Effects of Fe3+ and Antioxidants on Glycidyl Ester ... - ACS Publications

Article pubs.acs.org/JAFC

Effects of Fe3+ and Antioxidants on Glycidyl Ester Formation in Plant Oil at High Temperature and Their Influencing Mechanisms Weiwei Cheng,† Guoqin Liu,*,†,‡,§ and Xinqi Liu∥,⊥ †

School of Food Science and Engineering and ‡Guangdong Province Key Laboratory for Green Processing of Natural Products and Product Safety, South China University of Technology, Guangzhou 510640, China § College of Food Science and Engineering, Henan University of Technology, Zhengzhou 450001, China ∥ Beijing Advanced Innovation Center for Food Nutrition and Human Health and ⊥School of Food and Chemical Engineering, Beijing Technology and Business University, Beijing 100048, China ABSTRACT: This research investigated the effects of Fe3+ and antioxidants on the formation of glycidyl esters (GEs) and the free radical mediated mechanisms involving the recognition of cyclic acyloxonium free radical intermediate (CAFRI) for GE formation in both the plant oil model (palm oil, camellia oil, soybean oil, and linseed oil) system and the chemical model (dipalmitin and methyl linoleate) system heated at 200 °C. Results show that Fe3+ can promote the formation of GEs, which can be inhibited by antioxidants in plant oil during high-temperature exposure. Based on the monitoring of cyclic acyloxonium and ester carbonyl group by Fourier transform infrared spectroscopy, the promotion of Fe3+ and the inhibition of antioxidants (tertbutylhydroquinone and α-tocopherol) for GE formation occurred not only through lipid oxidation but also through directly affecting the formation of cyclic acyloxonium intermediate. Additionally, a quadrupole time-of-flight tandem mass spectrometry measurement was conducted to identify the presence of radical adduct captured by 5,5-dimethylpyrroline N-oxide, which provided strong evidence for the formation of CAFRI. Thus, one possible influencing mechanism can be that free radical generated in lipid oxidation may be transferred to dipalmitin and promote CAFRI formation. Fe3+ can catalyze free radical generation while antioxidants can scavenge free radical, and therefore they also can directly affect CAFRI formation. KEYWORDS: glycidyl esters, plant oil, iron ion, antioxidants, lipid oxidation, influencing mechanisms

■

INTRODUCTION In 2010, glycidyl esters (GEs) of fatty acids, a group of new potential food-processing contaminants, was first detected in edible oils and fats due to the regular overestimation of 3monochloropropane-1,2-diol (3-MCPD) esters of fatty acids analyzed by early indirect methods.1 Since then, plentiful literature has reported that GEs are mainly formed in the deodorization step during oil refining and therefore widely exist in refined oils, fats, and oil-based food, including infant formula.2−6 Although there is no available evidence for the detrimental effect of GEs on human and animal, the hydrolysate, glycidol, has been determined to be genotoxic and carcinogenic to rats7 and has already been categorized as a group 2A carcinogen by the International Agency for Research on Cancer (IARC).8 Therefore, to develop an efficient elimination and inhibition method for reducing GE levels in edible oils, further influencing factors and formation mechanisms should be investigated thoroughly. Up to now, it has been well-known that deodorization temperature and time, type of oil, and oil preparation method have significant effects on the formation of GEs.9 Furthermore, diacylglycerols (DAGs) and monoacylglycerols (MAGs) have been identified as the precursors of GE formation.9 Recently, our research group has published a comprehensive review of the formation, occurrence, analysis, and elimination method of GEs in refined edible oils.10 To sum up, there are four proposed reaction pathways for GE formation derived from DAGs and MAGs, all involving the intramolecular rearrangement via charge migration and differing by means of either the nature of © 2017 American Chemical Society

the intermediate or the leaving group. Two of the proposed mechanisms involve a common reactive intermediate, that is, cyclic acyloxonium ion, formed by either deacidification of 1,2DAGs or dehydration of MAGs (either 1-MAGs or 2-MAGs). The other 2 pathways consider a direct intramolecular rearrangement followed by elimination of fatty acid for DAGs (either 1,2-DAGs or 1,3-DAGs) or of water for 1-MAGs. Our previous study confirmed the presence of cyclic acyloxonium intermediate (CAI) during GE formation at high temperature using attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR), which was also earlier conducted to investigate the formation of CAI during 3-MCPD ester formation.11 Afterward, a novel free radical mediated mechanism for 3-MCPD ester formation from DAGs,12 triacylglycerols (TAGs),13 and MAGs14 was proposed, which was involved in the formation of cyclic acyloxonium free radical intermediates but not cyclic acyloxonium cations. Also, based on that, mitigating 3-MCPD ester formation by antioxidant and radical scavenger was also investigated in detail.15,16 However, the issue of whether the free radical mediated mechanisms are as well responsible for the formation of GEs remains still unclear. Lipid oxidation is one of the vital reactions resulting in the quality deterioration of edible oil and oil-based food and Received: Revised: Accepted: Published: 4167

