Under climate change, shifts in precipitation and temperature regimes are expected to impact ecosystem structure and function. These impacts may be determined by feedbacks between plants and their microbes, including both endophytic fungal symbionts and viral pathogens. These feedbacks may be altered by changes in precipitation and temperature which could change endophyte or virus rates of infection. To test for such effects, we manipulated temperature, precipitation, and endophyte infection in a grassland in northern Kentucky. Our experiment consisted of five blocks, each with a factorial manipulation of precipitation and temperature (total of 20 plots). Temperature was increased 3^oC, day and night, year-round, and precipitation plots received a 30% increase in long-term growing season mean annual precipitation. Plots consisted of a mixture of forage species including tall fescue (Schedonorus phoenix; 50/50 endophyte-infected and endophyte-free), Bermuda grass (Cynodon dactylon), hairy crabgrass (Digitaria sanguinalis), and Kentucky bluegrass (Poa pratensis).  After one growing season, fescue was tested for infection with three species of barley and cereal yellow dwarf viruses (B/CYDVs). B/CYDVs are aphid-vectored, host-generalist plant viruses that are widespread in natural and agricultural grasslands. Since endophytes produce alkaloids which can deter aphids from feeding, B/CYDV prevalence should be lower in endophyte-infected plants. However, endophyte infection can confer drought resistance to its host by increasing host water uptake and storage while reducing transpiration loss. This may increase aphid feeding under drought conditions which could increase the prevalence of B/CYDVs despite endophyte induced alkaloids. Thus how alterations to temperature and precipitation regimes will alter these interactions remains unclear.

Results/Conclusions

B/CYDV prevalence was significantly lower in tall fescue that were endophyte-free and grown under elevated temperature than any other treatment (endophyte*temperature interaction: p=0.02). This suggests an increase in virus prevalence in endophyte-infected plants under increased heat, possibly as a result of increased aphid feeding due to increased water content in endophyte-infected plants. Under increased heat, endophyte-induced drought resistance seems to outweigh herbivore deterring alkaloid production. The increase in B/CYDV infection in the heat treatment was largely driven by an increase in prevalence of the species BYDV-PAV (p=0.06). Overall, this work suggests virus prevalence will increase in plants protected by endophytes and experiencing elevated temperatures. This may lead to an increase in overall virus prevalence in grasslands. Ultimately, predicting the effects of climate change on grassland ecosystems may require that we understand how changes to precipitation and temperature regimes alter the strength and nature of multiple, co-occurring species interactions.