With the continual expansion of aquaculture of piscivorous fish there is a global requirement to reduce the dependence on wild caught fish for the generation of fishmeal and fish oil. There are a number of studies that have explored the transcriptional response to plant and vegetable oil replacements in fish diets, often when 100% replacement experimental diets are used results in reduced growth
[24, 25]. When both proteins and oils were replaced
 in rainbow trout microarray analysis of liver indicated processes such as protein metabolism and cell cycle being altered, however the reasons for these changes could either have been combined effect of lowered essential fatty acids or changes in amino acid profile. There is also growing body of work demonstrating the genotype diet interaction in rainbow trout
[25–27] and Atlantic salmon
, with the latter concluding transcription of key metabolic regulators that respond to plant based feeds depend on the fish’s genotype. Commercial salmon feed formulations partially replace fish meal with plant derived proteins, with varying effects on fish physiology and performance including fish growth and food conversion
[24, 29]. This paper examined the physiological and performance effects with transcriptional responses in mid intestine, liver and skeletal muscle in salmon to a post smolt feed containing high levels of plant derived proteins with low fishmeal content compared to a low plant protein-high fishmeal diet. The tissues examined play different physiological roles in the fish all of which may be affected by dietary changes as reflected in the different response profiles seen in these tissues. In common with a number of other transcriptomic studies on nutritional effects (where extreme dietary manipulations are avoided) the impact of the different diets appears quite subtle suggesting limited changes to physiological and metabolic pathways
. Although, in general, the scale of gene response is quite low, distinct changes in all tissues examined do indicate that there are responses occurring as a result of the dietary components.
When interpreting the findings it is important to take a comparative holistic approach whereby apparent down regulation of genes in fish fed one diet may in fact be the result of up regulation in fish fed the other dietary treatment and vice versa. In this paper the expression is analysed relative to the marine profile (MP) diet.
The occurrence of histological changes was very similar for fish fed either diet, with low frequencies observed for the majority of parameters measured. The main histological observation was a mild fatty change in liver, which is considered normal under intensive rearing conditions. This anomaly appeared to have no clinical effect, and was not prevalent for either dietary group. Although there was a higher incidence of peri‐biliar duct infiltration in liver of fish fed the MP diet no other histological changes were observed to indicate an inflammatory response. This reduced infiltration in the liver by the biliary ducts may be related to the altered immune gene expression in the liver. In mammals viral infections can result in biliary atresia
 and specifically can be related to immune related injury to bile ducts following infiltration of CD4+ T and production of interferon γ
. At this stage we are unable to tell if the reduction in the structures is a direct result of the nutritional components or due to the altered hepatic immune gene expression in between the two diets. The distal intestine is often the target organ for anti‐nutriental factors, in particular those present in SBM and pea protein concentrate
, inducing histological changes including shortening of simple and complex mucosal folds with widening of central stroma, inflammatory cell infiltration in the submucosa and lamina propria with a mixed leukocyte population
[33, 34]. In this study both dietary groups presented fish with distal intestine anomalies but with low frequencies, and no significant difference between diets. These results together with those reported by Sanden et al. (2005)
 indicate soybean products, including the alcohol soluble fraction, may be used within formulation constraints without inflicting gross histopathological anomalies in the intestine of Atlantic salmon.
Transcriptome changes in the mid intestine
The fish intestine has multiple functions which will be the first to respond to changes in nutritional intake; particularly digestion and absorption of nutrients and immune responsiveness to ingested pathogens, antigens and new antigens generated by the gut flora via gut associated lymphoid tissue
. The intestine contains three distinct regions: the proximal intestine containing pyloric caeca, the mid and distal intestine. Nutrient absorption occurs in all regions via enterocytes, by passive and facilitated diffusion and active transport
, with the highest rates of uptake in the proximal section
[37, 38]. The intestine of piscivorous fish can be particularly sensitive to plant derived ANFs and non-starch polysaccharides (NSP) in feed, resulting in altered permeability in trout feed 44% 44% SBM
[7, 39] and enteritis in the distal intestine of salmon given a high dietary inclusion of SBM
. The inflammation/enteritis may be similar to a hypersensitivity reaction
[33, 40]. Often these effects are temporary and quickly disappear when the intestinal tract is no longer exposed to the ANFs
[11, 33, 41].
Processing of plant products for fish feed is under continual improvement and some current plant derived protein concentrates have very low contaminating factors or botanical impurities. In addition knowledge of ANF containing plant feed materials has increased to the point where commercial feed formulations permit plant derived proteins at acceptable inclusion levels where no negative health effects or impacts on growth and performance occur. This was confirmed in the current study by the histology assessment where no gross morphological changes associated with plant ANFs were observed. In addition there was no difference in growth or feed utilisation efficiency during the feeding period, where a doubling of weight was achieved. However there were more subtle changes to biological processes that were not apparent during the classical physiological evaluation but were detected by global transcriptomic analysis. In this trial the mid intestine showed the greatest transcriptome response of the tissues studied, reflecting the sensitivity of the intestinal cells to dietary factors. Processes modified in the intestine were related to immune parameters, cell proliferation, apoptosis, protein metabolism, energy metabolism, transport and lipid metabolism.
