Effect of the antioxidant profile in the diet of farmed coho salmon ( Oncorhynchus kisutch ) on the nutritional value retention during frozen storage

Salmones coho (Oncorhynchus kisutch) fueron alimentados con una dieta enriquecida en antioxidantes sintéticos (butil-hidroxi-tolueno y etoxiquina, dieta I) y su efecto comparado con dos dietas enriquecidas en antioxidantes naturales (una mezcla rica en isómeros de tocoferol, dieta II; una mezcla de isómeros de tocoferol y extracto de romero, dieta III). Una vez sacrificado, el pescado fue conservado a –18 °C durante 18 meses. El empleo de la dieta II llevó a muestras congeladas con contenido superior en proteínas sarcoplásmicas y en γy δ-tocoferol al ser comparados con pescado previamente alimentado con la dieta I. Sin embargo, no se observaron diferencias a nivel de composición proximal, contenido en α-tocoferol y astaxantina y parámetros de color (L*, a*, b*). A nivel de composición en ácidos grasos, los individuos correspondientes a la dieta II reflejaron mayores contenidos en C22:6ω3 y monoinsaturados, pero menores en C20:5ω3 y saturados al ser comparados con sus homólogos de la dieta I.


INTRODUCTION
Marine foods have recently attracted a great interest from consumers as sources of nutritional components that have positive benefits for human health (Simopoulos, 1997).The lipid fraction is now the subject of a great deal of attention because of its high content of ω3 polyunsaturated fatty acids; among them, eicosapentaenoic acid (EPA, C20:5ω3) has been related to cardiovascular health protection, while docosahexaenoic acid (DHA, C22:6ω3) has been associated with the prevention of neurological disorders (Knoch et al., 2009, Valenzuela, 2009).
The unusual rapid rate of deterioration of seafood quality has been recognized for many years.Thus, freezing and frozen storage have largely been employed to retain sensory quality and nutrients.However, frozen marine species with both highly unsaturated lipid composition and pro-oxidant compounds have shown to suffer from enzymatic and non-enzymatic rancidity development (Erickson, 1997, Aubourg et al., 1999), leading to protein denaturation (Mackie, 1993) and loss in nutrients (Castrillón et al., 1996).
To extend the shelf life of frozen fatty fish species, many efforts have been directed to the employment of antioxidants.Since synthetic antioxidants have been reported to behave as carcinogen and mutating agents, more attention has been addressed to natural antioxidant addition.Accordingly, recent research has been focused on the employment of endogenous-type antioxidants (namely, tocopherols and organic acids) (Aubourg et al., 2004, Medina et al., 2009, Taheri et al., 2012) or natural antioxidants present in plant extracts (namely, polyphenol compounds) (Stodolnik et al., 2005, Lugasi et al., 2007, Tironi et al., 2010).
J. ORTIZ, M.A. LARRAÍN, N. PACHECO, J.P. VIVANCO AND S.P. AUBOURG at -18 °C in a freezing room.The frozen fish were taken for analysis on months 0, 3, 6, 9, 12 and 18 of storage at -18 °C.From each tank under study, five different individual fish were independently analyzed at each sampling time to achieve the statistical analysis (n = 5).Analyses were carried out on the white muscle included in the region from the end of the dorsal fin backwards towards the anus.

Proximate composition assessment
Moisture content was determined from the difference between the weight of fresh homogenized muscle (1-2 g) and the weight recorded after 16-18 hr at 100-102 °C, according to the AOAC (1990) method.Results were calculated as g 100 g -1 muscle.
Protein content was measured by the Kjeldahl method (AOAC, 1990), employing the 6.25 conversion factor.Results were calculated as g 100 g -1 muscle.
The lipid fraction was extracted by the Bligh and Dyer (1959) method, by employing a single-phase solubilization of the lipids using a chloroformmethanol (1:1) mixture.Quantification results were expressed as g lipid 100 g -1 muscle.
Ash content was measured according to the AOAC (1990) method by heating at 550 °C.Results were calculated as g 100g -1 muscle.
Total carbohydrates were estimated by rounding up, with the results expressed as g 100 g -1 muscle.
Sarcoplasmic protein extracts were prepared in a low-ionic strength extraction buffer and quantified according to Piñeiro et al. (1999).Results were expressed as g 100 g -1 muscle.

