Manipulating Atlantic salmon, Salmo salar, fillet fatty acids: Can we go from fish to plant and back again? Malcolm Jobling NFH, University of Tromsø, 9037 Tromsø, Norway e-mail:
[email protected] The fatty acid compositions of fish often bear a close resemblance of those of their food, and the fatty acids present in salmon fillets are influenced by the types of oils used in feed formulation.
General background Plant (vegetable) and marine fish oils differ in the types of fatty acids they contain (Table 1), so the fatty acid compositions of salmon fed plant and fish oils also differ (Figure 1).
Table 1. Summary of fatty acids characteristic of some plant (vegetable) and fish oils used in fish feeds. Fatty acids of the (n-6) and (n-3) series are the essential fatty acids, i.e. animals must obtain them via their diet. EPA [20:5 (n-3)] and DHA [22:6 (n-3)] are the major (n-3) highly-unsaturated fatty acids (HUFAs).
Characteristic fatty acids Plant oils 16:0 18:1 (n-9) 18:2 (n-6) 18:3 (n-3)
Fish oils Palm, Cottonseed
20:1 isomers Herring, Capelin
Rape, Olive, Sunflower, 22:1 isomers Herring, Capelin Palm Soya, Maize, Cottonseed, Anchovy, Sardine, 20:5 (n-3) Sunflower Menhaden Linseed
22:6 (n-3)
1
Marine fish oils
Plant (vegetable) oils are characterised by 18-carbon (18C) fatty acids of the (n-9), (n-6) and (n-3) series, whereas marine fish oils contain higher percentages of (n-3) HUFAs, particularly EPA and DHA. It is the (n-3) HUFAs that have been found to be beneficial for several aspects of human health. They are required for normal neural and sensory development in babies and small children, and may also reduce the risk of heart disease and some forms of cancer in adults. As a result of this there is currently a focus on the influences that plant (vegetable) oils and fish oils have on the compositions and sensory attributes of fish fillets.
Fillet fatty acids (%)
50 40 30 20 10 0
(n-6) Fatty acids
(n-3) Fatty acids
Plant oil
Fish oil
Figure 1. Plant (vegetable) and marine fish oils impart marked differences to the fatty acid profiles of Atlantic salmon, Salmo salar, fillets. Fish given feeds containing plant oils tend to have high fillet concentrations of (n-6) fatty acids, whereas salmon fed marine fish oils deposit large amounts of (n-3) HUFAs (particularly EPA and DHA) in the fillet fat.
2
Given this background it would an advantage to be able to predict the time-course of the change in the fillet fatty acid profile following the introduction of feeds with different types of oils. There are two inter-related research challenges linked to this problem Ø to find a model that describes how fillet fatty acid profiles change following introduction of a new type of feed Ø to be able to provide estimates of the time required for fillet fatty acid profiles to stabilise following the feed change
The means of change There are two main ways in which tissue fatty acid profiles may be changed following a change in feed oil composition; these may be termed wash-out and dilution. The dictionary definition of wash-out is to remove, eradicate or carry away. Thus, with wash-out a change in the fatty acid profile of a fillet would arise primarily as a result of metabolism and turnover (‘burning off’) of fatty acids mobilised from the existing stores. Dilution is defined as a reduction in strength that results from the addition of more material. As such, with dilution a change in fatty acid profile would arise because the existing fatty acid stores become diluted as the fish grow, and accumulate and deposit fatty acids obtained from the diet.
Incorporation of fatty acids into fish tissues is influenced by a number of factors. These include: Ø preferential incorporation of specific fatty acids into the tissue lipids Ø preferential metabolism ('burning off') of some fatty acids Ø modification of fatty acid structure (i.e. by chain elongation or desaturation).
3
If the factors relating to transformation and metabolism dominate following a change in the source of dietary fatty acids the prediction of tissue compositional changes will be extremely difficult. On the other hand, if these factors play a minor role in governing changes in tissue fatty acid composition following a switch from a feed containing plant (vegetable) oil to one containing fish oil, and dilution is most important, the change in fatty acid composition becomes relatively easy to model.
