Quality evaluation of rapeseed oil in Chinese traditional stir-frying

2.1 Chemicals

Solvents used for extraction and analysis included 95% HPLC-grade n-hexane and methanol, which were purchased from Merck Reagents Co., Ltd.. Other common analytical grade reagents, including ethanol, diethyl ether, acetic acid, chloroform, and sodium hydroxide, were obtained from Sinopharm Chemistry Reagents Co., Ltd.. Folin's-Ciocalteu reagent (purity, >97%), 5α-cholesterol, tocopherols (α-, β-, γ-, δ-), and eight types of trans-fatty acids (TFAs) standards (C₁₄–C₂₂) were purchased from Sigma-Aldrich Chemical Co.. Brassicasterol, campesterol, β-sitosterol, and Δ⁵-avenasterolwere obtained from Blue sky Biological Engineering Co., Ltd.. A mixture of fatty acid methyl ester standards was purchased from NU-CHEK-PREP. Canolol (purity, >98%) was purchased from BOCSCI Inc.

2.2 Sample preparation

High canolol rapeseed (HCR): Two hundred kilograms of rapeseed were heated in a commercial tunnel microwave oven (2,450 Hz, 55 KW) for 7 min and the whole transfer line was 10 m under the microwave machine. The microwaved rapeseeds were pressed using twin screw oil press machine (Wuhan, China), and the obtained rapeseed oil was degummed and deacidificated using physical adsorbent in order to get qualified frying oils. Low canolol rapeseed (LCR) was purchased from the factory with the hot pressing process. High canolol rapeseed (HCR) and low canolol rapeseed (LCR) were refined using physical methods.

2.3 Stir-frying and oil extraction

The stir wok (diameter, 32 cm, Supor) was washed and wiped before the frying experiments. All ingredients and oils were matched according to the prescribed ratio. To minimize human error, the same operator performed the entire process. Before stir-frying, the oils were heated within 40 s.

2.4 Determination of 2,6-dimethoxy-4-vinylphenol (canolol) content

Rapeseed oil sample (0.5 g) was added into a 10 ml colorimetric tube, followed by the addition of 1.5 ml hexane and 1.5 ml methanol: water (ratio, 4:1). The mixture was vortexed for 5 min at 706 g and centrifuged for 10 min at  2,823 g. The supernatant was repeat extracted twice as described above. The mixed extracts were passed through a 0.45-µm filter. Three replicates were used per experiment. Canolol was quantified using UPLC analysis according to Yang et al. (2014).

2.5 Determination of oxidative index

The peroxide value, p-anisidine value, and malonaldehyde content were measured according to the conventional method ISO 3960, ISO 6885–2006, and GB 5009.181–2009 standards, respectively.

2.6 Determination of cis and trans-fatty acid composition

Methyl esterification of the cis- and trans-fatty acids was performed according to Zribi et al. (2014). Cis- and trans-fatty acid compositions were analyzed using the Agilent 7890N GC instrument coupled with a flame ionization detector (Agilent Technologies Inc.). A nonpolar fused-silica capillary column (HP-88, 100 m × 0.25 mm × 0.20 μm) was used for gas chromatography (GC) analysis. Helium was used as a carrier gas at a flow rate of 1.5 ml/min. A split ratio was set as 1:30. The temperature program was as follows: Initial temperature set at 120°C was on hold for 1 min, after which it was increased to 180°C at 20°C/min; next, the speed was set as 5°C/min, and finally, the column temperature was increased to 220°C and held for 5 min. Both the detector and injector port temperatures were maintained at 260°C. The fatty acids were identified from retention time comparison versus mixed standards run under the same conditions. They were quantified according to percentage peak area ratio. The final results were expressed as the percentages of individual fatty acids.

2.7 Determination of total phenol content

Rapeseed oil (0.5 g) was added into a 10 ml colorimetric tube, followed by the addition of 5 ml 95% methanol and 0.5 ml Folin's phenol reagent, and mixed for 3 min. One milliliter saturated Na₂CO₃ solution was added, and the mixture was diluted to a constant volume of 10 ml using distilled water. The solution was stewed and allowed to react at room temperature for 60 min. The absorbance of the sample was measured at 765 nm. The result was expressed in sinapic acid equivalent (mg/100 g oil).

2.8 Determination of tocopherols

Tocopherol content in rapeseed oils was determined according to Zhang et al. (Zhang et al., 2017). Different tocopherols were identified by comparing their retention times to those of α-, β-, γ-, and δ-tocopherol standards. The final result was expressed as mg/100 g.

