Oxidation of lipid and protein in horse mackerel (Trachurus trachurus) mince and washed minces during processing and storage

Food Chemistry 114 (2009) 57–65

Contents lists available at ScienceDirect

Food Chemistry journal homepage: www.elsevier.com/locate/foodchem

Oxidation of lipid and protein in horse mackerel (Trachurus trachurus) mince and washed minces during processing and storage Sylvie Eymard, Caroline P. Baron, Charlotte Jacobsen * The Technical University of Denmark, National Institute of Aquatic Resources (DTU AQUA), Department of Seafood Research, Section for Aquatic Lipids and Oxidation, Building 221, Søltofts Plads, DK-2800 Kgs. Lyngby, Denmark

a r t i c l e

i n f o

Article history: Received 25 June 2008 Received in revised form 1 September 2008 Accepted 4 September 2008

Keywords: Omega-3 Washed fish mince Lipid oxidation Protein oxidation Surimi

a b s t r a c t Protein and lipid oxidation was followed during processing and storage of mince and washed minces prepared from horse mackerel (Trachurus trachurus). Briefly horse mackerel mince (M0) was washed with three volumes of water, mimicking the surimi production and different washed products were obtained: M1, M2 and M3, with one, two and three washing steps, respectively. The different products were characterised (i.e. lipid content, protein, water, iron, fatty acid profile and tocopherol content) and analysed for protein and lipid oxidation in order to investigate the impact of the washing steps on oxidation. Subsequently the different products were stored for up to 96 h at 5 °C and samples were taken out regularly for analysis. Lipid oxidation was investigated by measuring primary oxidation products (lipid hydroperoxides) and secondary oxidation products (volatiles). Protein oxidation was followed by determination of protein solubility, protein thiol groups and protein carbonyl groups using colorimetric methods as well as western blotting for protein carbonyl groups. Lipid and protein oxidation markers indicated that both lipid and protein oxidation took place during processing and the ranking for oxidation was as follows M0 < M1 < M2 6 M3 with M0 being significantly less oxidised than M3. Results indicated that washing creates an imbalance in the initial prooxidant-antioxidant equilibrium in the muscle tissue and contributes to the observed differences in the oxidative status of the four products obtained. In contrast, during storage of different products, lipid oxidation development was faster in M0 and the ranking was as follows M0 > M1 > M2 P M3. Lipid and protein oxidation developed simultaneously in different minces during storage, but it was not possible to determine at which level these two reactions were coupled. Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction Fatty fish contains high levels of long chain omega-3 polyunsaturated fatty acids (PUFA) such as eicosapentaenoic acid (EPA, C20:5 n-3) and docosahexaenoic acid (DHA, C22:6 n-3). Consumption of these fatty acids is believed to be beneficial to human health. For example, omega-3 PUFA are involved in brain development of child and they have protective effects against heart disease and some types of cancers (Shahidi & Miraliakkbari, 2004; Shahidi & Miraliakkbari, 2005). However, fish consumption, and consequently omega-3 PUFA intake is too low in many countries, particularly in the Western world. Therefore, nutritionists and public authorities recommend the populations to increase fish consumption. The increased demand for fish products has motivated the development of new seafood products. Fish can be consumed directly or after transformation by processes such as drying, smoking, salting, and washing. Surimi, a fish protein concentrate obtained by successive washings of the fish mince, is an intermedi* Corresponding author. Tel.: +45 45252559; fax: +45 45884774. E-mail address: [email protected] (C. Jacobsen). 0308-8146/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodchem.2008.09.030

ate product involved in the production of new seafood analogues such as crab sticks (Okada, 1992). Surimi is usually produced from relatively lean fish species such as Alaska pollack (Theragra chalcogramma). However, the content of omega-3 fatty acids in these surimi products is low and the overexploitation of lean species has led the industry to process more abundant species, such as small pelagic species, for producing surimi. The production of fish mince from fatty fish species has the potential to provide healthy fish products with appreciable levels of omega-3 fatty acids. Due to their high fat content and the high amount of sarcoplasmic proteins and dark muscle, surimi production has been investigated for fatty fish species in order to obtain surimi with no odour, light colour and good gel forming ability (Hultin & Kelleher, 2000). However, during fatty fish processing and storage, quality may decline as a result of several reactions affecting both the protein and lipid fractions and decreasing the nutritional and sensory properties of the product. Indeed, due to their high level of unsaturated long chain polyunsaturated fatty acids, fatty fish products are very susceptible to oxidation. This leads to the formation of free radicals and lipid hydroperoxides, primary products of oxidation which break down to secondary lipid oxidation compounds such as

58

S. Eymard et al. / Food Chemistry 114 (2009) 57–65

alcohols, aldehydes, and ketones. Among the consequences of these reactions are the development of unpleasant odour, rancid taste and discolouration (Frankel, 2005). Moreover, the compounds resulting from lipid oxidation can modify proteins by inducing cross linking, resulting in modifications of amino acids of nutritional interest and a decrease in protein functionality (solubility, hydrophobicity) due to protein denaturation (Pokorny, Elzeany, Luan, & Janicek 1976). In muscle cells, proteins are in high concentrations and they are also close to free radical initiators exposing them to oxidative reactions. Protein can be modified independently of lipids and according to Davies and Dean (1997) proteins are the primary targets of oxygen radicals in vivo. Proteins in muscle can be modified by reactive oxygen species (ROS) that include free radicals but also by non-radical species such as H2O2 and lipid hydroperoxides. These reactions can lead to radical formation, formation of amino acid derivatives, protein break down, and polymerisation and negatively impact the functional properties of the product such as its texture and gel forming capacity (Xiong, 2000). In contrast to lipid oxidation, which has been thoroughly investigated, protein oxidation and the link between the two reactions have received little attention in food matrices despite the fact that the protein fraction with its nutritional and technological properties plays an important role in food quality. The objective of this study was to follow the development of oxidation in both lipid and protein fractions during processing and storage of horse mackerel in order to determine which reaction developed first and to which extent both reactions were linked. By using the principle of surimi washing, it was possible to produce fish minces with compositional differences in their lipid and protein fractions and also to obtain fish minces with different oxidative level in order to compare lipid and protein oxidation development depending on the matrix properties.