The dilution model In a dilution model it is assumed that: Ø the initial fatty acid content becomes diluted as the fish grow and deposit increasing amounts of fat Ø dietary fatty acids are deposited in the tissues without influencing the metabolism or turnover of existing fatty acid stores Ø incorporation of dietary fatty acids into the fat occurs in the same way independent of the fatty acid compositions of the existing stores. A dilution model describes the change in the percentage of a given fatty acid in a tissue when increasing quantities of fatty acids are added. When there is a change to a new type of feed the fatty acid profiles of the ‘test’ fish change over time as more fatty acids are deposited in the tissues. The fatty acid profiles of these ‘test’ fish gradually come to resemble those of fish that have been fed this feed for prolonged periods (the latter are a ‘reference’ group that provide the basis for comparison).
The mathematical model The mathematical description of dilution is:
PT = PR + [(PI - PR)/(QT /QI)]
4
PT is the percentage of a fatty acid in the tissues of the ‘test’ fish at time T, PI is the initial percentage and PR is the percentage of the fatty acid in the ‘reference’ group. QI is the initial total amount of fatty acids (or total fat) present, and QT is the amount present at time T. The equation can be transformed to:
(PT - PR)/(PI - PR) = 1/(QT /QI) This means that a plot of (PT - PR)/(PI - PR) against QT /QI gives a curve that tends towards an asymptotic value at infinite QT /QI, giving a 'law of diminishing returns' relationship.
Does this type of curve describe the changes we observe when we change the types of oils in fish feeds? The model was first tested using data from a study of feed oils (plant vs. fish) and dietary fat concentrations (high, ca. 34% vs. low, ca. 22%) on growth and fatty acid profiles of several tissues (fillet, viscera and carcass) of Atlantic salmon parr and post-smolt. The changes in the percentages of three fatty acids, typical of plant oils, were examined following the introduction of a feed containing fish oil. The three fatty acids studied were 18:1, 18:2 (n-6) and 18:3 (n-3).
The curves fitted to the observed changes were almost identical to the ideal dilution model, so the results of this first test looked very promising. Figure 2 shows the plot for the pooled data of the three fatty acids [18:1, 18:2 (n-6) and 18:3 (n-3)] as an example.
5
1,0
(P T - P R )/(P I - P R )
0,8
(P T - P R )/(P I - P R ) = 0.995(Q T /Q I ) - 0.986 (R2 = 0.899; N = 180)
0,6
0,4
0,2
0,0 1
3
5
7
9
Q T /Q I
Figure 2. The ratios of % change [(PT - PR)/(PI - PR)] of the three fatty acids 18:1, 18:2 (n-6) and 18:3 (n-3) in relation to changes in total fat (QT /QI) in the fillet, viscera and carcass of Atlantic salmon following a change from feeds containing plant oil to ones containing fish oil. The circles indicate the observed values and the solid line is the fitted power curve regression. The calculated regression indicated in the figure is very close to the ideal dilution curve equation.
The test was then repeated using the changes in the percentages of the three fatty acids in the fillet (Figure 3). Once again the curve fitted to the observations was very similar to an ideal dilution curve.
6
1,0
-0.9377 (P T - P R )/(P I - P R ) = 0.9027(Q T /Q R )
0,8
2
(P T - P R )/(P I - P R )
(R = 0.923; N = 60)
0,6
18:1 18:2 (n-6) 18:3 (n-3)
0,4
0,2
0,0 1
3
5
7
9
Q T /Q I
Figure 3. The ratio of % change [(PT - PR)/(PI - PR)] of three fillet fatty acids [18:1, 18:2 (n-6) and 18:3 (n-3)] in relation to changes in total fat (QT /QI) in Atlantic salmon fillets following a change from feeds containing plant oils to ones containing fish oil. The small symbols are the observed values, the solid line is the fitted power curve regression, and the large grey-tone circles indicate values calculated using the ideal dilution curve equation.
