Additive genetic effects in interacting species jointly determine the outcome of caterpillar herbivory

Abstract

Plant-insect interactions are common and important in basic and applied biology. Trait and genetic variation can affect the outcome and evolution of these interactions, but the relative contributions of plant and insect genetic variation and how these interact remain unclear and are rarely subject to assessment in the same experimental context. Here, we address this knowledge gap using a recent host-range expansion onto alfalfa by the Melissa blue butterfly. Common garden rearing experiments and genomic data show that caterpillar performance depends on plant and insect genetic variation, with insect genetics contributing to performance earlier in development and plant genetics later. Our models of performance based on caterpillar genetics retained predictive power when applied to a second common garden. Much of the plant genetic effect could be explained by heritable variation in plant phytochemicals, especially saponins, peptides, and phosphatidyl cholines, providing a possible mechanistic understanding of variation in the species interaction. We find evidence of polygenic, mostly additive effects within and between species, with consistent effects of plant genotype on growth and development across multiple butterfly species. Our results inform theories of plant-insect coevolution and the evolution of diet breadth in herbivorous insects and other host-specific parasites.

Keywords: coevolution; genomic prediction; phytochemicals; plant–insect interaction; polygenic.

Fig. 1.

Main hypotheses tested about the contribution of plant and insect genetics to caterpillar performance were as follows: (i) Caterpillar performance is primarily affected by insect (L. melissa) genetics; (ii) caterpillar performance is primarily affected by plant (M. sativa) genetics; (iii) the genetics of the interacting species have similar effects on caterpillar performance and combine additively; (iv) the genetics of the interacting species have similar effects on caterpillar performance and combine epistatically; and (v) the null hypothesis that neither insect nor plant genetic variation has an appreciable effect on caterpillar performance. The illustration (by R. Ribas) shows an L. melissa caterpillar feeding on alfalfa, while being tended by ants; additional biotic or abiotic factors, such as the presence of mutualistic ants, also affect caterpillar performance in the wild (25), but are not a component of this study.

Fig. 2.

(A) Map of plant (M. sativa) and insect (L. melissa) common garden source populations. Symbol shapes denote source type—Lm, L. melissa; Ms-abs, M. sativa site without L. melissa butterflies; and Ms-pres, M. sativa site with L. melissa butterflies—and are colored to indicate different populations within taxa. The A, Inset illustration shows an adult L. melissa perched on M. sativa (illustration by R. Ribas). (B) Ordination of genetic variation via PCA for the M. sativa common garden plants. (C) Ordination of genetic variation via PCA for the L. melissa caterpillars from the rearing experiment. Points in B and C denote individual plants or caterpillars and are colored to match the map (A).

Fig. 3.

(A) Plot shows survival and development of L. melissa over the course of the rearing experiment. Colored regions denote the number of individuals that were living caterpillars, pupae, adults, or dead at each day posthatching. (B) Plots show pairwise correlations between L. melissa performance traits. Scatterplots are shown in the lower-triangle panels—each point denotes one individual—and Pearson correlations are reported in the corresponding upper-triangle panels. Traits are given along the diagonal panels: 8-d weight, 14-d weight, pupal weight, and truncated survival time. Scatterplots and Pearson correlations are based on residuals after controlling for confounding environmental effects (see Materials and Methods for details).

Fig. 4.

Genetic mapping of caterpillar performance. (A) Dot chart shows Bayesian estimates of the proportion of trait variation explained by M. sativa genetics (Ms), L. melissa genetics (Lm), or both combined (Ms+Lm) for each caterpillar-performance trait: W8d, 8-d weight; W14d, 14-d weight; Wpup, pupal weight; S8d, 8-d survival; S14d, 14-d survival; SPup, survival to pupation; SAdu, survival to adult; Stot, total survival time; and Stime, (truncated) survival time. Points and horizontal lines denote point estimates (posterior medians) and 95% equal-tail probability intervals, respectively. (B) Heatmap shows genetic correlations between pairs of caterpillar-performance traits based on M. sativa genetics (lower triangle) or L. melissa genetics (upper triangle). Manhattan plots in C and D show posterior inclusion probabilities (PIPs) for genotype–performance associations based on M. sativa and L. melissa SNPs, respectively. Points denote SNPs with different colors and symbols for different performance traits. Only SNPs with PIPs $\ge$ 0.01 are depicted. Horizontal lines at PIPs of 0.1 and 0.5 are included for reference.

Fig. 5.

Genetic mapping of caterpillar performance with epistasis. The dot chart shows Bayesian estimates of the proportion of trait variation explained by M. sativa genetics (Ms), L. melissa genetics (Lm), or both combined (Ms+Lm) for 8-d weight (W8d), 14-d weight (W14d), and pupal weight (Wpup). Points and horizontal lines denote point estimates (posterior medians) and 95% equal-tail probability intervals, respectively, for the proportion of trait variation explained by additive effects and pairwise epistatic effects. Vertical black lines denote point estimates (posterior medians) for the proportion of variation explained by additive genetic effects alone (as presented in Fig. 4A).

Fig. 6.

Genomic prediction of caterpillar performance. (A) Dot chart shows Pearson correlations between cross-validation genomic predictions of phenotypes and the observed values based on M. sativa genetics (Ms), L. melissa genetics (Lm), or both combined (Ms+Lm) for each caterpillar-performance trait: W8d, 8-d weight; W14d, 14-d weight; Wpup, pupal weight; S8d, 8-d survival; S14d, 14-d survival; SPup, survival to pupation; SAdu, survival to adult; Stot, total survival time; and Stime, (truncated) survival time. Points and horizontal lines denote point estimates (posterior medians) and 95% equal-tail probability intervals, respectively. For example, a large value on the x axis indicates a high correlation between observed performance values and predictions from genotype based on cross-validation. (B) Scatterplot of PVE vs. the Pearson correlation of genomic predictions from A. Each point denotes a trait and is colored to indicate values from M. sativa or L. melissa genetics. Colored lines are best fits from ordinary linear regression, and a dashed line denotes the 0 value on the y axis. (C) Dot chart similar to A, but for genomic predictions of phenotypes in a second common garden (the Gene Miller Life Science Garden) based on the models fit from the main garden. (D) Scatterplot of correlations between observed caterpillar-performance-trait values and genomic predictions of these values using cross-validation within the main garden vs. prediction for samples in the Gene Miller Life Science Garden based on the models fit for the main garden.

Fig. 7.

Associations between plant-trait polygenic scores and caterpillar-performance polygenic scores. Scatterplots show genetic correlations between plant chemistry and other plant traits and 14-d caterpillar weight (A) and survival to pupation (B) inferred from plant genetics as a function of the proportion of plant-trait variation explained by genetics (PVE). A dashed horizontal line denotes a genetic correlation of zero. C shows the variance explained by LASSO regression models of caterpillar-performance polygenic scores estimated from plant genetics as a function of polygenic scores for 1,750 plant chemistry traits and 10 nonchemistry traits. Black dots denote inferred values of r², and gray dots show similar estimates using randomized plant-trait polygenic scores (10 random datasets each). D and E show standardized regression coefficients (coef.) from the LASSO models for 14-d weight (D) and survival to pupation (E).