Effects of barley yellow dwarf virus on plant fitness components in perennial switchgrass
Recent metagenomic studies suggest that virus infections are common across the plant kingdom and within diverse ecosystems. However, there is surprisingly little information on the effects of infection on plant fitness, despite the fact that such data are essential for understanding the role of viruses in plant ecology and evolution. It is often assumed that virus effects will be negative given the pathogenic effects often seen in annual crops. But there are few studies of virus interactions with perennial plants, including native species. Field studies of fitness effects are even rarer. To address this deficit, we established a field experiment with switchgrass (Panicum virgatum), a long-lived North American prairie grass and potential bioenergy feedstock. We used two switchgrass types: a lowland cultivar (Kanlow: Oklahoma origin) and an upland cultivar (Shawnee: Illinois origin). In growth chambers, we inoculated seedlings of both cultivars with a field-collected Kansas isolate of Barley Yellow Dwarf virus (BYDV-PAV); mock-inoculated plants served as controls. After their infection status was verified, plants were transplanted into a naturalized field site in Kansas in competition with Festuca and other previously established grasses. We measured plant virus status, survival, size, and reproduction for two years (Yr 1: 2012; Yr 2: 2013).
Inoculation produced virus infection in both cultivars, as quantified by ELISA, but little apparent symptoms. Despite symptom absence, infection reduced plant biomass production in Yr 1 and altered flowering times in Yr 2. The two cultivars differed in some responses, illustrating the genetic specificity of biotic interactions. In Yr 1, both cultivars produced less biomass when virus-inoculated, with Kanlow more affected (-33%) than Shawnee (-6%). In Yr 2, virus remained detectable in 37.5% of the inoculated Shawnee but only 7.5% of inoculated Kanlow, and response patterns changed. Mean biomass of inoculated plants remained reduced, by -18% for Kanlow and -13% for Shawnee, but the treatment difference was no longer significant. In flowering, Shawnee typically begins earlier than Kanlow, as expected given its higher latitude origin, and this pattern was evident among mock-inoculated plants in Yr 2. However, inoculated Kanlow flowered earlier and Shawnee later than their respective mock-inoculated controls. The directions of these shifts in flowering time are likely to be suboptimal at the sites of cultivar origin. We conclude that virus effects may be hidden from ecologists, given the lack of visual symptoms on experimental plants, but merit further investigation.