Antioxidant content assessment
EQ was determined according to the He and Ackman (2000) method with direct extraction from salmon muscle with acetonitrile followed by HPLC separation using a reversed-phase C18 column.Quantification by fluorescence detection was achieved by means of an external standard curve.Results were calculated as mg EQ Kg -1 muscle.
BHT was determined according to the method described by Lundebye et al. (2010).Extraction from the muscle was carried out using acetonitrile followed by HPLC separation with fluorescence detection.Quantification was achieved by means of an external standard curve.Results were calculated as mg BHT kg -1 muscle.
Tocopherols were determined in the lipid extracts of the salmon muscle by HPLC analysis with fluorescence detection, following the standard method Ce 8-89 (AOCS, 1993).Tocopherols were identified and quantified using the external standards of the different molecules tested (α-, β-, γ-, and δ-tocopherol) (Merck, Darmstadt, Germany).Results were calculated as mg tocopherols kg -1 muscle.
Astaxanthin (AX) content was measured according to the Sheehan et al. (1998) method.Fish muscle was extracted from the salmon with acetone, Where aquaculture development is concerned, great attention has also been paid to the antioxidants present in cultivated fish feeds.Consequently, fish farmers have included a wide range of permitted synthetic (He and Ackman, 2000;Lundebye et al., 2010) and natural (Stéphan et al., 1995, Waagbø et al., 1993, Hamre et al., 1998) antioxidants in order to enhance the lipid stability of the corresponding processed food.
The present work focuses on the nutritional value of frozen farmed coho salmon (Oncorhynchus kisutch).A commercial diet enriched with butylatedhydroxytoluene (BHT) and ethoxyquin (EQ) (diet I) was fed to salmon fish and compared to two other diets, one of them including a tocopherolrich mixture (diet II) and the other, a tocopherolrosemary extract mixture (diet III).The basic aim of the study was to investigate the potential benefits of the natural antioxidants present in the diet for the nutritional value of the frozen product.

Salmon diets provided
The coho salmon individuals used in this study were cultivated in three different tanks by EWOS Innovation Research (Colaco, Puerto Montt, Chile) under the same experimental conditions reported by Ortiz et al. (2009).Thus, a diet with the following proximate composition was employed in all tanks: protein (43.0%), fat (29.0%), moisture (7.0%), ash (6.5%), crude fiber (1.3%) and carbohydrates (13.2%); the distribution of the fat composition in the diet into saturated (SFA), monounsaturated (MUFA) and polyunsaturated (PUFA) fatty acid groups was in all cases 32.5%, 27.0% and 40.2%, respectively.
In accordance with the objectives of this work, each of the three tanks was fed with a different antioxidant composition.Thus, the first tank was fed with an antioxidant mixture including 19.3, 3.0 and 22.4 mg of EQ, BHT and total tocopherols, respectively, per kg of feed (diet I); the antioxidant mixture provided to the second tank included 2.9, 0.0 and 101 mg of EQ, BHT and total tocopherols, respectively, per kg of feed (diet II); and the third tank was fed with an antioxidant mixture including 2.3, 0.0, 45, 18 and 13 mg of EQ, BHT, total tocopherols, carnosic acid and carnosol, respectively, per kg of feed (diet III).

Raw material, processing and sampling
Once the individual coho salmons had reached ca.2500 g weight, 30 fish per tank were withdrawn, sacrificed by a sharp blow to the head, the gills cut and bled in a water-ice mixture, beheaded, gutted and kept in ice for 24 h until they arrived at our laboratory.The fish were then frozen at -40 °C in individual low-density polyethylene bags, with hermetic sealing.After 3 days, the fish were stored in two different ways: diet effect and effect of frozen storage time.Comparison of means was performed using a least-squares difference (LSD) method.Statistica software (Statsoft Inc., Tulsa, OK, USA; version 6.0, 2001) was employed.

