A novel totivirus and piscine reovirus (PRV) in Atlantic salmon (Salmo salar) with cardiomyopathy syndrome (CMS)

Total RNA was extracted using the RNeasy Lipid Kit (QIAGEN AB, Oslo, Norway), from three specimens: two heart samples from fish from a CMS outbreak (peak phase with high mortality in the population) and one combined heart and head kidney sample from an experimentally challenged fish (nine weeks post inoculation; see details below and [10]). The RNA was DNase treated using TURBO DNA-free (Applied Biosystems/Ambion, Austin, TX, USA) and reverse transcribed/amplified using the QuantiTect kit (QIAGEN AB) according to manufacturer's instructions. Approximately 1 μg of RNA was used as template in each multiple displacement amplification (MDA), which was allowed to proceed for two hours. cDNA from the three reactions was combined in equal amounts and a library with a Multiplex Identifier tag was prepared according to the GS FLX Titanium General Library Preparation Method Manual (454 Life Sciences, a Roche company, Branford, CT, USA). The library was titrated and amplified using the large volume emulsion preparation protocol and sequenced using a Genome Sequencer FLX instrument and GS FLX Titanium chemistry (454 Life Sciences). Sequencing was done by the The Norwegian High-Throughput Sequencing Centre (NSC) at the University of Oslo.

Reverse transcription quantitative PCR (RT-qPCR) for quantification of viral PRV RNA in CMS field samples was performed as described earlier [14] using the minor groove binding (MGB) assay targeting the L1 genomic fragment. PRV RNA in head kidney and heart from 21 fish from four outbreaks of CMS was quantified and compared to assumed healthy farmed fish.

Samples from challenged fish and controls were obtained from 'Challenge group 1' described in a previous publication [10]. Briefly, 100 unvaccinated wild strain Atlantic salmon were intraperitoneally (i.p.) injected with 200 μl supernatant of tissue homogenate mix from six CMS diagnosed fish. Samples of heart, head kidney, spleen and liver were taken from 5 fish every third week 0 to 36 wpc (weeks post challenge), and directly frozen at -80°C. Total RNA was isolated using the RNeasy Mini Kit (QIAGEN AB). Each PRV RT-qPCR run contained 100 ng RNA in a volume of 12.5 μl and the PCR protocol for detection of PRV described in Palacios et al. [14] was used. Head kidney samples were also screened for the presence of IPNV (Infectious Pancreatic Necrosis Virus), SPDV (Salmon Pancreas Disease Virus) and nodavirus by RT-qPCR. The fish cultivation and research facilities were located in an area geographically separated from the endemic area of CMS in Norway.

A real-time PCR assay was designed for the totivirus found in the 454 sequence database. A protocol identical to the one described above was used with forward primer TTCCAAACAATTCGAGAAGCG, reverse primer ACCTGCCATTTTCCCCTCTT and the MGB probe CCGGGTAAAGTATTTGCGTC (all written in 5'-3' direction; Applied Biosystems, Life Technologies Corporation, Carlsbad, California, USA). The assay was used qualitatively to test both a panel of CMS outbreaks, control samples and fish from the experimental transmission.

In situ hybridizations were performed on heart sections as described earlier [14] using LNA probes (Exiqon, Vedbaek, Denmark) targeting PRV genomic segment L2 [14] and the totivirus RNA-dependent RNA polymerase gene (probe sequences CTTCCGCTACCATCCTGAGATA and TTTCTTCGACACACCTTTCCGC; 5'-3' direction). Viral RNA was visualized using aminoethylcarbazole (AEC). Heart sections of healthy fish served as controls.