Fish fed the PP diet showed higher expression of genes involved in inflammation suggesting a possible dysfunction in immune regulation. Our findings support previous studies on gut intraepithelial and systemic T cells in fish which showed rainbow trout IELs are rich in T cells
[10, 42]. Additionally the expression of TSC22D3, a regulator of T cell receptor mediated cell death, was found higher in PP, this protein may be induced by glucocorticoids
[43, 44] activated by components in the PP diet. Together these results support previous reports that trace levels of substances with allergenic properties may cause expression of genes indicative of a hypersensitivity reaction in the intestine
Interestingly genes involved in the inflammatory response were both higher and lower expressed in PP compared to MP. In particular genes involved in NF-κB pathway were induced such as the signalling adaptor molecule MyD88 and the inhibitory proteins of the IkB family, NF-kB1 p105 which sequesters NFκB in an inactive form in the cytoplasm
; inhibition of this pathway results in the production of proinflammatory cytokines
. MyD88 is also part of the signalling pathway that induces type I IFNs through the interaction of the MyD88–TRAF6–IRF7 complex
 had higher expression in fish fed the PP diet, this transcription factor may have activated interferon responsive genes in PP fed fish including Mx
, virus induced gene (vig) -2
 and a virus induced TRIM protein
. PP fed fish also had higher expression of (SOCS)-7, which functions to reduce inflammation
, potentially counteracting the expression of genes related to the inflammatory response in the intestine. Other genes involved in the innate immune response were expressed at lower level in fish fed PP. IL-8 is a chemokine that attracts neutrophils to a site of inflammation
, whereas IL-17D coordinates the clearance of extracellular bacteria and contributes to the pathology of many autoimmune and allergic conditions in Atlantic salmon
. The lower expression of these genes in fish fed the PP diet may indicate there was not a proinflammatory response to the PP diet compared to MP diet.
Several antioxidant genes were expressed higher in PP fed fish mid intestine indicating protection against oxidative damage. The free radicals could be either endogenously produced by immune cells or present in the diet. Alternatively a potential lower concentration of antioxidants in the PP feed, due to the lower fishmeal content, may require the antioxidant system within the fish to be increased accordingly. The overall low expression of heat shock protein mRNAs, suggests a limited stress response in the intestine. Additionally, two genes directly involved in apoptosis process (caspase-3 and14) were expressed at higher level in PP potentially indicating increased apoptotic activity of mid intestinal cells in these fish compared to MP fed fish.
The intestine has an extremely high rate of cellular turnover and hence generally high levels of protein synthesis and protein degradation. Protein metabolism genes relating to both synthesis and degradation were found generally to be higher in PP fed fish suggesting an increase in intestinal protein turnover. Genes related to both transcription and translations were at a higher level such as translation initiation factors, elongation factors and the ribosomal protein S6 kinase. Interestingly, a number of mRNAs encoding ribosomal proteins were at a lower expression in PP. This may relate to the stability of the mRNAs or multiple use of the ribosomal subunits during translation. In parallel to general increase in synthesis genes related to protein degradation were also at a higher level in PP such as cathepsins and components of the ubiqutin proteasome pathway
[56, 57]. The higher protein turnover is likely to be energy demanding and this is related to an increase in genes encoding proteins involved in oxidative energy metabolism. Together these changes in expression suggest modulation in control of protein turnover in the intestine with components of both synthesis and degradation being altered which may reflect an increased activity of the intestine in fish fed the PP compared to the MP diet.
Cellular membrane transport related genes were higher expressed in PP, which could suggest that salmonids are able to adaptively modulate the densities of transporters to match changes in diet composition. Lipid metabolism and transport were also affected even though the PP diet contained the same fish oil content as the MP diet. mRNAs encoding two apolipoproteins were higher in expression in PP fed fish reflecting a greater mobilization and transport of cholesterol and fatty acids in the intestine, possibly an adaptive response to the lower dietary cholesterol content in PP compared to MP diet. Fatty acid metabolism genes were higher expressed in the PP diet including both FAD5 and 6, a PUFA elongase and other genes related to cholesterol metabolism. FADs are critical enzymes in the biosynthesis of long-chain highly unsaturated fatty acids (HUFA) from shorter chain PUFA
[58, 59]. These lipid metabolism differences between PP and MP are of interest as the intestine is often over looked regarding these processes and cholesterol, even if the essential fatty acids are present at required levels, other factors including cholesterol may change, revealing the complex nature of the early digestion and modifications of nutrients in the mid intestine.
Transcriptome changes in the liver
The liver receives nutrients and compounds from the intestine and needs to respond to any substances that may have detrimental effects on the fish.