February 23, 2017 April 28, 2017 May 3, 2017 May 3, 2017 DOI: 10.1021/acs.jafc.7b00858 J. Agric. Food Chem. 2017, 65, 4167−4176

Article

Journal of Agricultural and Food Chemistry

solvent A was used to elute the nonpolar fraction that was regarded as model oil in this study. Additionally, the polar fraction remaining on the silica column was also eluted by 150 mL of solvent B. Last, the solvent in the nonpolar or polar eluent was preliminarily removed by rotary evaporator (Zhengzhou Yarong Instruments Co., Ltd., Zhengzhou, China) at 50 °C and then evaporated completely with a gentle nitrogen flow. Both fractions were stored in the nitrogen-filled bottle at −18 °C until analysis. Heating Experiment in Oil Model System. A series of different levels of TBHQ and Fe2(SO4)3 were added into model oil prepared in Preparation of Model Oil and mixed completely. The final concentrations were reached at 0.05, 0.12, and 0.18 mg/g for TBHQ, as well as 0.1, 0.2, and 0.4 mg/mL for Fe2(SO4)3. Alternatively, the reaction bottle (10.2 cm tall, i.d. 1.8 cm) was sealed by an ampule filling machine (Changsha Buyuan Pharmaceutical Machinery Co., Ltd., Changsha, China) under nitrogen aeration or exposed in air. Model oil without any addition of TBHQ or Fe2(SO4)3 was set as control. The heating treatment was conducted at 200 °C under air or nitrogen. After that, the samples were cooled to room temperature and stored in a nitrogen-filled glass bottle at 4 °C prior to further analysis. All experiments were performed in triplicate. Heating Experiment in the Chemical Model System. DPG and DOG, dissolved in hexadecane and prepared at a concentration of 1.0 mg/mL, were selected as the chemical model representing the precursors of GE formation in plant oil. The antioxidants (TBHQ and VE) were prepared at 0.005 mol/L in hexadecane. A 2 mg aliquot of pro-oxidant (Fe2(SO4)3) or 1.0 mL of antioxidant solution and 4 mL of pure ML were selectively added into the chemical model system to investigate their effect on GE formation. The formulas of model mixtures were exhibited in Table 1. After being blended thoroughly by

commonly occurs resulting from the oxidative cleavage of an electron-rich double bond in unsaturated fatty acid (UFAs). It has been well-known that lipid oxidation involves many types of free radicals for which it is possible to participate in the formation of CAI and even influence GE formation. Prooxidant and antioxidant can significantly affect the lipid oxidation via the corresponding promotion and inhibition of free radicals. Currently, the free radical mediated mechanisms of 3-MCPD ester formation have been identified before,12 but there is yet no report about the relationship between lipid oxidation/free radical and GE formation at high temperature. Therefore, our present study aimed to investigate the effect of Fe3+ and antioxidants on the formation of GEs and to further ascertain the free radical mediated mechanisms and the influencing mechanisms of lipid oxidation for GE formation in both model oil system and chemical model system. To exclude the interference of DAG and MAG discrepancy in different types of oils, model oil was selected as real oil system. A series of model experiments consisting of the addition of Fe2(SO4)3, tert-butylhydroquinone (TBHQ), and α-tocopherol (VE) were conducted at 200 °C with dipalmitin, diolein, or methyl linoleate as chemical models. FTIR was used to determine the cyclic acyloxonium and ester carbonyl groups, and a quadrupole time-of-flight mass spectrometer (Q-TOF MS) was used to identify the presence of radical adduct captured by 5,5-dimethylpyrroline N-oxide (DMPO).