2.9 Determination of phytosterols

Oil samples were saponified according to the method described by Azadmard-Damirchi, Nemati, Hesari, Ansarin, and Fathi-Achachlouei (2010). The 0.03 g oil sample (exact to 0.0001 g) was mixed thoroughly with 3 ml of 2 mol/L KOH in 95% ethanol in a glass tube and shaken in a water bath at 90°C for 15 min. After cooling the tubes, 2 ml of water and 1.5 ml of hexane were added and mixed vigorously. The mixture was then centrifuged at 5,000 rpm for 5 min, and the hexane layer containing unsaponifiables was separated for further analysis. The phytosterols were quantified using GC analysis according to Yang et al. (2013). The GC conditions for phytosterol identification are as follows: DB-5 capillary column (30.0 m × 320 μm × 0.10 μm) was used for analysis; the injector temperature was 260°C; the temperature program was as follows: hold at 60°C for 1 min, increased at 40°C/min to 310°C, and maintained for 6 min; the carrier gas was helium with flow rate of 2 ml/min; and analysis was performed in a split ratio of 10:1.

2.10 Oxidation induction time

The OSI of oil samples was evaluated using a Metrohm 743 Racimat apparatus (Metrohm Herisan, Switzerland). The oil samples (3.0 g) were heated at 110°C. This method is based on the hypothesis that continuous bubbling of air through the oil-containing sample at constant speed of 20 L/h can be used to measure the oxidation induction time of the oil. Results were expressed as hours.

2.11 Statistical analysis

Analysis of variance (ANOVA) and Duncan's multiple range tests were performed using the SPSS (Statistical Product and Service Solutions) 19.0 software version. The ANOVA test was performed for all experimental runs to determine the significance at 95% confidence interval. All experiments were performed in quintuplicate. The means and standard deviations for all values were computed in Origin 8.0.

3.1 Characterization of primary and secondary oxidation products after stir-frying

The temperature was monitored and shown in Figure 1. The oils were heated from 32 ℃ to about 190°C in 40 s. The temperature reduced rapidly when the food items were put into stir wok and then increased in the whole stir-frying process. Peroxide value is an important indicator reflecting the primary oxidation products of oil. Table 1 shows that peroxide values ranged from 2.1 mmol/kg to 2.7 mmol/kg, 2.3 mmol/kg, and 2.3 mmol/kg in HCR from shredded potato, tofu, and vegetable, respectively, indicating a slightly increasing trend. Peroxide values ranged from 1.6 mmol/kg to 2.7 mmol/kg, 2.5 mmol/kg, and 1.9 mmol/kg in the LCR of the above cooked foods, respectively. The oxidative rancidity of oils and the deterioration of the products are affected not only by the saturation degree of the oil, but also by other factors, such as temperature, light, oxygen, metal ions, water activity, and antioxidants. The high temperature of deep-frying is sufficient to disrupt the C-C and C-H covalent bonds of the acyl skeleton, forming various lipid hydrocarbyl radicals, which initiates the chain reaction of free radical oxidation (Goyary, Kumar, & Nayak, 2015). However, after a short stir-frying, the peroxide value was less than 6 mmol/kg, which was within the range of the recommended standard for oil. The effect of quick stir-frying on peroxide value is limited.

Oil is further decomposed to produce a series of aldehydes, ketones, and acids (O’Keefe and Pike 2010). The p-anisidine value is an index of the nonvolatile aldehyde content in oil, formed from fatty acid hydroperoxides, especially 2-alkenals and 2, 4-dienals. p-anisidine value and malondialdehyde (MDA) content is measures of the oxidative state of edible oil. P-anisidine value and MDA content was lower in HCR than in LCR. Table 1 shows that p-anisidine value increased highly and significantly (p < .05) from 3.2 to 6.5 and from 5.5 to 16.2 for HCR and LCR, respectively. The extent of increase was different for HCR and LCR, which may be because of their initial levels in uncooked oils. P-Anisidine value of more than 10 indicates severe deterioration (Neff et al., 2014). The canolol, total phenol, and total phytosterol contents showed significant correlation with the p-anisidine value of HCR samples (p < .05, r > .9). Similarly, Matthaus et al. (Matthäus et al., 2014) observed p-anisidine values of oil with different concentrations canolol were monitored in the range from 2 to 40 after frying 6 hr, with highly negative linear correlation between the amount of added antioxidant and the p-anisidine value. However, no significant correlation between the bioactive compound level and p-anisidine value was observed in LCR.

