologous strains of infectious spleen and kidney necrosis virus (ISKNV) genotype I infecting inland freshwater cultured Asian sea bass (Lates calcarifer) in Thailand

Infectious spleen and kidney necrosis virus (ISKNV) is a fish-pathogenic virus belonging to the genus Megalocytivirus of the family Iridoviridae. In 2018, disease occurrences (40-50% cumulative mortality) associated with ISKNV infection were reported in grown-out Asian sea bass (Lates calcarifer) cultured in an inland freshwater system in Thailand. Clinical samples were collected from seven distinct farms located in the eastern and central regions of Thailand. The moribund fish showed various abnormal signs, including lethargy, pale gills, darkened body, and skin hemorrhage, while hypertrophied basophilic cells were observed microscopically in gill, liver, and kidney tissue. ISKNV infection was confirmed on six out of seven farms using virus-specific semi-nested PCR. The MCP and ATPase genes showed 100% sequence identity among the virus isolates, and the virus was found to belong to the ISKNV genotype I clade. Koch’s postulates were later confirmed by challenge assay, and the mortality of the experimentally infected fish at 21 days post-challenge was 50-90%, depending on the challenge dose. The complete genome of two ISKNV isolates, namely KU1 and KU2, was recovered directly from the infected specimens using a shotgun metagenomics approach. The genome length of ISKNV KU1 and KU2 was 111,487 and 111,610 bp, respectively. In comparison to closely related ISKNV strains, KU1 and KU2 contained nine unique genes, including a caspase-recruitment-domain-containing protein that is potentially involved in inhibition of apoptosis. Collectively, this study indicated that inland cultured Asian sea bass are infected by homologous ISKNV strains. This indicates that ISKNV genotype I should be prioritized for future vaccine research.


Introduction
Asian sea bass (Lates calcarifer Bloch, 1790), also known as barramundi, is a species native to Thailand and widely distributed in the Indo-West Pacific region from the Persian Handling Editor: Kalpana Agnihotri * Pattanapon Kayansamruaj pattanapon.k@ku.th 1 3 Gulf to China, Taiwan, and northern Australia [14]. Asian sea bass is one of the most important marine cultured finfish in Australia and Asian countries including Indonesia, Singapore, Vietnam, and Thailand. Because of its significant economic potential, this fish species is expected by Thai governmental bodies and the private sector to be a major food fish along with tilapia (Oreochromis spp.) and hybrid catfish (Clarius gariepinus × C. batrachus). The total production of the Asian sea bass in Thailand was approximately 39,500 tons in 2018 [9]. Previously, Asian sea bass was cultivated mainly in the coastal area along the Gulf of Thailand and the Andaman Sea, where the fish can grow in open sea cages or earthen ponds with brackish/marine water. Presently, the main practice applied for Asian sea bass grow out has been shifted to inland freshwater pond systems, owing to the catadromous nature of the fish and advancements in the quality of pelleted feed, to align with the continual increase of consumer demand [14]. Asian sea bass farming in inland freshwater in Thailand is usually carried out in intensive systems. Although maximum yield can be expected with high stocking density, this aquaculture system poses a high risk for the emergence of infectious diseases due to the stressful conditions caused by the accumulation of metabolic waste, organic matter, and fluctuation in water quality. Mortality and morbidity related to the distribution of infectious diseases have been reported to be a major cause of economic losses for the Asian sea bass farming industry worldwide. According to the literature, Asian sea bass are susceptible to numerous pathogens, including bacteria, e.g., Streptococcus iniae, Vibrio alginolyticus, Vibrio harveyi, and Photobacterium damselae subsp. damselae [2,4,12] and viruses such as viral nervous necrosis virus, Lates calcarifer herpesvirus, Lates calcarifer birnavirus, and megalocytiviruses [6,36,38,39]. Among the potential viral pathogens, members of the genus Megalocytivirus (family Iridoviridae), such as red sea bream iridovirus (RSIV), turbot reddish body iridovirus (TRBIV), infectious spleen and kidney necrosis virus (ISKNV), and scale drop disease virus (SDDV, a distantly related member of the same genus) have been associated with several disease outbreaks in marine fish in many countries, e.g., red sea bream (Pagrus major), orange-spotted grouper (Epinephelus coioides), olive (Japanese) flounder (Paralichthys olivaceus), and turbot (Scophthalmus maximus). The clinical signs and lesions observed in megalocytivirus-infected fish can be very diverse, depending on pathogenic agent and host species. For example, fish infected with RSIV show lethargy, petechiae on gills, and enlargement of the spleen [26], ISKNV-infected fish exhibit a dark body color with pale gills, and red eyes [11], whereas SDDV infection in Asian sea bass is often associated with extensive scale loss and skin hemorrhage lesions [29,41]. Megalocytivirus infection (i.e., ISKNV and SDDV) has been also reported to cause high mortality (55-77%) in Asian sea bass in Thailand [29,48]. However, epidemiological information and molecular characterization of megalocytiviruses in Asian sea bass have been limited to marine environments, since inland freshwater culture is not yet widely practiced on a global scale. To date, only one incidence of coinfection with a pathogenic bacterium (Flavobacterium columnare) and SDDV has been reported in a freshwater system [29].
Recently, disease outbreaks have occurred in inland freshwater-based Asian sea bass grow-out farms located in the eastern and central parts of Thailand, with mortality rates ranging from 40 to 50%. The initial diagnoses indicated that a series of outbreaks may have involved ISKNV infection. Therefore, the aim of this study was to investigate the molecular characteristics of these ISKNV strains and their pathogenic role in Asian sea bass reared in a freshwater system. In addition, the genome sequence of this virus was examined using a metagenomic approach, and this, to our knowledge, is the first ISKNV genome from freshwater-cultured Asian sea bass ever sequenced.