■

MATERIALS AND METHODS

Table 1. Formulas of Nine Chemical Model Mixtures

Materials. Crude palm oil was a gift from Yihai Kerry (Tianjin) Investment Co., Ltd. (Tianjin, China), and it originates from Malaysia. Crude camellia oil was purchased from Yongxing Taiyu Camellia Oil Co., Ltd. (Chenzhou, China). Crude soybean oil was provided as a gift from China National Cereals, Oils and Foodstuffs Corporation (COFCO). Crude linseed oil was graciously provided by Gansu Abest Plant Oil Development Co., Ltd. (Jinchang, China). These crude oils were prepared by solvent extraction and free of GEs. Model palm oil (MPO), model camellia oil (MCO), model soybean oil (MSO), and model linseed oil (MLO) were prepared by column chromatography technology in the present study. sn-1,2-Dipalmitin (DPG, 98%), sn-1,2-diolein (DOG, ≥ 97%), 3-MCPD (98%), DMPO (≥97%), TBHQ (97%), hexadecane (≥99%), and VE (≥96%) were purchased from Sigma-Aldrich Co., Ltd. (St. Louis, MO). Methyl linoleate (ML, 95%) was purchased from Aladdin Reagent Co., Ltd. (Shanghai, China). d5-3-MCPD (99%) and phenylboronic acid (PBA) were bought from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). Analytical grade sodium thiosulfate, iron(III) sulfate (Fe2(SO4)3), n-hexane, glacial acetic acid, diethyl ether, petroleum ether (bp 30−60 °C), methyl tertiary butyl ether (MTBE), and anhydrous sodium sulfate were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). 100−200 mesh silica gel was obtained from Qingdao BANKE Separation Materials Co., Ltd. (Qingdao, China). All other chemicals in this work were of analytical grade. Preparation of Model Oil. The degumming, deacidification, and bleaching procedure of crude plant oils were conducted according to the same operation in our previous report,9 which aimed to obtain the refined-bleached (RB) plant oils. Subsequently, a column chromatographic method was carried out to prepare the corresponding model oil with the absence of polar fraction. The solvent A for the column was a mixture of petroleum ether and diethyl ether (87:13, v/v), and solvent B was diethyl ether. The silica gel (about 80 g) activated by 5% deionized water was loaded into a glass column fitted with a stopcock and a sintered disk with the aid of a slurry in solvent A. RB oil obtained above was treated with anhydrous sodium sulfate and activated carbon to remove the moisture and chlorinated substances, respectively. Subsequently, the RB oil was dissolved in 2 × 30 mL of solvent A and poured into the glass column. 150 mL of

chemical model system sample label

DPGa (mL)

A B C D E F G H I

10.0

DOGb (mL)

MLc (mL)

Fe2(SO4)3d (mg)

TBHQ (mL)e

VEf (mL)

10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0

4.0 4.0 4.0 4.0

2.0 1.0 1.0 2.0 1.0 1.0

a

1.0 mg/mL in hexadecane. b1.0 mg/mL in hexadecane. cPure ML. d Pure Fe2(SO4)3. e0.005 mol/L in hexadecane. f0.005 mol/L in hexadecane. vortexing, the mixture was heated at 200 °C for 1 h in a drying oven (Shanghai Keelrein Scientific Instruments Co., Ltd., Shanghai, China) and then stored in a nitrogen-filled glass bottle at 4 °C prior to GE analysis. All experiments were performed in triplicate. Analytical Methods. Peroxide Value (POV). POV was determined according to a modified iodometric method.17 Briefly, 5 g of oil sample was dissolved completely in 50 mL of solvent mixture consisting of acetic acid and isooctane (3:2, v/v), to which 0.5 mL of saturated KI solution was then added. After 1 min at ambient temperature with an occasional shake, 30 mL of deionized water was poured into the reaction cube, shaking vigorously to release all iodine from the isooctane layer, and then titration was performed against 0.1 mol/L sodium thiosulfate using 1% starch indicator until the blue-gray color just disappeared. All samples were determined in triplicate unless otherwise mentioned. Results were presented as mequiv of peroxide/ kg of oil. p-Anisidine Value (p-AV). A measure of p-AV was done based on AOCS Official Method Cd 18-9018 with some modifications. Briefly, a 2.0 g aliquot of oil sample was weighed into a 25 mL volumetric flask 4168