2. Materials and methods

horse mackerel mince (M0) obtained after the grinding, (2) mince after the first dewatering (M1), (3) mince obtained after the second dewatering (M2) and (4) mince obtained after the third washing/ dewatering and centrifugation (M3). Immediately after processing, products were divided in portions of 260 g and placed in aluminium boxes (18.5  13.5 cm) closed with a cardboard and aluminium cap. The boxes were stored at 5 oC for 12, 24, 60 and 96 h (T12, T24, T60 and T96). At each time point the content of the boxes was vacuum packed in plastic bags and stored at 80 °C until further analysis. For the T0, samples were vacuum packed in plastic bags and stored at 80 °C immediately after processing. 2.3. Water content The water content was determined by weighing after drying approximately 3 g of sample at 115 °C for 18 h according to the AOAC official method (1995). Analyses were performed in triplicate and results were expressed as g water per kg wet sample. 2.4. Total lipid content Total lipids (TL) were extracted in triplicate from 10 g of sample with methanol/chloroform (1/2, v/v) according to the Bligh and Dyer (1959) method. The lipid content was determined gravimetrically and the results were expressed as g lipid/100 g wet sample. 2.5. Iron content The total iron content of fish was determined using an atomic absorption spectrophotometer 3300 (Perkin Elmer, Waltham, MA, USA). Horse mackerel muscle sample (2.0 g) was mixed with 5.0 ml concentrated HNO3 and 2.0 ml H2O2 and the samples were digested using a microwave oven (CEM, MOS 81D, Matthews, NC, USA). After acid digestion, the samples were analysed according to the protocol from the Nordic Committee on Food Analysis (1991).

2.1. Reagents and standards 2.6. Tocopherol content Chemicals and standards were obtained from Sigma–Aldrich (St. Louis, MO, USA). Solvents were of analytical grade and from Lab–Scan (Slangerup, Denmark). Deionized water (Millipore system) was used throughout. Gel and standards for SDS–PAGE were from Invitrogen (Carlsbad, CA, USA). 2.2. Raw material, processing and storage conditions Horse mackerels (Trachurus trachurus) were caught in February 2006 at the cost of Spain, filleted manually and immediately frozen. Fish filets were packed in plastic bags and shipped frozen on dry ice. They were still frozen when they arrived at the laboratory where they were stored in plastic bags under vacuum in a 40 °C freezer until processing, which was carried out within one month after catch. All processing steps were performed in a 2 °C temperature-controlled room. Horse mackerel fillets (18 kg) were defrosted, handskinned and grounded through a plate with 3 mm holes using a grinder (ATOM TL22, ceg, Cardano al Campo, Italy). The obtained mince was washed three times in cold distilled water (2 °C) at a ratio 1:3 (w/v). Samples were slowly stirred for 30 s every 2 min for 4 min and then allowed to settle for 6 min. Total time spent at each washing step was 10 min. Following each washing, the mince was dewatered for approximately 5 min through a polyamide tissue (Sefar Nitex 03-100/44, Sefar AG, Heiden, Switzerland). After the last dewatering, the product was centrifuged 20 min at 5 °C and 24 500 g (Sorvall RC 5B Plus, SLA 1500, Dupont, Norwalk, CT, USA). The four products obtained during processing were: (1)

Tocopherol content was determined on the Bligh and Dyer lipid extracts by HPLC using an Agilent 1100 series HPLC (Agilent Technologies, Palo Alto, CA, USA), equipped with a fluorescence detector or a UV diode array detector. A fraction of the lipid extract was evaporated under nitrogen and redissolved in 2 ml n-heptane before injection of an aliquot (40 ll) on a Spherisorb s5w column (250 mm  4.6 mm) (Phase Separation Ltd., Deeside, UK). Elution was performed with an isocratic mixture of n-heptane/2-propanol (100: 0.4; v/v) at a flow of 1 /min. Detection was done using a fluorescence detector with excitation at 290 nm and emission at 330 nm and according to AOCS (1998, Official method). Results are expressed in ng tocopherol per g of lipid. 2.7. Phospholipid content of total lipid extracts Phospholipid content (PC) of the three total lipid (TL) extracts was measured using a colorimetric method (Stewart, 1980) based on the formation of a complex between phospholipids and ammonium ferrothiocyanate. Analyses were performed in triplicate. A standard curve was prepared with phosphatidylcholine in chloroform (5–50 lg/ml) and results were expressed as g phosphatidylcholine equivalents per kg of total lipid extract (g PC Eq kg1 TL). 2.8. Fatty acid composition of total lipid extracts Fatty acid composition of the three lipid extracts was determined by gas chromatography of fatty acid methyl esters (FAMEs).

S. Eymard et al. / Food Chemistry 114 (2009) 57–65

2.8.1. Preparation of methyl esters Fatty acids in the TL extract were trans-esterified to methyl esters using a base-catalysed transesterification followed by a Borontrifluoride-catalysed esterification according to AOCS (1998, Official method). The methyl esters were dissolved in n-heptane to a concentration of about 20 mg ml1. 2.8.2. Gas chromatographic analysis FAMEs (1.5 lg) were injected, in the split mode with a split mode ratio of 1:25, into an HP 5890 a gas chromatograph (Hewlett-Packard, Avondale, PA) equipped with an autosampler and a flame ionisation detector. The separation was carried out on an Omegawax 320 fused silica capillary column (30 m  0.32 mm, 0.25 lm; Supelco, Bellefonte, PA). The helium carrier gas flow was 21 cm s1. The injection and detection temperatures were 250 °C and 240 °C, respectively. The initial oven temperature was 160 °C, immediately raised by 3 °C min1 to 200 °C, held for 1 min, further raised by 3 °C min1 to 220 °C and held for 17 min. The total duration of the analysis was 40 min. Methyl esters were identified by comparison of the retention times of authentic standards. The peaks were integrated with PE Nelson software. Identified fatty acids were quantified by comparison of integrated areas and fatty acid compositions were expressed as area percentage of total fatty acids. 2.9. Lipid hydroperoxide measurement Lipid hydroperoxide content was determined in duplicate on the three total lipid extracts according to the method of Shantha and Decker (1994). Results were expressed in mg of peroxide equivalents per kg of total lipids (mg Eq peroxide kg1 TL) and mg of peroxide equivalents per kg of sample (mg Eq peroxide kg1).

