Functional variation in the gut microbiome of wild Drosophila populations

Abstract

Most of the evidence that the gut microbiome of animals is functionally variable, with consequences for the health and fitness of the animal host, is based on laboratory studies, often using inbred animals under tightly controlled conditions. It is largely unknown whether these microbiome effects would be evident in outbred animal populations under natural conditions. In this study, we quantified the functional traits of the gut microbiota (metagenome) and host (gut transcriptome) and the taxonomic composition of the gut microorganisms (16S rRNA gene sequence) in natural populations of three mycophagous Drosophila species. Variation in microbiome function and composition was driven principally by the period of sample collection, while host function varied mostly with Drosophila species, indicating that variation in microbiome traits is determined largely by environmental factors, and not host taxonomy. Despite this, significant correlations between microbiome and host functional traits were obtained. In particular, microbiome functions dominated by metabolism were positively associated with host functions relating to gut epithelial turnover. Much of the functional variation in the microbiome could be attributed to variation in abundance of Bacteroidetes, rather than the two other abundant groups, the γ-Proteobacteria or Lactobacillales. We conclude that functional variation in the interactions between animals and their gut microbiome can be detectable in natural populations, and, in mycophagous Drosophila, this variation relates primarily to metabolism and homeostasis of the gut epithelium.

Correlation and causation between the microbiome, Wolbachia and host functional traits in natural populations of drosophilid flies

“…Quality control was performed before and after trimming using fastqc . rsem (Li & Dewey, ) and bowtie 2 (Langmead & Salzberg, ) were applied, first, to generate a transcriptome reference (contigs >1,000 bp) for each Drosophila species using the three de novo transcriptomes from Bost, Martinson et al () and, then, to calculate the expected gene count (Supporting Information Dataset a). Expected gene counts were summed for each sample according to gene identity, KO, KEGG pathways and categories (correspondence between terms provided in Supporting Information Dataset b).…”

“…First, Procrustes analysis was performed on the ordination results of the microbiota (PCoA of the Bray–Curtis dissimilarity matrix of the ASV relative abundances) and transcriptome (PCA of the genes) using the functions procrustes and protest in the r package vegan . Second, a modified Kendall's τ b test ( mazeinda package) (Bost, Martinson et al, ) was applied for pairwise correlations between rsem for each host gene transcript and number of reads in bacterial modules and tested for significance with p ‐values adjusted for multiple comparisons with the Benjamini–Hochberg procedure, using the p.adjust function in r . The resultant q ‐value for each comparison is the minimum FDR obtained for all comparisons ordered by ascending p‐ values up to and including the focal comparison.…”

“…First, the taxonomic composition of their gut microbiota varies widely among individual flies (Martinson et al, ). Second, the gene expression profile of the flies has already been demonstrated as an informative index of host functional traits in this system, with evidence that transcript abundance of host genes associated with gut epithelium turnover is associated with metabolic functions of the gut microoorganisms, especially bacteria of the phylum Bacteroidetes (Bost, Martinson et al, ). Third, Wolbachia and Spiroplasma endosymbionts are prevalent in D. neotestacea (Jaenike, Stahlhut, Boelio, & Unckless, ), enabling study of interactions between these endosymbionts and the gut microbiota in natural populations.…”

“…This study was conducted on large numbers of individual flies of all three species from a single collection, with analysis of the microbiota and transcriptome of the whole body. This design enhanced the statistical power of the analysis, relative to previous published work on the microbiota of guts dissected from mycophagous drosophilids (Bost, Martinson et al, ) which used relatively small number of samples of pooled flies collected on multiple different occasions. Our use of whole flies additionally enabled us to include endosymbionts localized to organs other than the gut in the analyses.…”

“…In particular, we investigated the relationships between the endosymbionts Wolbachia and Spiroplasma in D. neotestacea with both other bacteria in these flies and the host gene expression patterns. Eukaryotic microorganisms were not quantified, following evidence that they make a quantitatively small contribution to the microbiome of these species (Bost, Martinson et al, ).…”