Gill health of freshwater farmed Atlantic salmon

The success of salmon aquaculture is a result of year-round production of consistent size and quality of product to market. This has been achieved through adoption of novel production and husbandry strategies such as Recirculation Aquaculture Systems (RAS) for freshwater smolt production. Differences in water chemistry and microbiota have been reported between RAS and traditional flow-through (FT) systems, nevertheless, the influence of production systems on microbiome-induced changes in gill health is unclear and may explain variable performance seen in farmed stocks. Apart from the rearing system, gill microbial communities can be influenced by life stage, treatment application, disease state and host genetics. This thesis investigated these factors using high-throughput data of gill morphology (histology) and 16S rRNA gene amplicon sequencing data, under commercial and semicommercial conditions in Australia and Ireland. The objective of this thesis was (i) to investigate the relationship between microbial community profiles and gill pathology during a production cycle of Atlantic salmon in two commercial hatcheries (ii) to examine the effects of a commercial Chloramine[1]T treatment on the skin and gill commensal microbial communities and gill histology in an Atlantic salmon FT hatchery (iii) to describe the prevalence and intensity of a natural infection by Candidatus Clavichlamydia salmonicola in the gills of two strains of Atlantic salmon reared in FT or RAS systems and (iv) to compare the gill microbiome of two genetic lines of Atlantic salmon in two commercially equivalent hatchery production systems with a common water source in freshwater and following transfer to seawater.
Initially, a comparison of gill health and microbiome was carried out throughout a single production cycle in two commercial hatcheries in Tasmania, Australia. The two hatcheries held first feeding fry to fingerling stage in RAS, but from parr to smolts stage fish were held in FT in Hatchery A and fish remained in RAS in Hatchery B. In both hatcheries, a higher prevalence of melanin deposits was detected in early time points (fingerling and parr stages) and correlated to a change in beta diversity of bacterial communities (R2=0.91, p adj= 0.002) and high prevalence of melanin in filaments was correlated to high abundance of Sphaerotilus sp., Pseudomonas sp., Nitrospira sp., Exiguobacterium sp., Deinococcus sp., and Comamonas sp. (p<0.05). The gill microbiome composition was significantly different between Hatchery A and B throughout the sampling time points (R2= 0.27, p adj = 0.001) and diversity was higher in Hatchery A compared to Hatchery B (richness, Shannon index and Simpson index). Chemical treatments are periodically used to control pathogenic or opportunistic bacteria in aquaculture production systems, but they may negatively impact non-target commensal bacteria, changing community diversity and resulting in microbial (or bacterial) dysbiosis. Application of Chloramine-T (Cl-T) in a commercial FT salmon hatchery decreased bacterial load on gills and water at 1 h post-treatment (p >< 0.05), The gill microbiome composition was significantly different between Hatchery A and B throughout the sampling time points (R2= 0.27, p adj = 0.001) and diversity was higher in Hatchery A compared to Hatchery B (richness, Shannon index and Simpson index) Chemical treatments are periodically used to control pathogenic or opportunistic bacteria in aquaculture production systems, but they may negatively impact non-target commensal bacteria, changing community diversity and resulting in microbial (or bacterial) dysbiosis. Application of Chloramine-T (Cl-T) in a commercial FT salmon hatchery decreased bacterial load on gills and water at 1 h post-treatment (p < 0.05), these loads returned to pre-treatment levels by 24 h. Microbial richness of bacterial communities on skin and gills varied at 1 h and 13 d compared to 0 h. Contrary to the gills, bacterial diversity on skin decreased after treatment (richness). Gill histology showed subepithelial oedema at 24 h and increased number of lamellar mucous cells at 7 d compared with 0 h. Although the initial investigations in Tasmanian indicated differences between commercial hatcheries, it was uncertain how much variation was related to the production system or if it was simply due to local water conditions or farm management. Therefore, further studies were undertaken in FT and RAS set up in parallel with a common water source that was managed at a semi[1]commercial scale at the Marine Institute, Co. Mayo, Ireland. This research described the prevalence and intensity of a natural epitheliocystis infection present in the gills of two strains of Atlantic salmon (an Irish strain and an Icelandic strain) reared in either FT or RAS and evaluate the gill microbiome in hatcheries and following marine transfer.
Specific PCR confirmed the presence of Candidatus Clavichlamydia salmonicola in both fish strains and its number of copies was correlated with the intensity of epitheliocystis lesions (χ2 = 47.87, p < .0001). A significant interaction between the hatchery system and fish strain on the prevalence of gill epitheliocystis was found both using histological (F _1,244 = 6.47, p = .01) and molecular methods (F _1,97 = 10.71, p = .001), similarly a significant interaction between fish strain and hatchery type on infection intensity was found using microscopy (F _1,244 = 6.47, p = .01) and qPCR (F _1,240 = 8.43, p = .004). Specifically, the prevalence and intensity of freshwater gill epitheliocystis were higher in FT-reared fish and the Irish strain was more affected than the Icelandic. A thorough assessment of the gill microbiome showed no significant effect of production system during freshwater rearing. Changes in the bacterial communities were driven by transfer to seawater, as gill microbial community composition was significantly different between pre-transfer and post-transfer (R2 =0.06 F _{1, 267} = 17.252 p= 0.001) and there was an increase in the microbiome diversity post-transfer (richness and Shannon index of alpha diversity). Hatchery rearing systems, chemical treatment application, life stage, disease state and fish genetics exert influence on gill health by shaping the bacterial community composition of gill microbiome and its association with gill histology. This research has reported the association of gill histology and gill commensal bacterial communities comparing two commercial hatcheries and two commercially equivalent RAS and FT. The results of this thesis showed the rearing system alone did not drive the gill microbiome under parallel RAS and FT with a shared water source and that genetic fish strain played a bigger role in the susceptibility of gills to epitheliocystis instead of driving the gill microbiome profile. The differences noted between commercial FT and RAS hatcheries most likely reflect water chemistry and microbiology as well as husbandry differences rather than being inherently linked to the hatchery system type