59

were performed in the mass range of 30–350 atomic mass unit with a repetition rate 2.2 scans s1. 2.10.4. Identification and quantification Compounds were identified by MS library searches and by comparing retention time and spectra with MS runs of external standards. For quantification purpose, a mixture of ten selected standards (1-hexanol, hexanal, heptanal, octanal, cis-4-hepten-1al, 2-penten-1-ol, 1-penten-3-ol, 2-pentenal, 2,4-heptadienal, 2-hexenal) diluted in ethanol in five sets of concentrations (0.01– 0.5 mg/g), were added directly on the Tenax tubes (1 ll) and thermal desorption and GC–MS analysis were performed under the conditions described above for the samples. Calibration curve were done for each standard using the HP Chem station software. 2.11. Sensory analysis An expert panel consisting of four assessors evaluated the odour of the samples. Firstly, each assessor evaluated the sample. Subsequently, the panel members discussed the different scores and agreed on a score for each attribute for the sample. A scale from 0 to 15 was used. The following odour descriptors were used: rancid, fatty/oily, cardboard, painting, sweet, sea, green, fat fish, hay, rubber. Only the results of rancid and rubber are reported here. At each session, six samples and one reference was evaluated. Each sample was presented two times. All samples were presented in a randomized order. 2.12. Free fatty acids Free fatty acids (FFA) contents were determined by titration (0.1 M, NaOH) of the TL extracts (10 g) after adding ethanol (15 ml) and using phenolphthalein as an indicator. The FFA content was calculated in % oleic acid (AOCS, 1998).

2.10. Determination of volatile compounds

2.13. Solubility of protein

2.10.1. Headspace sampling The volatile compounds were sampled by a dynamic headspace according to Refsgaard, Haahr, and Jensen (1999). The mince samples were cut in pieces and frozen in liquid nitrogen and reduced to a fine powder using a blender. To 5 g of the sample powder were added 10 ml of distilled water. The aqueous suspension of mince powder was placed at 37 °C for 30 min and purged with a nitrogen flow of 340 ml min1. The volatiles were collected on Tenax-GR traps (Chromapack, Bergen op Zoom, The Netherlands). All collections were made in triplicate.

Samples (0.5 g) were homogenised in 10 ml of 0.6 M KCl in 50 mM pH 7.4 tris–HCl buffer with a Polytron PT1200CL (Kinematica AG, Lucerne, Switzerland) for 1 min. The homogenate was centrifuged at 14,000g for 15 min at 5 °C (Sorvall RC 5B Plus, SS24 rotor, Dupont, Norwalk, CT, USA). The supernatant was diluted ten fold with 0.6 M KCl and protein determination was performed using the BCA kit (Pierce, Rockford, IL, USA). The solubility was expressed in mg of soluble protein/g of product.

2.10.2. Thermal desorption A Perkin–Elmer (Norwalk, CT) ATD-400 automatic thermal desorber system was used for thermally desorbing the collected volatiles from the Tenax traps. Helium was used as a carrier gas. The gas flow from the trap to the transfer line to the capillary column in the gas chromatograph (GC) was split in the ratio of 5.0 ml min1/1.3 ml min1.

Total sulphydryl contents were determined using 5,50 -dithiobis(2-nitrobenzoic acid) (DTNB) according to Ellman’s method (1959) with some modifications. A 0.5 g sample was homogenised in 10 ml of 0.05 M phosphate buffer pH 7.2 with a Polytron PT1200CL (Kinematica AG, Lucerne, Switzerland) for 30 s. Subsequently, 1 ml of the homogenate was mixed with 9 ml of 0.05 M phosphate buffer pH 7.2 containing 0.6 M NaCl, 6 mM EDTA, and 8 M urea. The mixture was centrifuged 15 min at 14,000g at 5 °C (Sorvall RC 5B Plus, SM34 rotor, Dupont, Norwalk, CT, USA). To 3 ml of the supernatant 0.04 ml of 0.01 M DTNB solution in 0.05 M sodium acetate were added and incubated at 40 °C for 15 min. A blank was prepared replacing the homogenate with 0.05 M phosphate buffer pH 7.2 containing 0.6 M NaCl, 6 mM ethylenediaminetetraacetic acid (EDTA), 8 M urea. The absorbance was measured at 412 nm (Shimadzu UV 160A, Kyoto, Japan) and the SH content was calculated using a molar extinction coefficient of 13,600 M1 cm1. Results were expressed in micromoles of SH per g of mince.

2.10.3. GC–MS The transfer line of the ATD was connected to a HewlettPackard (Palo Alto, CA) 5890 IIA gas chromatograph equipped with a HP 5972A mass-selective detector. A DB 1701 column (30  0.25, 1.0 lm; J1W Scientific, Folsom CA, USA) with a flow of 1.3 ml of helium/min and the following temperature program was used: 35 °C for 5 min, 35 °C to 90 °C at 3 °C min1, 90 °C to 240 °C at 10 °C min1, and finally hold at 240 °C for 4 min. The GC–MS transfer line temperature was kept at 280 °C. The mass-selective detector used ionisation at 70 eV in EI mode and 50 lA emission. Scans

