Effects of cooking techniques on fatty acid and oxylipin content of farmed rainbow trout (Oncorhynchus mykiss)

Abstract The aim of this study was to investigate the effect of various cooking techniques on the fatty acid and oxylipin content of farmed rainbow trout. Rainbow trout is an excellent source of long‐chain omega‐3 (n‐3) polyunsaturated fatty acids (PUFA) which have beneficial health effects. Fillets of 2‐year‐old farmed rainbow trout were baked, broiled, microwaved, or pan‐fried in corn (CO), canola (CaO), peanut (PO), or high oleic sunflower oil (HOSO). Fatty acids and oxidized lipids were extracted from these samples and their respective raw fillet samples. Fatty acid content was determined using gas chromatography and oxylipin content by mass spectroscopy. The values obtained from each cooking method were compared to those obtained from the respective raw fillets using paired t tests. PUFA content was not altered when samples were baked, broiled, microwaved, or pan‐fried in CO or CaO. Pan‐frying in PO reduced α‐linolenic acid (18:3n‐3), eicosadienoic acid (20:2n‐6), and dihomo‐γ‐linolenic acid (20:3n‐6), while pan‐frying in HOSO reduced 18:3n‐3, eicosapentaenoic acid (20:5n‐3), docosapentaenoic acid (22:5n‐3), docosahexaenoic acid (22:6n‐3), linoleic acid (18:2n‐6), 18:3n‐6, 20:2n‐6, 20:3n‐6, docosatrienoic acid (22:2n‐6), and adrenic acid (22:4n‐6) compared to raw fish. Cooking decreased the omega‐6 (n‐6) PUFA‐derived oxylipins, but caused no change in 20:5n‐3 or 22:6n‐3‐derived oxylipins of the fillets. In conclusion, pan‐frying was the only cooking method to alter the fatty acid content of the fillets, while observed changes in oxylipin content varied by cooking method. As the physiological impact of oxylipins is currently unknown, these results suggest that the cooking methods which optimize the consumption of n‐3 PUFA from rainbow trout are baking, broiling, microwaving, or pan‐frying in CO, CaO, or PO.

Oxylipins can be used as a measurement of lipid oxidation, indicating which fatty acids are most impacted by heat. Oxylipins are oxidized products of PUFA (Gabbs et al., 2015). Omega-6 (n-6)-derived oxylipins tend to have increased inflammatory and vasoconstrictive effects, while n-3-derived oxylipins tend to be anti-inflammatory (Calder, 2015;Gabbs et al., 2015;Ray et al., 2015). It is unknown how the method of preparation affects the oxylipin content of fish fillets and the potential health impact of those oxylipins. Answering these questions will determine if there is a superior method of preparation that maintains the beneficial lipids found in fish.
Farmed rainbow trout contain a high content of LCn-3 PUFA relative to other common food fish species such as tilapia and catfish (Weaver et al., 2008). Research on the impact of various cooking methods has been conducted on rainbow trout (Agren & Hanninen, 1993;Asghari et al., 2013;Gokoglu, Yerlikaya, & Cengiz, 2004;Tokur, 2007). However, the effect of pan-frying with various oils, microwaving, baking, and broiling on both the fatty acid and oxylipin profile of rainbow trout has not been investigated. Therefore, the aim of this work was to determine the effects cooking methods have on both the fatty acid and oxylipin content of farmed rainbow trout.

| Preparation of fillets
Farmed rainbow trout (Oncorhynchus mykiss) were hatched and reared at the USDA, ARS, National Center for Cool and Cold Water Aquaculture in Kearneysville, WV. The fish were raised in partially recirculated water (12-13°C) in 3 m 3 tanks and fed daily with a commercially available diet (Finfish G, Zeigler Bros, Inc., Gardners, PA) through automatic feeders. The feed was dispensed at or just below satiation levels and adjusted over time. After 2 years the trout were harvested in water containing a lethal dose of tricaine methanesulfonate (300 mg/L) (Western Chemical, Ferndale Washington). The fish were hand filleted immediately upon death and fillets were stored at −80°C until they were shipped on dry ice to the USDA, ARS Grand Forks Human Nutrition Research Center, Grand Forks, ND. Once the fillets were received, they were stored at −80°C until processing.
Fish fillets were moved from the −80°C freezer to a −20°C freezer.
After 3 days, the fillets were moved from the −20°C freezer into a 4°C refrigerator. The following day the fillets were removed from the refrigerator and were prepared. Each fillet was weighed and then cut into 55 g pieces longitudinally (dorsal-ventral) along the fillet. A 5 g piece, to serve as the raw sample, was cut from each of the 55 g fillet samples (from either the dorsal or ventral side), stored in a zip-lock bag, and placed into the −80°C freezer until processing. The remaining 50 g pieces were then stored in the 4°C refrigerator until cooked.

