Vertically and horizontally transmitted microbial symbionts shape the gut microbiota ontogenesis of a skin-mucus feeding discus fish progeny

Abstract

Fish gut microbial communities play key functions for their hosts, but their ontogenesis is poorly understood. Recent studies on the zebrafish suggest that gut symbionts are recruited naturally through horizontal transmission from environmental water. We used an alternative fish model, the discus (Symphysodon aequifasciata), to identify the main factors driving fish gut microbiota ontogenesis. The discus exhibits a unique parenting behavior: both discus parents vertically feed their fry with a cutaneous mucus secretion during three weeks post-hatching. We hypothesized that vertical microbial transmission via parental mucus feeding, along with horizontal transmission of environmental microbial symbionts, helps to shape the taxonomic structure of the discus fry gut microbiota. To assess this premise, we thoroughly documented the gut microbiota ontogenesis of a discus progeny during 100 days post-hatching. The V4 16S rRNA gene was sequenced to assess taxonomic structure of fry gut, parent mucus, and water samples. Our main results suggest that specific microbial symbionts both from the parents skin mucus and environmental water play important roles in shaping the structure of the fry gut microbiota.

Figure 1

Alpha diversity in whole larvae microbiota significantly decreases between 0–3 and 4–20 DPH, but significantly increases from 4–20 to 50–80 and 80–100 DPH. Non-parametric Shannon (np Shannon) diversity indexes of whole larvae and fry feces microbiota are represented through time in (a) (in a scatter plot) and (b) (in a boxplot, including significance letters of t-tests at p-value threshold of 0.05). The np Shannon diversity of other microbial niches is represented in (c). Chao diversity indexes of whole larvae and fry feces microbiota are represented through time in (d) (in a scatter plot) and (e) (in a boxplot, including significance letters of t-tests at p-value threshold of 0.05). The Chao diversity of other microbial niches is represented in (f). Note that the adult diet is not represented in this figure (neither in c nor f), as only one sample was collected from this microbial niche. Green markers represent samples taken before mucus feeding (0–3 DPH), blue markers represent samples taken during mucus feeding (4–20 DPH), yellow markers represent samples taken early after mucus feeding (50–80 DPH), and orange markers represent sample taken later after mucus feeding (80–100 DPH). Data layouts in (a and c) are best explained by logarithmic type trendlines. In (b and c), y-axis labels are the same than in (a). In (e and f), y-axis labels are the same than in (d). Since contrasted patterns of diversity were observed between 0–3 DPH to 4–20 DPH (decrease) and 50–80 DPH to 80–100 DPH (increase), the graphs in a and d are composites of two separate graphs that were merged together, to better illustrate the contrasted patterns from early to later developmental stages. The composite graphs in (a and d) were: one diversity scatter plot from 0–3 to 4–20 DPH, and another diversity scatter plot from 4–20 DPH to 50–80 and 80–100 DPH. The logarithmic trendlines were plotted independently on each composite graphs, before merging.

Figure 2

Convergence of whole larvae and fry feces microbial community towards adult-like gut microbiota. The bar charts in (a,b,c,d and e) respectively represent the relative abundances (%) of the most abundant 5 bacterial phyla, 10 bacterial classes, 10 bacterial orders, 15 bacterial families, and 15 bacterial genera in all samples, in regards to the number of DPH. In this figure, “AF” stands for “Adult feces” (or feces from the discus parents), “BDM” stands for “Breeding discus mucus”, “NBDM” stands for “Non-breeding discus mucus”, and “AD” stands for “Adult diet”. In (f), the relative abundances (%) of five OTUs from the Erysipelotrichaceae family is shown in regards to the number of DPH and to the different microbial niches studied. Note that in (f), the developmental stage from 4–20 DPH is decomposed in its five samplings at 6, 9, 13, 16, and 20 DPH to document more precisely the increased abundance of the Erysipelotrichaceae family during this key developmental stage. In (g), a Multi-Factorial Analysis (MFA) is shown including all the samples from the different niches (except the sample from Adult diet, since it distorted the representation of all the other samples). The MFA was constructed with the FactomineR package from R, using the relative abundance of the 50 most abundant OTUs in all samples. Confidence ellipses highlight differences at a 0.95 confidence threshold. PERMANOVAs between the three clusters were done on a Thetayc distance matrix, using the vegan package from R, with 10 000 permutations.

Figure 3

Microbial taxa from water and skin mucus are drivers of whole fry and fry feces microbiota. In (b,d,f and h) co-abundance based networks of the 50 most abundant OTUs, constructed with Cytoscape v. 3.2.1. Each dot (node) and its label represent an OTU defined at the lowest level possible. A link (edge) between two nodes highlights a Spearman correlation index >0.9 between the two taxa and a p-value < 0.0001 (corrected with Bonferroni). The size of each node is proportional to the number of edges that it is connected to. In (a,c,e and g) bar charts representing the recruitment index (see Methods section for details on calculation) from each of the environmental microbial reservoirs (water, parents’ mucus, adult diet), at the four life stages of the fry: before mucus feeding (0–3 DPH), during mucus feeding (4–20 DPH), early post-mucus feeding (50–80 DPH), and late post-mucus feeding (80–100 DPH). Recruitment from water is represented by blue bars, recruitment from skin mucus is red, recruitment from parental feces is grey, and recruitment from the adult diet is pink.

Figure 4

Taxonomic shift towards a core skin mucus microbiota on parents rearing fry. In (a) Multi-Factorial Analysis (MFA) based on the variations of relative abundance of bacterial taxa (identified at the lowest taxonomic level possible) in regards of the reproductive status of the discus: breeding or not-breeding. Confidence ellipses highlight significant differences at a 0.95 confidence threshold. In (b) boxplots representing the percentage of shared OTUs between both parents of the breeding couple before mucus feeding (MF), during mucus feeding, after mucus feeding, and between 20 random couples of non-breeding control fish. In (c) analysis of the taxonomic shifts that occurred in the skin mucus microbiota of parents when getting into reproduction phase. Red bars represent a decrease while blue bars represent an increase (in % of relative abundance). Each OTU was labeled at the lowest taxonomic level possible.