2.14. Determination of sulphydryl groups

60

S. Eymard et al. / Food Chemistry 114 (2009) 57–65

2.15. Determination of protein carbonyls 2.15.1. Spectrophotometric Protein carbonyls were measured as described by Levine, Williams, Stadtman, and Shacter (1994). A sample of mince (0.5 g) was homogenised in 10 ml tris-buffer (pH: 7.4, 50 mM, 1 mM EDTA) containing 0.01% BHT and 100 ll of the homogenate were precipitated with TCA. After centrifugation (12,000g, 3 min) the pellet was incubated with dinitrophenylhydrazine (DNPH) in 2 M HCl, in the dark for 30 min. For each sample a blank incubated in 2 M HCl and without DNPH was run in parallel. The samples were precipitated with TCA and the pellets were washed three times with 1 ml ethanol/ethyl acetate 1:1 (v/v). The pellet was re-dissolved in 6 M guanidine chloride in 20 mM KH2PO4. The carbonyl content was determined by measuring the absorbance at 370 and 280 nm. Results are expressed in nmoles carbonyl per mg of soluble protein. 2.15.2. Gel electrophoresis (SDS–PAGE) and protein carbonyls immuno-blot A mince sample (0.5 g) was homogenised in 4 ml tris-buffer (pH: 7.4, 50 mM, 1 mM EDTA). Protein concentrations, determined with the BCA Kit (Pierce, Rockford, IL, USA) and the mince homogenates were adjusted to 10 mg/ml. Subsequently, the homogenates (30 ll) were mixed 1:1 with 12% SDS and further diluted with 60 ll DNPH in 10% trifluoroacetic acid. The samples were incubated for 15 min in the dark at room temperature. The reaction was stopped by adding 60 ll neutralising solution (1.85 M Trizma-base, 28% glycerol and 0.1 M dithiothreitol). The samples were centrifuged for 3 min at 12,000g and 5 ll were loaded onto the gel (10% NuPage Bis–tris, Invitrogen (Carlsbad, CA, USA). After the run, one gel was stained with Coommasie Brilliant Blue G-250, and one gel was used for immuno-blotting. The proteins were transferred to polyvinylidene difluoride (PVDF) membrane 0.2 lM (Millipore, Billerica, MA, USA) using a Mini Cell SureLock, equipped with a XCell II blot module (Invitrogen A/S, Carlsbad, CA, USA) for 60 min. After transfer the membranes were blocked in 5% skim milk in tris-buffered saline (TBS) buffer (0.137 M NaCl and 20 mM tris–HCl, pH 8.0), and incubated with a 1:15,000 dilution of rabbit anti-DNP (DAKO Denmark AS, Glostrup, Denmark) in 1% skim milk, in TBS for 1 h. The membranes were washed in TBS and incubated in a 1:7500 dilution of the secondary antibody, peroxidase-conjugated swine anti-rabbit (DAKO). After washing in TBS the blot was developed using the ECL + kit (Invitrogen, Carlsbad, CA, USA). Chemiluminescence was detected on hyper-film ECL (Amersham Health A/S, Princeton, NJ, USA). 3. Results 3.1. Characterisation of the different minces at T0 Water content varied from 77% in M0 to 90% in M2 (Table 1). A major part of the lipids were removed during processing. Thus, the lipid contents were 2.4 and 1.6 g/100 g wet weight for M0 and M1, respectively, and decreased to 1.0 g/100 g wet weight for both M2 and M3 (Table 1). The proportion of saturated fatty acids in the lipid fraction was 25–26% of the total fatty acids for all minces whereas monounsaturated fatty acid proportions were highest in M1 (32.2%) and lowest in M3 (28.4%). Conversely, the highest polyunsaturated fatty acids content, mainly represented by EPA (C22:5n-3) and DHA (C22:6n-3), was obtained for M3 whereas M1 had the lowest value (Table 1). Phospholipid levels in the TL extract varied from 13.5% to 17.5% (Table 1) and the phospholipid fraction contained higher levels of polyunsaturated fatty acids (60%) compared to the neutral lipid fraction (23–25%) (Results not shown).

Table 1 Characterisation of the different minces obtained during processing of horse mackerel (n = 3)

Water content (g/100 g) Fat content (g/100 g wet w) Phospholipid (lg Eq PC/g lipid) Fatty acid compositiona (%) C14:0 C16:0 C18:0 R Saturated C16:1 R C18:1 R C20:1 R C22:1 R Monounsaturated C20:4n6 C20:5n3 C22:5n3 C22:6n3 R polyunsaturated PV (mEq peroxide/kg lipid) TBARS (mg Eq/g lipid) FFA (g/100 g lipids) Protein solubility (mg/g mince) Carbonyl content (nmoles/g protein) Sulfhydryl content (lmoles/ g protein) Tocopherol contentb (mg/kg mince) Iron content (lg/g mince) a b

M0

M1

M2

M3

77.1 ± 0.4 2.4 ± 0.3 13.5 ± 0.7

83.1 ± 0.4 1.6 ± 0.7 13.7 ± 0.7

89.6 ± 0.2 1.0 ± 0.1 18.0 ± 1.5

87.1 ± 0.4 1.0 ± 0.2 17.5 ± 0.6

3.7 15.0 5.4 25.6 3.3 14.7 4.9 7.6 32.2 1.1 6.5 2.5 22.4 36.9 2.8 ± 0.4 0.07 ± 0.0 4.5 ± 0.5 105.8 ± 5.2

3.4 15.0 5.6 25.5 3.0 14.5 4.6 6.9 30.7 1.2 6.5 2.6 24.2 38.9 9.8 ± 0.9 0.28 ± 0.03 7.7 ± 2.5 63.3 ± 1.5

3.4 15.6 5.8 26.2 2.9 14.3 4.4 6.6 30.0 1.2 6.6 2.6 24.8 39.6 48.6 ± 11.7 0.32 ± 0.05 8.1 ± 1.7 32.1 ± 1.2

2.7 15.14 5.92 25.17 2.66 14.0 4.2 6.24 28.45 1.27 6.56 2.67 26.27 41.13 50.7 ± 0.3 0.35 ± 0.03 11.2 ± 4.9 28.7 ± 2

1.7 ± 0.5

5.5 ± 2.6

4.2 ± 1.2

7.2 ± 1.7

99.1 ± 3.8

98.6 ± 2

124.4 ± 4.9

123.0 ± 6.8

4.75 ± 0.7

nd

nd

nd

17.8 ± 0.3

11.1 ± 0.3

9.4 ± 0.9

8.8 ± 0.6

Uncertainty: ±0.2%; only fatty acids with abundance >1% are shown. nd: not detected.

The level of a-tocopherol, a natural antioxidant, and the iron content, a strong prooxidant, were determined in the four mince products. a-tocopherol concentration in M0 was 4.7 mg/kg whereas no tocopherol was detected in the three other products M1, M2 and M3 (Table 1). During processing, iron content decreased from 17.8 lg/g in M0 to 8.8 lg/g in M3 (Table 1). Formation of primary (PV) lipid oxidation compounds both increased after each washing step. PV increased from 2.8 m Eq/kg lipid in M0 to 50.7 m Eq/kg lipid in M3. Free fatty acid concentrations also increased during processing, from 4.5% in M0 to 11.2% in M3 indicating that lipolysis was triggered during the washing steps (Table 1). Physico-chemicalpropertiesofproteinschangedduringwashingof the minces. The proportion of soluble proteins decreased during processingfrom106 mg/ginM0to29 mg/ginM3.Asobservedforlipidoxidation, protein oxidation markers such as carbonyl and sulphydryl groupscontentalsoincreasedduringprocessing.Carbonylcontentincreased from 2 nmol/g protein in M0 to 7 nmol/g protein in M3, while sulphydryl content increased from 100 lmol/g protein in M0 to 120 lmol/g protein in M3 (Table 1). In summary, lipid and protein oxidation markers indicated that both lipid and protein oxidation took place during processing. The ranking for oxidation was as follows M0 < M1 < M2 6 M3 with M0 being significantly less oxidised than M3. 3.2. Lipid oxidation during storage at 5 °C Sensory data indicated that at T0, the intensity of rancid and rubber attributes was lowest for M0 and highest for M3 (data not shown). The same was also the case for the descriptor ‘oily’. The intensity of rancid, oily and rubber attributes in M0 increased