MDA causes skin cancer in rats as it cross-links with amino groups of DNA. MDA can damage proteins or phospholipids via covalent bonding and cross-linking. Frying in oil that contains mutagens can lead to production of hazardous compounds. The MDA content is the measure of the degree of peroxidation of the plant cytoplasmic membrane. High MDA content is indicative of the high degree of peroxidation and serious damage to the plant cell membrane. MDA content of HCR and LCR increased from 3.4 μg/kg to maximum 6.3 μg/kg and 8.2 μg/kg (p < .05), respectively, after stir-frying (Table1). The increase for the LCR samples was higher than that for the HCR samples, which indicated that HCR suffered less oxidative deterioration after stir-frying. Lower p-anisidine value and MDA content in oil indicated lower extent of oxidation.

3.2 Changes in fatty acid composition after stir-frying

Heating can significantly affect the chemical composition of oils. The presence of unsaturated fatty acids (UFAs) in vegetable oils, including monounsaturated fatty acids (MUFAs) and polyunsaturated fatty acids (PUFAs), is associated with decreased risk of coronary heart disease (Bendsen, Christensen, Bartels, & Astrup, 2011).

The most predominant fatty acids in uncooked rapeseed oil are palmitic acid (4.20% and 5.31%), oleic acid (58.97% and 58.22%), linoleic acid (17.35% and 20.79%), linolenic acid (8.18% and 7.65%), arachidonic acid (3.89% and 2.12%), and erucic acid (5.18% and 3.55%) (Table 2). The results of this study are in agreement with those reported previously, which have shown that the major fatty acids in rapeseed oil included 5.9% of C16:0, 2.5% of C18:0, 48.4% of C18:1, 29.4% of C18:2, and 8.8% of C18:3 (Marinova et al., 2001). Significant differences (p < .05) were identified between the oils using multiple comparison analysis. TFAs were not detected after stir-frying, and simultaneously, no significant differences were observed in the level of fatty acids among the rapeseed oils. Linolenic acid content decreased after cooking from 8.18% to 7.97% in HCR and from 7.65% to 7.47% in LCR, respectively. Results of other studies showed that the linoleic acid content decreased progressively by 8.5% and 13.3% during frying at 185°C and 215°C, respectively (Felixa & Roman, 2009). The deterioration of linolenic acid content was more pronounced and was reduced by 24.0% and 47.1% during frying at 185°C and 215°C, respectively (Felixa & Roman, 2009). When the relative percentages of fatty acids were determined after stir-frying, a slight decrease (not significant, p > .05) of UFA content was observed (Table 2) due to low temperature. A considerable decrease was observed in the PUFA content, including those of linoleic and linolenic acids. Indeed, the amount of C18:1 slightly but significantly increased (p < .05) for all samples. According to Enríquez-Fernández et al. (2011) and Romero, Cueta, and Sánchez-Muniz (2000), the slight increase in oleic acid content might be because of linoleic and linolenic acid degradation. The C18:2/C16:0 ratio decreased for both HCR and LCR samples (Table 2) from 4.13% to 4.03% and from 3.94% to 3.87%, respectively. In fact, linoleic acids are usually used as indicators of the extent of fat deterioration because it is more susceptible to oxidation than saturated fatty acids. Therefore, the C18:2/C16:0 ratio is also used to indicate the degree of oxidative deterioration in frying oil (Flakelar et al. 2017). Overall, our results indicate that the oxidation resistance was higher in HCR than in LCR.

TFAs are always used for evaluating the quality of frying oil. With the exception of cis fatty acid, TFAs were not detected after comparison with standards poststir-frying. TFAs were not detected in any sample (Table 2). The TFAs concentration in frying oils can be determined from the extent of partial oil hydrogenation. The amounts of trans C18:1 in the fresh canola oils were less than the quantitative limit (QL). They were lesser than QL even after the tenth frying operation at 160, 180, and 200°C (Wakako et al., 2010). Trans C18:2 fatty acid accumulation in frying oil was statistically different from that in heated oil. High temperature (190°C) can increase the amount of trans-isomers during deep-frying in edible oils (Zribi et al., 2014). Cui et al. showed that the TFAs content in rapeseed oil with control sample (8.26) was slightly lesser than that of vegetable salad oil (8.33), stir-frying oil (10.76), and pan-frying oil (9.28) (Cui et al., 2017). TFAs were not formed in any of the samples in our study, possibly because the temperature of stir-frying did not reach the condition at which TFAs are produced, or the time of frying was not sufficient.

