Captivity and the co-diversification of great ape microbiomes


Increased sampling disrupts co-diversified lineages

The gyrB amplicon knowledge from wild apes (n = 130), captive apes (n = 72), and industrialized people (n = 16) generated on this examine and from revealed knowledge5, together with gyrB sequences extracted from metagenomic knowledge from over 9000 people worldwide20 (Supplemental knowledge 2), yield a complete of 7596 ASVs that typed to the Bacteroidales order. Of these, 6784 are restricted to a selected host species, and assembling these gyrB-ASVs into the largest doable monophyletic clades produces 356 well-supported clades whose constituents are current in 5 or extra host people (Fig. 1).

Fig. 1: Host-restricted clades of wild apes are missing in captive apes.

Phylogeny of host-restricted clades based mostly on gyrB-ASVs in the order Bacteroidales. In the phylogeny to the left, the three labeled and shaded clades (Bt1, Bt2, Bt3) correspond to the co-diversified lineages of Bacteroidaceae recognized in Moeller et al.5. Colors in the numbered columns that comply with point out the host-species supply of ASVs constituting every clade, with column 1 displaying the sources of all 356 clades current in no less than 5 people, column 2 displaying solely these 65 clades current in >25% of both wild or captive people of any host species, and column 3 displaying solely these 18 clades current in >25% of captive people of any host species. Circles in the subsequent two columns are shaded in response to host-species supply and sized to point the proportion of samples harboring the clade. For every of the clades distinguished in captive apes (column with white background), bacterial household and genus taxonomic assignments are color-coded in the closing column.

Given the variations in sampling amongst hosts, we thought of solely these host-restricted clades current in >25% of people of the numerous pattern sorts (i.e., captive chimpanzees, wild gorillas, industrialized people, and many others.). Our elevated sampling of wild apes and people reveals that the co-diversifying lineages reported in Moeller et al.5 symbolize solely a subset of Bacteroidales range current in wild apes and people. There are 26 host-restricted clades from wild apes that have been beforehand not acknowledged and that don’t coincide with the co-diversified lineages recognized by Moeller et al.5 (Fig. 1).

The addition of scores of host-restricted clades to the Bacteroidales phylogeny allowed re-examination of the lineages initially reported to co-diversify with great-ape species. Two of the authentic co-diversifying lineages now comprise a range of Bacteroidales sequences from people that disrupt the earlier sample of co-diversification (Fig. 2, Fig. S1). And in different circumstances, the addition of ASVs belonging to a unique host species creates mixed-host clades that disrupt the congruence between the host and bacterial phylogenies. For instance, in Bacteroidaceae lineage 2, the emergence of a host-restricted bonobo clade from inside a co-diversifying chimpanzee clade leads to the formation of two separate chimpanzee clades (Clade 2 in Fig. 2). Similarly, wild gorillas and wild chimpanzees harbor some intently associated ASVs, and this mixed-host clade (Clade 4 in Fig. S1) splits the two co-diversifying clades of Bacteroidaceae lineage 3. Thus, the newly recognized host-restricted clades that seem with elevated sampling are higher defined by bacterial diversification following a number of host-switch occasions.

Fig. 2: New ASVs inside the co-diversified lineage Bacteroidaceae 2 (Bt2 in Fig. 1).

Inset (A) at the higher left exhibits topology of co-diversified clades initially recognized by Moeller et al.5, with branches color-coded to indicate host species. The 5 main clades from this inset tree (Clades 1–5) are highlighted and labeled in the phylogeny (B), which incorporates the newly recognized gyrB-ASVs. Lineages and terminal ideas are color-coded to point the host-species supply of an ASV, and dashed traces correspond to ASVs recognized in captive apes. An emergent mixed-host clade can be labeled (Clade 6). Nodes with bootstrap assist >50 are indicated.