Disease history and sample collection
Fish farmers reported the occurrence of unknown diseases in Asian sea bass in seven freshwater grow-out farms located in Samut Sakhon, Samut Songkhram, and Chachoengsao provinces, Thailand, between February and November 2018. The diseased fish ranged from 20 to 30 g in weight. The cumulative mortality was reported by the operators to range from 40 to 50 percent. Moribund fish were euthanized by decerebration, and bacterial isolation was carried out onsite. Internal organs (kidney, spleen, and liver) were collected separately from each individual, preserved in 95% ethanol for PCR testing, and delivered on ice to the laboratory within 4 h. Ethanol-fixed tissues were maintained in a -20 °C freezer until further PCR assay, whereas fresh tissues for virus isolation were preserved at -80 °C. For histopathology, the collected tissue was immersed in 10% neutral buffered formalin at a ratio of 1:10 (w/v) for 24-36 h, followed by replacement with the same volume of 70% ethanol for long-term preservation. Preserved tissue samples were processed for standard histological analysis by dehydration, embedded in paraffin, sectioned, and stained with hematoxylin and eosin (H&E) [21]. The H&E-stained tissues were then examined under a light microscope equipped with a digital camera. The negative control used for further assays was an apparently healthy Asian sea bass obtained from a different location. The animal use protocol for this study was approved by the Institutional Animal Care and Use Committee, Faculty of Fisheries, Kasetsart University (permit ID: ACKU61-FIS-055).

Bacterial isolation and identification
Direct isolation from external lesions (gills and skin) and internal organs (spleen, kidney, and liver) was conducted using two different media comprised of (i) tryptic soy agar (TSA, Himedia, India) supplemented with 5% sheep blood using a generalized medium and (ii) Anacker and Ordal agar (AOA, tryptone 0.5g/L, yeast extract 0.5g/L, sodium acetate 0.2 g/L, beef extract 0.2 g/L, and agar 10 g/L) supplemented with 1 μg mL −1 tobramycin (Sigma-Aldrich, Singapore) using as a selective medium for Flavobacterium sp. [44]. Streaked plates were delivered to the laboratory and incubated at 28 °C until bacterial colonies were visible (24-48 h). Colonies were sub-cultured using the same kind of medium until a pure colony was obtained. Pure colonies of bacteria grown on either TSA or AOA medium were subjected to preliminary identification using Gram staining and primary biochemical assays including oxidase, catalase, oxidationfermentation, and motility tests. Bacterial taxonomy was determined to the genus level based on Cowan and Steel's manual [7].

DNA extraction from fish tissues
Screening of ISKNV infection was carried out for each individual fish collected in this study (n = 26). To extract DNA, the tissue sample was removed from the ethanol and homogenized using disposable polypropylene pestles. DNA was extracted using a Tissue Genomic DNA Extraction Mini Kit (Geneaid, Taiwan) according to the manufacturer's instructions. The DNA concentration was determined using a NanoDrop analyzer (Titertek Berthold, Germany) and stored at -20 °C.

