The structure of Brazilian Amazonian gut microbiomes in the process of urbanisation

Abstract

Shifts in subsistence strategy among Native American people of the Amazon may be the cause of typically western diseases previously linked to modifications of gut microbial communities. Here, we used 16S ribosomal RNA sequencing to characterise the gut microbiome of 114 rural individuals, namely Xikrin, Suruí and Tupaiú, and urban individuals from Belém city, in the Brazilian Amazon. Our findings show the degree of potential urbanisation occurring in the gut microbiome of rural Amazonian communities characterised by the gradual loss and substitution of taxa associated with rural lifestyles, such as Treponema. Comparisons to worldwide populations indicated that Native American groups are similar to South American agricultural societies and urban groups are comparable to African urban and semi-urban populations. The transitioning profile observed among traditional populations is concerning in light of increasingly urban lifestyles. Lastly, we propose the term "tropical urban" to classify the microbiome of urban populations living in tropical zones.

Fig. 1. Map and lifestyle patterns of the sampled populations.

a Location of each sampling site in relation to the South American continent and the Brazilian Amazon. b, c The village and typical Suruí (R) residence. d, e Tupaiú (R) home and armadillo being prepared for consumption. f–h A circular arrangement of Xikrin (R) households, the nearby Bacajá River and the preparation of the cassava flour.

Fig. 2. Within- and between-group diversity measures.

a Shannon index. b Unweighted Unifrac distances within groups. c Weighted Unifrac distances within groups. d PCoA of unweighted Unifrac distances. e Weighted Unifrac pairwise PERMANOVA differences among population pairs. f PCoA of weighted Unifrac distances. g Unweighted Unifrac pairwise PERMANOVA differences among population pairs. NS not significant, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.

Fig. 3. Core microbiome distribution.

a Abundance of core microbiome phyla per population. b Abundance of core microbiome families per population. c Abundance of core microbiome genera per population. d Network of the shared core microbiome genera in each population.

Fig. 4. Differentially abundant taxa among Brazilian populations as determined by ANCOM analyses considering a W > 0.9 cut-off significance value.

a Differentially abundant genera. b Differentially abundant families. c Differentially abundant orders. NS not significant, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.

Fig. 5. PCoA of Bray–Curtis distances.

a Distances between Amazonian Xikrin (R) and Belém (U) populations to other groups from the American continent. b Distances between Amazonian Xikrin (R) and Belém (U) populations to African groups from an urbanisation gradient in Cameroon. c Boxplots of PCoA scores along axis 1 for the American continent beta-diversity analysis. d Boxplots of PCoA scores along axis 1 for the African urbanisation gradient beta-diversity analysis. NS not significant, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.

Fig. 6. Functional predictions based on PICRUSt results.

a Relative abundance of level 2 KEGG pathways including all samples. b Differentially abundant KEGG pathways as determined by ANCOM (W > 0.9). NS not significant, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.