61

S. Eymard et al. / Food Chemistry 114 (2009) 57–65

during the first 24 h of storage and remained steady thereafter. However the intensity of these attributes for M2 and M3 remained steady or decreased during the storage period. M1 showed different patterns depending on the attributes (data not shown). Taken together these findings indicated that oxidative flavour deterioration during storage was more pronounced in M0 than in the other minces. Peroxide value progressed differently in the four products. PV increased significantly during the first 24 h of storage for M0 and during the first 12 h of storage of M1. Thereafter, the PV levelled out in these two minces although PV in M1 increased substantially between 24 h and 60 h. During the first 12 h of storage, PV decreased slightly in M2 and increased slightly in M3. Thereafter, PV levels remained steady in both minces (Fig. 1). The development of some volatile compounds (1-penten-3-ol, 2-penten-1-ol, hexanol, heptanal, hexanal, octanal, 2-heptenal, 4heptenal, 2-hexenal and 2,4-heptadienal) was followed during

M0

M1

M2

M3

PV (meq peroxide/kg product)

1.2 1 0.8 0.6 0.4 0.2 0 0

12

24

36

48

60

72

84

96

Storage time (hours) Fig. 1. Development of peroxide (PV) during storage at 5 °C of horse mackerel mince (M0), (M1), (M2) and (M3). Averaged values of triplicates; the error bars show the standard deviation.

a

storage of the minces at 5 °C. All of them followed the same tendency and only two of them: 1-penten-3-ol and 2, 4-heptadienal, which are typical for oxidation of omega-3 fatty acids are presented (Fig. 2). Concentrations of volatiles increased markedly during the first 12 h of storage of M1 and during the first 24 h of storage of M0, thereafter, they increased slightly or remained steady (Fig. 2). The level of volatile compound increased slightly or remained steady during the entire storage period for M2 and M3 (Fig. 2). In summary, all lipid oxidation markers indicated that the order of oxidation development was M0 > M1 > M2 P M3 during storage. 3.3. Protein alteration during storage at 5 °C As for lipid oxidation, protein alteration progressed differently in the four products. Sulphydryl contents decreased during storage for all products, indicating formation of disulphide bonds (Fig. 3). A strong decrease in sulphydryl content was observed during 24 h of storage of M0 and 12 h of storage for M1. After 96 h of storage, sulphydryl level had decreased by 37% and 50% from total sulphydryl measured at T0 for M0 and M1, respectively (Fig. 3). A slight decrease in sulphydryl group was observed during the entire storage period for M2 whereas a more pronounced decrease was obtained between 12 and 24 h of storage for M3 (Fig. 3). At T0, the protein carbonyl content was lowest in M0 (2 nmol/ mg), but protein oxidation developed very rapidly during storage at 5 °C and the carbonyl group level reached their maximum (8 nmol/mg) after 12 h of storage (Fig. 4). For the washed minces, a high carbonyl group level (4.2–7.2 nmol/mg) was detected at T0, and it did not increase significantly thereafter, except for M2 (Fig. 4). SDS–PAGE of the minces at T0 showed three heavy bands with molecular weight around 190 kDa, 45 kDa and 37 kDa (Fig. 5a). These bands correspond to the myosin heavy chain, the actin and probably the tropomyosin, respectively. At T0, the M0 lanes showed more bands compared to the other products, indicating removal of some proteins during processing. After 96 h of storage at 5 °C, several bands, between 100 kDa and 50 kDa, and around 20 kDa, disappeared or fainted and the intensities of the myosin

b 1-penten-3-ol

1600

200

Concentration (ng/g product)

1400

Concentration (ng/g product)

2,4-heptadienal

220

1200 1000 800 600 400

180 160 140 120 100 80 60 40

200 20 0

0 0

12

24

36

48

60

Storage time (hours)

72

84

96

0

12

24

36

48

60

72

84

96

Storage time (hours)

Fig. 2. (a) Formation of 1-penten-3-ol (a) and (b) of 2,4-heptadienal during storage at 5 °C of the horse mackerel mince M0 (e), M1 (D), M2 (s) and M3 (h). Averaged values of triplicates; the error bars show the standard deviation.

62

S. Eymard et al. / Food Chemistry 114 (2009) 57–65

M0

M1

M2

M3

SH content (micromoles/g product)

14 12 10 8 6 4 2 0 0

12

24

36

48

60

72

84

96

Time (hours) Fig. 3. Progress in sulphydryl (SH) contents during storage at 5 °C of the horse mackerel mince (M0), (M1), (M2) and (M3). Averaged values of three replicates; the error bars show the standard deviation.

T0

Carbonyls (nmol/mg protein)

12

12 h

96 h

10 8 6 4 2 0 M0

M1

M2

M3

Fig. 4. Progress in carbonyl contents during storage at 5 °C of the horse mackerel mince (M0), (M1), (M2) and (M3). Averaged values of three replicates; the error bars show the standard deviation.

and the actin bands were lower when compared with lanes loaded with the same minces at T0 (Fig. 5a). In order to determine which protein oxidised during processing and storage a Western blot against protein carbonyl groups was made. It revealed that at T0 high molecular weight proteins (above 200 kDa), myosin (200 kDa) and a protein band with a molecular weight of approximately 100 kD were oxidised (Fig. 5b). Some low molecular weight proteins were also oxidised in M2 and M3, but to a much lower extent. After 96 h of storage at 5 °C, low molecular weight proteins were oxidised to a greater extent when compared to T0. For M0, almost all the proteins seemed oxidised after storage, especially the actin and tropomyosin. These proteins were also oxidised in M1 but to a lower extent (Fig. 5b). Like for lipid oxidation, all protein oxidation markers indicated protein oxidation development during storage of the minces at 5 °C. The development of protein oxidation during storage of the minces could be ranked as M0 > M1 > M2 P M3 with M0 presenting the highest protein oxidation development. 4. Discussion 4.1. Effect of processing The washing process resulted in four products with qualitative and quantitative differences in their lipid and protein fractions

Fig. 5. (a) SDS–PAGE and (b) Western blot against protein carbonyls groups of horse mackerel minces at T0 and after 96 h of storage at 5 °C. With lanes 1, 10 and 11, molecular weight marker; lane 2, M0; lane 3, M1; lane 4, M2 and lane 5, M3 at T0 and after 96 h of storage at 5 °C for lane 6, M0; lane 7, M1; lane 8, M2 and lane 9, M3. For the blot lane 12, carbonylated BSA standard.