Identification of ISKNV by PCR
ISKNV-specific primers (Table 1) were used for screening by one-tube semi-nested PCR (snPCR) [11]. Each 25-µL PCR mixture contained 1X master mix (Go-Taq®-Green, Promega USA), 10 nM each working primer, and 100 ng of DNA template. The thermal cycling conditions were 94 °C for 3 min, followed by 35 cycles of 94 °C for 30 s, 65 °C for 30 s, and 72 °C for 1 min, and a final extension at 72 °C for 5 min. PCR products were analyzed by 1% agarose gel electrophoresis followed by staining with Red Safe (Chembio, UK) and visualized under UV light. The expected PCR products from ISKNV-positive samples were either a single amplicon of 164 bp, two amplicons of 517 and 164 bp, or three amplicons of 754, 517, and 164 bp, representing light, moderate, and heavy infections, respectively, as described previously [11]. As a positive control, we used a DNA template extracted from an ISKNV-infected sample that was kindly provided by Centex, Mahidol University. Healthy Asian sea bass and nuclease-free water without DNA template served as the internal and negative control, respectively.

DNA sequencing
Asian sea bass DNA extracts giving positive PCR results for ISKNV screening based on the one-tube snPCR were used for phylogenetic analysis. ISKNV-positive samples were selected randomly (only one sample per farm), and DNA was extracted from the liver. Two ISKNV genes, encoding for MCP and ATPase, were targeted for sequence analysis. Details about each primer set are shown in Table 1. The PCR conditions were described previously [20]. After agarose gel electrophoresis, the amplicons were purified using a Universal DNA Purification Kit (Tiangen, Beijing, China) according to the manufacturer's instructions. The purified DNA fragments were ligated into pGEM-T Easy Vector (Promega, WI, USA), and the construct was used to transform Escherichia coli JM109 competent cells as described by Russell and Sambrook [40]. Transformants were selected using Luria-Bertani agar (LB, Oxoid, UK) containing selective antibiotics and grown in LB broth prior to plasmid isolation using a NucleoSpin® Plasmid MiniPrep Kit (Macherey-Nagel, Germany). The extracted plasmid was submitted to a sequencing laboratory service (1st BASE Pte Ltd., Malaysia) for Sanger sequencing using pUC/M13 primers as described in Promega's pGEM-T Easy Vector manual.

Phylogenetic analysis
Low-quality bases were trimmed manually based on a chromatogram of the raw sequences. The trimmed sequences were then assembled into contigs using ContigExpress. A homology search was performed using Megablast against NCBI's nucleotide database for alignment and comparison to the other ISKNV strains. The MCP and ATPase genes of the ISKNV isolates from this study (n = 6, one sample per farm) were compared to those of other members of the genus Megalocytivirus, including ISKNV (n = 11-22), RSIV, (n = 22-43), and TRBIV (n = 7-14). The nucleotide sequences were aligned using ClustalW, and a phylogenetic tree was constructed using the maximum-likelihood method with the GTR+G+I substitution model and 1000 bootstrap replicates. Multiple sequence alignment and phylogenetic reconstruction were carried out using MEGA-X software [31].

Library preparation and next-generation sequencing
Two samples with possibly severe ISKNV infection (onetube snPCR yielded three distinct amplicons, as described in the previous section) from Samut Sakhon province were used for metagenomic shotgun sequencing. A Nextera XT Library Preparation Kit (Illumina, CA, USA) was used to construct a paired-end library from the extracted genomic DNA according to the manufacturer's instructions, and highthroughput sequencing was performed using an Illumina HiSeq system with a 150-bp read length. Library construction and sequencing were carried out using the service provided by Novogene (Beijing, China).

ISKNV genome reconstruction and annotation
Processing of raw reads and de novo assembly were conducted as described in our previous publication [28]. The adaptor sequences and low-quality reads were filtered out from raw reads using Trimmomatic v0.39 [3]. Then, hostderived reads were discarded by mapping trimmed reads against an Asian sea bass reference genome sequence (NCBI assemblies accession no. GCA_001640805.1) using the -x function in the Bowtie2 program [33]. The remaining non-host reads were subjected to de novo metagenome assembly using MEGAHIT v 1.2.9 with a minimum length of output contigs of 1000 bp [34]. The generated assemblies were submitted to the web server version of the Kaiju program to predict the taxonomic identity of each contig [37]. Contigs assigned to ISKNV were annotated using Prokka v1.14.0 with Viruses Annotation mode, and the complete ISKNV reference genome sequence (GenBank accession no. NC_003494.1) was selected as the annotation template. The names "ISKNV KU1" and "ISKNV KU2" were assigned to the ISKNV-like contigs observed in our two samples. Visualization of the ISKNV KU1 and KU2 sequences as circular genomes was performed by uploading the annotated genomes to the web server version of CGview (http:// cgview. ca/) [16].