DOI: 10.1021/acs.jafc.7b00858 J. Agric. Food Chem. 2017, 65, 4167−4176

Article


(09), and GE level was calculated as (assay A − assay B) × 0.67. All measures were conducted in triplicate. Determination of Cyclic Acyloxonium and Ester Carbonyl Groups by FTIR. The variation of cyclic acyloxonium and ester carbonyl groups was monitored using ATR-FTIR. The mixtures of DPG with Fe2(SO4)3 and VE, respectively, as well as single DPG, were heated at 200 °C for 15 min. After that, Fe2(SO4)3 or VE was removed by solvent extraction (n-hexane and methanol, respectively), centrifugation, and nitrogen blowing. The dried samples were separately pressed onto the KBr pellet and scanned immediately at wavenumbers ranging from 1500 to 1900 cm−1 in a VERTEX 70 spectrometer (Bruker Optics, Ettlingen, Germany). DSG at 25 °C was used as negative control. Data were processed using OPUS 7.2 software (Bruker Optics GmbH, Ettlingen, Germany). Determination of DMPO Radical Adducts by Q-TOF MS/MS. The determination of DMPO radical adducts was conducted according to the previous report by Zhang et al.12 with a minor modification. 2 mL of 1.0 mg/mL DPG in toluene was mixed with 2 mL of 5.0 mg/ mL DMPO, which was then heated at 200 °C for 15 min. Likewise, 4 mL of 0.5 mg/mL DPG−toluene solution was also heated at the same conditions, which was set as control. After both were cooled immediately to room temperature in the freezing chamber, the DPG solution was diluted 50-fold with toluene and then injected into the Bruker Maxis Impact Q-TOF mass spectrometer (Bruker Co., Bremen, Germany) combined with tandem mass spectrometry (MS/ MS) for analysis. Electrospray ionization positive mode (ESI+) was selected as the LockSpray source. High purity nitrogen was used for drying (4.0 L/min) and nebulization (0.6 bar). The drying heater was set as 180 °C, and the voltage of capillary was 3.5 kV. MS spectra were recorded at the range of 50−1000 mass/charge (m/z) and processed using Bruker Compass DataAnalysis 4.1 software (Bruker Co., Bremen, Germany). Statistical Analysis. All tests were carried out in triplicate independently. Data were reported as mean ± standard deviation (SD). Statistical significance was measured by Duncan’s multiple range tests using SPSS 22.0 software (IBM Corp. Armonk, NY, USA), and the difference between data was considered to be significant at P < 0.05.

and dissolved by isooctane, and then diluted to 25 mL. A Genesys 10S spectrophotometer (Thermo Scientific, Waltham, MA) was used to determine the absorbance of this above solution (Ab) at 350 nm, against isooctane as blank. Subsequently, 5 mL of oil−isooctane solution was mixed with 1 mL of 2.5 g/L p-anisidine reagent. After 8 min at 25 °C, the absorbance of this reaction solution (As) at 350 nm was measured, and the mixture containing 5 mL of isooctane and 1 mL of p-anisidine reagent was used as blank. p-AV was calculated based on the following formula: p-AV = 25 × (1.2As − Ab)/m, where As and Ab denote the absorbance of the reaction mixture and oil−isooctane solution, respectively, and m is the mass of oil. Additionally, according to the previous report,19 total oxidation value (TOTOX) was calculated as (2 × POV) + p-AV. All measures were performed in triplicate. Total Polar Component (TPC). The amount of TPC was determined by a gravimetric method. The polar fraction was isolated from model oil by column chromatography separation, as described in Preparation of Model Oil. After removing the elution solvent, TPC was then dried in an oven at 105 ± 2 °C and weighed accurately. All measures were performed in triplicate. TPC level was presented as the percentage of model oil. TPC Composition. High performance size exclusion chromatography (HPSEC) was utilized to analyze the composition of TPC based on the method reported by Cao et al.20 with a minor adjustment of sample preparation. TPC sample was prepared based on the procedure in Preparation of Model Oil and Total Polar Component (TPC). Then, this sample was dissolved in a mixture of n-hexane and isopropanol (10:1, v/v) to a concentration of 10 mg/mL. After being filtered through a 0.45 μm nylon membrane, this solution was transferred into the 1.5 mL test tubes and stored at −18 °C until further analysis. TPCs, consisting of oxidized triacylglycerols (oxTAGs), triacylglycerol polymers (TGPs), triacylglycerol dimers (TGDs), DAGs, MAGs, and free fatty acids (FFAs), were determined by high-performance liquid chromatography using a Waters 2695 system equipped with a refractive index detector and two tandem Ultrastyragel columns (300 mm × 4.6 mm, 5 μm; 100 and 500 Å), connected to a guard column (50 mm × 4.6 mm) (Waters Co., Milford, MA). A 10 μL aliquot of TPC sample was injected and eluted with tetrahydrofuran at a flow rate of 0.7 mL/min at the column temperature of 35 °C. oxTAGs, TGPs, TGDs, DAGs, MAGs, and FFAs in oil samples were quantified by the area normalization method. All measures were performed in triplicate. The results were presented as the percentage of model oil. GE Analysis. An indirect method for GE determination in oil sample using gas chromatography−mass spectrometry (GC-MS) was developed based on the AOCS Official Method (AOCS Cd 29c-13)21 and original DGF method C-III 18 (09)22 with some modifications. Briefly, two 100 mg oil samples were weighed into test tubes containing 100 μL of MTBE and 0.5 g of d5-3-MCPD, which were labeled as assay A and assay B. Then, 200 μL of 25 g/L sodium methoxide−methanol solution was added into the test tubes to liberate free 3-MCPD and glycidol from their bound forms, which is stop by the addition of 600 μL of 200 g/L acidic sodium chloride solution (assay A) and 600 μL of 600 g/L acidic sodium bromide solution (assay B), respectively. After 600 μL of n-hexane was added into both tubes and kept at room temperature with vigorous shaking for about 5 min, both tubes were centrifuged at 4000 rpm for 5 min. Subsequently, 200 μL of saturated PBA−diethyl ether solution was added to both obtained organic phases above and stood at 30 °C for 20 min; then both of these assays were dried using a soft nitrogen flow and then reconstituted in 500 μL of isooctane, which was stored at −18 °C until GC-MS analysis. 3-MCPD derivatives were separated and quantified in a Shimadzu QP2010Plus GC-MS system (Shimadzu, Tokio, Japan) equipped with a 30 m × 0.25 mm × 0.25 μm Agilent TG-5MS column. 1 μL of sample was injected in a splitless mode with high purity helium as the carrier gas at 1.18 mL/min. The oven temperature program and EI-MS conditions were as presented in our previous report.9 The quantitation of GEs was conducted according to the original DGF method C-III 18

