The newly described piscine reovirus (PRV) belongs to the Orthoreovirus/Aquareovirus group  within the Reoviridae family . Viruses of the two genera can infect large groups of animals and cause diseases with a broad range of manifestations; from inapparent to lethal. The reoviruses are often considered 'orphan' as they have not been definitely linked to particular diseases. PRV is widely distributed in farmed Atlantic salmon in sea water, and can be found in healthy as well as in diseased fish . It is thus not surprising that the virus was present in the inoculum used in our CMS experimental challenge. Totiviruses, on the other hand, are primarily known to infect unicellular organisms such as Saccharomyces cerevisiae and protozoans like Leishmania and Giardia spp. , but the host range has recently been found to include penaeid shrimp, insects and some totiviruses can also be grown in mammalian cell cultures . Totiviruses are naked dsRNA viruses and to our knowledge the closest link to our observation of a totivirus in fish is the penaeid shrimp infectious myonecrosis virus (PsIMNV) which causes necrosis in the skeletal muscle of shrimp .
In the present study, histopathological lesions in the heart consistent with CMS were observed from week 6 to the end (36 weeks) of the experimental infection. The PRV and totivirus infections appeared to be persistent, not only in immunologically important organs such as the head kidney and spleen that often harbor persistent and covert viral infections, but also in the heart, the organ where the major histopathological lesions of CMS are found. Neither HSMI nor CMS have been found in the fresh water phase of farmed Atlantic salmon, but both diseases occur after transfer to sea water. HSMI is normally observed 5 to 9 months after transfer while clinical CMS first appears after 12 to 18 months. Independent challenge trials, as well as epidemiological studies, have indicated that CMS has an infectious cause [9, 10]. In combination with the late onset of CMS relative to sea water transfer this indicates that a persistent infection is involved in the etiology. If an acute infection is the primary cause of the disease, it should typically have been observed also at earlier stages after sea water transfer. In the field, CMS is often observed as sudden death of large fish, caused by rupture of the atrium or sinus venosus and resultant cardiac tamponade . However, lesions characteristic of CMS develop over several months until the myocardial lesions are severe and widespread and the fish can no longer compensate for the compromised function. Both experimental infections and field observations strongly indicate that CMS is a chronic disease, culminating in sudden death, with moderately elevated mortality rates. In cultured shrimp, the totivirus PsIMNV causes a persistent, slowly progressive disease appearing as extensive necrotic areas in striated muscles, with presence of the virus in heart muscle ; a disease progression which has some resemblance to CMS of farmed salmon, characterized by degeneration and necrosis of the inner, spongious myocardium of the ventricle and the atrium .
Significantly lower levels of PRV were found in fish from field outbreaks of CMS than fish with a HSMI diagnosis. The totivirus gave a more compelling result where a perfect correlation was found between presence of the virus and disease. Additional non-infectious factors such as fast growth have been suggested to be important for development of CMS, and the highest mortality is usually seen among fish with the highest condition factor. Other contributing factors may include lack of exercise, nutritional factors, environmental factors, stress and co-infections with other known or unknown pathogens able to induce persistent infection with heart involvement. It has been proposed that CMS may be linked to previous viral infections, and an epidemiological link between IPNV outbreaks in early sea water phase and later outbreaks of CMS has been suggested [8, 24], but experimental transmission of CMS has been demonstrated in the absence of IPNV .
Our sequence data were derived from total RNA extracted from salmon sampled during field outbreaks of CMS and fish where CMS had been induced through experimental transmission. Viral reads were identified using nucleic/amino acid sequence similarity search tools and thus the absence of other candidates becomes significant. The possibility that another, causal virus, was transferred along with PRV and the totivirus in the experimental challenge cannot be excluded. Bioinformatics tools may not have identified the virus, or viral titers may have been too low for detection.
For PRV to be causally involved in two distinct diseases such as CMS and HSMI, the most likely explanation would be that there are different genotypes of the virus. The fish farming industry has many similarities with the poultry industry, such as the production of a large number of a single animal species in an industrial scale within a confined area. The avian reoviruses (ARV) of the Orthoreovirus genus are important pathogens of poultry and cause considerable economic loss in the industry. Like PRV, ARV is ubiquitous in production units and the etiology of specific diseases attributed to ARVs are difficult to reproduce experimentally . The pathogenicity of ARV strains differs considerably and the infections may not cause recognizable diseases or clinical signs at all. The gene encoding the neutralizing antibody inducing viral attachment protein σC, shows nucleotide variation in circulating field strains that has affected the efficacy of vaccines . The analogous protein in mammalian orthoreovirus, σ1, appears to affect tissue tropism and pathogenesis in mice . A hypothesis could be that the sequence diversity between PRV strains (Table 1) may account for the distinct appearances of HSMI and CMS. However, sequencing of the assumed σC/σ1 analogue of PRV from a number of CMS/HSMI outbreaks did not show any distinct pattern (data not shown).
The viral load and localization of totivirus in myocardial cells of heart lesions characteristic of CMS indicate that the totivirus is involved in the disease development. The load of PRV in fish with experimentally induced CMS and the persistence of the infection in the heart indicate that PRV may be opportunistically associated with the development of CMS, but we believe that the totivirus is a more likely causative candidate.