Introduction
The salmon louse (
Lepeophtheirus salmonis), is a marine ectoparasite feeding on mucus, skin and blood of salmonids in the northern hemisphere [
1,
2]. The salmon louse has a high reproductive capacity, and extensive farming of Atlantic salmon (
Salmo salar) has led to an increase in host availability and density [
2,
3]. Infestations of salmon lice are a serious problem for the salmon farming industry, with an estimated cost of €180 million each year [
4]. The infestations have also been suggested to have a detrimental effect on wild salmonids [
2,
3].
In the last few years, there has been a dramatic increase in the number of mononegaviruses discovered in arthropods, as new techniques for virus detection have been developed [
5‐
12]. The order
Mononegavirales consists of 11 families:
Rhabdoviridae,
Filoviridae,
Paramyxoviridae,
Pneumoviridae,
Bornaviridae,
Nyamiviridae,
Sunviridae,
Mymonaviridae,
Artoviridae, Lispiviridae, and
Xinmoviridae [
13,
14]. The genomes of the mononegaviruses have the gene order 3’-UTR – core protein genes – envelope protein genes – RNA-dependent RNA polymerase gene – 5′-UTR [
15]. For bornavirus genomes, this corresponds to the gene order 3′-UTR – nucleoprotein (N) gene – phosphoprotein (P) gene – matrix protein (M) gene – glycoprotein (G) gene – polymerase (L) gene – 5′-UTR [
16‐
18]. Within the phosphoprotein gene, there is also an overlapping open reading frame (ORF) encoding the X protein, which is involved in regulation of polymerase activity [
19,
20] and inhibition of type I interferon signalling and apoptosis [
21,
22]. In the family
Nyamiviridae, the genomes of the three viruses constituting the genus
Nyavirus (Nyamanini virus, Midway virus and Sierra Nevada virus) exhibit the gene order 3′-UTR – N gene – ORF II gene – P gene – ORF IV gene – G gene – L gene – 5′-UTR. The ORF II protein of nyaviruses is a negative regulator of the polymerase activity, and ORF II and ORF IV are suggested to form a two-complex matrix protein [
23]. The mymonaviruses are unique among the mononegaviruses because they encode the N protein in ORF II and have an ORF downstream of the L protein [
24].
In 2014, Økland and colleagues described two rhabdoviruses infecting salmon lice: Lepeophtheirus salmonis rhabdovirus No 9 (LSRV-No9) and Lepeophtheirus salmonis rhabdovirus No 127 (LSRV-No127). These viruses are present in the glandular tissue of the louse and have a high prevalence in all developmental stages. Viral RNA is also present in the skin of the salmon surrounding the site where chalimi were attached, but the viruses do not replicate in selected fish cell cultures [
9]. The viruses do not significantly affect the developmental rate, survival or fecundity of the salmon louse. However, infected lice appear to induce a dampened inflammatory response in salmon compared to virus-free lice [
25]. Virus-free salmon louse strains have been established through RNAi-mediated treatment of the viruses, and studies have indicated that LSRV-No9 is transmitted both vertically and horizontally [
26]. Recently, a related rhabdovirus genome was described from
Caligus rogercresseyi: Caligus rogercresseyi rhabdovirus Ch-01 (CrRV-Ch01). CrRV-CH01 clusters phylogenetically with the two other caligid rhabdoviruses to form the newly created genus “
Caligrhavirus” (awaiting ratification by the ICTV) within the family
Rhabdoviridae. CrRV-Ch01 differs from LSRV-No9 and LSRV-No127 by having an additional ORF with unknown function [
27]. Here, we describe the genome, phylogeny, tissue tropism and prevalence of a third putative virus from
L. salmonis, Lepeophtheirus salmonis negative-stranded RNA virus 1 (LsNSRV-1), which shows similarities to artoviruses.
Discussion
The family
Mononegavirales currently consists of 11 families [
14]. LsNSRV-1 clusters phylogenetically with members of the
Artoviridae, a family containing seven other arthropod viruses, including PpNSRV-1. Artoviruses have five ORFs with gene lengths similar to those presented here for LsNSRV-1. ORF IV and ORF V share similarities with mononegaviral G proteins and L proteins, respectively. The possible functions of ORF I-III of artoviruses have not yet been examined.
