The phyllosphere microbiome of host trees contributes more than leaf phytochemicals to variation in the Agrilus planipennis Fairmaire gut microbiome structure

Abstract

The microbiome composition of living organisms is closely linked to essential functions determining the fitness of the host for thriving and adapting to a particular ecosystem. Although multiple factors, including the developmental stage, the diet, and host-microbe coevolution have been reported to drive compositional changes in the microbiome structures, very few attempts have been made to disentangle their various contributions in a global approach. Here, we focus on the emerald ash borer (EAB), an herbivorous pest and a real threat to North American ash tree species, to explore the responses of the adult EAB gut microbiome to ash leaf properties, and to identify potential predictors of EAB microbial variations. The relative contributions of specific host plant properties, namely bacterial and fungal communities on leaves, phytochemical composition, and the geographical coordinates of the sampling sites, to the EAB gut microbial community was examined by canonical analyses. The composition of the phyllosphere microbiome appeared to be a strong predictor of the microbial community structure in EAB guts, explaining 53 and 48% of the variation in fungi and bacteria, respectively. This study suggests a potential covariation of the microorganisms associated with food sources and the insect gut microbiome.

Figure 1

Map representing sampling sites including the LCBD value per site. For each site LCBD values of bacteria (purple) and fungi (orange) are represented in percentages illustrated in a pie chart. The white star on the colour indicates the significant contribution of the corresponding community to the local β-diversity. Thus, the bacterial community was found significantly contributing to the local β-diversity in six sites (A06, A13, A14, A22, A28, and A32), Holm-corrected LCBD p values, whereas the fungal community was found significantly contributing to the local β-diversity in three sites (A18, A29, and A35), Holm-corrected LCBD p values.

Figure 2

Taxonomic profiles of bacterial (A) and fungal (C) communities associated with the adult EAB gut and those associated with the leaves of the host trees. The taxonomic profile is based on the presence absence of ASVs in each biotope. The right part of the figure shows the number of ASVs shared and unshared between the two habitats for bacteria (B) and fungi (D).

Figure 3

Species richness and diversity indices in the insect gut and on the leaves computed from raw data for communities of bacteria (A) and fungi (B). Values at the top of the panels: Wilcoxon signed-rank test statistics and significance: Significance codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘NS’ > 0.05.

Figure 4

Variation partitioning of the gut bacterial community among three predictor matrices: bacteria associated with leaves (Hellinger-transformed), fungi associated with leaves (Hellinger-transformed), and dbMEM spatial eigenfunctions generated from the geographic coordinates of the sampling sites (Fraxinus trees). The selected explanatory variables (leaves) are represented in blue, and the response variables (insects’ gut) in red. Thus, the three explanatory matrices (bacteria, fungi, and geographic coordinates) are represented in individual RDA analyses. The adjusted R-square (adj.R²) corresponds to the R² adjusted to the model containing all variables. The figure below the RDAs represents the partitioning variation analyses of the bacterial community associated with adult EAB gut.

Figure 5

Variation partitioning of the gut fungal community among three predictor matrices: bacteria associated with leaves (Hellinger-transformed), fungi associated with leaves (Hellinger-transformed), and dbMEM spatial eigenfunctions generated from the geographic coordinates of the sampling sites (Fraxinus trees). The selected explanatory variables (leaves) are represented in blue, and the response variables (insects’ gut) in red. Thus, the three explanatory matrices (bacteria, fungi, and geographic coordinates) are represented in individual RDA analyses. The adjusted R-square (adj.R²) corresponds to the R² adjusted to the model containing all variables. The figure below the RDAs represents the partitioning variation analyses of the fungal community associated with adult EAB gut.

Figure 6

Global representation showing variation partitioning analysis of microbial communities associated with leaves on the microbial communities associated with adult EAB gut. The host tree descriptors that have been found as explanatory matrices (bacteria, fungi, and geographic coordinates) are represented on the left portion (for the variation of the EAB gut fungal community), and on the right portion (for the variation of the EAB gut bacterial community). The dash box indicates the explanatory variable (properties of the host tree), and the values indicated in the circle correspond to the percentage of explanation. *After computing a forward selection (alpha < 0.05) prior to multiple regressions. The cellulose content significantly explained the species richness observed in the fungal community associated with the adult EAB gut.

Figure 7

Step-by-step representation of the two ‘variance partitioning’ analyses, including the variables datasets concerned. The circles show the response dataset, and each rectangle corresponds to a dataset of predictor variables. The response variables were transformed before the variance partitioning analysis.