Proximate composition
Table 1 indicates the results obtained for the proximate composition of frozen salmon muscle.The data are shown as value ranges since none of the chemical constituents provided a definite tendency (increasing or decreasing scores) throughout the frozen storage period.Additionally, no differences (p > 0.05) could be assessed among fish individuals corresponding to the three different diets previously provided.
Proximate composition profile obtained agreemeent with the previous research carried out on farmed coho salmon individuals (Perea et al., 2008;Vinagre et al., 2011).The present results were, however, different from those obtained for the same species under wild condition; thus, wild individuals showed higher moisture contents and lower lipid values, while protein and ash contents remained quite constant (Vinagre et al., 2011).
The analysis of the sarcoplasmic protein fraction showed some differences related to the previous feeding (Table 2).Thus, individuals corresponding to diet II provided higher mean values than their counterparts from diet I; such differences were significant at months 6, 12 and 18.An increase in the protective action on sarcoplasmic protein fraction can be concluded as a consequence of employing the tocopherol-rich diet under the present conditions.Additionally, a marked sarcoplasmic protein content decrease was obtained for all kinds of samples as a result of the storage time.
Protein damage during the frozen storage of fish has partly been attributed to interaction with the oxidized lipids produced throughout frozen storage (Mackie, 1993;Castrillón et al., 1996).In a parallel study to the present one (Ortiz et al., 2009), individuals corresponding to diet II showed a higher and the extract was dried under nitrogen flux and dissolved in the mobile phase (20% ethyl acetate and 80% methanol/water, 9/1).The absence of 9Z-and 13Z-isomers was confirmed.Results were calculated as mg all-E-AX kg -1 muscle.

Fatty acid composition analysis
Lipid extracts were converted into fatty acid methyl esters (FAME) using acetyl chloride and analyzed by gas liquid chromatography (Perkin-Elmer 8700 chromatograph), employing a fused silica capillary column SP-2330 (0.25 mm i.d.× 30 m, Supelco Inc., Bellefonte, PA, USA) (Aubourg et al., 1995).The temperature program was as follows: increased from 145 to 190 °C at 1.0 °C min -1 and from 190 °C to 210 °C at 5.0 °C min -1 ; held for 13.5 min at 210 °C.The carrier gas was nitrogen at 10 psi and detection was performed with a flame ionization detector (FID) at 250 °C.A programmed temperature vaporizer injector was employed in the split mode (150:1) and was heated from 45 to 275 °C at 15 °C min -1 .Peaks corresponding to FAME were identified by comparing their retention times with those of standard mixtures (Qualmix Fish, Larodan, Malmo, Sweden; FAME Mix, Supelco, Inc.).Peaks were automatically integrated; C19:0 fatty acid was used as internal standard for quantitative purposes.The content of each fatty acid was expressed as g 100 g -1 total fatty acids.

Instrumental color analysis
A color analysis (CIE 1976 L*, a*, b*) was performed by employing a tristimulus colorimeter (Hunter Labscan 2.0/45) (Ortiz et al., 2008).Measurements were made directly on the salmon muscle by employing a quartz cuvette.For each sample analysis, color scores were obtained as mean values of four measurements obtained by rotating the measuring head 90° between duplicate measurements per position.

Statistical analysis
The data obtained were subjected to the ANOVA method (p < 0.05) to explore differences accounted for a 0-12.3ppm range (Hwang et al., 1995).A negligible amount was also obtained in farmed cod fed with a diet including such synthetic antioxidant (Lundebye et al., 2010); however in the same experiment, BHT was detected in Atlantic salmon (Salmo salar), rainbow trout (Oncorhynchus mykiss) and halibut (Hippoglossus hippoglossus) in the range of 1.7-3.9mg kg -1 muscle.
Concerning EQ, a trace presence was obtained in most samples in this study, with all values ranging below 0.03 mg kg -1 muscle.A higher content in individuals corresponding to diet I was not attained, while no differences could be depicted with frozen storage time (p > 0.05).Previous research has investigated the biological fate of EQ in salmonid species (He and Ackman, 2000;Lundebye et al., 2010).Both studies revealed that EQ could be converted into de-ethylated EQ (hydroxy-trimethyl-dihydroquinoline) and oxidized metabolites (dihydro-trimethyl-quinolone and dihydroethoxy-trimethylquinoline), so that low values of EQ were obtained in the final product.rancidity stability and lower oxidized taste scores than their counterparts previously fed with the diet including a rich mixture of synthetic antioxidants.The present results concerning a higher retention of sarcoplasmic protein content in individuals corresponding to diet II could be explained on the basis of this lower oxidation development.

