May 17th, 2018
Advances in DNA sequencing have dramatically increased the rate at which new viruses are being identified. However, the host associations of viruses that have only been identified via sequencing are often open to question.
In 2013, DNA sequences derived from mysterious viruses called ‘cycloviruses’ were detected in patients with severe neurological disease on two continents. In Vietnam, a virus named "Cyclovirus-VN" was identified in association with severe brain infections. Around the same time, a closely related virus was detected in Malawi - again via sequencing - in association with cases of unexplained paraplegia.
The authors of these studies emphasised that a causal link with disease had yet to be determined. Somewhat inevitably though, the impression has been created - at least in some quarters - that cycloviruses are emerging viruses of humans and domestic animals.
"Newly emerging viruses such as Cycloviruses, which are causing neurological problems in children in Asia, are also emerging in sewage and are spreading.". The Guardian
What is certainly true is that cycloviruses continue to be detected in human samples, as well in samples derived from other sources. In most respects, however, they remain a mystery - so are these viruses really a threat to humans?
Cycloviruses are members of the virus family *Circoviridae* (circoviruses). It is now well established that DNA sequences derived from circoviruses - called endogenous circoviral elements (CVe) - occur sporadically in animal genomes. These sequences are thought to be generated via rare events in which DNA derived from the genomes of ancestral circoviruses ended up being incorporated into the germline of host species.
CVe are highly informative with respect to virus evolution because - like fossils - they preserve information from the distant past. For example, they reveal what ancient circoviruses looked like on a molecular level, and which species have been infected with circoviruses in the past.
Earlier this year, we reported the results of a study in which we screened vertebrate genomes for CVe. More recently, we expanded this screen to include invertebrate species. This led us to identify CVe in insect genomes that are apparently derived from cycloviruses.
In collaboration with researchers at the Field Museum of Natural History in Chicago we were able to independently confirm the presence of cyclovirus-derived EVEs in the germline of the elongate twig ant (Pseudomyrmex gracilis), establishing that cycloviruses infect insects (or at the very least, have done in the past).
Next, we used phylogenetic methods to investigate the evolutionary relationships between CVe and all known circoviruses. This allowed us to examine the host associations of circoviruses and CVe in the light of their evolutionary relationships.
When we did this, we noticed that if we only focussed on (i) CVe and (ii) circoviruses with well-established host associations, then the overall pattern of host association was quite well conserved across major branches of the circovirus 'tree'. So for example, bird genomes contain CVe that we assume are derived from an ancestral bird circovirus that - as we previously demonstrated - existed over 38 million years ago. In our tree we saw these CVe clustering closely with beak and feather disease virus a contemporary circovirus of birds.
By contrast, when circoviruses that have only been identified via sequencing were considered, host associations were chaotically distributed across the tree. This was particularly noticeable in the cyclovirus branch of the tree. One example of this was the close grouping of CVe from the genomes of arthropods (e.g. ants and mites) with cycloviruses sequences derived from bat and human samples.
There are a number of possible explanations for the findings of our study, but perhaps the most parsimonious, looking at all the evidence, is that cycloviruses are arthropod viruses that commonly contaminate vertebrate samples. Notably, we found that cyclovirus-derived CVe occur in the genomes of two distinct mite species.
Although we may not like to contemplate it, most surfaces, including our own bodies, are replete with multitudes of tiny arthropods such as dust mites. We can be fairly certain that these animals have viruses too, and it is not too hard to imagine some DNA from these viruses finding its way into samples by one means or another. A tiny particle of dust is probably all that is required.
With respect to this, it is worth noting that all cycloviruses have been identified via metagenomic methods. No cycloviruses have ever been grown in cell culture. In fact, the only confirmed host associations we have for these viruses are with arthropods, via the CVe identified in our study.
Although our findings raise doubts about the role of cycloviruses as human pathogens, it is important to keep an open mind here. Especially since we have really only just begun to explore the cyclovirus diversity.
I have a lot of trouble believing that a non-vectored virus that normally infects insects would replicate in humans. But perhaps some cycloviruses infect insects while others infect birds and mammals, and perhaps this second group can cause indeed disease in humans. Further mapping of cyclovirus distribution and diversity will no doubt be instructive.
Since there is always a risk of being misled by contamination when identifying viruses via sequencing-based approaches, host associations of viruses identified via sequencing should be viewed with caution wherever they are found to strongly contradict well established host associations among well-defined virus groups. Our study shows the utility of endogenous viral elements for establishing host associations when using genomic data to identify and characterise novel viruses. To facilitate these kinds of investigations we are collating all of the endogenous viral sequences identified in our research (as well as any associated auxiliary data) in an openly accessible online repository called DIGS-for-EVEs.
Dennis TPW, Flynn PJ, de Souza WM, Singer JB, Moreau CS, Wilson SJ, and RJ Gifford
Journal of Virology (in press)
Insights into circovirus host range from the genomic fossil record. preprint
Dennis TPW, de Souza WM, Marsile-Medun S, Singer JB, Wilson SJ, and RJ Gifford
Virus Research S0168-1702(17)30904-8
The evolution, distribution and diversity of endogenous circoviral elements in vertebrate genomes. view