Divergent Viruses Discovered in Arthropods and Vertebrates Revise the Evolutionary History of the Flaviviridae and Related Viruses

Abstract

Viruses of the family Flaviviridae are important pathogens of humans and other animals and are currently classified into four genera. To better understand their diversity, evolutionary history, and genomic flexibility, we used transcriptome sequencing (RNA-seq) to search for the viruses related to the Flaviviridae in a range of potential invertebrate and vertebrate hosts. Accordingly, we recovered the full genomes of five segmented jingmenviruses and 12 distant relatives of the known Flaviviridae ("flavi-like" viruses) from a range of arthropod species. Although these viruses are highly divergent, they share a similar genomic plan and common ancestry with the Flaviviridae in the NS3 and NS5 regions. Remarkably, although these viruses fill in major gaps in the phylogenetic diversity of the Flaviviridae, genomic comparisons reveal important changes in genome structure, genome size, and replication/gene regulation strategy during evolutionary history. In addition, the wide diversity of flavi-like viruses found in invertebrates, as well as their deep phylogenetic positions, suggests that they may represent the ancestral forms from which the vertebrate-infecting viruses evolved. For the vertebrate viruses, we expanded the previously mammal-only pegivirus-hepacivirus group to include a virus from the graceful catshark (Proscyllium habereri), which in turn implies that these viruses possess a larger host range than is currently known. In sum, our data show that the Flaviviridae infect a far wider range of hosts and exhibit greater diversity in genome structure than previously anticipated.

Importance: The family Flaviviridae of RNA viruses contains several notorious human pathogens, including dengue virus, West Nile virus, and hepatitis C virus. To date, however, our understanding of the biodiversity and evolution of the Flaviviridae has largely been directed toward vertebrate hosts and their blood-feeding arthropod vectors. Therefore, we investigated an expanded group of potential arthropod and vertebrate host species that have generally been ignored by surveillance programs. Remarkably, these species contained diverse flaviviruses and related viruses that are characterized by major changes in genome size and genome structure, such that these traits are more flexible than previously thought. More generally, these data suggest that arthropods may be the ultimate reservoir of the Flaviviridae and related viruses, harboring considerable genetic and phenotypic diversity. In sum, this study revises the traditional view on the evolutionary history, host range, and genomic structures of a major group of RNA viruses.

FIG 1

Phylogenetic tree (unrooted) of the family Flaviviridae and jingmenviruses based on conserved domains in the NS3 (upper panel) and NS5 (lower panel). Viruses discovered in the present study are marked with solid black circles. Based on their host types, the virus names are shaded red (vertebrate only), yellow (vertebrate and arthropod), green (plant), blue (nonarthropod invertebrates), or black (arthropod only). The names of previously or newly defined genera/phylogenetic groups are labeled. The trees shown here were inferred using an ML method, although highly congruent topologies were obtained using a Bayesian approach. Asterisks indicate well-supported nodes by both the approximate likelihood-ratio test (aLRT) in the ML method (>0.9) and by posterior probabilities in the Bayesian approach.

FIG 2

Phylogenetic tree of the hepacivirus and pegivirus group based on the NS3 (left panel) and NS5B (right panel) alignments and midpoint rooted for clarity only. Viruses discovered in the present study are marked with solid black circles. The names of previously defined genera are labeled to the right of the phylogenies. The trees shown here were inferred using an ML method, although highly congruent topologies were obtained using a Bayesian approach. Asterisks indicate well supported nodes by both the approximate likelihood-ratio test (aLRT) in the ML method (>0.9) and by posterior probabilities in the Bayesian approach.

FIG 3

Comparison of polyprotein structures among the family Flaviviridae and jingmenviruses, which includes 14 newly discovered viruses and 5 prototypical members of the Flaviviridae described previously. A unified length scale is used for all of the polyproteins. Within each protein, regions containing the key domains for serine protease, RNA helicase, and RdRp are colored blue, orange, and green, respectively. Predicted or known transmembrane domains (gray vertical bars) and cleavage sites for the host signalase (red arrow), viral protease (brown arrow), and other proteases (black arrow) are indicated.

FIG 4

Variation in levels of abundance for the four segments of the jingmenviruses. Abundance is measured as the frequency of transcripts for TCLA and as transcripts per million (TPM) for the remaining viruses. The abundance of JMTV segments are quantified within two libraries: ticks R. microplus (11) and ticks (16), respectively.