Presence of synthetic antioxidants
Synthetic antioxidants (BHT and EQ) content was analyzed in salmon muscle throughout the frozen storage.In the case of BHT, this compound could not be quantified satisfactorily since values ranged below 0.01 mg kg -1 muscle in all cases.Previous research has shown that BHT provided in the diet would be highly retained in the fish liver (Hollas et al., 2008).A study concerning the synthetic antioxidant contents in cultured fish in Taiwan showed that a negligible content was present in the fish muscle, while fish liver   , 1998).Higher values of γ-and δ-tocopherol found in individuals corresponding to diet II can be explained as being influenced by the antioxidant composition of the diet provided; these higher values would be in agreement with the lower oxidation development found in such individuals in the above-mentioned parallel experiment (Ortiz et al., 2009).

Astaxanthin content and color changes
Color plays an important role in the appearance, presentation and acceptability of seafoods, especially in those related to salmonid species (Benzce Rørå et al., 2005;Choubert and Baccaunaud, 2006).AX is well known as the main pigment responsible for the pink color of salmonid fish species, so its retention during processing should be very important to guarantee consumer acceptance and retain the commercial value of the product.
AX content values were included in the 6-12 mg kg -1 muscle range (Table 3).Differences among individuals corresponding to the various diets could hardly be found, so that a definite effect of the diet on the AX content could not be inferred.In a previous study (Jensen et al., 1998), the addition of α-tocopherol in the diet did not affect the AX retention in rainbow trout (O.mykiss) muscle either.In addition, a definite tendency with frozen storage time could not be concluded in the actual research for the AX content in the different kinds of samples.Under the present experimental conditions, it can be concluded that diets enriched with synthetic antioxidants as well as those enriched with natural ones have led to an AX content retention in the frozen salmon muscle.This AX content retention agrees with previous investigation concerning the same species under similar frozen conditions, previously fed with a diet enriched with synthetic antioxidants (Vinagre et al., 2011).However, and contrary to the present results, an AX content decrease during frozen storage has been described for Atlantic salmon (S. salar) (Christophersen et al., 1992) and rainbow trout (O.mykiss) (Jensen et al., 1998).
Color changes in salmon muscle were investigated by means of physical (CIE L*, a* and b*) analysis.According to Table 3, no significant differences could be observed among samples, so that no effect of dietary antioxidants provided to salmon could be concluded on any of the color parameters under study.This result agrees with the fact that AX content did not show practically any differences as a result of the antioxidants provided in the diet.
In the case of the L* value, an increasing tendency with frozen time could be concluded for all kinds of samples.A marked increase in this parameter has already been mentioned as a result of different technological treatments on salmonid species such as frozen storage (Christophersen et al., 1992), vacuum-packaging refrigerated (4 °C) storage (Choubert and Baccaunaud, 2006) and hydrostatic high-pressure (Amanatidou et al., 2000).