The nucleoprotein of mononegaviruses is most commonly encoded by ORF I, with the exception of mymonaviruses where ORF I encodes a possible membrane protein and the nucleoprotein is encoded by ORF II [
24]. The hypothetical protein encoded by LsNSRV-1 ORF I shares no characteristics with any known viral nucleoproteins, but it does show sequence similarity to several hypothetical ORF I proteins from unclassified mononegaviruses and PpNSRV-1. The LsNSRV-1 ORF I protein also contains a possible late domain, YPDL, corresponding to the YXXL late domain of membrane and Gag proteins of arenaviruses, paramyxoviruses and retroviruses [
57‐
60]. Late domains are often proline-rich and are usually found in membrane proteins interacting with proteins of the endosomal sorting complexes required for transport (ESCRT) machinery, thus facilitating virion budding [
61]. Such domains are also found in the nucleoproteins of arenaviruses, filoviruses, paramyxoviruses and retroviruses, where they are described to function as accessory factors for virion budding [
61‐
64]. Whether the LsNSRV-1 ORF I protein primarily functions as a nucleoprotein or a membrane protein, the predicted late domain and the presence of an additional proline-rich region at the C-terminus suggest that the protein is involved in virion budding.
The phosphoprotein of mononegaviruses is a multifunctional protein acting as a cofactor for the RNA-dependent RNA polymerase complex [
65]. For most mononegaviruses, the phosphoprotein is encoded by ORF II. ORF II may also encode other proteins in addition to the phosphoprotein [
15,
19,
20,
66,
67]. For Nyamanini virus, the ORF II protein has been suggested to function as a matrix in a complex with the ORF IV protein [
23]. The ORF II protein of LsNSRV-1 shows no resemblance to any of these proteins, and its putative function remains unknown. The ORF III of Nyamanini virus encodes an approximately 400-aa-long protein that functions as a polymerase cofactor [
23]. The putative phosphoproteins of all three members of the genus
Nyavirus are predicted to contain two coiled-coil regions at the N- and C-terminal ends of the protein [
7]. The hypothetical ORFIII protein of LsNSRV-1 shows no sequence similarity to these proteins. Nevertheless, given that the LsNSRV-1 ORF III protein also has two predicted coiled-coil regions and is similar in size to the Nyamanini virus ORF III protein, combined with the fact that its gene is in the same position, it is likely that the LsNSRV-1 ORF III protein has a function similar to that of the putative phosphoproteins of nyaviruses. However, the domains
xPFSAP
x and
xLDRLF
x could represent the two late domains PT/SAP and LXXLF found in the matrix proteins of arenaviruses, filoviruses, rhabdoviruses and the Gag proteins of retroviruses [
59,
68‐
70]. Thus, the LsNSRV-1 ORF III protein could also be a matrix protein involved in virion budding.
Based on sequence analysis of the hypothetical LsNSRV-1 ORF IV protein, its genome position, and the presence of a signal peptide and a transmembrane region, ORF IV is predicted to encode the G protein. The sequence similarity of the hypothetical LsNSRV-1 ORF V protein to other polymerases and the presence of several conserved domains related to the function of the polymerase indicate that ORF V encodes an RNA-dependent RNA polymerase.
Nucleorhabdoviruses and dichorhaviruses (family
Rhabdoviridae), nyaviruses and bornaviruses replicate in the nucleus [
23,
71,
72], and for bornaviruses, the nucleolus has been identified as the site of replication [
73]. The nucleocytoplasmic trafficking of the ribonucleoprotein (RNP) complex is mediated by viral proteins possessing NLSs and NESs [
74]. In bornaviruses, NLSs are found in the nucleoprotein, phosphoprotein, the non-structural protein p10, and the polymerase [
75‐
78]. NLSs have also been reported to be present in the nucleoproteins and phosphoproteins of nucleorhabdoviruses and an unclassified
Culex tritaeniorhynchus rhabdovirus [
72,
79‐
81]. One leucine-rich domain in the nucleoprotein and one methionine-rich domain in the phosphoprotein of bornaviruses have been identified as NESs [
82,
83]. Leucine-rich NESs have also been described in the C protein and nucleoprotein of morbilliviruses and the phosphoprotein of rabies virus [
84‐
86]. Our analysis suggests the presence of NLSs in the ORF III protein and the polymerase, and one NES in the ORF I protein of LsNSRV-1, suggesting that of LsNSRV-1 replicates in the nucleus. There are examples of viruses with proteins exhibiting NLSs and NESs that replicate in the cytoplasm. The NLSs and NESs of both morbillivirus nucleoprotein and rabies virus phosphoprotein mediate nucleocytoplasmic trafficking of the protein, and both are involved in blocking of the IFN response [
85‐
88]. However, the presence of both genomic RNA and viral mRNA of LsNSRV-1 in the nucleolus and the low efficiency of viral knockdown observed after treatment of lice with dsRNA targeting LsNSRV-1 ORFI indicate that this virus most likely replicates in the nucleus. While the cytoplasmic LsRVs have previously been shown to be entirely removed from lice by RNAi with only half the concentration of dsRNA used in this study [
26], the presence of a nuclear reservoir of LsNSRV-1 might prevent efficient clearance of the virus by RNAi. Given that the virus particles have not been observed and that we were not able to cultivate the putative virus, one could argue that the virus is endogenous and that this prevents dsRNA-mediated removal of the virus. However, a viral genome incorporated into the
L. salmonis genome with no exogenous phase should only be present in the cytoplasm as mRNA, and not as both mRNA and genomic RNA, as demonstrated by
in situ hybridization. Moreover, the successful ligation and complementary termini of the putative viral genome strongly suggest that it is not incorporated in the host genome.