including their pro- and antioxidant status. During processing blood, soluble proteins and lipids were to some extent removed. The neutral lipids, which are storage lipids, were more easily removed during processing than the membrane phospholipids, which is in accordance with previous studies (Eymard et al., 2005). Consequently, the phospholipids, rich in omega-3 PUFA, concentrated in the end product (Table 1) giving a mince with good nutritional properties. However, phospholipids are very sensitive to oxidation, which increased significantly during processing of fatty fish minces especially after the second wash (Table 1). Nonetheless, the content of lipid and the development of lipid oxidation were very similar for the last two products (M2 and M3). Three washes are commonly used for fatty fish processing, but the process has been reported to be effective with only two washes when raw materials have low lipid contents (Toyada, Kimura, Fujima, Noguchi, & Lee, 1992) as also observed in the present study. Others have observed that the last washing step resulted in the removal of lipid oxidation products giving a final product with a lower PV and also TBARS (Eymard et al., 2005). This was not clearly observed here, perhaps because the resultant oxidation products were

S. Eymard et al. / Food Chemistry 114 (2009) 57–65

immediately removed by the washing process, giving similar products for M2 and M3. As observed for the lipid fraction, the protein fraction was also modified during processing. Soluble proteins were removed and myofibrillar proteins concentrated in the end products (M3). Protein oxidation occurred during processing as observed by the significant increase in carbonyl content. Immuno-blot against protein carbonyl groups showed that myosin and other high molecular weight proteins (150 kDa and above) were already oxidised at T0 in all four products including the unwashed mince. Similar results were also obtained in others studies conducted on fatty fish, which showed that a significant proportion of high molecular weight proteins were already oxidised early post mortem (Kjaersgard, Norrelykke, Baron, & Jessen, 2006). The entire fillets, including both the red and the white muscles, were processed. The high proportion of red muscle and consequently the high level of haem proteins (haemoglobin and myoglobin) in the fillet explained the high level of iron in M0. However, iron loss was most pronounced after the first washing and was likely due to removal of haemoglobin and myoglobin. This finding is in accordance with results obtained after washing of sardine and mackerel (Chaijan, Benjakul, Visessanguan, & Faustman, 2004). Some iron remained in the mince (M3) after the third wash, suggesting that haeme proteins were not completely removed during processing. This may be due to the fact that the haem, which is very hydrophobic, are incorporated into membranes (Richards, Dettmann, & Grunwald, 2005). In addition, proteins possess some iron chelating properties as reported by Villiere, Viau, Bronnec, Moreau, and Genot (2005) and therefore iron itself could be chelated by muscle proteins giving a residual iron level in the washed minces. As iron and haem proteins are strong prooxidants in muscle foods their removal or inactivation is a crucial step for the oxidative stability of fatty fish mince. In vivo, the cells are protected against oxidation by several antioxidants. These natural antioxidant, such as tocopherols, which are present in the raw material (M0), were removed or/and degraded during washing as also found by Sannaveerappa, Carlsson, Sandberg, and Undeland (2007). Therefore, it is clear that processing such as washing creates an imbalance in the initial prooxidant-antioxidant equilibrium in the muscle tissue and may therefore contribute to the differences in the oxidative status of the four products obtained at T0. Hence, in order to limit oxidation development and to obtain high quality washed fatty fish mince it was previously suggested that it is important to start with good quality raw material, to control temperature and oxygen incorporation during processing (Eymard et al., 2005), but also to evaluate the impact of each washing step on the oxidative stability of the product. Our results support the importance of these guidelines. Moreover, the results also indicate that incorporation of antioxidant in the washing water for processing fatty fish might be a possible strategy to prevent oxidation (Bakir, Hultin, & Kelleher, 1994) and that this topic deserves further investigation. 4.2. Lipid oxidation during storage The development of lipid oxidation during storage seemed to depend on the composition of the mince matrix and its initial oxidative status. The results suggest that the mince with the highest lipid, prooxidant and antioxidant contents was linked to the highest rate of lipid oxidation. However, Larsson, Almgren, and Undeland (2007) indicated that lipid content did not determine the haemoglobin initiated lipid oxidation in washed fish minces. Lipid content has probably an impact on oxidation as lipid is the substrate for the reaction, but other parameters such as the nature and composition of lipid, the type and concentration of proteins, antioxidants and prooxidants are also important factors.

63

The level of lipid hydroperoxides in M0 was low compared to the washed minces, indicating that lipid oxidation had not been initiated in this product at T0. This could be due to the protective effect of tocopherols, initially present in M0, which decreased rapidly to 0.1 mg/kg mince after 12 h of storage. In contrast, oxidation had progressed significantly at T0 in the washed minces revealing that processing/washing promote lipid oxidation more than just mincing. During storage the rate of lipid oxidation was higher in M0 compared to the washed minces. It can be speculated that for M2 and M3 lipid oxidation was so advanced at T0 that the steady state – where formation and decomposition of oxidation products as well as interactions of oxidation compounds with other components are equivalent – was reached faster. For all the minces, the different lipid oxidation parameters including sensory analysis generally showed the same pattern during storage and the delay between the development of primary and secondary lipid oxidation products was not observed. This delay may have been observed with more sampling points at the beginning of the storage period (first 12 h storage). Others have also reported that the first 24 h were crucial in the development of lipid oxidation during storage of washed fish muscle (Larsson et al., 2007). All the parameters studied have an impact on lipid oxidation, and it is not possible to conclude from this study if it is the initial level of oxidation, the lipid content and composition, or the concentration of pro- and antioxidants that determined the rate of lipid oxidation in the different minces. Mathematical modelling of the effect of the different parameters could be a way to further understand the different mechanisms of oxidation in washed mice and the rate of the different reactions involved. Such modelling has been performed for simple systems such as liposomes (Mozuraityte, Rustad, & Storrø, 2006) but is urgently needed for more complex systems such as washed mince products. 4.3. Protein oxidation during storage Protein oxidation developed rapidly during storage of M0 whereas it remained steady during storage of the washed minces. Other studies have showed no increase in protein carbonyl groups during frozen or ice storage of rainbow trout (Baron, Kjrsgård, Jessen, & Jacobsen, 2007) while others reported an increase in carbonyl groups during ice or frozen storage (Passi, Cataudella, Tiano, & Littaru, 2005). Since the haem protein content in M0 is high and since haeme protein is known to also induce protein oxidation it is possible that the increase in protein carbonyl during storage could be catalysed by haem proteins. In salted herring, Hb has been postulated to be responsible for protein cross-linking (Andersen, Andersen, & Baron, 2007). The decrease in protein solubility observed already at T0 and the observed aggregate on the blot at high molecular weight are likely to be the result of protein cross-linking. The decrease in free thiol groups was very pronounced in the first 24 h for M0, whereas it was slower during the entire storage period for the washed products. These findings suggest that protein aggregation had already happened during processing, which is also in agreement with a high protein carbonyl level already at T0. The immuno-blot indicated that high molecular weight proteins were already oxidised in all products at T0, whereas low molecular weight proteins oxidised during storage. As suggested by several authors (Kjaersgard et al., 2006; Stagsted, Bendixen, & Andersen, 2004) myosin appears to be the first target of oxidation while low molecular weight protein seemed to oxidise later. It has been reported that protein oxidation is highly dependent on the structure of the protein and peptides and that random coiled protein are more susceptible to oxidation than globular proteins (Dalsgaard, Otzen, Nielsen, & Larsen, 2007). It is therefore possible