Tocopherol presence
The Tocopherol analysis is included in Table 2.The presence of β-tocopherol was found to be negligible in most cases, according to its previously reported low deposition ability (Sigurgisladóttir et al., 1994).
The analysis of α-tocopherol content indicated that no significant differences could be attributed to the diet applied when comparing the different kinds of frozen samples.However, a general α-tocopherol loss tendency was observed as a result of the frozen storage time; this decrease was found significant at the end of the experiment in all cases.Previous research also accounts for α-tocopherol content losses in frozen fish muscle such as Atlantic salmon (S. salar) (Hamre et al., 1998), rainbow trout (O.mykiss), horse mackerel (Trachurus trachurus) (Medina et al., 2009) and channel catfish (Ictalurus punctatus) (Brannan and Erickson, 1996); these losses were attributed to α-tocopherol consumption in order to retard lipid oxidation in fish muscle throughout the frozen storage.
Concerning the γ-tocopherol presence, lower mean values were observed in individuals corresponding to diet I when compared to their counterparts from diets enriched with natural antioxidants; differences were found significant at months 0, 3, 12 and 18.A higher retention of this tocopherol molecule is concluded to be produced in individuals belonging to both diets including relatively high levels of natural antioxidants.Additionally, higher mean values were also obtained for samples corresponding to diet II when compared to their counterparts from diet III; however, differences were not found to be significant.Concerning the effect of the frozen storage time, a mean value analysis leads to a general γ-tocopherol content decrease; this decrease was found significant in individuals corresponding to diet I at the end of the experiment.Such results agree with a previous investigation on frozen channel catfish (I.punctatus) (Brannan and Erickson, 1996) fillets.On the contrary, no loss in γ-tocopherol content during the frozen storage of Atlantic salmon (S. salar) was observed (Hamre et al., 1998).
The content of δ-tocopherol showed higher mean values throughout the whole experiment in individuals corresponding to diet II; differences were found significant at months 0, 9, 12 and 18.Consequently, a marked effect of dietary antioxidants on the content of this tocopherol compound is inferred.Large fish-to-fish differences were found in most cases, so that a definite tendency of δ-tocopherol content throughout the frozen storage time was not attained for any kind of sample under study.In agreement with the present results, previous research did not lead to content losses for this tocopherol molecule during the frozen storage of Atlantic salmon (S. salar) (Hamre et al., 1998).
Tocopherols are known as endogenous antioxidants that can act as scavengers of free radicals, so that protection against the very early stages of lipid oxidation would be favored (Jensen J. ORTIZ, M.A. LARRAÍN, N. PACHECO, J.P. VIVANCO AND S.P. AUBOURG of polymerized Schiff bases and fluorescent compounds has been proven (Undeland et al., 2003).As in the present research, previous studies did not find changes in b* value in salmonid species for the frozen storage (Tironi et al., 2010).

Fatty acid analysis
In all cases, C18:1ω9 and C16:0 fatty acids were found to be the most abundant, followed by C22:6ω3, C16:1ω7 and C20:5ω3.A similar profile was obtained by Perea et al., (2008) and reviewed by Vinagre et al., (2011) for the same farmed salmon species.Concerning EPA content (Table 4), a higher mean value was obtained in samples corresponding to diet I, being significantly higher at time 0 when compared to fish individuals corresponding to both diets enriched with natural antioxidants.The opposite result was obtained for DHA presence (Table 4); its mean value was in all cases lower in individuals corresponding to diet I, However, L* values remained under 70 in all cases of the present study throughout the storage period, which has been depicted as the permitted border line value for salmonid species (Amanatidou et al., 2000).
Concerning a* and b* parameters, some differences with frozen storage time could be concluded for the three kinds of salmon samples.However, a definite tendency could not be inferred in any case.In all kinds of samples, a* value remained above 13, which has been recognized as the permitted border line value for salmonid species (Amanatidou et al., 2000).
Previous investigations on fish species have proven a general a* value decrease as a result of different technological treatments such as hydrostatic high-pressure (Ashie et al., 1996), frozen storage (Tironi et al., 2010) and vacuumpackaging (Gobantes et al., 1998).Concerning b* value, this parameter has been directly related to lipid oxidation development; thus, an important relationship between b* value and the formation When the fatty acid groups are considered, the following general decreasing content sequence was obtained: PUFA > MUFA > SFA (Table 4).This similarity of fatty acid group composition among kinds of samples can be explained on the basis that the same fatty acid composition was present with scores significantly lower in the 0-6-month period when compared to individuals from diet II.In all cases, both fatty acids (EPA and DHA) hardly provided differences as a result of the frozen storage time in any of the different kinds of samples under study.Further research focusing on the employment of diets enriched with natural antioxidants (diets II and III) is to be continued.As a first step, optimization of the tocopherol-rich mixture is to be achieved in order to obtain an enhancement of shelf life time, sensory acceptance and nutritional value of the frozen product.
in the three diets.In this sense, previous research accounts for a marked effect of the fatty acid composition of the diet on the fatty acid composition of the product (Stéphan et al., 1995;Waagbø et al., 1993;Benzce Rørå et al., 2005).Additionally, dietary fish oils have shown an enhancement of the lipid oxidation development during frozen storage when compared to diets including non-marine oils (Stéphan et al., 1995;Waagbø et al., 1993); this result has been explained as a result of a higher PUFA presence in fish fed with marine oils.
According to Table 4, SFA analysis showed higher mean values for individuals corresponding to diet I; such differences were found significant at the 12-18-month period, while no differences were inferred between samples corresponding to both diets enriched with natural antioxidants.An opposite result could be concluded from the MUFA presence; thus, higher mean values were found in individuals corresponding to diets II and III, with such differences significant at months 0, 12 and 18. Related to PUFA content, mean values were found higher throughout the whole experiment in individuals corresponding to the diet enriched with synthetic antioxidants, although significant differences were only observed at months 3 and 6 when compared to individuals belonging to diet III.
A majority of the Western population does not consume adequate levels of ω3 fatty acids through natural dietary sources, such as fish.Consequently, a great interest has been accorded recently to the ω3/ω6 ratio of foods included in the human diet.Its value has shown a great effect on the development of certain health problems (Knoch et al., 2009, Valenzuela, 2009), with a recommended ratio near 1/6 (ω3/ω6) (Simopoulos, 1997).In the present work, the ω3/ω6 ratio (Table 4) did not provide differences as a consequence of the dietary antioxidants provided or the frozen storage time.In all cases and throughout the whole storage period, ω3/ω6 values obtained can be considered very positive for maintaining a balanced ω3/ω6 ratio.

