The Koala (Phascolarctos cinereus) faecal microbiome differs with diet in a wild population

Abstract

Background: The diet of the koala (Phascolarctos cinereus) is comprised almost exclusively of foliage from the genus Eucalyptus (family Myrtaceae). Eucalyptus produces a wide variety of potentially toxic plant secondary metabolites which have evolved as chemical defences against herbivory. The koala is classified as an obligate dietary specialist, and although dietary specialisation is rare in mammalian herbivores, it has been found elsewhere to promote a highly-conserved but low-diversity gut microbiome. The gut microbes of dietary specialists have been found sometimes to enhance tolerance of dietary PSMs, facilitating competition-free access to food. Although the koala and its gut microbes have evolved together to utilise a low nutrient, potentially toxic diet, their gut microbiome has not previously been assessed in conjunction with diet quality. Thus, linking the two may provide new insights in to the ability of the koala to extract nutrients and detoxify their potentially toxic diet.

Method: The 16S rRNA gene was used to characterise the composition and diversity of faecal bacterial communities from a wild koala population (n = 32) comprising individuals that predominately eat either one of two different food species, one the strongly preferred and relatively nutritious species Eucalyptus viminalis, the other comprising the less preferred and less digestible species Eucalyptus obliqua.

Results: Alpha diversity indices indicated consistently and significantly lower diversity and richness in koalas eating E. viminalis. Assessment of beta diversity using both weighted and unweighted UniFrac matrices indicated that diet was a strong driver of both microbial community structure, and of microbial presence/absence across the combined koala population and when assessed independently. Further, principal coordinates analysis based on both the weighted and unweighted UniFrac matrices for the combined and separated populations, also revealed a separation linked to diet. During our analysis of the OTU tables we also detected a strong association between microbial community composition and host diet. We found that the phyla Bacteroidetes and Firmicutes were co-dominant in all faecal microbiomes, with Cyanobacteria also co-dominant in some individuals; however, the E. viminalis diet produced communities dominated by the genera Parabacteroides and/or Bacteroides, whereas the E. obliqua-associated diets were dominated by unidentified genera from the family Ruminococcaceae.

Discussion: We show that diet differences, even those caused by differential consumption of the foliage of two species from the same plant genus, can profoundly affect the gut microbiome of a specialist folivorous mammal, even amongst individuals in the same population. We identify key microbiota associated with each diet type and predict functions within the microbial community based on 80 previously identified Parabacteroides and Ruminococcaceae genomes.

Keywords: Dietary specialist; Eucalyptus; Folivore; Formylated phloroglucinol compounds (FPCs); Marsupial; Microbiome; Plant secondary metabolites (PSMs).

Figure 1. PCoA of β diversity of koala faecal bacterial communities from Cape Otway.

Scatterplots from (A) weighted and (B) unweighted UniFrac matrices for the combined Cape Otway koala population with diets comprising E. viminalis and E. obliqua from 2013 and 2015. PERMANOVA analysis indicated no influence of collection year on the (A) weighted UniFrac data (pseudo F₁ = 0.44, PERMANOVA P = 0.81). A significant influence was detected when we analysed the (B) unweighted UniFrac data (pseudo F₁ = 1.76, PERMANOVA P = 0.01). When diet × collection years was analysed an influence was detected on community structure ((A) weighted UniFrac, pseudo F₂ = 10.07, PERMANOVA P = 0.0001) and presence/absence of microbial taxa ((B) unweighted UniFrac, pseudo F₂ = 2.98, PERMANOVA P = 0.0001).

Figure 2. PCoA of β diversity based on UniFrac matrices from the Cape Otway 2013 koala collection.

Scatterplots from (A) weighted and unweighted (B) UniFrac matrices for koalas with diets of E. viminalis and E. obliqua from the 2013 collection year. PERMANOVA assessment of the weighted and unweighted UniFrac matrices from the 2013 collection year, indicated that diet was a strong driver of both microbial community structure (relative abundance), and of microbial presence/absence during 2013 (A) weighted UniFrac, pseudo F₁ = 5.88, PERMANOVA P = 0.0001; (B) unweighted UniFrac, pseudo F₁ = 1.89, PERMANOVA P = 0.0001.

Figure 3. PCoA of β diversity based on UniFrac matrices from the Cape Otway 2015 koala collection.

Scatterplots from (A) weighted and (B) unweighted UniFrac matrices for koalas with diets comprising E. viminalis and E. obliqua from the 2015 collection year. The influence of diet was significant in 2015 (A) weighted UniFrac, pseudo F₁ = 8.89, PERMANOVA P = 0.0001; (B) unweighted UniFrac, pseudo F₁ = 2.76, PERMANOVA P = 0.0001.

Figure 4. Taxonomic bar charts of the most abundant genera in faecal microbiomes of koalas feeding from E. viminalis and E. obliqua.

Taxonomic bar charts show the dominance of Parabacteroides and Ruminococcaceae in faecal microbiomes of koalas eating E. viminalis and E. obliqua, respectively, in (A) 2013, and the dominance of Bacteroides, Parabacteroides (E. viminalis) and Ruminococcaceae (E. obliqua) in (B) 2015.

Figure 5. Estimated diet composition of koalas eating E. viminalis and E. obliqua.

Bar chart showing koala identity and percentage of diet species eaten for (A) 2013 (n = 14 koalas) and (B) 2015 (n = 15 koalas) koala faecal samples.