| Cooking of fish
Cooking techniques were based on previously published methods (Al-Saghir et al., 2004;Asghari et al., 2013;Raatz et al., 2011). For each cooking method, samples were cooked in triplicate. As soon as the desired temperature was attained by all cooking methods, the samples were cooled and frozen at −80°C until analyzed. frying pan (National Presto Industries, Inc., Eau Claire, WI). The frying pan was set to 177°C (350°F). When this temperature was reached, the 50 g sample, with a thermometer placed at the center of the fillet, was added to the pan in the center of the oil. The fillet was flipped over to the other side midway through cooking and cooked until the core temperature attained 63°C (145°F) for 15 s (The National Restaurant Association, 2010).

| Baking
A conventional oven was preheated to 177°C (350°F). A 50 g sample, with a thermometer placed at the center of the fillet, was placed in a 5.5″ × 4.5″ (14 cm × 11.4 cm) bake and serve container (Pactive Pressware, Columbus, OH) and baked until the core temperature reached 63°C (145°F) for 15 s (The National Restaurant Association, 2010).

| Oven broiling
An oven rack was placed 4.5″ (11.4 cm) away from the heating source.
The oven was then preheated to 260°C (500°F). A 50 g sample, with a thermometer placed at the center of the fillet, was placed in an 8″ × 8″ (20 cm × 20 cm) Pyrex glass container which had 3 g of CO spread on the bottom. The sample was placed into the oven, 4.5″ (11.4 cm) away from the heat source, and was cooked until the core temperature reached 63°C (145°F) for 15 s (The National Restaurant Association, 2010). The fillet was flipped over to the other side midway through cooking.

| Microwaving
A 50 g sample was placed in a 5.5″ × 4.5″ (14 cm × 11.4 cm) bake and serve container (Pactiv Pressware, Columbus, OH). Wax paper was used to cover the container. The sample was then placed into a 1,200-W high-power microwave (Panasonic, lot # NN-SA661S). The sample was cooked in short intervals until the core temperature reached 63°C (145°F) for 15 s (The National Restaurant Association, 2010).

| Lipid extraction
Frozen fillet samples were pulverized in liquid nitrogen. Lipids were extracted using a modified Folch method (Folch, Lees, & Sloane Stanley, 1957). Briefly, 50 mg of pulverized fish were weighed into a 6-ml test tube, combined with 2.5 ml of chloroform (Sigma-Aldrich, St.  to 225°C at a rate of 8°C/min, and held at 225°C for 4 min once the temperature was reached (Masood, Stark, & Salem, 2005). The acquired data were analyzed using a Dionex Chromeleon 7.2 Chromatography data system (Thermo Fisher Scientific, Waltham, MA). Analysis of the oils used for pan-frying followed the same gas chromatography method. Table 1 presents the analyzed fatty acid content of the oils used.

| Extraction of lipid oxidation products
Oxidized lipids were extracted from tissue as described previously (Brose, Thuen, & Golovko, 2011;Golovko & Murphy, 2008;Raatz et al., 2011). Briefly, tissue was pulverized under liquid nitrogen into a homogenous powder. The pulverized samples (~50 mg) were sonicated for two cycles, 7 s each with a power output of 50J (  Data expressed as g/100 g; n = 1. 5(S)-hydroxy-eicosatetraenoc acid (HETE)-d 8 as internal standards. To release esterified prostanoids and fatty acid monohydroxides (MHFA) from phospholipids, the samples were incubated for 1 hr at room temperature with soluble phospholipaseA 2 (sPLA 2 ; ~0.9 μmole/min of total activity, Cayman Chemical Co, Ann Arbor, MI). Prostanoids and MHFA were extracted with acetone liquid/liquid extraction by adding 2 ml acetone and 800 μl saline (0.9% NaCl). The samples were centrifuged (2,000g; 10 min) and the supernatant was transferred to a new tube. MHFA were extracted from the supernatant using 3 × 2 ml hexane. Following the MHFA extraction, prostanoids were extracted from the same supernatant by acidification of supernatant with formic acid to pH = 3.5 (30 μl of 2 mol/L formic acid), and extraction with 2 ml of chloroform. The chloroform extract containing prostanoids was transferred to glass screw top tubes which were previously si-