The pyrosequencing gave a total of 556,947 reads with average length of 240 bases. Data are archived at the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) under accession SRP002117. In order to perform computational subtraction of the reads, removal of ribosomal RNA reads or assembly of read contigs would normally be performed, but due to the possibility of having a significant number of chimerical reads (the QuantiTect kit includes a ligation step were the reversely transcribed cDNA fragments are ligated to facilitate MDA) such filtering of results was not done. Instead, all reads were analyzed using Mega BLAST and a database comprising all nucleotide sequences annotated as viral in the nr database at NCBI was compiled. Mega BLAST searches were conducted using word size 28 and for each read the match that gave the highest bit score was recorded. Results were tabulated and showed a large number of hits to a relatively small number of entries in our viral sequence database. Many of the sequences that showed a high degree of similarity to our reads were obvious errors in the nr database (incorrectly annotated/assembled sequences, vector sequences etc.) and could be discarded. Only four groups of sequences gave credible matches; eight reads were (imperfect matches) to Chum salmon reovirus (accession number AF418297; [15]), two matched Atlantic salmon swim bladder sarcoma virus (acc. no. DQ174103; [16]) and 1,084 reads matched the PRV genome with a very high degree of sequence similarity. A 2,914 basepair sequence contig comprising approximately 2,600 reads from what appeared to be a novel virus could also be constructed. Alignment of conserved amino acid sequence motifs in the RNA-dependent RNA polymerase (RdRP) of the Totiviridae indicated that the virus was related to this virus family [17]. Using protein BLAST with default parameters and 290 amino acids of RdRP, the closest match was a gag-pol protein (27% identity; 45% positive) from a totivirus infecting Giardia lamblia[18]. Partial RdRP was control-sequenced using Sanger chemistry and used as target for an MGB RT-qPCR assay and for in situ hybridization (GenBank accession number HQ401057).

PRV reads were assembled using the software Sequencher (version 4.5; Gene Codes Corporation Ann Arbor, MI, USA) and revealed approximately 85% coverage of the PRV genome. This was further increased to 94% coverage using Sanger sequencing of PCR amplified fragments (data not shown). The genomic segments showed considerable variation in coverage, both regarding the relative part of the segments covered by reads and the depth of coverage (Table 1). In particular, segments S1 and S2 showed relatively poor coverage. Most of the longer fragments had reads that indicated the presence of missense mutations.

Screening of samples from field outbreaks of CMS revealed only a modest increase in viral loads compared to the elevated PRV titers seen in HSMI fish (Additional File 1, Table S1 and [14]). In order to investigate further the potential role of PRV in the experimental challenge we quantified the viral loads in various organs throughout the experiment. A general increase with time in PRV RNA levels in all organ samples was found by RT-qPCR 0 to 12 wpc (Figure 1). Six wpc, all organ samples of all analyzed fish were positive. The levels of PRV RNA peaked in head kidney 12 wpc (Ct 18.2) followed by spleen 18 wpc (Ct 16.3). In heart and liver, the PRV levels were in general lower than in the spleen and head kidney, and remained stable from 15 wpc and onwards with the highest levels detected 33 wpc (Ct 25.2) and 15 wpc (Ct 27.7), respectively. Heart control samples taken 3, 6 and 9 wpc (n = 2/time point) were PRV negative and the fish population was negative for SPDV and nodavirus. Low and steady levels of IPNV (37 positives of 67 samples, mean Ct ~30) were detected in the head kidney from 6 wpc and onwards.

Testing of samples from field outbreaks of CMS showed a perfect correlation between the presence of totivirus RNA and CMS (Table 2). All the control groups, including HSMI outbreaks, were negative. Based on the observation that there appeared to be a consistent relationship between the presence of virus and disease, results were scored qualitatively (Table 2). Control fish from the experimental challenge were negative and by week six post challenge, all fish tested were positive for the virus (Table 2).

The histopathological changes in the transmission experiment have previously been described in detail [10]. Briefly, the severity of atrial lesions peaked at 12 wpc, and remained at this level throughout the study. Ventricular spongy layer lesions were found in ~50% of the challenged fish from 12 wpc and onwards, and epicardial changes were mild and decreased towards the end of the study. Staining for PRV RNA by in situ hybridization was seen in the spongious as well as the compact myocardium, and in endocardial cells of areas with distinct lesions and cellular infiltration. Staining for totivirus RNA was seen in myocardial cells associated with characteristic lesions in the spongious myocardium (Figure 2). Control sections were negative.