RNA interference as a control strategy for the honey bee parasite, Varroa destructor, and the pathogen it vectors, Deformed wing virus

Varroa destructor mites (varroa hereafter) are devastating ectoparasites of the European honey bee (Apis mellifera) and can vector several viral pathogens to bees while feeding. Varroa is most closely associated with the honey bee pathogen Deformed wing virus (DWV). Together, varroa and DWV have detrimental and multifaceted effects on honey bee health and are considered one of the leading causes of colony losses around the world. Currently, beekeepers rely on a few chemical-based pesticides for controlling varroa infestations in their beehives and mitigating the negative effects of DWV. However, some varroa pesticides can have negative effects on bees and reports of resistance to some available pesticides have been increasing among varroa populations. Therefore, development of effective, specific and sustainable treatments for controlling varroa and DWV is important for improving honey bee health as well as providing beekeepers with alternative management strategies.

RNA interference is a highly conserved intracellular mechanism of gene regulation and antiviral defence in eukaryotes, which has widely been proposed as a promising strategy to control pests and pathogens through administration of specific double stranded RNA (dsRNA) sequences. This thesis focuses on investigating the effectiveness of RNAi as a novel control method for varroa and DWV using pest and pathogen-specific dsRNA sequences.

To test effects of DWV-specific dsRNA on DWV loads and symptoms in parasitised and non-parasitised honey bees reared from larvae I conducted laboratory experiments with mini-hives consisting of adult bees, larvae and varroa mites. Bees parasitised during their development had significantly higher DWV loads compared to non-parasitised bees (PERMANOVA, p < 0.010). However, DWV-dsRNA did not reduce DWV loads (PERMANOVA, p = 0.66) or symptoms of mini-hive reared bees (ANOVA, 0.49). RNA-sequencing (RNA-seq) was used to assess the impact of DWV-dsRNA on viral loads and gene expression in brood-parasitising varroa mites. Varroa mites from DWV-dsRNA treated mini-hives did not show an elevated RNAi response or reduced DWV loads. Next, I tested effects of the same DWV-dsRNA sequences at a higher concentration on DWV loads in adult bees reared in either varroa-free or varroa-infested cages. Similarly, bees from varroa-free (PERMANOVA, p = 0.473) and varroa-infested (PERMANOVA, p = 0.151) cages fed DWV-dsRNA did not show significantly reduced DWV loads. Analysis of DWV loads in actively parasitising mites from varroa-infested cages found no significant effects of dsRNA on DWV loads in mites (PERMANOVA, p = 0.200). Interestingly, varroa-infested bees fed GFP-dsRNA (as a non-specific dsRNA control) showed significantly higher DWV loads compared to bees fed sugar water (PERMANOVA, p = 0.036). For varroa-infested cages, bee survival was also negatively affected by GFP-dsRNA and DWV-dsRNA (Cox proportional-hazards model, p < 0.001). These results suggest that effects of dsRNA on honey bees may be influenced by varroa parasitism.

GreenLight Bioscience Inc. (Medford, MA, USA) has recently developed a dsRNA product specific in sequence to a V. destructor mRNA encoding the calcium signalling protein calmodulin (VdCaM-dsRNA) as a potential biopesticide for varroa. Testing this product in field trials conducted in the U.S reduced varroa mite infestation in honey bee colonies. Here, I used laboratory mini-hives of bees and varroa mites to test the effects of two concentrations of VdCaM-dsRNA on varroa reproduction, survival and gene expression. Varroa were sampled from brood cells of developing pupae either 5 days after cell capping or just prior to bees emerging from brood cells, as a measure for time since mite exposure to VdCaM-dsRNA. While VdCaM-dsRNA did not have a significant effect on survival of adult female varroa mites, varroa reproduction was significantly and substantially reduced in high and low dose VdCaM-dsRNA treated mini-hives (PERMANOVA, p < 0.001). Calmodulin expression was only significantly reduced in VdCaM-dsRNA treated mites from brood cells 5 days post capping (Kruskal-Wallis, p < 0.001), while effects of VdCaM-dsRNA on calmodulin expression appeared to have waned in mites from emerging bees (PERMANOVA, p = 0.543). Differential gene expression analysis and gene ontology results suggest that VdCaM-dsRNA induced reduction in calmodulin-mediated calcium signalling dysregulates non-canonical Wnt and notch signalling pathways that are critical for embryogenesis. Finally, given that current varroa pesticides are used to mitigate negative effects of DWV in honey bee colonies I investigated short-term effects of VdCaM-dsRNA on DWV loads and symptoms in mini-hive reared bees. Varroa RNA-sequencing data was also used to assess viral community abundance in mites, with a focus on DWV loads. For mites sampled from brood cells of emerging bees, DWV loads were significantly lower in VdCaM-dsRNA high dose treated mites compared to non-specific dsRNA treated mites (ANOVA, p = 0.040). Mites sampled from brood cells 5 days after cell capping showed no effect of dsRNA treatment on DWV loads. While DWV loads in bees was not significantly affected by VdCaM-dsRNA, DWV loads followed a strong bimodal distribution in dsRNA treated bees, possibly indicating some interacting factors that influence effects of dsRNA on DWV loads in bees.

Results from this thesis show that VdCaM-dsRNA is a highly promising biopesticide for controlling varroa mite infestation levels by significantly reducing mite reproduction in honey bee hives. Moreover, using RNA-seq to investigate the underlying molecular mechanism of VdCaM-dsRNA provided a deeper understanding to molecular pathways important for varroa reproduction and suggests that calmodulin-mediated calcium signalling plays an important role in regulating embryo developmental processes. On the other hand, using DWV-dsRNA to control DWV loads in bees and varroa mites was less successful possibly due to sequence divergence between the DWV-dsRNA sequences used and the DWV variants circulating in New Zealand honey bees and varroa. Possible DWV-dsRNA targets specific to New Zealand DWV variants are suggested, which will be useful for testing this hypothesis and better understanding the role of RNAi as a potential DWV control strategy.