We Are What We Eat: A Stoichiometric and Ecometabolomic Study of Caterpillars Feeding on Two Pine Subspecies of Pinus sylvestris

Abstract

Many studies have addressed several plant-insect interaction topics at nutritional, molecular, physiological, and evolutionary levels. However, it is still unknown how flexible the metabolism and the nutritional content of specialist insect herbivores feeding on different closely related plants can be. We performed elemental, stoichiometric, and metabolomics analyses on leaves of two coexisting Pinus sylvestris subspecies and on their main insect herbivore; the caterpillar of the processionary moth (Thaumetopoea pityocampa). Caterpillars feeding on different pine subspecies had distinct overall metabolome structure, accounting for over 10% of the total variability. Although plants and insects have very divergent metabolomes, caterpillars showed certain resemblance to their plant-host metabolome. In addition, few plant-related secondary metabolites were found accumulated in caterpillar tissues which could potentially be used for self-defense. Caterpillars feeding on N and P richer needles had lower N and P tissue concentration and higher C:N and C:P ratios, suggesting that nutrient transfer is not necessarily linear through trophic levels and other plant-metabolic factors could be interfering. This exploratory study showed that little chemical differences between plant food sources can impact the overall metabolome of specialist insect herbivores. Significant nutritional shifts in herbivore tissues could lead to larger changes of the trophic web structure.

Keywords: herbivory; metabolomics; plant-insect; processionary moth; scots pine; secondary metabolites; stoichiometry.

Figure 1

Variable and case plot of the principal component (PC) 1 of the principal component analysis (PCA) including all elemental, stoichiometric, and metabolomic data of the dataset of caterpillars (Caterpillar-Dataset). Only elemental, stoichiometric, and identified metabolomic variables are shown in the variable plot. Each bar represents the loading value for each variable on PC1. Elemental and stoichiometric variables are shown in red. Metabolite groups are shown in different colors: dark blue, saccharides (carbohydrates); green, amino acids; light blue, nucleobases/nucleosides; orange, organic acids related to tricarboxylic acid cycle; violet, phenolic compounds; and dark red, other. Caterpillars feeding on Pinus sylvestris ssp. iberica (C-iberica) are shown in blue triangles and caterpillars feeding on Pinus sylvestris ssp. nevadensis (C-nevadensis) are shown in red circles. Asterisks and crosses next to elemental, stoichiometric and metabolomic variables denote statistical significance (* p < 0.05) and marginal significance (^† p < 0.1), respectively.

Figure 2

Relative difference (%) of elemental and stoichiometric variables between caterpillar groups (C-iberica relative to C-nevadensis) and between pine subspecies for each folivory level (iberica NATs relative to nevadensis NATs; iberica AT.ABs relative to nevadensis AT.ABs; iberica AT.NABs relative to nevadensis AT.NABs). Positive values indicate higher elemental concentration or stoichiometric ratio value in C-iberica or iberica trees while negative values indicate higher elemental concentration or stoichiometric ratio value in C-nevadensis or nevadensis trees. Asterisks and crosses denote statistical significance (* p < 0.05) and marginal significance (^† p < 0.1), respectively.

Figure 3

Pearson lineal correlations between total carbohydrate abundance versus C, C:N, C:P, and N:P ratios, and total amino acid abundance versus N, C:N, C:P, and N:P for both caterpillar groups (C-nevadensis and C-iberica) separately (n = 12) and together (n = 24). Coefficient of determination (R²), statistic Fisher value (F), and p value (p) are shown. Data corresponding to caterpillars feeding on Pinus sylvestris ssp. iberica (C-iberica) are shown in blue triangles and caterpillars feeding on Pinus sylvestris ssp. nevadensis (C-nevadensis) are shown in red circles. Lineal correlations for C-iberica and C-nevadensis are shown in blue and red, respectively. Lineal correlations for both C-iberica and C-nevadensis are shown in black. Correlations with P > 0.1 are represented by dashed lines.

Figure 4

Variable (a) and case plot (b) of the principal component (PC) 1 vs. the PC2 of the principal component analysis (PCA) including all elemental, stoichiometric and metabolomic data of the dataset of caterpillars and pines (Pine-Caterpillar-Dataset). Only elemental, stoichiometric and identified metabolomic variables are shown in the variable plot (a). The location of a variable (element, stoichiometric ratio or metabolite) on the variable plot plane represents the loading value on PC1 vs. PC2 for such specific variable. Elemental and stoichiometric variables are shown in red. Metabolite groups are shown in different colors: dark blue, sugars (carbohydrates); green, amino acids; light blue, nucleobases/nucleosides; orange, organic acids related to tricarboxylic acid cycle; violet, phenolic compounds; dark red, other. Case plot (b) shows caterpillars feeding on Pinus sylvestris ssp. iberica (C-iberica) in blue circles and caterpillars feeding on Pinus sylvestris ssp. nevadensis (C-nevadensis) in red circles. Needle samples of attacked branches (AT.ABs) and non-attacked branches from attacked trees (AT.NABs) are shown in dark and light blue for Pinus sylvestris ssp. Iberica, and dark and light red for Pinus sylvestris ssp. nevadensis. Different letters next to each colored arrow denote significant difference between groups of samples after Tukey’s HSD post-hoc test (p < 0.05). One-way ANOVA plot (c) contrasting the metabolomic distances along the PC2 of the PCA between caterpillars and pines. Different letters indicate significant difference after Tukey’s HSD post-hoc test (p < 0.05). Not attacked trees are not represented in the PCA. Detailed discussion of the effects of the caterpillar on pines metabolism has been already published elsewhere [23,36].