Host-associated and Environmental Microbiomes in an Open-Sea Mediterranean Gilthead Sea Bream Fish Farm

Abstract

Gilthead seabream is among the most important farmed fish species in the Mediterranean Sea. Several approaches are currently applied to assure a lower impact of diseases and higher productivity, including the exploration of the fish microbiome and its manipulation as a sustainable alternative to improve aquaculture practices. Here, using 16S rRNA gene high-throughput sequencing, we explored the microbiome of farmed seabream to assess similarities and differences among microbial assemblages associated to different tissues and compare them with those in the surrounding environment. Seabream had distinct associated microbiomes according to the tissue and compared to the marine environment. The gut hosted the most diverse microbiome; different sets of dominant ASVs characterized the environmental and fish samples. The similarity between fish and environmental microbiomes was higher in seawater than sediment (up to 7.8 times), and the highest similarity (3.9%) was observed between gill and seawater, suggesting that gills are more closely interacting with the environment. We finally analyzed the potential connections occurring among microbiomes. These connections were relatively low among the host's tissues and, in particular, between the gut and the others fish-related microbiomes; other tissues, including skin and gills, were found to be the most connected microbiomes. Our results suggest that, in mariculture, seabream microbiomes reflect only partially those in their surrounding environment and that the host is the primary driver shaping the seabream microbiome. These data provide a step forward to understand the role of the microbiome in farmed fish and farming environments, useful to enhance disease control, fish health, and environmental sustainability.

Keywords: Aquaculture sustainability; Fish microbiome; Gilthead seabream; Mediterranean Sea.

Fig. 1

Panel A: barplot showing the prokaryotic community composition (as relative abundance) at the phylum and class level (for Proteobacteria only). Taxa with an average relative abundance across all samples < 1% were aggregated as “Others”. “Unknown” includes all those reads that did not match any known taxa. Panel B: nonmetric multidimensional scaling (NMDS) ordination of community composition of fish and environmental microbiomes based on Bray–Curtis dissimilarity matrix. Panel C: richness calculated for the different types of samples; asterisks indicate the occurrence of significant differences as calculated by the Kruskal–Wallis test (p < 0.05); non-significant comparisons are not reported in the plot. Richness in environmental samples also significantly differed from those observed in fish-associated microbiomes (Kruskal–Wallis p < 0.05). The fish figure has been created in BioRender (https://biorender.com/)

Fig. 2

Bubble chart reporting the proportion and taxonomic identification of the most abundant ASVs (avg. > 0.5% across the dataset) observed in each fish and environmental microbiome. The dimension of each bubble is proportional to the log of the relative abundance of the ASV

Fig. 3

Network-based analysis to display and analyze how ASVs are partitioned between samples. In the network diagram, there are two kinds of “nodes” represented, ASV-nodes (grey) and sample-nodes (colored according to the type of sample: red, gut; yellow, gill; cyan, skin; pink, fillet; dark blue, seawater, green, sediment). The network has been calculated excluding singletons. Sample nodes’ size is proportional to their degree, i.e., the number of connections with ASVs nodes. Edges (i.e., lines between nodes) connect ASVs to the samples where they are present; shorter distances between the sample and ASV nodes reflect larger numbers of sequences from that ASV found in the connected sample. Finally, ASVs occurring in the center of the network are more evenly distributed between all the samples, while those on the edges occur preferentially in certain samples