64

S. Eymard et al. / Food Chemistry 114 (2009) 57–65

that globular proteins, due to their conformation were more protected against oxidation and consequently were oxidised later during storage as showed on the immuno-blot for the mince after 96 h. According to the SDS–PAGE pattern, the different minces had the same protein composition, but according to the immuno-blot, their susceptibility to oxidation differed. It seems that protein oxidation is not only depending on the nature of the protein, but also on its location, conformation, accessibility, and exposure to the prooxidants in the matrix. The more the mince was processed the less the protein seemed to be oxidised on the blot. The conformational properties of the proteins might have changed during processing and resulted in the exposure or concealment of sensitive residues leaving them more or less accessible to oxidation during storage depending on the minces. As indicated by the loss in protein solubility at T0, some proteins might also have interacted during processing, making them less sensitive to oxidation during storage. The last two washed minces (M2 and M3) had the highest carbonyl contents at T0, but the immuno-blot showed that fewer proteins were oxidised after 96 h of storage compared to M0 and M1. However, immuno-blots are not quantitative, and it is possible that fewer proteins were oxidised, but that they were oxidised more heavily. In addition, carbonylated proteins might have reacted further with other components or be degraded, whereby they cannot be revealed by the immuno-blot. The carbonyl level of the two last products (M2 and M3) did not increase during storage and, some authors have also showed a decrease in carbonyl during prolonged storage at 20 °C, suggesting interactions between protein carbonyl and other cell constituents (Baron et al., 2007). A study performed on trout showed disappearance of oxidised protein on the immuno-blot with prolonged incubation (Baron, Pham, & Jacobsen, 2006). Therefore, protein carbonyl formation should be seen as one step in the protein oxidation processes, but cannot always be used solely as a stable marker of protein oxidation. It is clear that a better understanding of protein oxidation mechanisms and the development of more sensitive methods will reveal to which extent protein oxidation impact fish and fish product quality both during processing and storage. 4.4. Lipid and protein interactions Lipid and protein oxidation developed simultaneously in the different minces and it seemed that the reactions were linked, but it was difficult to determine at which level. The two reactions have the same catalysts, they can develop independently or in parallel, but they can also interact with each other. It is well known that radicals, hydroperoxides and secondary compounds resulting from lipid oxidation react with proteins leading to the loss of protein functionality (Gardner, 1979). Jarenback and Liljemark (1975) demonstrated that linoleic acid hydroperoxides can interact with cod myofibrillar proteins inducing denaturation and precipitation of the myosin. Aldehydes groups formed during lipid oxidation can interact with amino groups of proteins to form Schiff base products (Chio & Tappel, 1969). In contrast, the effect of protein oxidation products on lipid is not documented. Østdal, Davies, and Andersen (2002), showed in model system that protein radicals could attack free fatty acids, which resulted in an increased level of primary oxidation products. In more complex system such reactions have not been investigated. Lipid and protein oxidation have the same catalysts and are interlinked, which makes them difficult to distinguish from each other in complex matrices. However, it is suggested that the type of modification on the protein or the lipid depends on the nature of the radical attacking the target (Davies & Dean, 1997). Therefore, it can be speculated that more research will lead to the development of very specific markers of

