Fig 1.
Atta colombica leaf-cutter ants remove waste material from their nest (A) and dispose of it in a single refuse pile on the rainforest floor (B). The most conspicuous insects in the leaf-cutter ant refuse pile were juvenile Panchlora cockroaches (C), which for this study were collected from the top 5 cm of the refuse pile (arrow in B). Adult Panchlora cockroaches (D) deposit young in the refuse pile and their young consume and thrive on the nutrient rich material, as evidenced by a refuse packed gut and enlarged fat body (E). Individually dissected guts were sectioned into the foregut, midgut, and hindgut (F). Photo credit: Robert Cullen (A), Erin A. Gontang (B/E), Justin C. Touchon (C/D).
Table 1.
Sequencing statistics of the juvenile Panchlora foregut, midgut, and hindgut community metagenomes.
Fig 2.
Phylum level classification of 16S rRNA gene amplicons (A) and metagenomic sequence data (B) for the juvenile Panchlora cockroach gut. Partial list of bacterial KEGG category (C) and CAZy (D) distribution for dominant bacterial community members (genera representing > 1% of the total metagenomic sequence); the heat maps illustrate the percent of the metagenomic sequence data annotated to the particular KEGG category (0–18%) or CAZy family grouping (0–0.014%), normalized by the total number of proteins binned to the genus. For the total bacterial KEGG category distribution for the foregut, midgut, and hindgut, and the total KEGG category and glycosyl hydrolase (GH) distribution of dominant bacterial community members, see S1 Fig.
Fig 3.
Phylum level classification of all assembled contigs (A) and the genus level classification for the Gammaproteobacteria (B), Firmicutes (C), and Bacteroidetes (D) of the Panchlora foregut, midgut, and hindgut; n is the total number of base pairs assembled and used to construct the graph. Genus level classification is shown for groups that make up more than 12% of each gut section. The metagenomic sequences provide evidence that microbial community composition changes and diversity increases along the alimentary tract.
Fig 4.
Class level classification of 16S rRNA gene amplicons for the Panchlora foregut, midgut, and hindgut, as compared to the Atta colombica garden top, garden bottom, and averaged refuse piles (A); n is the total number of 16S amplicons used to construct the graph. UniFrac unweighted principal coordinate analyses (PCoA) (B) and weighted PCoA (C) for the Panchlora foregut, midgut, and hindgut, and A. colombica garden top, garden bottom, and three different refuse piles. The 16S rRNA gene amplicons and weighted UniFrac analysis illustrate the shift in the Panchlora gut community composition along the alimentary tract and show that the cockroach gut communities are distinct from the A. colombica garden and refuse piles.
Fig 5.
Non-metric multidimensional scaling (nMDS) based on the cosine similarity index of the Clusters of Orthologous Groups (COG) of 21 microbiomes representing the foregut, midgut, and hindgut of the Panchlora cockroach, the fungus gardens and refuse piles of Attini ants, and the guts of a grasshopper, a moth, beetles, honeybees, and termites.
Metagenome information for each microbiome is available in the S5 Dataset.