Virus discovery, dynamics, and disease in a wild Drosophila community

Abstract

The fruit fly Drosophila melanogaster remains a key model system for the study of insect-virus interactions. Its tractable life history and associated genetic toolkit have aided in the discovery of many of the key invertebrate immune defences against viruses. Recent metagenomic sequencing studies have identified over 130 naturally-occurring viruses of the Drosophilidae. This could enable the study of insect-virus ecological and evolutionary dynamics in this native multi-host, multi-virus community. In particular, the use of naturally-occurring host-virus combinations allows the study of ‘typical’ wild co-evolutionary dynamics between insects and their viruses, which might help us to understand the evolution of insect-vectored viruses with economic and public health impacts. However, proper parameterisation of these models of insect-virus co-evolution requires data from long-term studies of wild host-virus communities, which do not currently exist. In this thesis, I make the first attempt to quantify the dynamics of wild Drosophilid virus communities at both local, and continent-wide scales. I begin by using metagenomic sequencing to describe the viruses present in populations of Drosophila in South-east Scotland, identifying 17 new RNA viruses. I characterise the host range of 41 new and previously described Drosophila viruses in this system, finding that over 90% of these viruses infect multiple host species. I then examine how ten Drosophila viruses vary in prevalence in the wild using repeated, spatially and temporally structured sampling of a total of 2227 flies of 15 species, over three years. I find that prevalence varies between viruses, and within viruses, can vary across host species, and collecting season. Co-infection is not a rarity in this system, as over 30% of flies are infected with multiple viruses. Expanding my spatial scale of analysis, I then characterise patterns of Drosophila melanogaster DNA virus prevalence and genetic diversity across Europe. I find that DNA viruses show varying levels of diversity, and large-scale spatiotemporal predictability. Finally, I experimentally examine the ability of Drosophila viruses to transmit across species, and quantify their impact on host fitness. I find that some Drosophila viruses can readily transmit across species without the need for systemic injection. I identify four Drosophila viruses which significantly reduce host lifespan, and three which significantly reduce offspring production, indicating that viral infection may be a substantial fitness burden on wild flies. Together, these data increase and demonstrate the utility of the Drosophila model for community-level studies of host-virus interactions.