I was prompted to this post by the breathless and much-hyped response to the discovery – the repeated discovery should I say; there was an earlier one that gets glossed over – of a multicomponent flavirus-like virus, this time in mosquitoes.
The actual report was published here: it is a well-done study, describing
“…a genetically distinct, segmented virus isolated from mosquitoes that also exhibits homology to viruses in the familyFlaviviridae and that appears to be multicomponent …, with each genome segment separately packaged into virions”
The authors say
“Although multicomponent genomes are relatively common among RNA viruses that infect plants and fungi, this method of genome organization has not previously been seen in animal viruses [my emphasis]“
…which is why there’s all the hype, of course: claiming the virus “…breaks the rules of infection” is simply incorrect, because it is in fact related to very well characterised single-component ssRNA+ viruses of arthropods and mammals – flaviviruses – and infects its mosquito host exactly as these do, except with its genome in separate particles. Which makes it similar to quite a few plant viruses, several of which are, incidentally, probably evolutionarily related to viruses infecting insects – but more later.
Thus, a claim like “…a new study published Thursday is making researchers rethink how some viruses could infect animals” is simply hype. But it is a sort of hype familiar to plant virologists, who after all showed that multicomponent viruses (=viruses with multipartite genomes packaged in separate particles) existed over 50 years ago – and who also showed that gene silencing was a factor in plant resistance to viruses long before their better-funded animal-researching colleagues got in on the act, but that is another story.
The way in which multicomponency was discovered with plant viruses is interesting: it relied on the fact that plants can respond with local lesions – qualitatively the same as plaques in bacterial or animal cell lawns – to mechanical infection, and that this can be used an an accurate assay of virus titre, as for phages or animal viruses (see here). It became evident, though, that certain plant viruses produced significantly steeper lesion vs dilution curves than were expected from “one-hit” kinetics, where infection with a single virus particle sufficed to cause a lesion.
This is best shown by a plot like the one below, modified from REF Matthews’ Virology, 3rd Edn, attributed to Lous van Vloten-Doting from 1968. This shows the curves obtained from accurate and painstaking local lesion assays with the single-component Tobacco necrosis virus (TNV), and the multicomponent Alfalfa mosaic virus (AMV): both are ssRNA+ and have isometric particles, but TNV has a single-component genome, and AMV a tripartite genome packaged in 3 particles.
The insect virus investigators did much the same thing:
“We used a similar approach to assay the nature of segment packaging for GCXV using cell culture plaques instead of leaf lesions. The dose-response curve for GCXV differed significantly from expectations for a single-component virus (i.e., the number of plaques decreased more quickly than expected with dilution of the inoculant)…we used our dose-response curve to estimate the presence of 3.27 ± 0.37 distinct GCXV particles required for plaque formation”
…but with the addition of rapid sequencing techniques not available in 1968, to show that indeed, the different segments were 5 distinct pieces of ssRNA, 3 mono- and 2 tricistronic (=3 ORFs), with the 2 largest monocistronic pieces being similar to flavivirus NSPs and the 3 smallest not encoding anything similar to sequences in the databases. Four RNAs were essential for infectivity, while the smallest appeared dispensable. Particles formed during infection of cultured cells were enveloped and 30-35 nm in diameter, considerably smaller than flavivirus virions.
This is a very interesting finding, although not unique: similar viruses were previously found in ticks in 2014, when the authors claimed that:
“To our knowledge, JMTV is the first example of a segmented RNA virus with a genome derived in part from unsegmented [flavi]viral ancestors“
They were also wrong: there are a number of viruses for which this could have been said years ago, like the picornavirus superfamily-related comoviruses of plants. These have two-component genomes which both encode polyproteins, one with non-structural and the other with structural ORFs. In fact, an evolutionary precursor to such viruses could be the more closely picornavirus-related dicistroviruses of insects, which have a classic picornavirus precursor polyprotein ORF split into two, with the structural protein ORF at the 3′ end and the regulatory or non-structural polyprotein at the 5′ end.
I got into this because it irked me mildly that such a fuss was being made of a second group of animal-infecting multicomponent ssRNA viruses, when the multicomponent plant virus precedent and history was VERY well established – but then got more interested when speculation started about what advantage multicomponency could confer on a virus.
I have thought for years that people discussing this generally have it backwards: it’s not that having a divided genome in separate particles offers advantage(s), it’s that it is not a DISadvantage in some circumstances – and particularly where there is no selection against the state.
A reason that multicomponency HAS been seen quite frequently with plant viruses could be that mechanically-transmitted viruses can reach VERY high concentrations in infected plants, and even obligately vector-transmitted viruses (eg: the bicomponent ssDNA begomoviruses, multicomponent ssDNA nanoviruses) reach quite high concentrations in the phloem tissue to and from which they are transmitted, compared to viruses in vertebrates.
This is also true for viruses of arthropods compared to vertebrate viruses: dicistroviruses in aphids can reach concentrations that are comparable to those of viruses like TMV in plants, to the point that aphids inject enough virus into plants that our lab originally mistook Rhopalosiphum padi virus for a plant virus. Moreover, plant virus virions often aggregate into quasi-crystalline arrays which can be hard to separate and which are even visible inside insect vectors, thus virtually guaranteeing that >1 virion will be present in any inoculum, even if significantly diluted.
This is most definitely NOT the case for vertebrate viruses, even where the same virus infects both an arthropod and a vertebrate host: the titre in the latter is guaranteed to be orders of magnitude lower, largely due to a more sophisticated immune system keeping viraemia in check. Thus, high inoculum concentrations relative to vertebrate viruses, and a tendency to aggregate, mean there is no DISadvantage inherent in multicomponency.
Having said this, there may be advantages to having a multicomponent genome: one such is presented in a recent article by Sicard et al. (2013), (thanks, @LauringLab and @DiagnosticChick!) in a study of the ssDNA nanovirus Faba bean necrotic stunt virus (FBNSV), which has an 8-component genome of ~1 kb/segment, encapsidated in 8 virions. They proposed:
“…that the differential control of gene/segment copy number may represent an unforeseen benefit for multipartite viruses, which may compensate for the extra costs induced by the low-frequency segments”
Thus, multicomponent viruses may achieve the sorts of gene dosage control only possible in viruses with larger genomes, by virtue of having multiple genome components rather than control elements which add genomic bulk.
Another possible advantage that I recall being touted by plant virological luminaries is the ease of reassortment compared to recombination: this is exemplified by the reo- and orthomyxoviruses, albeit in vertebrates, where they are constrained by having to have all genome components in the same capsid to guarantee infectivity.
I think Vincent Racaniello is correct in the breathless article I quoted in opening, where he is quoted as saying
“There’s so much we don’t know about viruses…We should always expect the unexpected.”
Absolutely. And I think it’s a safe bet that a LOT more multicomponent viruses will be found in arthropods – and even in some vertebrates, to which they will have been transmitted by arthropods. Because that’s the link between many of these viruses: an evolutionary history that involves plants and arthropods, or arthropods and other animals, at an early stage of life on land. Because that’s all there was for advanced eukaryotes, early on: primitive vascular plants, insects that preyed on them and on each other, and protists.
ViroBlogy 