Epistasis and an oligogenic architecture underlie the strength of a plant-herbivore interaction

Abstract

1. Interaction strength between consumers and resources is a fundamental driver of eco-evolutionary dynamics, and its genetic architecture is key to predicting the pace of such dynamics. Our understanding of this architecture is limited for three reasons: (i) defense phenotypes do not necessarily translate to interaction strength; (ii) interaction strength is difficult to measure in ways tied to consumer-resource theory; and (iii) association studies typically prioritize detecting large-effect loci rather than characterizing whether interactions are polygenic (many small-effect genes) or oligogenic (few large-effect genes). 2. We address these gaps by examining how genomic variation in the plant Arabidopsis thaliana influences interactions with the aphid herbivore Lipaphis pseudobrassicae in a greenhouse experiment. We grew 390 accessions spanning the species' natural range and measured aphid per-capita population growth as our interaction-strength metric. To test whether a previously identified plant defense gene (AOP2) acts additively or epistatically with other defense loci, we integrated our estimates of aphid population growth with public data on defense chemistry and gene expression. To determine whether the genetic architecture of interaction strength was polygenic or oligogenic, we used a Bayesian sparse linear mixed model that partitions these two components. 3. AOP2 showed epistasis with two other defense genes (MAM1 and GS-OH), indirectly modifying aphid population growth via the production of a specific defense compound (glucosinolate 2-hydroxy-3-butenyl). More broadly, interaction strength showed an oligogenic architecture: top single nucleotide polymorphisms (SNPs) occurred at low minor allele frequencies (< 3%) and grouped into six linkage disequilibrium blocks, including candidate genes outside canonical plant defense pathways. These six blocks explained 47% of the genetic variance in interaction strength, with only 3% additional variance attributable to polygenic effects. 4. Taken together, these results show that plant genomic variation modifies herbivore interaction strength through epistasis among defense genes and rare alleles at few loci. This oligogenic architecture imposes multiple constraints on the potential for rapid evolution, highlighting the need to integrate genomic effects into eco-evolutionary models.