■

RESULTS AND DISCUSSION The relationship between GE formation and lipid oxidation and radical scavenging in plant oil was investigated and demonstrated in the present work. Previously, we obtained a result that 200 °C is a turning point for the formation rate of GEs in laboratory-scale deodorization, and thus it was used as the heating temperature. To eliminate the interference of polar fractions containing the precursors of GE formation, that is, DAGs and MAGs, model oil without polar fractions was prepared and used as the research subject. Based on the fatty acid composition of plant oil that is essentially and closely linked to lipid oxidation, palm oil, camellia oil, soybean oil, and linseed oil were selected in this study. The fatty acid composition of their corresponding model oil is shown in Figure 1. The effects of Fe3+, TBHQ, and VE on GE formation were investigated in both model oil and chemical model systems via GC-MS and FTIR. At last, in order to ascertain their influencing mechanisms, the free radical in model systems was captured by DMPO and identified by Q-TOF MS/MS. The results presented in this work will provide evidence for the free radical mediated GE formation mechanism and show that lipid oxidation/free radical promotes the formation of GEs in plant oil. Further, this work will lay a foundation for the development of efficient elimination methods of GEs. Evolution of GE Level, POV, TOTOX, and TPC Level in Plant Oil Model at High Temperature. It is well-known that high temperature not only is largely responsible for the formation of GEs23,24 but also can easily lead to lipid oxidation. 4169


Article


MLO were heated directly at this temperature, GE levels were rather low and varied insignificantly during heating (data not shown). This observation is explained based on the previous finding that TAG is not the precursor of GE formation.24 Therefore, 4% DPG was added into the model oil to ensure the feasible formation of GEs in model oil. GE level, POV, TOTOX, and TPC level were monitored every 30 or 60 min during heating, and the results are shown in Figure 2. From Figure 2A, GE levels in all tested model oils increased significantly with time. Interestingly, GE level in MLO was rather higher than that in other model oils. Based on the fact that the highest GE content was detected in palm oil and even palm oil based fat.,1,25−27 it was interesting to find that the formation of GEs in MPO was lowest in the tested model oils, which suggested the great contribution of TPC involving precursors to the formation of GEs. To our knowledge, this is the first report that when plant oil possesses the same amount of precursors, such as DAG, the type of plant oil seems responsible for the difference of GE levels in different tested oils. Lipid oxidation is one of the most common thermal reactions at high temperature, which is largely related to the degree of unsaturation of plant oil. It was well expected that MLO had the highest POV and TOTOX, followed by MSO, and then MCO and MPO (Figure 2B and Figure 2C), which

Figure 1. Fatty acid composition of model palm oil (MPO), model camellia oil (MCO), model soybean oil (MSO), and model linseed oil (LO). SFA, saturated fatty acid; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid.

