In silico, in vitro and in vivo characterization of host-associated Latilactobacillus curvatus strains for potential probiotic applications in farmed Atlantic salmon (Salmo salar)

Abstract

Salmon aquaculture is the fastest growing animal protein production system in the world; however, intensive farming leads to poor weight gain, stress, and disease outbreaks. Probiotics offer the potential to enhance growth performance and feed efficiency in Atlantic salmon, as well as immunostimulate fish against common pathogens, benefitting farmers and consumers with more efficient production. Here, we isolated and identified 900 native microbial isolates including 18 Lactobacilli from the farmed salmon intestines. Based on whole-genome sequencing and phylogenetic analysis, the Lactobacillus candidates belonged to Latilactobacillus curvatus (L. curvatus) species and formed two distinct phylogenetic groups. Using bioinformatics and in vitro analyses, we selected two candidates L. curvatus ATCC PTA-127116 and L. curvatus ATCC PTA-127117, which showed desirable safety and probiotic properties. The two L. curvatus candidates were evaluated for safety and efficacy (higher final weight) in Atlantic salmon alongside spore-forming Bacilli isolated from salmon, poultry, and swine. All the tested candidates were safe to salmon with no adverse effects. While we did not see efficacy in any Bacillus supplemented groups, compared to untreated group, the group administered with the two L. curvatus strains consortium in feed for seven weeks in freshwater showed indicators of improvement in final body weight by 4.2%. Similarly, the two L. curvatus candidates were also evaluated for safety and efficacy in Atlantic salmon in saltwater; the group administered with the two L. curvatus strains consortium in feed for 11 weeks showed indicators of improvement in final body weight by 4.7%. Comprehensive metabolomics analyses in the presence of different prebiotics and/or additives identified galactooligosaccharide as a potential prebiotic to enhance the efficacy of two L. curvatus candidates. All together, these data provide comprehensive genomic, phenotypic and metabolomic evidence of safety and desirable probiotic properties as well as indicators of in vivo efficacy of two novel endogenous L. curvatus candidates for potential probiotic applications in Atlantic salmon. The in vivo findings need to be confirmed in larger performance studies, including field trials.

Figure 1

Phylogenetic relationship of Latilactobacillus curvatus, Latilactobacillus sakei and Latilactobacillus fuchuensis strains using 92 core genes. The phylogenetic relationship was explored using UBCG v3.0 and a maximum likelihood tree was inferred using GTR + CAT model. L. reuteri ATCC PTA-126788 was used as an outgroup.

Figure 2

Antimicrobial susceptibility of Bacillus and Lactobacillus strains. MIC (μg/mL) values for each antibiotic tested against the respective genus are shown. Nine medically important antibiotics at a concentration range of 0.06–32 μg/mL were tested, and the respective antimicrobial susceptibility cut-off concentrations required for that genus are shown at the bottom of each panel. *NR = not required by EFSA. The results are representative of 3 independent experiments.

Figure 3

Effect of probiotic supplementation on the weights of salmon following daily administration in feed for 45 days in freshwater. (A) Timeline of experimental events. (B) Body weights and specific growth rates (SGR, %/day) for each group following daily administration of the respective probiotic candidates in feed for 45 days. Horizontal bar denotes mean. *P = 0.0172 for NCP vs TP3, ANOVA with Dunnett’s test (n = 70). TP, Test Product; NCP, negative control product.

Figure 4

Effect of probiotic supplementation on the weights of salmon following daily administration in feed for 75 days in saltwater. (A) Timeline of experimental events. (B) Mean body weights ± standard error as well as specific growth rate (SGR, %/day) for each group following daily administration of the respective probiotic candidates in feed for 75 days. *P = 0.041 for Co vs P1, ANOVA with Dunnett’s test (n = 64). P1, Test Product 1-Saltwater; P2, Test Product 2-Saltwater; Co, Negative Control Product-Saltwater.

Figure 5

Principal component analysis (PCA) of feature abundance changes across media additives compared to media controls. Each marker in the figure represents one of three replicates in the corresponding treatment (shown in different colors). Numbers in parenthesis indicate the variance explained by each of the principal components. The histogram on the bottom represents the distribution of samples from each of the two strains along the first principal component. GLC, glucose. The data are representative of 3 independent experiments.

Figure 6

Features showing at least tenfold difference in abundance with different media additives compared to a control condition. (A) The number of features with higher than tenfold increase or decrease in abundance in media supplemented with different prebiotics and/or additives compared to a glucose media control for strain PTA-17. Error bars represent the standard error of the mean across replicates (n = 3). (B) Like (A), but for strain PTA-16. In both panels, prebiotics and/or additives are sorted according to the number of metabolites with increased abundance.