Search

Alpha mannose-oligosaccharide (MOS) prebiotics are widely deployed in animal agriculture as immunomodulators as well as to enhance growth and gut health. Their mode of action is thought to be mediated through their impact on host microbial communities and their associated metabolism. Bio-Mos is a commercially available prebiotic currently used in the agri-feed industry, but studies show contrasting results of its effect on fish performance and feed efficiency. Thus, detailed studies are needed to investigate the effect of MOS supplements on the fish microbiome to enhance our understanding of the link between MOS and gut health. To assess Bio-Mos for potential use as a prebiotic growth promoter in salmonid aquaculture, we have modified an established Atlantic salmon in vitro gut model, SalmoSim, to evaluate its impact on the host microbial communities. The microbial communities obtained from ceca compartments from four adult farmed salmon were inoculated in biological triplicate reactors in SalmoSim. Prebiotic treatment was supplemented for 20 days, followed by a 6-day washout period. Inclusion of Bio-Mos in the media resulted in a significant increase in formate (P = 0.001), propionate (P = 0.037) and 3-methyl butanoic acid (P = 0.024) levels, correlated with increased abundances of several, principally, anaerobic microbial genera (Fusobacterium, Agarivorans, Pseudoalteromonas). DNA metabarcoding with the 16S rDNA marker confirmed a significant shift in microbial community composition in response to Bio-Mos supplementation with observed increase in lactic acid producing Carnobacterium. In conjunction with previous in vivo studies linking enhanced volatile fatty acid production alongside MOS supplementation to host growth and performance, our data suggest that Bio-Mos may be of value in salmonid production. Furthermore, our data highlights the potential role of in vitro gut models to complementin vivo trials of microbiome modulators.

Additional file 8: Figure S2. Specificity of the primers that target Lactobacillus and Mycoplasma genus. The results in figure summarise bacterial genus targeted by Lactobacillus (Fig. 1 A, B, C) and Mycoplasma (Fig. 1 D, E, F) specific primer set. It shows that of all genus captured by Lactobacillus primer pair 98% were Lactobacillus in fish 1, 78% in fish 2, and 65% in fish 3. While of all genus captured by Mycoplasma primer pair 95% were Mycoplasma in fish 1, 79% in fish 2, and 56% in fish 3.

Additional file 14: Figure S8. In vivo phenotypic fish performance fed on two different feeds. Figure visually represents different phenotypic performance data of fish (n=32 per feed) fed on two different feed. A Atlantic salmon length in centimetres; B Atlantic salmon length in weight in kilograms; C Atlantic salmon percentage carcass yield; D Atlantic salmon gonad weight in grams; E Atlantic salmon gutted weight in kilograms; F Atlantic salmon liver weight in grams. Blue box plots represent data from salmon (n=32) fed on Fish meal free diet, and red represents Atlantic salmon fed on Fish meal diet (n=32).

Additional file 15: Figure S9. VFA production within different SalmoSim compartments fed on different feeds. Figure represents 11 volatile fatty acid production within SalmoSim system fed on Fish meal and Fish meal free diets within different gut compartments. Y axis represents the concentration of specific volatile fatty acid (mM) while the X axis represents each gut compartment (stomach, pyloric caeca, midgut). Red colour denoted Fish meal and blue – Fish meal free diets. The lines above bar plots represent statistically significant differences between different feeds and gut compartments. The stars flag the levels of significance: one star (*) for p-values between 0.05 and 0.01, two stars (**) for p-values between 0.01 and 0.001, and three stars (***) for p-values below 0.001.

Additional file 11: Figure S5. Stability within SalmoSim system, within different biological replicates and different gut compartments, calculated by using generalised UniFrac values for pairwise beta diversity analysis. The figure represents microbial stability within the SalmoSim system (data separated by different biological replicates and gut compartments) as the pairwise beta diversity comparison between different sampling time points (days), calculated by using generalised (50%) UniFrac as a distance measure. A small p-value indicates that the two time points are statistically different, and p>0.05 indicates that two time points are not statistically different. The colour key illustrates the p-value: red end of spectrum denoting low p values (distinct compositions between time points) and dark green indicating high p values (similar compositions between timepoints).

Additional file 9: Figure S3. Stability within SalmoSim system calculated by using unweighted and weighted UniFrac values for pairwise beta diversity analysis. The figure represents microbial stability within the SalmoSim system (data from all gut compartments combined) as the pairwise beta diversity comparison between different sampling time points (days), calculated by using A unweighted (0%) and B weighted (100%) UniFrac as a distance measure. A small p-value indicates that the two time points are statistically different, and p>0.05 indicates that two time points are not statistically different. The colour key illustrates the p-value: red end of spectrum denoting low p values (distinct compositions between time points) and dark green indicating high p values (similar compositions between timepoints).

Additional file 1. Supplementary methods.

Additional file 12: Figure S6. Stability within SalmoSim system calculated by using different UniFrac values for pairwise beta diversity analysis. The figure represents microbial stability within the SalmoSim system (data from all gut compartments and two different technical replicate runs combined) as the pairwise beta diversity comparison between different sampling time points (days), calculated by using unweighted (0%), generalised (50%) and weighted (100%) UniFrac as a distance measure. A small p-value indicates that the two time points are statistically different, and p>0.05 indicates that two time points are not statistically different. The colour key illustrates the p-value: red end of spectrum denoting low p values (distinct compositions between time points) and dark green indicating high p values (similar compositions between timepoints).

Additional file 10: Figure S4. Measured value (qPCR, ammonia and protein concentrations) stability within different SalmoSim compartments fed on Fish meal and Fish meal free diets. The figure summarises the Estimated Marginal Means output for each mixed-effect linear model (Model 1) run with different values measured in different SalmoSim compartments (qPCR measurements, ammonia and protein concentrations) identifying the difference between different time points during the first (system fed on Fish meal diet) and last 20 days (system fed on Fish meal free diet) of validation experiment. A small p-value indicates that the two time points are statistically different, and p>0.05 indicates that two time points are not statistically different. The colour key illustrates the p-value: red end of spectrum denoting low p values (low correlation between time points) and dark green indicating high p values (no differences between timepoints).

Additional file 7: Figure S1. Physiochemical conditions measured within different real Atlantic salmon gut compartments. 1A-1C measured physicochemical conditions within real salmon (n=3) gut compartments: pH (1A), temperature (°C, 1B), dissolved oxygen (mg/L, 1C).

Additional file 13: Figure S7. Calculated alpha-diversity metrics within different gut compartments of real salmon and SalmoSim fed on Fish meal and Fish meal free diets. Figure visually represents different alpha diversity outputs within different gut compartments of real salmon in red and SalmoSim in yellow (stable time points: 16, 18 and 20 fed on Fish meal, and 36, 38 and 40 fed on Fish meal free diet) fed on Fish meal and Fish meal free diets. A visually represents effective richness (number of OTUs), B represents effective Shannon diversity. The lines above bar plots represent statistically significant differences after feed change. The stars flag the levels of significance: one star (*) for p-values between 0.05 and 0.01, two stars (**) for p-values between 0.01 and 0.001, and three stars (***) for p-values below 0.001.