Structure of Chimpanzee Gut Microbiomes across Tropical Africa

Abstract

Understanding variation in host-associated microbial communities is important given the relevance of microbiomes to host physiology and health. Using 560 fecal samples collected from wild chimpanzees (Pan troglodytes) across their range, we assessed how geography, genetics, climate, vegetation, and diet relate to gut microbial community structure (prokaryotes, eukaryotic parasites) at multiple spatial scales. We observed a high degree of regional specificity in the microbiome composition, which was associated with host genetics, available plant foods, and potentially with cultural differences in tool use, which affect diet. Genetic differences drove community composition at large scales, while vegetation and potentially tool use drove within-region differences, likely due to their influence on diet. Unlike industrialized human populations in the United States, where regional differences in the gut microbiome are undetectable, chimpanzee gut microbiomes are far more variable across space, suggesting that technological developments have decoupled humans from their local environments, obscuring regional differences that could have been important during human evolution. IMPORTANCE Gut microbial communities are drivers of primate physiology and health, but the factors that influence the gut microbiome in wild primate populations remain largely undetermined. We report data from a continent-wide survey of wild chimpanzee gut microbiota and highlight the effects of genetics, vegetation, and potentially even tool use at different spatial scales on the chimpanzee gut microbiome, including bacteria, archaea, and eukaryotic parasites. Microbial community dissimilarity was strongly correlated with chimpanzee population genetic dissimilarity, and vegetation composition and consumption of algae, honey, nuts, and termites were potentially associated with additional divergence in microbial communities between sampling sites. Our results suggest that host genetics, geography, and climate play a far stronger role in structuring the gut microbiome in chimpanzees than in humans.

Keywords: climate; diet; host genetics; parasites; prokaryotes; tools.

FIG 1

Map (from https://open.africa/) of the 29 sites included in this study, showing either forest, forest mosaic, savanna mosaic, or savanna habitat types and the ranges of the four main geographic regions of chimpanzees (from reference 133). Insets show the variation among sites (points) in consumption of algae, honey, nuts (hard-shelled drupes), and termites (orange = consumed, gray = not consumed). In many cases, these items are accessed using tools (algae [6 of 7 sites], honey [10 of 13 sites], nuts [5 of 5 sites], termites [9 of 12 sites]).

FIG 2

Heatmaps showing site mean percent relative abundances for 14 bacterial families with >1% mean relative abundance in at least one site (a) and parasite percent prevalence at each site (number of samples with parasite/total samples at site × 100) (b). Each panel is sorted from top to bottom in order of abundance (a) or prevalence (b). Note that there were 225 prokaryote families in the data set and if they were all included, columns would sum to 100. Different numbers after genera or families denote different ASVs; Trichomonadidae_12_51 represents two highly correlated ASVs that were combined. For more information about parasite taxonomy, see Table S3 at https://doi.org/10.6084/m9.figshare.14390426. Asterisks in panel a denote taxa with significantly different mean relative abundances among regions (Kruskal-Wallis, Bonferroni, *, P < 0.05; **, P < 0.01; ***, P < 0.001), and asterisks in panel b indicate taxa with significantly different probabilities of occurrence among regions (logistic regression, *, P < 0.05; **, P < 0.01; ***, P < 0.001). N-C, Nigeria-Cameroon.

FIG 3

Principal coordinate analysis of chimpanzee site-level genetic distance [F′_ST/(1 − F′_ST)] (a), site-level vegetation distance (Bray-Curtis) (b), prokaryote (bacteria and archaea) community dissimilarity (Bray-Curtis) (c), and ubiquitous parasite community dissimilarity (Jaccard) (d). Numbers in the bottom corners represent the percent variation explained by principal coordinate 1 (PC1) and PC2, respectively. Point colors and convex hulls delineate the four geographic regions. n = 32, 27, 560, and 560 for a, b, c, and d, respectively. Overall, much of the structure in microbiomes is accounted for by genetic distance; note that the clustering in panel c matches the clustering in panel a rather than panel b. To see climate distances by region, see Fig. S2 at https://doi.org/10.6084/m9.figshare.14607735.

FIG 4

Prokaryote Bray-Curtis dissimilarity as a function of geographic distance (in kilometers), and vegetation dissimilarity (Bray-Curtis) for the whole data set and within the West, Central, and East regions. Within-region analyses for Nigeria-Cameroon are not shown here because there are only two sites. Statistics are from Mantel tests with Pearson r values. To aid in visualization, quadratic (geography) and linear (vegetation) models are displayed. To view similar figures in relation to climate distance, rather than geographic distance, see Fig. S2 at https://doi.org/10.6084/m9.figshare.14607735. To view similar figures in relation to genetic distance, see Fig. S3 at https://doi.org/10.6084/m9.figshare.14607738.

FIG 5

Relationship between geographic distance (in kilometers), climate distance (Euclidean distance), sex (FF = female-female comparisons, FM = female-male comparisons, MM = male-male comparisons), and prokaryote (bacteria and archaea) community composition in chimpanzees and humans. For chimpanzees, a quadratic (geography) and linear (climate) trendline is shown, while for humans a dashed, nonsignificant, linear trendline is shown. Blue points and error bars represent means and standard deviations, respectively. Statistics are from Mantel tests (geography and climate) and analysis of variance (ANOVA) (sex). Note that while there were significant effects of sex on dissimilarity, and male-male comparisons were more dissimilar than female-female and female-male comparisons in chimpanzees, but less dissimilar in humans, the extremely small effect sizes in both species suggests sex does not structure the gut microbiome in either species.