Phylogenetic and functional diversity among Drosophila-associated metagenome-assembled genomes

Abstract

Host-associated microbial communities can mediate interactions between their hosts and biotic and abiotic environments. While much work has been done to document how microbiomes vary across species and environments, much less is known about the functional consequences of this variation. Here, we test for functional variation among drosophilid-associated bacteria by conducting Oxford Nanopore long-read sequencing and generating metagenome-assembled genomes (MAGs) from communities associated with six species of drosophilid flies collected from "anthropogenic" environments in North America, Europe, and Africa. Using phylogenetic analyses, we find that drosophilid flies harbor a diverse microbiome that includes core members closely related to the genera Gilliamella, Orbus, Entomomonas, Dysgonomonas, and others. Comparisons with publicly available bacterial genomes show that many of these genera are associated with phylogenetically diverse insect gut microbiomes. Using functional annotations and predicted secondary metabolite biosynthetic gene clusters, we show that MAGs belonging to different bacterial orders and genera vary in gene content and predicted functions, including metabolic capacity and how they respond to environmental stressors. Our results provide evidence that wild drosophilid flies harbor phylogenetically and functionally diverse microbial communities. These findings highlight a need to quantify the abundance and function of insect-associated bacteria from the genera Gilliamella, Orbus, Entomomonas, and others on the performance of their insect hosts across diverse environments.IMPORTANCEWhile much attention has been given to catalogue the taxonomic diversity intrinsic to host-associated microbiomes, much less is known about the functional consequences of this variation, especially in wild, non-model host species. In this study, we use long-read sequencing to generate and analyze 103 high-quality metagenome-assembled genomes from host-associated bacterial communities from six species of wild fruit fly (Drosophila). We find that the genomes of drosophilid-associated bacteria possess diverse metabolic pathways and biosynthetic gene clusters that are predicted to generate metabolites involved in nutrition and disease resistance, among other functions. Using functional gene predictions, we show that different bacterial lineages that comprise the insect microbiome differ in predicted functional capacities. Our findings highlight the functional variation intrinsic to microbial communities of wild insects and provide a step towards disentangling the ecological and evolutionary processes driving host-microbe symbioses.

Keywords: Drosophila; MAGs; Orbaceae; insects; metagenomics; microbiome; symbiosis.

Fig 1

Bacterial sequence reads derived from whole-organism extraction and Nanopore MinION sequencing represent diverse taxa belonging to 23 classes (A). Differences in the relative abundances of microbial taxa among samples were affected by the number of sequences classified as bacterial (B; dim 1) and by species-level differences in the microbial community (B; dim 2). In panel A, individual names include details of the species (dhyd = D. hydei; drep = D. repleta; dimm = D. immigrans, zind = Z. indianus, ztaro = Z. taronus, and ztsac = Z. tsacasii), location (USA, UK, or STP [São Tomé and Principe]), and sex (fem = female, mal = male). Two individuals in the dataset were F1 offspring from wild-caught females and are indicated with an asterisk (*) in panel A.

Fig 2

Phylogenetic classifications of drosophilid-derived MAGs. (A) Phylogenetic relationships among Drosophila MAGs from maximum-likelihood placement by pplacer as implemented in GTDB-TK’s ‘classify_wf’ pipeline against the GTDB-Tk reference tree. Phylogenetic relationships among Drosophila MAGs from the three most abundant orders—Enterobacterales (B), Pseudomonadales (C), and Bacteriodales (D)—are generated using GTDB-Tk’s ‘de_novo_wf’ pipeline, highlighting genera represented by multiple MAGs within each order. In panels B to D, details on the host species, individual sample identifier, and location are reported in tip names (details as in Fig. 1).

Fig 3

Core bacterial genera are shared among Drosophila species (A) and across geographic regions we sampled (B). Bacterial genera (number inside circles) and the number of MAGs assembled for each genus (beside generic names) are given for each scenario of overlap. Bacterial genera highlighted in orange (Entomomonas, Gilliamella, and Pseudomonas_E) were only assembled from D. immigrans hosts. Note that numbers of MAGs belonging to Dysgonomonas and Orbus include MAGs assembled from Z. indianus (see Fig. 2), and D. repleta were grouped with D. hydei for this analysis.

Fig 4

Phylogenetic relationships among focal MAGs assembled from drosophilid flies (bold type) and publicly available genomes included in GTDB-Tk’s ’de_novo_wf’ pipeline. (A) Enterobacterale MAGs from drosophilid hosts are closely related to genomes from the genera Gilliamella, Orbus, and Frischella that have been assembled from diverse honey bee (Apis) and bumble bee (Bombus) hosts. (B) Drosophilid MAGs from the genus Dysgonomonas are closely related to diverse Dysgonomonas genomes assembled from insect, mammal, and environmental sources (sources not shown in panel B). (C) Drosophilid MAGs from the genus Entomomonas are closely related to genomes from Entomomonas sequenced from the eastern honey bee (Apis cerana) and the house cricket (Acheta domesticus).

Fig 5

MAGs belonging to different bacterial genera differ in their functional gene content. (A) COG categories that show variation in enrichment among MAGs. (B) KEGG pathways that are enriched in greater than 80% of the MAGs assembled from Gilliamella or Orbus (order: Enterobacterales) but were enriched in fewer than 50% of other MAGs in our dataset. (C) As in panel B, but with overrepresented enrichment within genera Acinetobacter, Entomomonas, or Pseudomonas (order: Pseudomonadales). (D) As in panel B, but with overrepresented enrichment within the genus Dysgonomonas (order: Bacteriodales). In panel A, the letters on the x-axis are the one-letter identifiers for COG categories. Categories F, G, H, I, and P are involved in “metabolism”; K and L are involved in “information storage and processing”; and N, O, T, U, and W are involved in “cellular processing and signaling.”

Fig 6

Drosophila MAGs differ in the secondary metabolite biosynthetic gene clusters (BGCs) they possess. (A) Summary of biosynthetic products for MAGs from the Bacteriodales, Pseudomonadales, and Enterobacterales, including details of the host species and location the flies were collected from. Across MAGs, we identified 177 biosynthetic gene products belonging to 18 types of BGC (B). The relative abundance of BGC types was significantly different among the three focal orders of bacteria included in this analysis (C).

Fig 7

Abundances of biosynthetic gene cluster (BGC) classes (i.e., “Type of BGC”) that were the most common within the focal bacterial genera Orbus, Gilliamella, and Entomomonas.