According to our previous work,9 the experimental temperature was set at 200 °C in this study. When MPO, MCO, MSO, and

Figure 2. Evolution of GE levels (A), POV (B), and TOTOX (C), as a function of time (0.5−3.0 h) in tested model oils with the addition of 4% DPG heated at 200 °C, and evolution of TPC contents (D) in tested model oils heated at 200 °C. Data are mean ± standard deviation (SD) (n = 3). 4170


Article

Journal of Agricultural and Food Chemistry Table 2. Changes of TPC Composition in MPO, MCO, MSO, and MLO during Heating at 200 °C up to 3 h TPC composition (mg/g oil) type of oil a

MPO-1 MPO-2 MPO-3 MCO-1 MCO-2 MCO-3 MSO-1 MSO-2 MSO-3 MLO-1 MLO-2 MLO-3 a

oxTAG 0.56 1.94 2.43 0.69 1.12 1.64 0.98 1.24 1.66 1.39 3.04 4.95

± ± ± ± ± ± ± ± ± ± ± ±

0.02 0.04 0.05 0.02 0.03 0.08 0.03 0.02 0.05 0.02 0.06 0.05

TGP b

0.61 1.23 2.82 0.88 1.62 3.01 1.24 2.07 3.05 1.43 1.93 2.51

± ± ± ± ± ± ± ± ± ± ± ±

0.03 0.04 0.06 0.02 0.03 0.05 0.03 0.03 0.08 0.04 0.05 0.03

TGD 1.73 3.65 5.44 2.08 3.13 4.16 2.39 4.62 7.18 2.72 4.10 5.25

± ± ± ± ± ± ± ± ± ± ± ±

FFA

0.04 0.07 0.09 0.05 0.06 0.05 0.04 0.06 0.08 0.05 0.06 0.03

0.28 0.51 0.87 0.37 0.65 0.99 0.17 0.76 1.16 0.35 0.76 1.16

± ± ± ± ± ± ± ± ± ± ± ±

0.01 0.01 0.02 0.01 0.00 0.02 0.00 0.01 0.02 0.00 0.01 0.02

MAG 0.06 0.09 0.14 0.04 0.06 0.10 0.05 0.09 0.15 0.06 0.12 0.15

± ± ± ± ± ± ± ± ± ± ± ±

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

DAG 0.43 0.65 0.96 0.50 0.87 1.29 0.37 0.74 1.08 0.41 0.79 1.08

± ± ± ± ± ± ± ± ± ± ± ±

0.02 0.01 0.02 0.01 0.01 0.03 0.01 0.01 0.04 0.00 0.01 0.02

MPO was heated at 200 °C for 1 h. bMean ± SD (n = 3).

accounted for the difference of unsaturated fatty acid in tested oils (Figure 1). These results seemingly coincided with GE formation in tested oils (Figure 2A). However, it has also been reported that conventional heating enabled the increasing DAG level in edible oil.28 In the present study, as shown in Figure 2D, it was indeed found that TPC contents increased linearly with time up to 3 h, regardless of the type of oil. This finding was in agreement with the results reported by Correia et al.29 Additionally, TPC content in tested oil was ranked as MLO > MSO > MCO > MPO, which seemed also to be positively related to GE formation. Accordingly, these facts raised a question: what led to the discrepancy of GE levels among tested oils, lipid oxidation or TPC formation? Evaluation of Contribution of Formed TPC, ProOxidant, and Antioxidant to the Formation of GEs in Plant Oil Model at High Temperature. As depicted depicted in Figure 2, both TPC content and degree of lipid oxidation seemingly increased with the same trend with GE formation in tested oils. To further explain this phenomenon, first, TPC, consisting of oxTAG, TGP, TGD, FFA, MAG, and DAG, was analyzed every hour for a period of 3 h, and the results are presented in Table 2. In spite of the linear increase of TPC contents in all tested oils (Figure 2D), DAG and MAG levels were relatively low compared to oxidized/polymerized products, that is, oxTAG, TGP, and TGD, suggesting that oxidative polymerization was the main chemical reaction in heated oil exposed in air, as reported by Choe and Min.30 Previously, Craft et al.31 have reported that GE content increases exponentially when DAG level exceeds 3−4% of oil but varied slightly with below 3% of DAG level. Thus, DAG and MAG (altogether 0.042−0.139% of oil) from pyrolysis of TAG contributed mildly to the formation of GEs during heating. In addition, as shown in Figure 3, the total levels of DAG and MAG almost did not present significant difference (P > 0.05) in tested oils heated for the same time. These findings seemingly indicated that the remarkable increase of TPC content did not necessary contribute to GE formation and the discrepancy of GE levels in different tested oils (Figure 2A) compared with our previous report.9 Subsequently, the second experiment was to confirm the effects of pro-oxidants and antioxidants on the formation of GEs in four tested oils. For this sake, Fe2(SO4)3 and TBHQ were added into the model oil to promote or inhibit the lipid oxidation evaluated by POV and TOTOX levels (data not shown). Meanwhile, GE level was detected in these model oils

