The primate gut mycobiome-bacteriome interface is impacted by environmental and subsistence factors

Abstract

The gut microbiome of primates is known to be influenced by both host genetic background and subsistence strategy. However, these inferences have been made mainly based on adaptations in bacterial composition - the bacteriome and have commonly overlooked the fungal fraction - the mycobiome. To further understand the factors that shape the gut mycobiome of primates and mycobiome-bacteriome interactions, we sequenced 16 S rRNA and ITS2 markers in fecal samples of four different nonhuman primate species and three human groups under different subsistence patterns (n = 149). The results show that gut mycobiome composition in primates is still largely unknown but highly plastic and weakly structured by primate phylogeny, compared with the bacteriome. We find significant gut mycobiome overlap between captive apes and human populations living under industrialized subsistence contexts; this is in contrast with contemporary hunter-gatherers and agriculturalists, who share more mycobiome traits with diverse wild-ranging nonhuman primates. In addition, mycobiome-bacteriome interactions were specific to each population, revealing that individual, lifestyle and intrinsic ecological factors affect structural correspondence, number, and kind of interactions between gut bacteria and fungi in primates. Our findings indicate a dominant effect of ecological niche, environmental factors, and diet over the phylogenetic background of the host, in shaping gut mycobiome composition and mycobiome-bacteriome interactions in primates.

Fig. 1. Gut mycobiome composition differs based on subsistence strategy across different primates.

a Alpha diversity analysis showing lower fungal richness in US humans, compared with other primate populations. b Principal coordinate analysis based on Bray-Curtis distances showing different fungal community composition across different primate groups. Each symbol represents mycobiome composition, at the ASV level, in the fecal samples of an individual primate. A Bray-Curtis distance dendrogram (hierarchical clustering, average distances) based on average fungal ASV abundances showed similarities between phylogenetically distinct primate groups. Each group with similar mycobiome composition is shown to reflect PAM clustering (Supplementary Fig. 2). c Relative abundance of top 20 fungal families and their mean distribution among different primate groups are shown in a Bubble plot. Color code represents membership of each family to a specific Fungal Phylum. Differentially abundant taxa (Supplementary Table 1) were identified using Indicator Species Analysis and the families marked with stars are representative of a given primate group. The identifiers used for primate groups are Western Lowland Gorilla: WLG, Mangabey: Agile Mangabeys and Chimps: Chimpanzee. In boxplots, center values indicate the median; bounds of box represents lower/upper quartiles; whiskers show inner fences.

Fig. 2. Heatmap distribution of 30 most abundant fungal genera.

Heatmap showing the relative distribution of the 30 most abundant fungal genera based on their normalized relative abundances in each group. Color distribution of individual fungal genera is reported in color key based on normalized Z scores. The primate groups were arranged as per the hierarchical clustering shown in Fig. 1b. Cumulative relative abundance of individual genera across groups is shown in the separate bar plot. Genera highlighted in green and deepskyblue belong to the Ascomycota and Basidiomycota phyla, respectively. Tropic mode and guild of these most abundant genera were selected from the total assigned trophic mode and guild for each ASV. To make FUNGuild assignments more accurate, only ASVs ranked as “Probable” and “Highly Probable” hits were considered for final classification. ASVs ranked as “Possible” were not considered. These categories are shown at the bottom of the heatmap with different colors and numbers. The dotted boxes along with the asterisk are drawn to show statistical significance (based on species indicator analysis, indval > 0.3, and p < 0.05) of each fungal genera in the respective primate group. The identifiers used for primate groups are US-H: US-Human, CCO: Captive Chimpanzees Ostrava, CWLG: Captive Western Lowland Gorilla, MG: Mountain Gorilla, Ban-H: Bantu-Human, BaA-H: BaAka-Human, C: Chimpanzee, M: Agile Mangabey, and WLG: Western Lowland Gorilla.

Fig. 3. Combined gut bacteriome and mycobiome community composition differ across different primate species.

a Principal coordinate analysis based on Bray-Curtis distances showing significant differences in combined bacteriome and mycobiome community composition across different primate groups. Each symbol represents the combined bacteriome and mycobiome composition at the ASV level in fecal samples of an individual primate species. A Bray-Curtis distance dendrogram (hierarchical clustering, method = average) based on mean abundances. b Distances to centroid show higher inter-individual variations in mycobiome community composition as compared to bacteriome across different primate groups. Fold changes along with the statistical significance based on wilcoxon rank-sum tests (p < 0.05, highlighted by *) are provided at the top of each boxplot (c) Correspondence between bacteriome and mycobiome composition as shown by Procrustes analysis on Bray-Curtis distances (Protest corr = 0.78, p = 0.001 and mantel test, r = 0.63, p = 0.001). D Procrustes distances between bacteriome and mycobiome composition in each primate group show greater correspondence in traditional human populations. The identifiers used for primate groups are US-H: US-Human, CCO: Captive Chimpanzee-Ostrava, CWLG: Captive Western Lowland Gorilla, MG: Mountain Gorilla, Ban-H: Bantu-Human, BaA-H: BaAka-Human, and WLG: Western Lowland Gorilla. In boxplots, center values indicate the median; bounds of box represents lower/upper quartiles; whiskers show inner fences.

Fig. 4. Complexity of bacteriome-mycobiome co-occurrence networks varies among different primate groups.

a Co-occurrence networks constructed using all significant correlations (compositionally corrected corrs r > +/−0.6 and p < 0.01) showed clear distinctions in network complexity between human and nonhuman primates. Size of the node shows the number of connections of each fungal ASV. Color of nodes represents bacterial and fungal ASVs shown in Fig. 4a. Each symbol represents an individual primate species. Edge color represents negative (red lines) and positive (green lines) correlations respectively. Bacterial fungal co-occurrence network attributes (b) Neighborhood connectivity and (c) Degree shows a smaller number of bacterial fungal connections in all humans and in western lowland gorillas. Different letters denote significant differences according to Kruskal–Wallis tests. In boxplots, center values indicate the median; bounds of box represents lower/upper quartiles; whiskers show inner fences. The identifier used for primate groups are Mangabey: Agile Mangabeys and Chimps: Chimpanzee.

Fig. 5. Associations between bacterial taxa and representative fungal genus of individual primate groups.

Compositional based correlations between all ASVs of significantly discriminating (species indicator analysis, p < 0.01) fungal genus and bacterial ASVs to show co-occurring (pairs connected using citrus color) and co-exclusive (pairs connected using light salmon pink) fungal bacterial pairs. Color code in the bottom of the plot shows the enrichment of fungal genus in respective primate groups. Shape of the node represents bacterial and fungal ASVs. Edge thickness represents correlation strength.