protein and lipid oxidation at the sub-molecular leading to a better understanding of the interaction between protein and lipid during oxidation of fish and fish products. Acknowledgements The authors would like to thank Lis Berner and Inge Holmberg for excellent technical assistance, Isabel Medina (IIM/CSIC in Spain) for supplying the horse mackerel and the expert sensory panel at DTU Aqua for performing the sensory analysis. We would also like to thank Grethe Hyldig for fruitful discussions on the sensory analysis. This research was supported by a Marie Curie Intra-European Fellowships within the 6th European Community Framework programme. References Andersen, E., Andersen, M. L., & Baron, C. P. (2007). Characterization of oxidative changes in salted herring (Clupea harengus) during ripening. Journal of Agricultural and Food Chemistry, 55, 9545–9553. AOAC (1995). Official method of analysis (16th ed.). Arlington, VA, USA: AOAC. AOCS (1998). Official methods, and recommended practices (5th ed.). USA: AOCS Champaign IL. Bakir, H. M., Hultin, H. O., & Kelleher, S. D. (1994). Gelation properties of fatty fish processed with or without added sodium-chloride, cryoprotectants and antioxidants. Food Research International, 27, 443–449. Baron, C.P., Pham, K.P., & Jacobsen, C. (2006). Oxidation of sarcoplasmic proteins from rainbow trout feed different diet. In Proceedings of Institute of Food Technologist, IFT 2006, Orlando, Florida, USA. Baron, C. P., Kjrsgård, I. V. H., Jessen, F., & Jacobsen, C. (2007). Protein and lipid oxidation during frozen storage of rainbow trout (Oncorhynchus mykiss). Journal of Agricultural and Food Chemistry, 55, 8118–8125. Bligh, E. G., & Dyer, W. J. (1959). A rapid method of total lipid extraction and purification. Canadian Journal of Biochemistry and Physiology, 37, 911–917. Chaijan, M., Benjakul, S., Visessanguan, W., & Faustman, C. (2004). Characteristics and gel properties of muscles from sardine (Sardinelle gibbosa) and mackerel (Rastrelliger kanagurta) caught in Thailand. Food Research International, 37, 1021–1030. Chio, K. S., & Tappel, A. L. (1969). Synthesis and characterization of the fluorescent product derived from malonaldehyde and amino acids. Biochemistry, 8, 2821–2826. Dalsgaard, T. K., Otzen, D., Nielsen, J. H., & Larsen, L. B. (2007). Changes in structures of milk proteins upon photo-oxidation. Journal of Agricultural and Food Chemistry, 55, 10968–10976. Davies, M. J., & Dean, R. T. (1997). Radical-mediated protein oxidation. New York: Oxford University Press Inc. Ellman, G. L. (1959). Tissue Sulfhydryl Groups. Archives of Biochemistry and Biophysics, 82, 70–77. Eymard, S., Carcouët, E., Rochet, M. J., Dumay, J., Chopin, C., & Genot, C. (2005). All stages are critical for development of lipid oxidation during manufacturing of horse mackerel surimi. Journal of the Science of Food and Agriculture, 85, 1750–1756. Frankel, E. N. (2005). Lipid oxidation. Bridgewater, England: P.J. Barnes and Associates. Gardner, H. W. (1979). Lipid hydroperoxide reactivity with proteins and amino acids: A review. Journal of Agricultural and Food Chemistry, 43, 651–656. Hultin, H. O., & Kelleher, S. D. (2000). Surimi processing from dark muscle fish. In J. W. Park (Ed.), Surimi and surimi seafood (pp. 59–78). New York: Marcel Dekker, Inc. Jarenback, L., & Liljemark, A. (1975). Ultrastructural changes during frozen storage of cod. III: Effects of linoleic acid and linolenic acid hydroperoxydes on myofibrillar proteins. Journal of Food Technology, 10, 437–452. Kjaersgard, I. V. H., Norrelykke, M. R., Baron, C. P., & Jessen, F. (2006). Identification of carbonylated protein in frozen rainbow trout (Oncorhynchus mykiss) fillets and development of protein oxidation during frozen storage. Journal of Agricultural and Food Chemistry, 54, 9437–9446. Larsson, K., Almgren, A., & Undeland, I. (2007). Hemoglobin-mediated lipid oxidation and compositional characteristics of washed fish mince model systems made from cod (Gadus morhua), herring (Clupea harengus), and salmon (Salmo salar) muscle. Journal of Agricultural and Food Chemistry, 55, 9027–9035. Levine, R. L., Williams, J. A., Stadtman, E. R., & Shacter, E. (1994). Carbonyl assays for determination of oxidatively modified proteins. Method in Enzymology, 233, 346–357. Mozuraityte, R., Rustad, T., & Storrø, I. (2006). Pro-oxidant activity of Fe2+ in oxidation of cod phospholipids in liposomes. European Journal of Lipid Science and Technology, 108, 218–226. Nordic Committee on Food Analysis (1991). Metals, determination by atomic absorption. Spectrophotometry in foodstuffs. Method No. 139. Oslo, Norway. Okada, M. (1992). History of surimi technology in Japan. In T. C. Lanier & C. M. Lee (Eds.), Surimi technology (pp. 3–22). New York: Marcel Dekker, Inc.

S. Eymard et al. / Food Chemistry 114 (2009) 57–65 Østdal, H., Davies, M. J., & Andersen, H. J. (2002). Reaction between protein radicals and other biomolecules. Free Radicals in Biology & Medicine, 33, 201–209. Passi, S., Cataudella, S., Tiano, L., & Littaru, G. P. (2005). Dynamics of lipid oxidation and antioxidant depletion in Mediterranean fish stored at different temperatures. BioFactors, 25, 241–254. Pokorny, J., Elzeany, B. A., Luan, N., & Janicek, G. (1976). Nonenzymatic browning. 15. Effect of unsaturation on browning reaction of oxidized lipids with protein. Zeitschrift fur Lebensmittel-untersuchung Und-forschung, 161, 271–272. Refsgaard, H. H. F., Haahr, A. M., & Jensen, B. (1999). Isolation and quantification of volatiles by dynamic headspace sampling and mass spectrometry. Journal of Agriculture and Food Chemistry, 47, 1114–1118. Richards, M. P., Dettmann, M. A., & Grunwald, E. (2005). Pro-oxidative characteristics of trout hemoglobin and myoglobin: A role for released heme in oxidation of lipids. Journal of Agriculture and Food Chemistry, 53, 10231–10238. Sannaveerappa, T., Carlsson, N.-G., Sandberg, A.-S., & Undeland, I. (2007). Antioxidative properties of press juice from herring (Clupea harengus) against haemoglobin (Hb) mediated oxidation of washed cod mince. Journal of Agricultural and Food Chemistry, 55, 9581–9591. Shahidi, F., & Miraliakkbari, H. (2004). Omega-3 (n-3) fatty acids in health and disease: Part 1 – Cardiovascular disease and cancer. Journal of Medicinal Food, 7, 387–401.

65

Shahidi, F., & Miraliakkbari, H. (2005). Omega-3 (n-3) fatty acids in health and disease: Part 2 – Health effects of omega-3 fatty acids in autoimmune diseases, mental health, and gene expression. Journal of Medicinal Food, 8, 133–148. Shantha, N. C., & Decker, E. A. (1994). Rapid, sensitive, iron-based spectrophotometric methods for determination of peroxide values of food lipids. Journal of the Association of Official Analytical Chemists, 77, 421–424. Stagsted, J., Bendixen, E., & Andersen, H. J. (2004). Identification of specific oxidatively modified proteins in chicken muscle using a combined immunologic and proteomic approach. Journal of Agricultural and Food Chemistry, 52, 3967–3974. Stewart, J. C. (1980). Colorimetric determination of phospholipids with ammonium ferrothiocyanate. Analytical Biochemistry, 104, 10–14. Toyada, K., Kimura, I., Fujima, T., Noguchi, S. F., & Lee, C. M. (1992). The surimi manufacturing process. In T. C. Lanier & C. M. Lee (Eds.), Surimi technology (pp. 79–112). New York: Marcel Dekker Inc. Villiere, A., Viau, M., Bronnec, I., Moreau, N., & Genot, C. (2005). Oxidative stability of bovine serum albumin- and sodium caseinate-stabilized emulsions depends on metal availability. Journal of Agricultural and Food Chemistry, 53, 1514–1520. Xiong, Y. L. (2000). Protein oxidation and implications for muscle food quality. In E. Decker, C. Faustman, & C. J. Lopez-Bote (Eds.), Antioxidants in Muscle Foods (pp. 85–112). New York: John Wiley & Sons, Inc.