Figure 3. Changes of total contents of DAGs and MAGs in MPO, MCO, MSO, and MLO during heating at 200 °C up to 3 h. All data are mean ± SD (n = 3). Different letters at the same time represent significant difference between total levels of DAG and MAG in different types of model oil (P < 0.05).

heated at 200 °C for 1 h. The results are shown in Figure 4. It is well-known that Fe3+ can initiate oil peroxidation by catalyzing the generation of alkyl radical.32,33 From Figure 4A, GE levels in tested model oils heated at 200 °C for 1 h increased significantly with the addition of Fe2(SO4)3 (P < 0.05). However, when EDTA-2Na, a chelating agent, together with Fe2(SO4)3 was added to model oil, GE level was significantly lower than that without EDTA-2Na (P < 0.05), regardless of the type of model oil (Figure 4A). Thus, Fe3+ did indeed facilitate the formation of GEs in oil during high-temperature exposure. TBHQ, an artificial antioxidant, is used widely in edible oil to protect it from oxidative deterioration. It was well expected that the formation of GEs was inhibited by TBHQ, based on the decreasing GE level with growing additional TBHQ in tested model oils (Figure 4B). Additionally, due to different fatty acid composition in different model oils (Figure 1), the different susceptibility of the model oils to lipid oxidation was presented. However, it was found that there was the same response on GE content with increasing Fe3+ and TBHQ amounts in different model oils, possibly suggesting that Fe3+ and TBHQ might directly affect the formation of GEs in model oils, not or relatively slightly by lipid oxidation. 4171


Article


Figure 4. Effects of Fe2(SO4)3 (A), TBHQ (B), and isolating oxygen (C) on GE formation in tested model oils with addition of 4% DPG heated at 200 °C for 1 h. aEquimolar quantities of Fe3+ and EDTA-2Na were added into model oils. All data are mean ± SD (n = 3). Values labeled by a different letter are significantly different for the same model oil (P < 0.05).

However, if model oil was heated at high temperature under nitrogen, GE level was significantly lower than that under air (Figure 4C). Correspondingly, the oxidation level of the former oil was higher than that of the latter one (data not shown). Taken together, these findings also gave a possible hypothesis whether there was a positive relationship between GE formation and lipid oxidation in plant oil. However, it also led to the other question whether Fe3+, TBHQ, and cutting off oxygen had direct effects on the formation of GEs but not through lipid oxidation. To illustrate this, further study was carried out as shown in the following section. Effects of Pro-Oxidant and Antioxidant on the Formation of GEs in Chemical Model at High Temperature. Based on our previous work,9 DPG was selected as the chemical model for GE formation, and ML was used to simulate the UFAs for lipid oxidation in this research. Also, no loss of TBHQ occurred in the airtight system during heating at 200 °C. The effects of Fe3+, TBHQ, and VE on the formation of GEs were assessed in the model system heated at 200 °C for 1 h. The results are presented in Figure 5. When the mixture of DPG and ML (C) was heated, GE level was significantly higher than that in only DPG (A) (P < 0.05), which suggested that ML favored the formation of GEs. It has also been well reported that the oxidation of ML occurs readily at high temperature. 34 Therefore, these findings confirmed the promotion of lipid oxidation to GE formation in plant oil at

Figure 5. GE levels in different chemical model systems heated at 200 °C for 1 h. All data are mean ± SD (n = 3). Values labeled by different letters are significantly different (P < 0.05). aExpressed as DPG or DOG.

