Herbivorous insects have evolved many mechanisms to overcome plant chemical defences, including detoxification and sequestration. Herbivores may also use toxic plants to reduce parasite infection. Plant toxins could directly interfere with parasites or could enhance endogenous immunity. Alternatively, plant toxins could favour down-regulation of endogenous immunity by providing an alternative (exogenous) defence against parasitism. However, studies on genomewide transcriptomic responses to plant defences and the interplay between plant toxicity and parasite infection remain rare. Monarch butterflies (Danaus plexippus) are specialist herbivores of milkweeds (Asclepias spp.), which contain toxic cardenolides. Monarchs have adapted to cardenolides through multiple resistance mechanisms and can sequester cardenolides to defend against bird predators. In addition, high-cardenolide milkweeds confer monarch resistance to a specialist protozoan parasite (Ophryocystis elektroscirrha). We used this system to study the interplay between the effects of plant toxicity and parasite infection on global gene expression. We compared transcriptional profiles between parasite-infected and uninfected monarch larvae reared on two milkweed species. Our results demonstrate that monarch differentially express several hundred genes when feeding on A. curassavica and A. incarnata, two species that differ substantially in cardenolide concentrations. These differentially expressed genes include genes within multiple families of canonical insect detoxification genes, suggesting that they play a role in monarch toxin resistance and sequestration. Interestingly, we found little transcriptional response to infection. However, parasite growth was reduced in monarchs reared on A. curassavica, and in these monarchs, several immune genes were down-regulated, consistent with the hypothesis that medicinal plants can reduce reliance on endogenous immunity. K E Y W O R D SAsclepias, cardenolides, immunity, Lepidoptera, RNA-Seq, secondary metabolites

Many plants express induced defenses against herbivores through increasing the production of toxic secondary chemicals following damage. Phytochemical induction can directly or indirectly affect other organisms within the community. In tri-trophic systems, increased concentrations of plant toxins could be detrimental to plants if herbivores can sequester these toxins as protective chemicals for themselves. Thus, through trophic interactions, induction can lead to either positive or negative effects on plant fitness. We examined the effects of milkweed (Asclepias spp.) induced defenses on the resistance of monarch caterpillars (Danaus plexippus) to a protozoan parasite (Ophryocystis elektroscirrha). Milkweeds contain toxic secondary chemicals called cardenolides, higher concentrations of which are associated with reduced parasite growth. Previous work showed that declines in foliar cardenolides caused by aphid attack render monarch caterpillars more susceptible to infection. Here, we ask whether cardenolide induction by monarchs increases monarch resistance to disease. We subjected the high-cardenolide milkweed A. curassavica and the low-cardenolide A. syriaca to caterpillar grazing, and reared infected and uninfected caterpillars on these plants. As expected, monarchs suffered less parasite growth and disease when reared on A. curassavica than on A. syriaca. We also found that herbivory increased cardenolide concentrations in A. curassavica, but not A. syriaca. However, cardenolide induction in A. curassavica was insufficient to influence monarch resistance to the parasite. Our results suggest that interspecific variation in cardenolide concentration is a more important driver of parasite defense than plasticity via induced defenses in this tri-trophic system.

13

Humoral and cellular immune responses provide animals with major defences against harmful pathogens. While it is often assumed that immune genes undergo rapid diversifying selection, this assumption has not been tested in many species. Moreover, it is likely that different classes of immune genes experience different levels of evolutionary constraint, resulting in varying selection patterns. We examined the evolutionary patterns for a set of 91 canonical immune genes of North American monarch butterflies (Danaus plexippus), using as an outgroup the closely related soldier butterfly (Danaus eresimus). As a comparison to these immune genes, we selected a set of control genes that were paired with each immune for approximate size and genomic location. As a whole, these immune genes had a significant but modest reduction in Tajima's D relative to paired‐control genes, but otherwise did not show distinct patterns of population genetic variation or evolutionary rates. When further partitioning these immune genes into four functional classes (recognition, signalling, modulation, and effector), we found distinct differences among these groups. Relative to control genes, recognition genes exhibit increased nonsynonymous diversity and divergence, suggesting reduced constraints on evolution, and supporting the notion that coevolution with pathogens results in diversifying selection. In contrast, signalling genes showed an opposite pattern of reduced diversity and divergence, suggesting evolutionary constraints and conservation. Modulator and effector genes showed no statistical differences from controls. These results are consistent with patterns found in immune genes in fruit flies and Pieris butterflies, suggesting that consistent selective pressures on different classes of immune genes broadly govern the evolution of innate immunity among insects.

Aphids, like most animals, mount a diverse set of defenses against pathogens. For aphids, two of the best studied defenses are symbiont-conferred protection and transgenerational wing induction. Aphids can harbor bacterial symbionts that provide protection against pathogens, parasitoids and predators, as well as against other environmental stressors. In response to signals of danger, aphids also protect not themselves but their offspring by producing more winged than unwinged offspring as a way to ensure that their progeny may be able to escape deteriorating conditions. Such transgenerational wing induction has been studied most commonly as a response to overcrowding of host plants and presence of predators, but recent evidence suggests that pea aphids (Acyrthosiphon pisum) may also begin to produce a greater proportion of winged offspring when infected with fungal pathogens. Here, we explore this phenomenon further by asking how protective symbionts, pathogen dosage and environmental conditions influence this response. Overall, while we find some evidence that protective symbionts can modulate transgenerational wing induction in response to fungal pathogens, we observe that transgenerational wing induction in response to fungal infection is highly variable. That variability cannot be explained entirely by symbiont association, by pathogen load or by environmental stress, leaving the possibility that a complex interplay of genotypic and environmental factors may together influence this trait.