Extensive retroviral diversity in shark

Retroviruses infect a wide range of vertebrates and cause many notorious diseases, such as AIDS and cancers. However, much remains unknown about the diversity of retroviruses in basal vertebrate species. In particular, only several retroviruses have been identified in fishes, including Snakehead retrovirus, walleye dermal sarcoma virus, walleye epidermal hyperplasia virus, and Atlantic salmon swim bladder sarcoma virus [1-4]. Retrovirus employs a unique replication strategy, which requires reverse transcription of its RNA genome into DNA and integration of viral DNA into the host chromosomes. Occasionally, retroviruses infect germ line cells, and the resulting integrated retrovirus, known as endogenous retrovirus (ERV), becomes vertically inherited as a host genomic locus. Over time, some retroviral insertions are fixed in the host population. ERVs provide important insights into the ecology and evolutionary history of retroviruses.

Cartilaginous fishes (Chondrichthyes) are the most basal class of vertebrates from which retrovirus has been reported [5]. Here, I analyzed the recently available genome sequence of the elephant shark (Callorhinchus milii), a high-quality genome assembly covering approximately 94% of the C. milii genome, for retroviral insertions [6]. The tBLASTn algorithm with various representative retroviral Pol protein sequences was employed to screen the elephant shark genome for candidate ERV sequences. To distinguish ERVs from other LTR-retrotransposons, I used a strict criterion: only the retroviral Pol protein homolog sequence with a downstream Env protein homolog is defined as an ERV element. After initial identification of ERVs, the BLASTn algorithm was used to identify short ERV fragments. My genome-scale screening procedure identified three complete or nearly complete ERV insertions (within the C. milii genome scaffolds 2, 324, and 2324, respectively; Additional file 1: Dataset 1) and many short ERV fragments in the elephant shark genome. I designate these retroviral elements “C. milli ERVs” (CmiERVs).

To assess the relationship between CmiERVs and other retroviruses, CmiERV and representative retroviral Pol protein sequences (Additional file 2: Table S1) were aligned using MUSCLE [7]. The ambiguous regions in the Pol protein alignments were removed using Gblocks 0.91b and then manually edited [8]. Phylogenetic analyses were performed using MrBayes 3.1.2 [9]. My phylogenetic analysis shows that these CmiERVs form three distinct lineages (Figure 1). Lineage I CmiERVs cluster with the retroviruses isolated from the snakehead fish (Ophicephalus striatus), while lineage II and III CmiERVs cluster with the epsilonretroviruses isolated from the walleye (Sander vitreus) and amphibians. CmiERV lineage I and lineages II/III are only distantly related to each other. These CmiERV lineages are likely to result from three independent retroviral invasion events. My analysis provides clear evidence there is extensive retroviral diversity in the elephant shark. It is possible that exogenous retroviruses related to CmiERVs identified here are still circulating in the elephant shark and possibly other Chondichthyes.

Previously, a single ERV sequence was identified in the lemon shark (Negaprion brevirostris), which is closely related to human ERV; this ERV was thought to have a cross-transmission origin [5]. However, I find that CmiERVs cluster together with retroviruses of fish origin. The phylogenetic pattern is compatible with the hypothesis of an ancient marine origin of retroviruses [15]. Chondrichthyes are the most basal class of vertebrates from which retrovirus has been identified; no retrovirus is identified in earlier-diverging vertebrate lineages, the lampreys (Cephalaspidomorphi) and the hagfish (Myxini) [5]. It follows that these CmiERV elements are likely to represent “primitive” retroviruses. However, the possibility that these elephant shark retroviruses originated from cross-transmission from other fishes cannot be formally excluded.

To date, only a limited number of exogenous/endogenous retroviruses have been identified in fishes [1-5,14]. My results reveal the unexpectedly extensive retroviral diversity of the elephant shark. The initial candidate ERVs were identified based on a strict criterion – whether there is a downstream Env protein homolog following the Pol protein homolog. This approach is conservative, given that the retroviral Env protein evolves rapidly and its similarity to other retroviral Env proteins will erode over a long time. On the other hand, the ERVs identified using this approach are authentic retroviruses. It is likely that there are additional ERV insertions that were not detected. Also, it should be noted that only a small proportion of retroviruses could leave endogenous copies in their host genomes [16]. Therefore, I believe the actual diversity of retroviruses is more extensive in the elephant shark. Further analysis of ERV in basal vertebrates would improve our understanding of the diversity and evolution of retroviruses.