CONCLUSIONS
A comparative study on the nutritional value of frozen farmed coho salmon was achieved, taking into account the effect of employing diets enriched with natural antioxidants.As a result, the employment of a tocopherol-rich mixture (diet II) has led to frozen samples showing higher contents of sarcoplasmic proteins and γ-and δ-tocopherols when compared to their counterparts previously fed with a diet enriched with synthetic antioxidants (diet I).However, no effect of the antioxidant profile could be depicted in the proximate composition, α-tocopherol and astaxanthin contents or color (L*, a*, b*) parameters.Concerning the fatty acid composition, individuals corresponding to diet II showed higher C22:6ω3 and MUFA contents, but lower C20:5ω3 and SFA contents when compared with their counterparts belonging to diet I. Differences found between frozen EFFECT OF THE ANTIOXIDANT PROFILE IN THE DIET OF FARMED COHO SALMON (ONCORHYNCHUS KISUTCH)…

*
Mean values of five (n = 5) replicates; standard deviations are indicated in brackets.For each parameter and for each frozen time, values followed by different letters (z, y) express significant (p < 0.05) differences as a result of the diet employed.For each parameter and for each diet, values preceded by different letters (a, b, c) express significant (p < 0.05) differences as a result of the frozen time.No letters are indicated in the case of no significant differences (p > 0.05).** Diet compositions as expressed in the Materials and Methods section.EFFECT OF THE ANTIOXIDANT PROFILE IN THE DIET OF FARMED COHO SALMON (ONCORHYNCHUS KISUTCH)… et al.

Table 1 Proximate composition (g 100 g -1 muscle)* of frozen coho salmon previously fed with different diets**
* Value ranges obtained throughout the frozen storage.Values were obtained from five (n = 5) replicates at each sampling time in fish corresponding to each diet.**Diet composition as expressed in the Materials and Methods section.J. ORTIZ, M.A. LARRAÍN, N. PACHECO, J.P. VIVANCO AND S.P. AUBOURG

Table 3 Astaxanthin content (mg kg -1 muscle) and color (L*, a* and b* values) assessment § in frozen coho salmon previously fed with different diets § §
Mean values of five (n = 5) replicates; standard deviations are indicated in brackets.For each parameter and for each frozen time, values followed by different letters (z, y) express significant (p < 0.05) differences as a result of the diet employed.For each parameter and for each diet, values preceded by different letters (a, b, c) express significant (p<0.05)differences as a result of the frozen time.No letters are indicated in the case of no significant differences (p > 0.05).** Diet compositions as expressed in the Materials and Methods section.EFFECT OF THE ANTIOXIDANT PROFILE IN THE DIET OF FARMED COHO SALMON (ONCORHYNCHUS KISUTCH)… *

Table 4 Fatty acid analysis (g 100g -1 total fatty acids)* in frozen coho salmon previously fed with different diets**
. ORTIZ, M.A. LARRAÍN, N. PACHECO, J.P. VIVANCO AND S.P. AUBOURG individuals belonging to both diets enriched with natural antioxidants were scarce, only accounting for higher δ-tocopherol and DHA contents in individuals corresponding to diet II. J