As a final check the concordance between the predicted percentages of the fatty acids (from the dilution model) and those found in the fillets 98 days after the feed switch (observed values) was examined (Figure 4). Values were very similar, giving a regression that was extremely close to a 'line of equality' (a regression that has a slope of 1 and passes through the origin).
7
% Predicted from dilution model
25
Predicted = 0.9654 Observed + 0.2036 (R
2
= 0.996; N = 12)
20
15
10
18:1 18:2 (n-6) 18:3 (n-3)
5
0 0
5
10
15
20
25
% Observed after 98 days of feeding
Figure 4. Plot showing the degree of concordance between the fillet percentages of 18:1, 18:2 (n-6) and 18:3 (n-3) predicted by the dilution model and the percentages of the same fatty acids observed in the fillets 98 days after a change from feeds containing plant oil to ones containing fish oil. The observed values are the mean±SD for six fish per sample.
Conclusions from the modelling The results provide evidence that Ø a simple dilution model can be used to describe the change in fillet fatty acid profile of Atlantic salmon following a change in feed oil type Ø fatty acid catabolism and turnover ('burning-off') play only a minor role in governing the compositional changes Ø preferential incorporation and structural modification, i.e. elongation and desaturation, of fatty acids are of limited importance in modulating the fillet fatty acid profiles of salmon that are feeding and growing well.
8
Applications and implications The dilution model can be used to make predictions about the changes in fatty acid compositions of salmon fillets that occur following a change from feeds containing one type of oil to another (e.g. from plant oil to marine fish oil). However, for the predictions to be accurate quite a lot of information is needed: Ø The % of the fatty acids present in the fillet at the time of the feed change, i.e. PI Ø The % of the same fatty acids present in the fillet fat of the 'reference' fish, i.e. PR Ø An estimate of QT /QI, which requires information about: Ø the growth, or size increase, of the fish from the time of the feed change until the time of harvest Ø possible changes in the mass of the fillet relative to body mass, i.e. fillet yield Ø possible changes in the fat % in the fillet over the feeding period.
The size increase of the fish from the time of the feed change to harvest will be predetermined (e.g. growth from 1 kg ® 5 kg), so will not create problems for the estimation. Both fillet yield and fillet fat % change with increasing fish size, so these factors must be taken into account when making predictions. There is probably sufficient information about these changes in farmed Atlantic salmon to enable 'reference' (literature) values to be used for making predictions.
Fillet yield increases with growth of small salmon, but in farmed salmon above ca. 2 kg in body weight the yield remains fairly stable at about 55% (54-58%).
9
Atlantic salmon deposit increasing amounts of body fat as they grow, and the fat % in the fillet and cutlet increase as the fish increase in body size (Figure 5). The rate of increase tends to be higher in small salmon than in larger fish, so fat % approaches a plateau when salmon reach large size.
Cutlet (NQC) fat (%)
25
20
15
10
5
0 0
3
6
9
12
Fish weight (kg)
Figure 5. Fat % in the cutlets (NQC = Norwegian Quality Cut) of farmed Atlantic salmon of different sizes (body weights, kg).
The reliance upon 'reference' values will be adequate for most practical purposes, when a 'rough-and-ready' estimate of the change in fillet fatty acids will suffice. The use of 'reference' values would also reduce the numbers of analyses needed each time a prognosis (prediction) is made.
10
Is the dilution model also suitable to describe fillet fatty acid changes in other fish species? Atlantic salmon is a 'medium-fat' fish. The fillets of wild salmon often contain 4-8% lipid, but the fillets of large, farmed fish may have about 20% fat. The salmon fillet contains a lot of storage fat, and the fatty acid composition of the fat generally follows that of the feed quite closely.
On the other hand, some species of fish, termed 'lean' fish (