Overall, we discover that with further sampling, most lineages beforehand described as co-diversifying not current topologies in step with strict co-speciation and now comprise a combination of each host-restricted and mixed-host clades. Although three separate statistical assessments offered important assist for co-diversification between the hominid phylogeny and Bacteroidaceae lineage 2 (PACo, p < 0.001; Parafit, p = 0.001; HCT, p < 0.001), these assessments additionally reached the similar degree of assist when the host phylogeny was randomized (PACo, p < 0.001; Parafit, p = 0.001; HCT, p < 0.001). Cumulatively, these assessments set up that gyrB-ASVs assort into host-restricted clades extra typically than anticipated by likelihood—however the proven fact that random host timber additionally produce important host-restricted associations signifies that distance-based statistical assessments are unreliable for figuring out whether or not the topology of host-restricted clades is in step with co-diversification.

Loss and sharing of wild ape host-restricted ASVs in captivity

We initially got down to decide whether or not the co-diversified bacterial lineages current in wild apes persist in captive apes, which might each reveal the constancy of transmission regardless of main modifications in way of life and geography, and lend assist to the view that these bacterial lineages are vital to, and seemingly co-evolved with, their host species. However, most host-restricted gyrB clades in wild-ape species, each these beforehand recognized by Moeller et al.5 and by this examine, are absent from captive apes (Fig. 1). In the single case the place we observe a wild chimp-host-restricted clade of wild chimpanzees that can be current in captive apes, it’s not confined to chimpanzees however present in each captive chimpanzees and orangutans. Unlike wild apes, captive great apes largely harbor human-restricted, mixed-host, and unique-to-captive-apes gyrB clades which might be shared broadly amongst host-species.

To decide whether or not the transmission of gyrB-ASVs between captive apes and people are recapitulated throughout broader taxonomic groupings, we examined the distribution of 16S-ASVs amongst wild apes (n = 330), captive apes (n = 87), industrialized (n = 140), and non-industrialized people (n = 134) utilizing microbiome composition knowledge generated by this examine in addition to by different revealed research2,3,17,21,22,23 (Supplemental knowledge 1). Relative to the gyrB dataset, a far better proportion of Bacteroidales 16S-ASVs are recognized as mixed-host (Fig. S2), which seemingly displays the lack of ability of the V4 area of the 16S rRNA gene to differentiate amongst intently associated bacterial strains. However, regardless of the decrease proportion of host-restricted ASVs in the 16S dataset, we’re nonetheless in a position to study their distribution throughout captive apes to check whether or not patterns noticed in the gyrB dataset are constant when a broader taxonomic range of microbial taxa is analyzed.

Paralleling the host distributions noticed with fine-grained gyrB knowledge, we discover that host-restricted 16S-ASVs which might be confined to a selected ape species in the wild are largely absent from the microbiomes of captive apes (Fig. 3B) (Kruskal–Wallis: wild bonobo vs. captive apes, df = 1, H = 127.1, p < 0.001; wild chimp vs. captive apes, df = 1, H = 144.8, p < 0.001; wild gorilla vs. captive apes, df = 1, H = 151.3, p < 0.001). Of the few wild-ape host-restricted 16S-ASVs that persist in captivity, the majority are current in a number of captive species. There are solely three 16S-ASVs, two in gorillas and one in chimpanzees, which might be solely current in wild and captive conspecifics and no different host species.

Fig. 3: Compositional modifications in captive great apes are related to the loss of host-restricted ASVs from wild apes and the acquire of host-restricted ASVs from people.

Average relative abundances of (A) bacterial phyla (B) host-restricted 16S-ASVs, mixed-host 16S-ASVs, and unique-to-captive apes 16S-ASVs in the intestine microbiomes of captive apes, wild apes, and people from industrialized and non-industrialized societies, (C) bacterial genera, and (D) host-restricted 16S-ASVs, mixed-hosted 16S-ASVs, and unique-to-captive apes 16S-ASVs inside the genera proven in (C). Note that common relative abundances of a number of genera listed in (C) and of the host-restricted 16S-ASVs, mixed-hosted 16S-ASVs, and unique-to-captive apes 16S-ASVs inside these genera present parallel will increase in all captive ape species affecting convergence in microbiome compositions.