Phylogenomic analysis and genome distance
The ISKNV KU1 and KU2 sequences were aligned to the ISKNV reference genome sequence using MAUVE progressive alignment [8]. The genome segments were rearranged manually to correspond to the ISKNV reference genome. To determine genetic distance, a multiple alignment of the new sequences with those of other megalocytiviruses was performed using the FFT-NS-i method in the MAFFT v7 online service [25], and MEGA X was used to calculate the distance based on the maximum-compositelikelihood model. A phylogenomic network was generated using SplitsTree4 based on the alignment of the megalocytivirus genome sequences [22].

Identification of orthologous groups
OrthoFinder [13] was used to identify possible orthologs among four ISKNV genome sequences, including two ISKNV strains from this study (KU1 and KU2) and two closely related strains (RSIV-Ku and the ISKNV reference strain). The coding sequences (CDSs) of strain RSIV-Ku and the reference genome were obtained from GenBank under the accession nos. KT781098 and NC_003494, respectively. The OrthoFinder pipeline automatically categorizes the proteins from tested subjects into orthologous groups (also called orthogroups) based on sequence similarity. In this study, proteins of ISKNV KU1 and KU2 predicted as 'non-orthologous' comparing to the reference strain and RSIV-Ku were subjected for further protein BLAST analysis.

Propagation of ISKNV using the GF cell line
ISKNV isolate KU1 was used for virus isolation. One gram of liver and spleen preserved at -80°C was pooled and homogenized in 10 ml of L15 medium (Gibco, CA, USA), followed by centrifugation at 9600 × g for 30 min at 4 °C. After centrifugation, the supernatant was collected and filtered using a sterile 0.22-µm membrane filter. The filtrate (0.5 mL) was then inoculated onto a monolayer grunt fin (GF) cells in a 5-mL flask for 2 hours, and the medium was replaced with fresh L15 medium supplemented with 10% fetal bovine serum (Gibco, CA, USA). The flasks were then incubated at 25 °C and observed under a microscope daily for 10 days to monitor the cytopathic effect (CPE). The virus was harvested by centrifugation of the cell culture supernatant at 1000 × g for 5 minutes at 4°C. Cell debris was discarded, and the supernatant containing ISKNV was collected and preserved in a -80 °C freezer until used in the challenge experiment.
The viral copy number of ISKNV in viral suspension was determined using a qPCR assay specific for ISKNV [27]. One hundred forty microlitres of the supernatant was used for DNA extraction by the phenol-chloroform method. The qPCR reaction consisted of 1X iTaq Universal Probes Supermix (Bio-Rad), 3 µL of DNA template, 900 nM each forward and reverse primer (Meg-MCP160F and Meg-MCP349R) and 250 nM Meg-MCP239P probe in a total volume of 20 µL. The qPCR conditions included initial denaturation at 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s and 64 °C for 1 min. Both virus detection protocols were performed using a Bio-Rad CFX Connect RealTime PCR System, and the product size was 190 bp.

Pathogenicity test
Healthy Asian sea bass (n = 150) were used in the experimental challenge test to fulfill Koch's postulates. The fish, with an average weight of 22 g, were purchased from a local grow-out farm. Five fish were sampled randomly to verify their ISKNV-free status, using splenic tissue in a virusspecific one-tube snPCR as described above. ISKNV-free fish were used for pathogenicity tests after a week of acclimatization in a 3,000-L tank. Fish were divided into three groups (50 per group) comprising one control group and two challenge groups. The control group received sterile cell culture medium, whereas the challenge groups were injected intraperitoneally with 0.2 ml of virus suspension (low dose [10 -2 dilution] or high dose [undiluted]). Injected fish was transferred to aerated 500-L tanks (50 per tank) with the water temperature and dissolved oxygen maintained at 30-32 °C and 5-7 ppm, respectively. Freshly dead and moribund fish were removed from the tank as soon as they were noticed. The liver, spleen, and kidney of affected fish were collected for histopathological investigation and ISKNV screening using one-tube snPCR. The daily mortality was recorded for 21 days, and a Kaplan-Meier survival curve with a logrank statistical test was made using the IBM SPSS Statistics 25 program.