3.3 Degradation of bioactive components in the rapeseed oils after stir-frying

Internal bioactive compounds or exogenous antioxidants are used to improve the thermal stability of oil during processing as they can inactivate pro-oxidant metals, scavenge free radicals, or quench singlet oxygen depending on their functional group. The total phenol and canolol level of rapeseed oil was increased by microwaving or roasting, as detected in the sample preparation obtained from an oil factory. Oil stability during food processing can be easily determined from the performance of a frying experiment and analysis of relevant parameters associated with the quality of the used frying oil. The initial content of total phenols in HCR and LCR oils was 188.3 mg/100 g and 111.4 mg/100 g, respectively (Figure 2). After frying potatoes, the content of total phenols decreased to 88.9 mg/100 g and 23.0 mg/100 g, respectively, with a percentage of weight loss of 53.2 and 79.3%, respectively. The total polyphenols of HCR and LCR were reduced to 72.0 and 17.1 mg/100 g, respectively, during the frying of vegetables (Figure 2) and the percentage weight loss were 61.7% and 84.6%, respectively. While frying tofu, the content of total phenols decreased to 74.4 and 25.9 mg/100 g, respectively, with a percentage weight loss of 60.4% and 76.7%. Therefore, after frying three raw materials, the total phenol content of HCR was 70.5–88.9 mg/100 g, whereas for LCR it was of 17.1–25.9 mg/100 g. It is obvious that the total phenol concentration of HCR was significantly higher than that of LCR.

2,6-Dimethoxy-4-vinylsyringol is a newly identified phenolic compound in rapeseed oil that is formed by thermal decarboxylation of the sinapic acid naturally occurring in rapeseed seeds after roasting or microwaving (Yang et al., 2014). Thermal treatment of rapeseed increased the canolol content, and temperature considerably influenced the amount of canolol generated in rapeseed. However, it may be converted into polymeric canolol or be degraded during cooking. During potato frying, the level of canolol in HCR and LCR decreased from 528.9 to 237.7 mg/kg and from 250.5 to 69.4 mg/kg, respectively, with a loss rate of 55.0% and 72.3%, respectively (Figure 3). The degradation rate of canolol was 64.5% for HCR and 81.8% for LCR in tofu frying and 60.5% for HCR and 61.5% for LCR in vegetable frying, respectively. Therefore, the level of canolol in HCR after frying different raw materials was in the range of 187.8–237.7 mg/kg, with a loss of 55–64.5%. The percentage of weight loss of canolol in LCR was in the range of 61.5–81.8%, while the concentration after cooking was in the range of 45.6–96.4 mg/kg. Canolol content correlated significantly with induction time for both HCR and LCR. It correlated significantly with p-anisidine value in HCR and with MDA in LCR. Interestingly, Matthäus et al. (2014) obtained a similar result, showing that the degradation rate of canolol decreased with increase in the initial concentration of canolol-enriched extracts Hosseini et al. (2016) found lower content in the oil at higher frying temperature.

There are four forms of tocopherols in rapeseed oil, including α-, β-, γ-, and δ-tocopherols. Tocopherols play an important role as antioxidants, as they prevent oxidation reactions by donating hydrogen atoms to peroxyl radicals (Normand, Eskin, & Przybylski, 2001; Saguy & Dana, 2003). Two types of tocopherols, α-tocopherol and γ-tocopherol, were detected in the two rapeseed oils. The main isomer of tocopherol in rapeseed oil is γ-tocopherol, which accounts for almost two-third of the total amount of tocopherol. These preliminary observations are consistent with the results of previous studies showing that α-tocopherol (130 mg/kg) and γ-tocopherol (320 mg/kg) were detected in raw rapeseed oil (Barrera-Arellano, Ruiz-Mendez, Velasco, Marquez-Ruiz, & Dobarganes, 2002).

Figure 4 shows that the contents of α- and γ-tocopherols were 182.1 and 455.5 mg/kg in HCR and 136.9 and 347.5 mg/kg in LCR. After vegetable frying, α-, and γ-, tocopherol content decreased to 82.5 and 351.3 mg/kg in HCR and to 45.4 and 133.7 mg/kg in LCR, respectively. Results suggested the content of both α- and γ-tocopherols decreased significantly after frying different raw materials, although the content of α- and γ-tocopherols in HCR was significantly higher than those in LCR. The γ-tocopherol content of fresh canola oil used was 214.0 mg/kg, which was similar to the 347.0 mg/kg of tocopherol reported by Felixa and Roman (2009). We also observed that approximately 31% of the total tocopherols remained after frying at 185°C. Loss in natural tocopherol content in oils with different degrees of unsaturation, heated via the two heating periods to 180°C, ranged from 450 mg/kg to 359 mg/kg and to 145 mg/kg after 10 cycles of heating (Barrera-Arellano et al., 2002). Compared to that after baking or boiling, the instability of tocopherols after stir-frying is possibly because of the heat-sensitive nature of the tocopherols (Saguy & Dana, 2003). The rates of degradation of α-tocopherols and γ-tocopherols were more than 40% during stir-frying of three types of dishes. The decrease in tocopherol content was accompanied by increase in MDA in LCR (p < .001); however, MDA levels did not correlate with the change in tocopherols in HCR.

