Comparative Analysis of Fecal Microbiomes From Wild Waterbirds to Poultry, Cattle, Pigs, and Wastewater Treatment Plants for a Microbial Source Tracking Approach

Abstract

Fecal pollution in coastal areas is of a high concern since it affects bathing and shellfish harvesting activities. Wild waterbirds are non-negligible in the overall signal of the detectable pollution. Yet, studies on wild waterbirds' gut microbiota focus on migratory trajectories and feeding impact on their shape, rare studies address their comparison to other sources and develop quantitative PCR (qPCR)-based Microbial Source Tracking (MST) markers to detect such pollution. Thus, by using 16S rRNA amplicon high-throughput sequencing, the aims of this study were (i) to explore and compare fecal bacterial communities from wild waterbirds (i.e., six families and 15 species, n = 275 samples) to that of poultry, cattle, pigs, and influent/effluent of wastewater treatment plants (n = 150 samples) and (ii) to develop new MST markers for waterbirds. Significant differences were observed between wild waterbirds and the four other groups. We identified 7,349 Amplicon Sequence Variants (ASVs) from the hypervariable V3-V4 region. Firmicutes and Proteobacteria and, in a lesser extent, Actinobacteria and Bacteroidetes were ubiquitous while Fusobacteria and Epsilonbacteraeota were mainly present in wild waterbirds. The clustering of samples in non-metric multidimensional scaling (NMDS) ordination indicated a by-group clustering shape, with a high diversity within wild waterbirds. In addition, the structure of the bacterial communities was distinct according to bird and/or animal species and families (Adonis R ² = 0.13, p = 10^-4, Adonis R ² = 0.11, p = 10^-4, respectively). The Analysis of Composition of Microbiomes (ANCOM) showed that the wild waterbird group differed from the others by the significant presence of sequences from Fusobacteriaceae (W = 566) and Enterococcaceae (W = 565) families, corresponding to the Cetobacterium (W = 1427) and Catellicoccus (W = 1427) genera, respectively. Altogether, our results suggest that some waterbird members present distinct fecal microbiomes allowing the design of qPCR MST markers. For instance, a swan- and an oystercatcher-associated markers (named Swan_2 and Oyscab, respectively) have been developed. Moreover, bacterial genera harboring potential human pathogens associated to bird droppings were detected in our dataset, including enteric pathogens, i.e., Arcobacter, Clostridium, Helicobacter, and Campylobacter, and environmental pathogens, i.e., Burkholderia and Pseudomonas. Future studies involving other wildlife hosts may improve gut microbiome studies and MST marker development, helping mitigation of yet unknown fecal pollution sources.

Keywords: NGS; enteric pathogens; environmental pathogens; fecal pollution; microbial source tracking; microbiome; qPCR; wild waterbird.

FIGURE 1

UpSetR visualization of interactions between the obtained ASVs within the whole dataset. The grid along the bottom is used to identify interaction sets (analogous to a Venn diagram). Heavily colored and connected blue dots in the grid indicates that the key group shown on the left has contributed to the interaction set shown on the top. The number of ASVs per group and the size and taxonomy (at the phylum level) of each interaction set are represented by horizontal bars on the left and vertical bars on the figure above, respectively.

FIGURE 2

Boxplots illustrating the alpha diversities with variations intra- and inter-groups for Chao1 richness estimator (A) and Shannon diversity index (B) computed from the ASV contingency table. WW, wastewater.

FIGURE 3

Distribution of predominant bacterial genera in wild waterbird samples according to relative abundance obtained by the gene encoding 16S rRNA. Bacterial community compositions were grouped by wild waterbird families: (A) wild Anatidae, (B) Laridae, (C) Haematopodidae and Scolopacidae (wader birds), and (D) Phalacrocoracidae and Hydrobatidae (cormorants and storm petrels, respectively). Stacked bar plots represent the sequence abundances of the 17 most abundant genus-level taxa identified in the fecal samples. Percent sequence abundances given as the number of reads matching a given bacterial family per total reads for that sample.

FIGURE 4

Distribution of the 18 most predominant bacterial phyla in wild waterbird samples according to relative abundances obtained by the gene encoding 16S rRNA. Bacterial community compositions were grouped by wild waterbird families: (A) wild Anatidae, (B) Laridae, (C) Haematopodidae and Scolopacidae (wader birds), and (D) Phalacrocoracidae and Hydrobatidae (cormorants and storm petrels, respectively). Stacked bar plots represent the sequence abundances of the 18 most abundant phylum-level taxa identified in the fecal samples. Percent sequence abundances given as the number of reads matching a given bacterial family per total reads for that sample.

FIGURE 5

Non-metric multidimensional scaling (NMDS) plots based on Bray–Curtis (A, stress = 0.177; B, stress = 0.179) distance metrics in relation to the whole dataset (A) or the wild waterbird groups (B). Colors represent host classes. Ellipses represent 95% confidence intervals of centroids of each point.

FIGURE 6

Heatmaps based on the number of reads of a selection of (A) 3 genera harboring known MST markers and (B) 37 bacterial genera harboring potential pathogens, and derived from the whole dataset at the bird or livestock animal family levels including wastewaters (WW).