| UPLC separation and MS analysis
Separation for both the hexane fraction and the chloroform fraction were 39% solvent B. The gradient was slightly modified from a previously described method to improve separation of MHFA (Brose, Baker, & Golovko, 2013). The initial conditions were held for 0.5 min, solvent B was increased to 40.5% over 6.88 min, then increased to 70% over 1.62 min, further increased to 75% over 3 min, and finally increased to 98% over 1.5 min. Solvent B was held at 98% for 5.3 min. Solvent B was then returned to the initial conditions over 0.2 min and held for 2 min.
For MS/MS analysis, a triple quadrupole mass spectrometer (Xevo TQ-S, Waters) with electrospray ionization operated in negative ion mode was used. The capillary voltage was 0.71 kV and the cone voltage was 30 V. The desolvation temperature was 350°C and the source temperature was 150°C. The desolvation gas flow was 1,000 L/hr, the cone gas flow was 150 L/hr, and the nebulizer gas was at 5.0 Bar.
MassLynx V4.1 software (Waters) was used for instrument control, acquisition, and sample analysis.
The analytes were monitored in MRM mode using the mass transitions and collision energies presented in Table 2. Prostanoids were quantified using PGE 2 -d 9 as an internal standard, while MHFA were quantified using 15-S-HETE-d 8 as the internal standard.

| Statistical analysis
Data (normalized to 100 g of fillet) are reported as mean ± standard deviation. Paired t tests were used to compare the fatty acid content of the raw samples to the corresponding cooked samples. p values ≤.05 were considered statistically significant. All analyses were done using SAS version 9.4 (SAS Institute, Inc., Cary, NC).

| Saturated fatty acids
SFA content was not altered when the samples were pan-fried in CaO.

| Oxylipin content
The effect of cooking on total oxylipin content of the fillets was dependent on both the cooking method and oxylipin under examination ( Figure 1) Data expressed as mg/100 g fish (wet weight) and represented as mean ± SD, n = 3. *p ≤ .05.
T A B L E 3 (Continued) oxylipins were reduced with different cooking methods (Table 4).
Microwaving and baking decreased the ARA-derived oxylipin prostaglandin F 2α (PGF 2α ), while oven broiling resulted in a decrease of PGE 2 and PGF 2α . When the fillets were fried in CaO, there was no significant difference in oxylipin content relative to the raw fillets. Pan-frying in CO resulted in a decrease of PGE 2 and PGF 2α . When pan-fried in PO, PGF 2α , 15-keto-PGE 2 , and 13-HODE. Frying in HOSO resulted in a decrease of oxylipins 11-HETE, 9-HODE, and 13-HODE.