Phytosterols are important bioactive compounds of vegetable oils after tocopherols and phenol. Extensive studies have shown that phytosterols are involved in preventing atherosclerosis. Increasing phytosterol consumption is an effective way of containing hypercholesterolemia as it decreases cholesterol concentration in the blood.

Table 3 shows a progressive decrease in the content of phytosterols in HCR and LCR after frying of three different raw materials. The total amount of phytosterols in those two raw oils was 597.24 and 427.44 mg/100 g, respectively. The total values were 480.57 ~ 496.05 mg/100 g and 350.1 ~ 361.51 mg/100 g after stir-frying in HCR and LCR, respectively, suggesting that the extent of phytosterol loss was lower than those of total phenols, canolol, and vitamin, with about 80% retention rate. Loss of phytosterols during stir-frying is less, which may be due to its excellent stability at high temperature compared to that of canolol, phenols, and tocopherols. In addition, the phytosterols in rapeseed oil mainly include brassicasterol, campesterol, β-sitosterol, and Δ5-avenastrerol. Among these, β-sitosterol is the most abundant, followed by campesterol, and Δ5-avenasterol is the least abundant in HCR and LCR. β-Sitosterol loss was about 16% after stir-frying in both HCR and LCR. A similar degradation rate was observed in these two rapeseed oils. This may be attributed to the fact that many important components exhibit antioxidant activity during high temperature cooking. According to Rudzińska, Korczak, and Wąsowicz (2005), the oxidation products of campesterol, stigmasterol, and β-sitosterol include epimers of 7-hydroxy, 5,6-epoxy, and 7-ketotriols. Previous studies have shown that the content of phytosterols decreased significantly during deep-frying of French fries in rapeseed oil (ca. 60%) (Marinova et al., 2012).

3.4 Relationship between OSI and bioactive compounds

Comparison of micronutrients showed that the canolol, total phenolic, and phytosterol contents in HCR were significantly higher than those in LCR before and after cooking different raw materials. The degradation rate indicated slight differences in oxidation stability among the three raw materials. Among bioactive compounds with antioxidant properties, the total phenolic content showed significant (p < .05) differences, with initial concentrations of 188.3 mg/100 g for HCR and 111.4 mg/100 g for LCR. Consequently, the oxidative stability, as determined using the Rancimat method, also differed significantly (p < .05), with 11.8 hr for HCR and 9.5 hr for LCR (Table 1). Oxidative induction time decreased sharply after stir-frying vegetables, being 7.3 hr and 3.2 hr for HCR and LCR, respectively. Analysis of the correlation between different active components of rapeseed oil, p-anisidine value, MDA content, and oxidation induction time showed that 2,6-dimethoxy-4-vinylphenol correlated significantly with oxidation induction time. The oxidative stability of oils and fats depends not only on their fatty acid composition, but also on the presence of fatty bioactive constituents, such as phytosterols and phenols (Rudzińska et al., 2005). Generally, HCR showed better cooking oxidation stability than LCR. This is similar to the results of a recent study showing that the canolol-rich extracts were able to retard the degradation rate of α- and γ-tocopherol during frying (Hosseini et al., 2016).

However, the results of correlation analysis were different in the two rapeseed oils (Table 4). In HCR samples, total phenols, canolol, total tocopherols, and phytosterols showed a significant positive relationship with OSI. Although total phenols, canolol, and phytosterols showed significant negative correlation with p-anisidine value, they showed significant correlation with MDA content in LCR. Bioactive compounds contribute significantly to oil oxidation stability. Anisidine is formed from fatty acid hydroperoxides, especially 2-alkenals and 2,4-dienals. However, MDA is a degradation product of polyunsaturated fatty acid peroxide. The MDA in fried food products is mainly derived from oil oxidation. Longer duration of frying substantially increased MDA level at 170°C (Normand et al., 2001). In LCR, the total tocopherol content correlated significantly with MDA levels, but not in HCR. Canolol possess stronger antioxidant properties than tocopherols (Shrestha, Stevens, et al., 2013). In the rapeseed oils with high canolol content, canolol as a major antioxidant plays an important role in frying-associated oxidation resistance. The good correlation with oxidative induction time illustrates its antioxidative character.