Clinical manifestations in naturally infected fish
The observed losses in grow-out Asian sea bass varied from 40 to 50% within 1-2 weeks after the disease was noticed. The moribund fish showed clinical signs such as lethargy and anorexia, while external lesions including darkened body color, pale gills, skin hemorrhage, and ascites were also observed. Internally, swollen spleen and hemorrhage of the liver and trunk kidney were the most abundant lesions (Supplementary Fig. S1). Histopathological manifestations included severe hemorrhage, inflammatory cell infiltration, accumulation of melanocytes, extensive necrosis of the skin and gills, and moderate tubular degeneration in the trunk kidney (Fig. 1A). Interestingly, the typical microscopic lesions associated with viral infection, namely, eosinophilic inclusion, and the hallmark histological changes associated with ISKNV infection, i.e., basophilic hypertrophied cells, were also detected in gill, liver, and kidney tissue.
Bacterial isolation was attempted from necrotic gills and skin lesions using the Flavobacterium-selective medium AOA, whereas generalized medium, TSA, was used for internal organs with apparent lesions. Bacterial colonies grown on AOA medium were yellowish in color and showed rhizoid morphology. Details about the collected fish samples and the observed lesions, together with the diagnostic results, are summarized in Supplementary Table S1.

Identification of ISKNV
Primary screening for ISKNV infection in Asian sea bass samples was done using a one-tube snPCR targeting the MCP gene. Out of 26 fish collected in this study, 20 were positive for ISKNV (6 out of 7 farms). Despite the fact that the samples were from the same farm, the degree of infection, as interpreted by the number of amplicons, varied among the positive samples. The electrophoresis photographs are shown in Supplementary Figure S2. According to the original article describing the semi-quantitative nature of the one-tube snPCR [11], relatively high, moderate, and low viral copy numbers would result in three, two, and one amplicon, respectively, appearing in agarose gel electrophoresis.

Phylogenetic analysis based on the MCP and ATPase genes
The MCP and ATPase genes were amplified from six ISKNV-positive samples (one per farm) ( Supplementary  Fig. S3). The nucleotide sequences of the almost complete MCP (1,275 bp) and complete ATPase (720 bp) genes were identical among the ISKNV isolates of this study. Therefore, we submitted the MCP and ATPase gene sequences from only a single ISKNV isolate to the Gen-Bank database under the accession nos. MW269579 and MW269580, respectively. BLAST analysis of the MCP and ATPase gene sequences indicated that the ISKNV isolate from this study was identical to several ISKNV strains available in NCBI's nucleotide sequence database, including strains AFIV-16 (MK689685) and RSIV-Ku (KT781098). Phylogenetic trees based on the MCP and ATPase genes of ISKNV, RSIV, and TRBIV (members of the genus Megalocytivirus) are shown in Figure 2. In the MCP-based tree, the ISKNV isolates clustered into two different subclades, corresponding to genotypes I and II, and the isolates from this study were all in genotype I. In contrast, there was no clear sub-clustering among the ISKNV isolates in the ATPase-based tree, which could be due to the smaller number of sequences.

Genome features
Asian sea bass samples from farms E and F (n = 2) in Samut Sakhon province with a relatively high viral load (as shown by one-tube snPCR) were selected for metagenomic shotgun sequencing. Non-host reads were assembled de novo into 1,482 and 1,030 contigs, respectively. Taxonomic identification using the Kaiju web server indicated that, from each sample, only the longest contig was identified as ISKNV. The length (111 kb) and GC content (54.8%) of these contigs were almost identical to those of the ISKNV reference genome sequence. These virus  Table 2. Both ISKNV genome sequences were submitted to the GenBank database under the accession numbers MT128666 and MT128667. Their genome map is shown in Figure 3.

Phylogenomic analysis
The neighbor-net network analysis was able to differentiate 15 members of the genus Megalocytivirus into three distinct clusters, namely, the RSIV, TRBIV, and ISKNV groups (Fig. 2C). The genome sequences of ISKNV KU1 and KU2 were almost identical to those of the ISKNV reference strain and the RSIV-Ku strain, with minimal genome distances of 0.02% and 0.06%, respectively. In contrast, the genome distance between distinct clusters can be as high as 5.5-9.4%. A reticulated pattern was clearly observed among the ISKNV and RSIV groups, which indicated possible genetic recombination within this cluster.

