Host genotype alters endophyte effects in viral infected grasses

Background/Question/Methods

Variation in commodity exchange, in which mutualists exchange goods and services with their hosts, can influence not only the host-mutualist interaction, but interactions between hosts and other associated biota such as parasites. Thus, in order to fully understand host-parasite interactions, we may also need to consider mutualists. We tested this hypothesis using three common organisms, tall fescue (Schedonorus phoenix), foliar endophytic fungi (Neotyphodium coenphialum), and Barley yellow dwarf virus-PAV. This virus is aphid-vectored and infects hundreds of grass species in both natural and agricultural ecosystems. N. coenphialum is a vertically transmitted intercellular fungus that forms obligate associations with grasses and produces alkaloids. We ran a greenhouse experiment factorially manipulating virus and endophyte infection within two fescue genotypes (KY 31 and PDF). For KY 31, endophyte-free and common toxic endophyte-infected individuals were utilized.  For the PDF genotype, we also evaluated fungal genotype effects by utilizing the common toxic endophyte and a novel endophyte (AR 584) that does not produce the mammalian active alkaloid, ergovaline. Since some alkaloids may deter aphid vectors from feeding, we hypothesized that virus susceptibility would be lower in endophyte-infected plants. Additionally, endophytes may compete with virions, leading to decreased viral production and infection severity in endophyte-infected plants.

Results/Conclusions

We found a fescue genotype difference such that those plants from the KY 31  genotype had double the relative viral titers than the more recently derived PDF genotype (F_1,65=15.646, p<0.0001). Independent of host genotype, endophyte infection decreased relative viral titer by 25% compared to endophyte-free plants (F_1,65=4.016, p=0.031). Plant genotype and endophyte infection both also influenced aboveground biomass. Specifically, PDF plants were always larger than KY 31 plants (F_1,105=11.93, p=0.0008). However, endophyte-infected PDF plants had a larger shoot biomass than endophyte-free plants. This pattern was reversed in KY 31, where endophyte-infected plants had a smaller shoot biomass than their endophyte-free counterparts (genotype x endophyte interaction: F_1,105=5.914, p=0.017). While we found significant main effect differences in the total number of aphid vectors found on plants after three days for virus infection (p=0.003), endophyte infection (p=0.0598), and plant genotype (p=0.0106), there were no interactive effects of these treatments. This indicates that endophyte infection does not change vector performance to decrease host susceptibility to virus infection. Overall, this work suggests that to understand the full impact of virus infection in grass hosts, we may need to not only consider the mutualists with which grasses associate but the genotype-dependency of that association as well.