Arboviruses rely on horizontal transmission, mainly through feeding and infection of the arthropod’s vertebrate host [
89,
90]. The dampened salmon immune response and higher parasitic success of lice infected with LSRV-No9 and LSRV-No127 suggest that these viruses have adapted to promote horizontal transmission. Like LSRVs, LsNSRV-1 is present in several glands that have ducts ending in cuticular pores on both the ventral and dorsal side of the salmon louse [
91]. Viral RNA of LsNSRV-1 is also present in the gut and salivary glands. This could allow viral particles to be excreted and thus enable horizontal transmission. However, LsNSRV-1 has not been found in substantial amounts in the skin of salmon, and there is no evidence of replication in salmon. Vertical transmission of viruses in arthropods mainly relies on maternal transmission, though these viruses are also dependent on horizontal transmission in order to persist in the host population [
89,
90,
92‐
94]. Vertical transmission from both males and females has currently only been reported in sigmaviruses, a possible reovirus, and PpNSRV-1. These viruses have been shown to persist in the host population without horizontal transmission [
12,
95,
96]. Due to the presence of the viral genome of LsNSRV-1 in the genital products of both sexes of the salmon louse, and in the developing embryos and newly hatched nauplii, it is likely that LsNSRV-1 is transmitted vertically. The dense staining of LsNSRV-1 RNA in the vas deferens and spermatophore sac also indicate that the virus may be transmitted horizontally from males to females via seminal fluids as shown for LSRV-No9 [
26]. Interestingly, a large variation in the amount of viral RNA was seen in the offspring of dsRNA-treated females, despite the relatively stable knockdown of viral RNA in adult females. Since a large variation in viral RNA levels was seen in the adult males as well, it is possible that LsNSRV-1 was transmitted vertically from the males to their offspring. Unfortunately, the experimental setup did not allow us to distinguish which male fertilized which female, and future production or identification of LsNSRV-1-free louse strains is needed to confirm such vertical transmission.
Understanding the role of the viruses infecting
L. salmonis could be vital for the control of this parasite. Indeed, LSRV-No9 and LSRV-No127 infection enhances the parasitic success of
L. salmonis [
25]. LsNSRV-1 does not seem to infect salmon, as viral RNA was only present in the skin in small amounts, and it was not possible to cultivate the virus in the fish cell cultures that were tested. The close coexistence of salmon lice and salmon frequently exposes the viruses infecting salmon lice to potential new hosts. The host range of a virus is generally dependent on multiple genes encoding structural or non-structural proteins. Mutation, recombination or reassortment of these genes may facilitate a change in the host range of the virus [
97]. Such events are probably very rare [
98], and host shifts are most often observed between closely related hosts [
99]. However, all arboviruses have undergone an interphyletic host shift at some point in time, and it has also been shown by Li et al. [
100] that the plant pathogen tobacco ringspot virus underwent an interkingdom host shift to be able to infect and replicate in the honey bee
Apis mellifera. It is therefore possible that viruses in the blood-feeding salmon lice could pose a risk to Atlantic salmon. Surveying and characterization of the virome of salmon lice could thus be of value for the fish farming industry. Clearly, more research is needed to clarify the effect of LsNSRV-1 on its host and to assess the risk of a host shift to Atlantic salmon.