Instead of harboring strains which might be current of their wild conspecifics, a big fraction of the captive-ape microbiome consists of 16S-ASVs which might be in any other case restricted to people, in step with a sample of colonization by human-associated strains. In truth, the proportion of human-restricted 16S-ASVs noticed in captive apes doesn’t differ considerably from that in people (Fig. 3B) (Kruskal–Wallis, df = 1, H = 6.4, p = 0.093). Even amongst these bacterial genera which might be commonest in wild apes and people (Fig. 3C), the compositional shifts in captive apes are brought on by a rise in human-restricted and mixed-host 16S-ASVs (Fig. 3D).

It is feasible that the identification of ASVs as being host-restricted outcomes from restricted sampling, and, due to this fact, the findings that captive apes lack of wild-ape ASVs and possess human ASVs are affected by the relative sampling of wild apes, captive apes, and people. However, this isn’t the case: Firstly, the identification of ASVs as host-restricted isn’t more likely to be an artifact of sampling provided that they show an identical relationship between imply relative abundance and prevalence throughout host samples as mixed-host ASVs (Fig. S3). Host-restricted ASVs are extra frequent in the Bacteroidetes phylum than in the comparatively much less plentiful Firmicutes phylum (df = 16, X2 = 255.93, p < 0.0001) (Fig. S4), indicating there’s a taxonomic sample to host-restriction relatively than a random distribution. Secondly, if captive apes have been harboring unidentified strains which might be current of their wild ape conspecifics, we count on to watch many unique-to-captive-ape ASVs which might be restricted to a selected captive host species. However, ASVs noticed solely in captive apes aren’t extra more likely to be restricted to a selected host species or website (Fig. S5). Lastly, if the frequency of human-restricted ASVs in captive apes have been an artifact of under-sampling wild apes, we would count on that further sampling of wild-ape populations would shift some human-specific ASVs to mixed-host ASVs. However, we discovered that wild apes normally possess mixed-host ASVs which might be current solely in different ape species, and that captive apes are likely to harbor mixed-host ASVs which might be current each in people and in wild apes (Fig. 2), indicating that further sampling of wild apes isn’t more likely to uncover troves of further 16S-ASVs which might be disproportionately shared with people.

No proof of host-species filtering in captive great apes

A particular function of the current examine is the capability to match the intestine microbiomes of the similar captive great-ape species from a number of areas, permitting us to find out the extent to which microbiome compositions regulate to every host species in captivity. If host species differentially filter bacterial strains, captive apes of the similar host species are anticipated to harbor extra related microbiome compositions after controlling for zoo website and enclosure. Among captive apes, zoo website explains 31% of the variation in microbiome composition assessed by Bray-Curtis distance (PERMANOVA, df = 4, F = 10.6, p < 0.001, r2 = 0.32; betadisper, df = 4, F = 5.3, p = 0.004), enclosure explains a further 18% of the variation in microbiome composition (PERMANOVA, df = 5, F = 4.9, p < 0.001, r2 = 0.18; betadisper, df = 9, F = 2.3, p = 0.024), and host species isn’t considerably related to microbiome composition after controlling for website and enclosure. Because we embrace knowledge from research that employed various strategies to extract and sequence samples, we examined the diploma to which examine supply contributed to the similarity of microbiome composition (after controlling for website and enclosure) however discover no important impact.

