by Merry Youle
A double stranded RNA (dsRNA) viral genome, introduced into a host cell, is met by formidable host defenses. The very presence of dsRNA in a eukaryotic or prokaryotic cell announces a viral infection and elicits effective responses, ranging from silencing of the viral mRNAs to apoptosis. Despite that, there are successful dsRNA viruses throughout the biosphere. By 2000, eight families with close to 200 "species" were known to infect bacteria, fungi, plants, and animals. The broad diversity of their hosts notwithstanding, all dsRNA viruses share the same secret to success: they bring their capsid into the cell along with their genome to serve as a safe compartment where they transcribe and replicate their genome. Their dsRNA is never exposed.
For the virus whose virion is but a simple protein capsid, it is the entire virion that enters and persists intracellularly. For others that have additional outer layers of protein and/or a membrane, those layers are removed during cell entry and the inner capsid alone enters the cytoplasm. The outer layers vary greatly from group to group, presumably reflecting adaptations to particular hosts or modes of transmission, while the proteins of the inner capsid, as well as its architecture, are highly conserved among all dsRNA viruses. Doesn't this suggest a common ancestry?
This strategy poses particular challenges, not the least of which is how do you transport something as large as a virion across the cell membrane. Also, since a dsRNA genome is not a suitable template for protein translation or for cellular replicases, these viruses have to bring their own RNA-dependent RNA polymerase (RdRp) with them. The capsid itself has to be selectively porous, allowing nucleotides to enter and RNA transcripts to exit.
Picture a capsid containing both a dsRNA genome and an RdRp that has arrived in the cytoplasm of a susceptible host. Its first job is to transcribe its dsRNA genome inside the capsid to yield genome-length positive sense transcripts that are immediately extruded through holes in the capsid. The transcripts serve two functions, the first being as mRNAs for the synthesis of viral proteins. Secondly, as the capsid proteins accumulate and self-assemble in the cytoplasm, some of the ssRNA transcripts are encapsidated to become the genomic RNA. The final step takes place inside the newly-assembled virions when the ssRNA is used as the template for synthesis of the negative sense strand by the RdRp in virio. This elegant and efficient replication scheme is shared by all known dsRNA viruses.
Quite a feat, really, transcribing a packaged RNA genome in such confined quarters. And for viruses with segmented genomes, such as the bluetongue virus (family Reoviridae), the task is even more daunting. They must simultaneously and repetitively transcribe 19 kb of dsRNA divided among ten different molecules inside an 80 nm capsid. Moreover, this virus also methylates and caps each mRNA before extrusion.
A simpler, well-studied example is the LA virus (family Totiviridae) found in most strains of yeast. Its non-segmented genome is a mere 4.6 kb and contains only two overlapping genes. One of them encodes the major capsid protein, Gag, while the other encodes the RdRp. The RdRp is synthesized as a Gag fusion protein, i.e. as a Gag protein covalently joined to an RdRp tail. The 40 nm capsid is built from 120 Gag protein molecules, two of which have RdRp tails and are located at five-fold vertices. Examination of the capsid structure by X-ray crystallography at 3.4 ? resolution revealed that the RdRp tail faces the interior of the capsid. The transcriptase, being firmly attached to the capsid structure, is stationary while the RNA — approximately 1.3 µm long for even this small genome — is translocated past it. The newly-synthesized strand is extruded through 15 ? diameter holes that are large enough for ssRNA but too small to pass either dsRNA or cellular enzymes.
(We can't resist mentioning that a bluetongue virus capsid contains not two but ten copies of the RdRp, one to process each genome segment. Here, too, the RdRps are located at the five-fold vertices. Since there are but 12 such vertices per capsid, this appears to limit dsRNA viruses to a maximum of 12 genome segments.)
Another LA virus trick: Unlike some other dsRNA viruses, they do not cap their extruded mRNAs, thus leaving them, one would suppose, vulnerable to degradation by the cell. Instead they play a numbers game. The Gag proteins provide an enzymatically active site on the outer capsid surface that removes the caps from host mRNAs, thus flooding the cell's mRNA degradation system with decapped cellular mRNAs. This lets enough of their own mRNAs escape destruction.
Lastly, LA viruses take the usual dsRNA defensive strategy one step farther. Not only does their dsRNA genome never leave the capsid, but their capsids never exit the cell! These non-infectious "viruses" exploit the promiscuous mating habits of their S. cerevisiae hosts and move from cell to cell via the cytoplasmic mixing that accompanies mating. This virus-host relationship is stable and seemingly copacetic; the viruses do not alter their host's phenotype or rate of growth, but rely on some host factors for their own replication.
If you're the kind of person who wants biological entities to adhere to an inviolate set of defining rules, I suppose you could argue that the LA virus isn't a real virus. But if you are that kind of person, many viruses are bound to give you a headache.
Mertens, P. (2004). The dsRNA viruses Virus Research, 101 (1), 3-13 DOI: 10.1016/j.virusres.2003.12.002
Castón JR, Trus BL, Booy FP, Wickner RB, Wall JS, & Steven AC (1997). Structure of L-A virus: a specialized compartment for the transcription and replication of double-stranded RNA. The Journal of cell biology, 138 (5), 975-85 PMID: 9281577