Orthology between ISKKNV KU1, KU2, and reference strains
OrthoFinder categorized a total of 499 genes, obtained from four distinct ISKNV strains (KU1, KU2, RSIV-Ku, and reference strain), into 124 orthogroups. Among these, 88 orthogroups (70.9%) were presented in all ISKNV strains. The RSIV-Ku and reference strains possessed 13 and 18 genes that were not assigned to any orthogroup (unique genes). Seven orthogroups were found exclusively in the strains KU1 and KU2 but were absent in the other two ISKNV strains. The strain KU1 also carried two genes that are unique to its genome. Most of these KU1-and KU2-specific genes were predicted to encode hypothetical proteins and were similar to genes of other viruses in the genus Megalocytivirus, including angelfish iridovirus AFIV-16, scale drop disease virus, Banggai cardinalfish iridovirus, and red sea bream iridovirus. There was one protein from the KU1 (QQZ00456) and KU2 (QQZ00673) strains that were almost identical to the caspase recruitment domain (CARD) containing protein of angelfish iridovirus AFIV-16. Details about the nine genes present only in strains KU1 and KU2 are shown in Table 3, together with the BLAST protein analysis results.

Virus isolation
In vitro replication of ISKNV was carried out by propagating the virus in the GF cell line. ISKNV KU1 was selected as the source of the virus seed. At 25 °C, the onset of CPE was observed at 5 days post-inoculation (dpi), when focal degeneration of GF cells was observed as a limited area of cell detachment and cell shrinkage, suggesting pyknosis ( Fig. 4, top-right panel). At 7 dpi, vacuolization and pyknosis were observed throughout the confluent monolayer, indicating severe degeneration of host cells (Fig. 4, bottomright panel). In contrast, no CPE was observed in GF cells propagated with sterile L15 medium until the end of the culture period of 10 days. Cell culture supernatant was harvested from both control and virus-infected cells at 7 dpi, and genomic DNA was extracted. This DNA was used as a template in ISKNV-specific one-tube snPCR, and specific amplification was observed only with the ISKNV-infected cells (Supplementary Fig. S4). ISKNV-specific qPCR indicated that the viral copy number in the viral stocks used for the challenge assay was 1.45 × 10 6 copies/mL.

Viral pathogenicity assay
Healthy Asian sea bass reared in freshwater were challenged intraperitoneally with GF-cell-grown ISKNV at a dosage of 2.9 × 10 5 (high dose) or 2.9 × 10 3 (low dose) copies/fish. The onset of mortality for the high-dose and low-dose groups occurred at 5-and 8-days post-challenge (dpc), respectively. The daily mortality in both groups was 1-3 fish per day, except at 18 and 19 dpc, when 5-8 fish in the high-dose group died. In the low-dose group, no more deaths occurred after19 dpc. The cumulative mortality at the end of the experiment (21 dpc) was 90% for the high-dose group (45 out of 50 fish), which was significantly higher than for the low-dose group (50%, 25 out of 50 fish [p < 0.05]). The experimentally infected fish tested positive in the one-tube snPCR. These infected fish samples yielded three specific amplicons, suggesting a relatively high viral load in the liver (Supplementary Fig. S5). Two fish in the control group (4%) died a day after mock infection, which could have been due to injection injury, since the one-tube snPCR showed negative results. The survival curve is shown in Figure 5.
The experimentally infected fish showed clear pathological changes similar to those observed in naturally infected fish. Externally, moribund fish exhibited a darkened body color and with pale gills. Microscopically, hematopoietic tissues showed apoptosis (pyknotic nuclei) and cytoplasmic inclusion bodies, tubular degeneration was found in the kidney, and gill lamella contained showed hypertrophic basophilic cells (Fig. 1C and D).