We additionally examine the relative results of shared geography, host species, and enclosure on the fraction of shared 16S-ASVs (i.e., Sorenson distance, which disregards the relative abundance of ASVs). Individual apes which might be neither of the similar host species nor residing at the similar zoo website share a prodigious 30% of their 16S-ASVs (Fig. 4). This excessive diploma of sharing amongst captive apes is indiscriminate of species project: there isn’t any important affiliation between host species and 16S-ASV sharing after excluding people residing in the similar enclosure (permutation t-test, df = 1, t = 2.2, p = 0.12), related to what’s noticed when making use of Bray-Curtis distance metrics. Shared geography is related to a slight improve in the proportion of shared 16S-ASVs amongst people (permutation t-test, df = 1, t = 16.3, p = 0.012); nonetheless, this improve is due largely to the in depth sharing of 16S-ASVs amongst gorillas and chimpanzees at the Houston Zoo (Fig. S6). As anticipated, captive apes in the similar enclosure (that are invariably the similar species) exhibit the highest proportions of shared ASVs (±50%) (Fig. 4), far exceeding the affect of shared geography or host-species (permutation t-test, all comparisons, df = 1, t > 15.9, p = 0.012).

Fig. 4: Cohabiting captive apes exhibit the highest ranges of ASV sharing.

Proportions of shared 16S-ASVs (Sørenson similarity index) decided by in captive apes in relation to geography and host species membership.

16S microbiome composition of great ape species converge in captivity

Based on our sampling of great apes from a number of areas, captivity disrupts intestine microbiomes in an identical method throughout host species (PERMANOVA, df = 8, F = 47.9, p < 0.001, r2 = 0.35; betadisper, df = 8, F = 44.0, p < 0.001). The intestine microbiomes of captive apes are extra just like these of different captive host species than to their wild ape counterparts (Fig. 5; Supplemental knowledge 5), a convergence as a result of parallel will increase in the relative abundance of the Bacteroidetes (Kruskal–Wallis: captive vs. wild chimp, df = 1, H = 46.5, p < 0.001; captive vs. wild bonobo, df = 1, H = 16.5, p = 0.002, captive vs. wild gorilla, df = 1, H = 29.0, p < 0.001) and Spirochaetes (Kruskal–Wallis: captive vs. wild chimp, df = 1, H = 28.1, p < 0.001; captive vs. wild bonobo, df = 1, H = 16.5, p = 0.002, captive vs. wild gorilla, df = 1, H = 35.9, p < 0.001) and decreases in the relative abundances of Actinobacteria (Kruskal–Wallis: captive vs. wild chimp, df = 1, H = 62.7, p < 0.001; captive vs. wild bonobo, df = 1, H = 23.6, p < 0.001, captive vs. wild gorilla, df = 1, H = 56.0, p < 0.001) (Fig. 3A).

Fig. 5: Gut microbiome convergence in captive great apes based mostly on 16S-amplicon sequencing.

Gut microbiome compositions of captive great apes, wild great apes, and people partitioned by species and way of life, and visualized by non-metric multidimensional scaling (NMDS) based mostly on Bray-Curtis distances (PERMANOVA, df = 8, F = 47.9, p = 0.001, r2 = 0.35; betadisper, df = 8, F = 44.0, p = 0.001).

The discount in the abundance of Actinobacteria in captive apes is accompanied by a discount in actinobacterial range; nonetheless, whole bacterial range in captive apes, although various by zoo website, is just like that of wild apes (Fig. S7). All captive ape species have elevated abundances of six bacterial genera, together with a number of genera of Ruminococcaceae which might be frequent in the human microbiome (Fig. 3C; Supplemental knowledge 6). Based on all metrics examined, captive ape microbiomes are extra just like these of people residing in non-industrialized societies (Fig. 5 and Supplemental knowledge 5; Bray-Curtis distance; Fig. S8, Jaccard, unweighted and weighted UniFrac distances). The similarity is due, partly, to the prevalence of Treponema, that are not often current in people sampled from industrialized societies (Fig. 2B). Captive apes additionally exhibit a loss in the relative abundances of microbial genera which might be distinctive to and distinguish the microbiomes of wild ape species (Fig. S9). For occasion, Acinetobacter is attribute of the microbiome of wild gorillas however is absent from captive gorillas. These compositional shifts are constant throughout captive-ape host species sampled from a number of websites and research (Fig. S10). Overall, captivity homogenizes great ape intestine microbiomes such that people exhibit a loss of microbial taxa attribute of wild apes and a acquire of taxa distinguished in people.