Virus-host interactions determine virus transmissibility and virulence, and underlie coevolution that shapes interesting biological phenomena such as the genetic architecture of host resistance and host range. Characterization of the virus factors that exert selective pressure on the host, and the host genes which underlie resistance and adaptation against viruses will help to define the mechanistic pathways embroiled in host-virus coevolution. In this thesis, I describe the viral causes and host consequences of host-virus coevolution. These include genomic signatures consistent with antagonistic coevolution in antiviral RNA interference pathway genes such as high rates of positive selection and polymorphism, loci that underlie genetic variation in resistance to virus infection, and apparent conflict between NF-κB signalling and DNA virus infection. The RNA interference (RNAi) pathway is the most general innate immune pathway in insects, underlined by the observation that many viruses encode suppressors of RNAi (VSRs). The relationship between RNAi and VSRs has garnered attention as a plausible battleground for host-virus antagonistic coevolution, and genomic patterns in Drosophila support this hypothesis. However, genomic patterns in the N-terminal domain of the key RNAi effector gene, Argonaute-2, have not been described. In Chapter 2, I sequence the Argonaute-2 N-terminal domain using PacBio long-read sequencing technology to describe variation within and across Drosophila species, and test whether this variation is associated with resistance to Drosophila C Virus. The RNAi pathway evolves adaptively in Drosophila, but this has not been formally extended across invertebrate species. In Chapter 3, I quantify rates of adaptive protein evolution and describe evidence for selective sweeps in RNAi pathway genes using population genomic data from 8 insect and nematode species. These analyses indicate that RNAi genes involved in suppression of transposable elements and defence against viruses evolve rapidly across invertebrates, and I identify genes with signatures of elevated adaptation in multiple insect species. Host genes that underlie host-virus interactions have been described in RNA virus infection of Drosophila, however substantially less attention has focussed on the host response to DNA viruses, primarily because no DNA viruses have been isolated from Drosophila. In Chapter 4, I describe the isolation of Kallithea virus, a Drosophila dsDNA nudivirus, and characterise the host response to infection and genetic variation in resistance. I find that Kallithea virus infection causes early male-specific lethality, a cessation of oogenesis, and induction of undescribed virus-responsive genes. Further, I describe genetic variation in resistance and tolerance to Kallithea virus infection, and identify a potential causal variant for virus-induced mortality in Cip4. Insect viruses commonly encode viral suppressors of RNAi, however there are a multitude of antiviral immune mechanisms besides RNAi which may select for viral-encoded inhibitors. In Chapter 5, I describe the requirement for RNAi and NF-κB in immunity against Kallithea virus, and map gp83 as a virus-encoded inhibitor of NF-κB signalling. I find that gp83 inhibits Toll signalling at the level of, or downstream of NF-κB transcription factors, and that this immunosuppressive function is conserved in other nudiviruses.

The coconut rhinoceros beetle, an economically important pest of coconut and oil palms, is effectively managed by application of its natural pathogen, the Oryctes nudivirus (OrNV), which act as a bioinsecticide. While this approach offers an environment-friendly alternative to chemical pesticides, the current method of production in infected larvae suffers from inconsistencies in virus productivity and purity. While the anchorage-dependent DSIR-HA-1179 insect cell line has been identified as a susceptible and permissive host for OrNV and therefore would be suitable for the in vitro mass production of the virus, no attempts have been made toward the mass production of the virus, because of the technological challenges that working with DSIR-HA-1179 cells represent. Thus, the main objective of this research was to develop processes for the in vitro production of OrNV in the DSIR-HA-1179 cell line. Knowledge of the growth kinetics and metabolic properties of the host cell line in a chosen culture medium, as well as the selection of an appropriate infection strategy, form the basis for the rational development of bioreactor-based virus production processes. However, characterization of these properties in the DSIR-HA-1179 cell line has been virtually precluded, due to its strongly adherent growth characteristics and the lack of a reliable method to accurately dissociate and count cells grown in monolayers. Using TrypLE™ Express enzyme, a technique allowing the precise counting of cells was developed. The cell line was adapted to grow in four serum-supplemented culture media: TC-100, IPL-41, Sf-900 II and Sf-900 III, which were then individually screened for cell growth and virus production in 25 cm2 attached T-flask cultures. TC-100 supplemented with 10% fetal bovine serum was chosen as a suitable culture medium, based on its capacity