Prevalence patterns and co-infection dynamics of fungal and viral pathogens in a native perennial grass stand along a nitrogen gradient

Michael P Ryskamp, Matthew A Carey and Carolyn M Malmstrom, Plant Biology, Michigan State University, East Lansing, MI, Michael P Ryskamp, Matthew A Carey and Carolyn M Malmstrom, Ecology, Evolution, and Behavior, Michigan State University, Carolyn M Malmstrom, AgBioResearch, Michigan State University

Background/Question/Methods
Nitrogen availability influences plant – pathogen interactions in wild plant communities. However, little is known about N effects on partitioning of host resources among different pathogen groups that co-infect hosts. We asked how does nitrogen addition affect prevalence and co-infection frequency for fungal and viral pathogens of switchgrass (Panicum virgatum)? If pathogens are N limited, then we expected to see prevalence increase with N. Alternatively, negative effects of (co-)infection could result in stand-thinning of disease at high N, where competition is more intense.
We conducted surveys of three phytopathogens, Switchgrass mosaic virus (SwMV), rust (Puccinia sp.), and anther smut (Tilletia sp.), in a ten-year-old switchgrass stand with eight N addition treatments (0—196kg N/ha). SwMV is a recently discovered wild virus that is leaf-hopper vectored. Rust and smut are fungal phytopathogens that are mostly wind- and water-dispersed. Unlike rust, which colonizes leaf and stem tissue, smut is a sterilizing infection that forms spore-filled galls in florets. These pathogens are common in switchgrass plots at the biological station where we conducted our fieldwork. We surveyed 640 individual tillers from independent plants (80 tillers x 8 N levels), and we measured end-of-season tiller size for a subsample at low, medium, and high N.

Results/Conclusions
The overall prevalences of the two fungal pathogens were similar (~20%), but showed contrasting responses to nitrogen. Rust prevalence rose from 9% to 35% with increasing nitrogen, while smut prevalence fell from 32.5% to 10.8% (p<0.0001). In contrast, virus prevalence was low (4%), and did not vary with N (p=0.062). Co-infection occurred in 18% of infected tillers, but was unrelated to N. Co-infection with smut and SwMV occurred more frequently than would be predicted by chance (p=0.04), and was 5x more common than SwMV co-infection with rust.
The low prevalence of SwMV was surprising because the virus was 5—10x more prevalent in nearby plots at the field station (with no/limited N addition). Together with our tiller trait data, our results suggest that rust is directly/indirectly limited at low N. Smut and SwMV infected tillers were 5—25% shorter than rust-infected tillers, and increased competition at high N may be resulting in stand-thinning of disease. Co-infection dynamics between smut and SwMV may also be limiting the accumulation of SwMV in stand. Our study highlights the need for a holistic understanding of plant – microbe interactions, one that includes the ubiquitous yet often overlooked wild plant viruses.