| DISCUSSION
We hypothesized that cooking method would have no effect on the fatty acid content of the fillets. From this study, we found that panfrying was the only cooking method that impacted the fatty acid content. Previous studies evaluating the effect of cooking methods on the fatty acid profile of various fish species have also found that panfrying has the largest impact (Sioen et al., 2006). We observed that pan-frying in HOSO significantly decreased total n-3 PUFA, total n-6 PUFA, EPA, DHA, and total SFA levels. These results were similarly observed in a study conducted by Ansorena et al.
(2010) on salmon which showed that pan-frying in SO significantly changed the lipid profile even though the total fat content was unchanged. They found the amount of total SFA, MUFA, and PUFA were significantly lower in the fried sample than the raw sample. In addition, the individual fatty acid content was significantly altered. There were slight, but significant, decreases in the levels of EPA, DHA, 16:0, and 18:3n-3, which our study also observed.
Pan-frying in CO resulted in minimal changes in the fatty acid profile. Another study observed similar results with no changes in SFA, MUFA, PUFA, EPA, and DHA levels when salmon was fried in CO (Al-Saghir et al., 2004). These observations likely resulted due to the composition of the fish fillet and of the oil. As illustrated in Table 1, CO has a general FA distribution of PUFA > MUFA > SFA with the predominant FA being 18:2n-6 (Hosseini, Ghorbani, Meshginfar, & Mahoonak, 2016). Both farmed rainbow trout and farmed salmon contain a high level of PUFA (Weaver et al., 2008). Therefore, there would likely be minimal movement of fatty acids. Our study, like previous studies, showed no significant change in the fatty acid composition of the fillet when pan-fried in CO.
When pan-fried in CaO no differences were found in the fatty acid profile of the fried compared to the raw fillet. Agren and Hanninen (1993) found that pan-frying rainbow trout in CaO resulted in small F I G U R E 1 Changes in oxylipin content of fillets cooked by various methods compared to raw fillets. Data expressed as ng/g of wet weight (gww) and represented as mean ± SD. *p ≤ .05. Black columns represent raw samples, gray columns represent cooked samples. Ba, baking; Br, broiling; FCaO, pan-fried in canola oil; FCO, pan-fried in corn oil; FPO, pan-fried in peanut oil; FHOSO, pan-fried in high oleic sunflower oil; Mi, microwaving; PO, peanut oil increases of 18:1n-9, 18:2n-6, and 18:3n-3. These results differ from ours because the skin of the trout was kept on during the cooking process in the other study which is thought to prevent the transfer of lipids between the fillet and the culinary oil, but still allowed the loss of moisture (Agren & Hanninen, 1993).
To our knowledge, this is the first study to observe the change in fatty acid content when pan-frying fish in PO. As illustrated in Table 1, the predominant fatty acids in PO are 18:1n-9 and 18:2n-6 (Hosseini et al., 2016). The movement of fatty acids from the fillet to the oil would favor the movement of PUFA. The significant decrease of 18:3n-3 in the fillet was likely due to the movement of this fatty acid to the culinary oil.
We hypothesized that cooking would increase lipid oxidation and thus the content of oxylipins due to the exposure to high temperatures. We observed decreases in the content of some oxylipins of cooked versus raw fillets, depending on the cooking technique. Our previous work found similar results after baking salmon (Raatz et al., 2011). In addition, other studies found that pan-frying had minimal thermal oxidation of the lipids in the fillet (Al-Saghir et al., 2004;Nieva-Echevarría et al., 2016). The decrease in oxylipins content after cooking therefore was likely due to decrease in fatty acid and preexisting oxylipins content, as the fillets were minimally exposed to thermal oxidation.
Baking, oven broiling, microwaving, and pan-frying in CO decreased the prostanoid content, the likely result of decomposition of the prostanoids. As oxylipins are derived from fatty acids, the observation of no change in oxylipin content when pan-frying in CaO reflects the unaltered fatty acid content. Pan-frying in PO resulted in a decrease in prostanoids and 13-HODE. This was likely due to decomposition of the prostanoids and perhaps as a result of decreased n-6 PUFA in the fillet.
The change in fatty acid content when pan-frying in HOSO complements the changes in oxylipin content. Total HODE (9-HODE, 13-HODE) levels decreased likely due to the decrease in 18:2n-6 in the fillet as these oxylipins are derived from this fatty acid. The decrease in 20:4n-6 likely explains the observed decrease in 11-HETE. The other eicosanoids decreased, but not significantly due to the large standard deviations.
This study provides new insight into how the fatty acid and oxylipin composition of rainbow trout is impacted by various cooking methods. There remain, however, a few limitations to this study. A primary limitation is the large standard deviations of our results. This can be rectified in the future by analyzing a greater number of fillet samples. Another reason for the large standard deviations could be due to real variation in the fatty acid content of the fillets used. Across fillet variation was controlled as much as possible by using fish hatched and reared in the same environment and cutting samples from similar parts of the fillets. Even so, there still could have been within fillet variation in the composition of the fish due to the location of where the samples were taken from each of the fillets (Fjellanger, Obach, & Rosenlund, 2001;Testi, Bonaldo, Gatta, & Badiani, 2006). Another limitation of this study was that the culinary oils used were not analyzed after the fish was prepared; therefore, the true movement of the fatty acids was unknown. Previous research has shown that there is an interaction between the fatty acids of the fillet and the culinary fat used (Nieva-Echevarría et al., 2016;Sioen et al., 2006). Even with these limitations, this study is novel because it not only observed the changes in fatty acid composition of farmed rainbow trout, but also the oxylipin content.
Future research is needed with a larger sample size in order to confirm the results of this study. In addition, research is needed to determine the health impact of oxylipin compounds in the fillet.

| CONCLUSIONS
In summary, pan-frying was the only cooking method that resulted in a significant change in the overall fatty acid profile of the farmed rainbow trout. Changes in the oxylipin content varied by cooking method, however, the physiological impact of these changes is currently unknown. Therefore, cooking methods of rainbow trout that would optimize the consumption of n-3 PUFA are baking, broiling, microwaving, or pan-frying in CO, CaO, or PO as these cooking methods did not significantly impact the fatty acid profile of the fillets.