Transcriptome and proteomic studies on salmonids show nutritionally related modifications in both liver mRNAs and proteins due to feeding status
[60, 61] and diet composition
The PP fed fish showed a significant difference to fish fed MP for genes related to immune function with a lower expression of innate factors such as lectins and hepcidin. Acquired immune system components were also found at lower level including T cell receptors, MHC I and II and components of the TGF-β pathway. TGF-β has an important role in the maintenance of T-cell
 and B-cell homeostasis
 by regulating cell proliferation process and apoptosis in these cells. This result, with the higher expression of genes involved in cell death such as CIDE-3 and angiopoietin-related protein 4, indicates that apoptosis may be a mechanism induced by the PP diet salmon liver. This is not surprising as apoptosis plays a central role in the differentiation and maintenance of the liver
. A balancing effect on the apoptotic TGF-β pathway is seen in the induction of several genes encoding proteins related to cell proliferation (such as HTRA1 serine protease and annexin A3). In particular, HTRA1 serine protease inhibits signalling mediated by TGF-β family proteins
, playing an important role in contrast to cell death, whereas annexin A3 has a role in the signalling cascade during liver regeneration
. Other researchers have found the immune system to be altered following vegetable oil replacement in salmon
 and in sea bass
Genes involved in oxidative stress response (MPV17 protein, amine oxidase and HSP 70 kDa protein) were higher in liver of fish fed PP compared to MP. This is interesting to note as increases in antioxidant genes were also noted in a salmon diets that had marine oil replaced with vegetable oils
 in the liver. In rainbow trout HSP expression in liver was increased following SBM rich replacement diets
[9, 62, 63, 71], the induction of these genes may indicate a diet-induced stress response in fish fed the PP diet. Moreover, during general high protein turnover to deal with misfolded proteins
 as may be the case of fish fed PP diet.
It is interesting to note that few lipid related metabolic genes were found significantly different in the liver. Vigilin, a protein implicated in both biosynthesis and metabolism of lipids and steroids, facilitates removal of excess cholesterol from cells
 and secondly apolipoprotein A IV which facilitates transport of cholesterol to the liver were both expressed at higher level in fish fed PP compared to MP. The high expression of apolipoproteins in fish fed diets containing high levels of plant proteins has previously been observed
[9, 74] and is most likely indicates reduced cellular cholesterol in fish fed the PP diet due to decreased dietary cholesterol, associated with low fishmeal content, and/or in response to trace levels of phytoestrogens
 and phorbol esters
 co purified with the plant proteins.
Transcriptome changes in the skeletal muscle
Genes related to processes such as protein metabolism, energy metabolism, cell proliferation, apoptosis and immune function were all significantly different in PP fed fish compared to MP. Transcripts related to protein metabolism such as ribosomal protein mRNAs, transcription and translation initiation factors were generally lower in PP fed fish compared to MP fed fish indicating lower protein synthesis. In parallel, protein degradation related genes were also less, for example the ubiquitin proteasome pathway and lysosomal peptidase proteins, together these would suggest a lower protein turnover activity in the muscle tissue in PP group. Both protein synthesis and degradation are highly energy demanding processes
 and the indication of lower protein turnover, may suggest reduced energy wastage
[61, 78, 79]. This idea is strengthened by lower expression of COX2 and COX8 and other genes encoding proteins involved in glycolysis and gluconeogenesis in fish fed PP. Genes involved in cell proliferation were also expressed at lower levels in PP fed fish indicating further energy saving. Together these changes in biological processes may indicate efficient metabolic activity following feeding of the plant protein enriched diet.
Biological tissues with high metabolic rate require mechanisms to deal with free oxygen radicals, on the other hand those tissues where the metabolic rate is reduced, for example when protein turnover is decreased a reduction in oxidative stress response may be observed as was observed in this study. Additionally a number of HSPs 30, 47 and a heat shock transcription factor 1a were all at lower levels in PP reflecting the reduced protein turnover and requirement of stabilising many newly translated proteins. Genes involved in cell proliferation and related to cell death including two caspases (caspase-8 and 14) were also lower expressed in fish fed PP compared to the MP diet. The induction of bh3 interacting domain death agonist, a pro-apoptotic member of the Bcl-2 protein family
 and the suppression of caspases is in accordance with apoptosis of skeletal muscle in mammals where the Bcl2/bax system was found crucial for muscle apoptosis, whilst the caspase activity appeared inhibited
Differential expression of a number of immune related genes, particularly a decrease in interferon responsive genes including γ-ip
 and interferon-induced protein 35
 were also expressed at lower level in PP. Other pro-inflammatory agents including platelet-activating factor receptor was also at a lower level PP, which regulates several pro-inflammatory functions such as chemokine and eicosanoid receptors
. An antimicrobial peptide beta defensin 1which is a central component of the non-specific defences
 was also found lower in PP. Relating to the adaptive immune factors, TCRβ, MHCI and CD80 were all at lower levels in PP. Although we have observed differences in genes related to immune function in skeletal muscle of fish fed PP, the low level of inflammation and the subsequent immune response observed in the intestine did not cause a large immune shift in the muscle tissue. Instead, the lower expression of such genes, may allow fish fed PP to spend less energy resources on immune function for use in growth