Broad patterns in host physiological traits predict key epidemiological traits

Background/Question/Methods Predicting the influence of host communities on vectored, generalist disease dynamics has become a critical ecological goal. Two key epidemiological traits used to predict disease dynamics are pathogen transmission and host tolerance to infection. We hypothesize that these epidemiological traits are causally determined by a few, general host physiological traits. Plant phenotypes can be mapped onto a continuum that represents a rapid resource acquisition - resource conservation trade-off. At one end are quick return (QR) phenotypes, which have short-lived leaves with high photosynthetic capacity (Amax), tissue nitrogen concentration (TN), and specific leaf area (SLA). At the other end, are slow return (SR) phenotypes, which have long-lived leaves of low Amax, TN, and SLA. Relationships between host phenotypes and insect herbivores, which comprise 80-90% of plant virus vectors, are well known: QRs are generally susceptible to herbivory (i.e., herbivores prefer nitrogen rich, soft tissue) but tolerate damage (i.e., damage to cheap, high resource return tissue is not extremely costly). In contrast, SRs are generally less susceptible to herbivory (i.e., nitrogen poor, tough leaves are resistant to herbivory) but are intolerant of herbivory (i.e., damage to expensive, low resource return tissue is costly). Direct tests linking QR-SR phenotypic gradients to epidemiological traits, however, are lacking. Pathogen transmission is positively correlated with vector preference and increased tolerance can increase vector populations, further increasing transmission. We therefore predicted that our host species phenotypes would vary along the SR-QR continuum and that pathogen transmission and disease tolerance would be greater in QR phenotypes. We conducted controlled greenhouse experiments using six California grass species and an aphid vectored generalist plant pathogen (barley yellow dwarf virus PAV; BYDV). We quantified host phenotypes and conducted experimental BYDV inoculations using aphid vectors. Results/Conclusions Our species fell along a QR-SR continuum. QR phenotypes supported greater vector populations. QR phenotypes were also more likely to become infected when inoculated with the virus (i.e., they were more susceptible). Further, QR phenotypes produced more infected vectors (i.e., they were more competent). Finally, infection caused stronger reductions of total mass in SR phenotypes (i.e., QR phenotypes were more tolerant to infection). In total, our results suggest that recently discovered broad patterns in host phenotype may predict key epidemiological traits and that pathogen transmission may be greatest in QR dominated communities.