Discussion
In this study, ISKNV was identified in diseased Asian sea bass cultured in an inland freshwater system. The most common histopathological lesions in ISKNV-infected Asian sea bass collected in this study, i.e., severe necrosis and the appearance of basophilic inclusion bodies in gills, liver, and kidney, was consistent with the pathognomonic lesions of megalocytivirus infection (hypertrophy/megalocytosis in gill and liver) reported previously [32,45]. In this study, the histopathological manifestations in the spleen, kidney, and liver also suggested viral tropism for these hematopoietic organs which, to some extent, may result in immune incompetence/suppression and increased susceptibility to opportunistic infections. Coinfection with Flavobacterium columnare and SDDV, another member of the genus Megalocytivirus, in grown-out Asian sea bass was observed in our previous investigation [29]. Considering that coinfection with ISKNV and potential bacterial pathogens such as Aeromonas sp., Streptococcus sp., and Flavobacterium sp. was identified in three out of seven disease incidences in this study, it is feasible that simultaneous infection is relatively common in cases of natural disease in Asian sea bass. Simultaneous infections, also called concurrent infections, occur frequently in farmed fish and might be more common than single infections [10]. In our case, we speculated that simultaneous infection by ISKNV and bacteria could influence the large diversity of clinical appearances, which range from inapparent to severe, e.g., extensive hemorrhage. In this study, it was not clear whether ISKNV or a bacterial pathogen was the primary pathogen, and their relative contributions in vivo require further investigation.
In addition to Asian sea bass, ISKNV has been reported in farmed Nile tilapia and ornamental fish in Thailand as well [1,10,47]. The clinical appearances of these ISKNVinfected fishes also similar to those observed in this study. The ISKNV from ornamental fish were genetically classified as genotype I [1] based on the MCP gene sequence, similar to the ISKNV strains in this study. In previous studies, disease outbreaks were observed in the same region (central Thailand) where the ISKNV-positive Asian sea bass samples were collected in this study. Therefore, the possibility of cross-species transmission between these freshwater-farmed fishes, although not yet documented officially, should be of concern (particularly in the area with a history of outbreaks), since sharing of water resources among farms cannot be easily avoided.
ISKNV has also been identified in various marine fish species in other Southeast Asian countries, including Indonesia, Vietnam, Malaysia, and Singapore [11,23,39,46]. Most of the ISKNV isolates found in these countries were genotype I, like in this study, indicating that this ISKNV genotype is widespread in Southeast Asia. According to the original article describing ISKNV genotyping [15], ISKNV can be classified into three genotypes (I, II, and III) based on the diversity of the MCP gene. These genotypes were later assigned to three clusters comprising the RSIV, ISKNV, and TRBIV groups in the same virus species, as described by the International Committee on Taxonomy of Viruses (ICTV, https:// talk. ictvo nline. org/ ictv-repor ts/ ictv_ online_ report/ dsdna-virus es/w/ irido virid ae/ 615/ genus-megal ocyti virus). To date, ISKNV genotype II has been reported in orbiculate batfish (Platax orbicularis), Banggai cardinalfish (Pterapogon kauderni), and marble goby (Oxyeleotris marmorata) in Indonesia, Japan, the USA, and China [32,43,49], while Asian sea bass mortality associated with ISKNV genotype II infection has been reported in Southern China and Vietnam [11,50]. According to the phylogenetic analysis conducted in this study, the genetic diversity among the current ISKNV isolates was rather low, as all of the samples were classified as genotype I and shared 100% sequence identity in both the MCP and ATPase genes. The MCP, ATPase, and DNA polymerase genes are generally used to determine genetic relationships between megalocytivirus due to their evolutionary conservation [30,32]. However, a relationship between genetically similar strains cannot be inferred, at least for the current collection of virus isolates, relying on these conserved genes alone. To date, in-depth epidemiological information regarding the ISKNV genotype distribution in Thailand and neighboring countries in Southeast Asia is scarce. Thus, an efficacious 'regional' vaccine against ISKNV in Southeast Asia cannot be developed unless more epidemiological data are collected. For further investigations, other molecular markers, such as the four ankyrin repeat domains [30], could be added to the phylogenetic comparisons which would allow intra-genotype diversity to be analyzed on a finer scale.
Isolation of ISKNV from infected Asian sea bass specimens using the GF cell line was successful in this study. Application of GF cells for the propagation of ISKNV simultaneously with nervous necrosis virus was described in our previous publication [24]. However, the results of this study were slightly inconsistent with those of another publication in terms the onset of CPE [48]. In the previous study, CPE was not observed clearly until 14 dpi, whereas it was observed already at 5 dpi in this study and was quite pronounced at 7 dpi. This discrepancy might have been due to the difference in the initial inoculation dose, since, in this study, we intentionally selected a specimen with a potentially high viral titer (as indicated by one-tube snPCR). Recently, the alternative cell line GS-1, originating from orange-spotted grouper (Epinephelus coioides) fibroblasts, was reported to be susceptible to ISKNV infection, allowing the titer to reach 10 5.2 TCID 50 /ml within 7 days. However, a Neighbornet split graph generated from a whole-genome alignment. The genetic distance between genomes is shown as percent substitutions per nucleotide (labeled in red with a dashed line). The ISKNV recovered in this study is indicated by a black circle. The numbers at the nodes of the tree represent bootstrap values, and the scale represents the rate of substitution per nucleotide. ◂ direct comparison of the ISKNV replication kinetics in these two cell lines has not been reported [19].