for achieving a high cell yield and OrNV production. The cell line metabolism was characterized with respect to nutrient consumption and metabolites production in this culture medium. Glucose, along with glutamine were found to be the nutrients that were consumed faster and to a greater extent, while other amino acids were not consumed to a significant degree. The production of metabolites was characterized by non-production of lactate and ammonia, and production of alanine, as a non-toxic alternative to ammonia. The influence of cell density (CD) at time of infection (TOI) and multiplicity of infection (MOI) on OrNV production was evaluated in T-flask cultures that were infected at different CDs at the TOI and a range of MOIs. The CD at TOI was found to significantly influence OrNV yields, while MOI influenced the dynamics of infection. The cell density effect was found to exist for the DSIR-HA-1179/OrNV system with the progressive decline in cell-specific yield beginning at low cell densities. It was found that in order to maximize OrNV volumetric yield, a combination of MOI and CD at TOI should be selected that allows to keep the maximum cell density reached by the infected culture within a range between 5.0 and 7.0 x 105 viable cells/ml. The roller bottle system was evaluated for its potential to scale-up DSIR-HA-1179 cell growth and OrNV production, and culture parameters were optimized for the improvement of cell and virus yields. An inoculum density of 3.3 x 105 cells/ml and culture volume of 60 ml resulted in the highest cell yield of 1.5 x 106 cells/ml, in 490 cm2 roller bottles. It was found that an optimal infection strategy for roller bottle cultures, which represented the most efficient use of viral inoculum, involved infecting cells at a density of 5.0 x 105 cells/ml and at a MOI of 1. The resulting OrNV volumetric yield of 2.5 x108 TCID50/ml, improved significantly the viral yields obtained in attached T-flask cultures infected under similar conditions (6.8 x 107 TCID50/ml). The microcarrier system was also evaluated for culturing DSIR-HA-1179 cells and producing OrNV in spinner flask bioreactors. Three types of microcarriers (Cytodex-1, Cytodex-3 and Cultispher-G microcarriers) were screened for their ability to support DSIR-HA-1179 growth. Cells attached to Cytodex-1 and 3, but failed to attach to Cultispher-G microcarriers. The final cell density reached in microcarrier culture was dependent on bead type and concentration, and the cell to bead ratio. At an optimal bead concentration of 1 mg/ml and cell to bead ratio of 30, cells grew to a maximum density of 1.7 x 106 cells/ml on Cytodex-1, but only to 1.3 x 106 cells/ml on Cytodex-3 microcarriers. Since it supported higher cell yields, Cytodex-1 was chosen to study the kinetics of OrNV production in this system. Microcarrier cultures infected at a cell density of 5.0 x 105 cells/ml and a MOI of 1, produced OrNV at 1.4 x 108 TCID50/ml, which was higher than the yield obtained in T-flask cultures infected under similar conditions. A framework of knowledge on the physiology, metabolism and growth kinetics of the DSIR-HA-1179 insect cell line has been developed in this thesis. In addition, the feasibility of using roller bottles and microcarrier systems for the in vitro production of the virus has been ascertained. It is envisaged that these findings will contribute to the future development of a large-scale industrial process for the production of the OrNV biopesticide.

Chez la guêpe parasitoïde Venturia canescens, des particules virales dépourvues d'ADN appelées VLP (pour "Virus-like Particules") sont produites spécifiquement dans les ovaires et tapissent le chorion des oeufs qui sont injectés dans la chenille hôte. Les VLP ont une fonction immunosuppressive pour l'hôte parasité et permettent ainsi la survie des oeufs du parasitoïde. Ces VLP résultent de l’intégration d’un nudivirus dans le génome de l’ancêtre de la guêpe, nudivirus qui a été ensuite domestiqué pour former des liposomes viraux capables de véhiculer dans l’hôte des protéines de virulence d'origine cellulaire. L’étude réalisée au cours de cette thèse a eu pour objet, d’une part, d'étudier les mécanismes de domestication virale qui ont conduit au virus symbiotique endogène actuel nommé VcENV (pour V. canescens endogenous nudivirus) et d’autre part, d'apporter des éléments de réponse sur le processus de morphogénèse et le mode d'action parasitaire des VLP. / Viral particles devoid of DNA called VLPs (for Virus-Like Particles) are specifically produced in the ovaries of the parasitoid wasp Venturia canescens and line the chorion of the wasp’s eggs injected into the host caterpillar. VLPs are immunosuppressive and allow parasitoid eggs survival. These VLPs result from the integration of a nudivirus into the wasp ancestor genome, nudivirus which was then domesticated to form viral liposomes capable of carrying, into the host, virulence proteins of cellular origin. The aim of the study carried out during this thesis was, first, to analyze the viral domestication mechanisms that led to the current endogenous symbiotic virus called VcENV (for V. canescens endogenous nudivirus) and secondly to provide some answers on VLPs morphogenesis process and parasitic mode of action.