The pathogenicity assay conducted in this study was able to fulfill Koch's postulates, showing that ISKNV is pathogenic to freshwater-reared Asian sea bass. The results suggested that the pathogenicity of ISKNV and onset of mortality depend on the infection dose (Fig. 5). The cumulative mortality rates were similar to those seen in previous investigations (90, 77, and 85.89%), as was the onset of mortality (5-9 dpc) [48,50]. It should be mentioned that the fish examined in this study (juvenile, 22 g weight) were larger than those in the previous report (fingerling, 3.5 g weight) [48], indicating that ISKNV is able to cause high mortality in Asian sea bass at various life stages. Regarding clinical signs, it is worth mentioning that the experimental animals exhibited only the typical darkened body and pale gills, in agreement with those described in the recent study [50]. Other obvious external lesions, such as scale loss, muscle necrosis, and hemorrhage, were observed only in naturally infected fish. This difference in clinical manifestations could be due to coinfection with bacterial pathogens in natural cases (Supplementary Table S1). This emphasizes that diagnosis of field outbreaks should be performed with caution, and multiple approaches, including pathogen isolation, PCR, and histopathology, should be applied when possible to get a comprehensive understanding of the disease scenario.
Recently, we showed that metagenomic shotgun sequencing was able to produce a draft genome sequence of SDDV directly from the infected specimen, without a need for culturing the virus [28]. In this study, the same analytical approach was implemented, and the complete genome sequences of ISKNV KU1 and KU2 were reconstructed with an acceptable coverage depth of 47-56×. A genomescale phylogenetic network (Fig. 2C) showed a reticular pattern within the ISKNV and RSIV groups, implying that genetic recombination had occurred between the members of these groups. In fact, one of the ISKNV group members, RSIV-Ku, was similar to GSIV-K1 in 7% of its genome,  suggesting that this strain is an ISKNV/RSIV recombinant [42]. Here, possible recombination between ISKNV KU1/ KU2 and other megalocytiviruses was screened using the RDP4 program [35], but no evidence of recombination was found in their genomes (data not shown). Orthology analysis showed a surprisingly large number of unassigned orthogroups (unique genes) among the genomes of ISKNN KU1/ KU2, RSIV-Ku, and the reference strain. It is predicted that strain RSIV-Ku, a natural recombinant virus, may possess numerous unique genes, since its genome has a 7.8-kb-long region resembling RSIV genotype II rather than ISKNV [42]. In the case of the ISKNV reference strain, 18 genes were not assigned to any orthologous groups, although the core genome similarity in comparison to ISKNV KU1/ KU2 was as high as 99.98%. This could be explained by differences in the genome annotation methods used for CDS prediction in the ISKNV reference strain and KU1/KU2 genomes. Protein-encoding sequences of the ISKNV reference strain were identified by querying sequences through a protein domain database [17], whereas a program based on an unsupervised machine learning algorithm (Prokka) was employed in the case of ISKNV KU1/KU2. Among the non-orthologous genes present in ISKNV KU1/KU2, a caspase recruitment domain (CARD)-containing protein was identified. This protein was highly similar to those found in angelfish iridovirus AFIV-16, which also belongs to the ISKNV genotype I [27], isolated from the angelfish Pterophyllum scalare in Southeast Asia. CARDs are well-known interaction motifs involved in regulation of inflammation and apoptosis [18]. CARD proteins have been demonstrated in vitro to have inhibitory effects on apoptosis in grouper iridovirus, a member of the genus Ranavirus of the family Iridoviridae [5]. However, the role of the CARD protein in the molecular pathogenesis of ISKNV and whether it is involved in inhibition of apoptosis, remain to be elucidated. In summary, homologous strains of ISKNV genotype I were identified as causative agents of mass mortality in freshwater-cultured Asian sea bass in eastern and central Thailand by a combination of histopathology, molecular analysis, and pathogenicity assays. The complete genome sequences of two ISKNV isolates, KU1 and KU2, were determined using a metagenomics approach. The genome information, as well as the virus archive collected in this study, could be useful for evolutionary analysis and selection of potential genotype I vaccine candidates for the sustainable prevention of ISKNV outbreaks in the future. Availability of data and material This manuscript has data included as electronic supplementary material.

Conflict of interest
The authors declare no conflict of interest.
Ethical approval The animal use protocol followed in this study was approved by the Institutional Animal Care and Use Committee, Faculty of Fisheries, Kasetsart University (permit ID: ACKU61-FIS-055).