high temperature. However, it was interesting that GE level in the DOG system (B) was significantly lower than that in the DPG system (A) (Figure 5), suggesting that the self-oxidation 4172


Article

Journal of Agricultural and Food Chemistry of precursor reduced exceptionally the formation of GEs. The reason for this result might be that the oxidative reaction took place more readily than the formation reaction of GEs for the same reactant, and GEs with UFAs were more likely to degrade compared to saturated GEs at high temperature. Further study with regard to the competitive mechanism of both reactions in the DOG system is being performed in our group and will be reported in the future. If the mixture of DPG and ML was regarded as the control sample, the addition of Fe2(SO4)3 (D) and TBHQ (E) resulted in an acceleration and inhibition effect, respectively, which was in agreement with the results obtained above in model oil system. To further understand the effect of antioxidant on GE formation, VE, a natural antioxidant with a lower radical scavenging activity than TBHQ,35 was also added into the model oil with the same amount as TBHQ. As expected, the inhibition efficiency of VE, which reduced GEs by 16.81% based on the mixture of DPG and ML (C), was rather lower than TBHQ, which was 31.47%. These observations suggested that the formation of GEs might be inhibited by means of resisting lipid oxidation, that is, scavenging free radical, during hightemperature exposure. Additionally, to ascertain whether Fe3+, TBHQ, and VE could also directly affect the formation of GEs without the occurrence of oxidation, some other experiments were conducted in DPG model system without ML. From Figure 5, in comparison with DPG model system (A), Fe3+ (G) could also favor the formation of GEs, as well, TBHQ (H) and VE (I) exhibited an efficient inhibition of GE formation, and the inhibition efficiency was ranked as TBHQ > VE. These results were consistent with that in the presence of ML and therefore indicated that Fe3+, TBHQ, and VE might also directly act on the formation of GEs, which could involve the free radical mechanisms based on the promotion or inhibition mechanisms of these substances for lipid oxidation as depicted above. Monitoring of Cyclic Acyloxonium and Ester Carbonyl Groups in Chemical Model System Using ATR-FTIR. As depicted above, pro-oxidants and antioxidants could influence the formation of GEs not only through lipid oxidation but also through the direct effect on its formation pathway (for example, accelerated or inhibited the formation of CAI) in plant oil. Accordingly, to further investigate whether and how the prooxidant and antioxidant affect the formation reaction of GEs, cyclic acyloxonium and ester carbonyl groups were monitored in DPG system using ATR-FTIR. Due to the absence of ester carbonyl group in VE and without formation of cyclic acyloxonium group at high temperature,16 VE was just selected as the antioxidant in FTIR test. From Figure 6, two different carbonyl absorbance bands centered at 1709 and 1733 cm−1, representing two different chemical environments for the ester carbonyl groups (at sn-2 and sn-1 position, respectively),11 were observed in the spectra of DPG at 25 °C. Upon heating to 200 °C, just a carbonyl band centered at 1736 cm−1 was observed and the absorbance band of cyclic acyloxonium group centered at 1640 cm−1 appeared. Fe2(SO4)3 and VE were respectively added into two same DPG model systems and heated at 200 °C for 1 h. Compared with the bands of carbonyl band and cyclic acyloxonium group of heated DPG system without any addition, for the former, the carbonyl band at 1736 cm−1 was weaker but the band of cyclic acyloxonium group was stronger; for the latter, on the contrary, the DPG system heated together with VE still exhibited two carbonyl peaks and stronger one at 1733 cm−1 as well as weaker band of cyclic acyloxonium group. These findings suggested that Fe3+ could accelerate the

Figure 6. FTIR spectra of DPG, DPG heated with Fe2(SO4)3, and VE at 200 °C for 15 min compared with that of DPG at 25 °C.

formation of cyclic acyloxonium intermediate, which was nevertheless slowed by VE. Previously, it has been reported that a free radical, determined using electron spin resonance (ESR) spectrometry, was formed in sn-1,2-stearoylglycerol (DSG) system during heat exposure, and a higher temperature seemingly resulted in an increasing ESR signal.12,16 As a consequence, these observations led to an interesting hypothesis that the formation of GEs involved a free radical mediated mechanism and, more likely, occurred via a free radical intermediate. Monitoring of DMPO Radical Adducts in Chemical Model Using Q-TOF MS/MS. To confirm the occurrence of a free radical intermediate during GE formation at high temperature, Q-TOF MS and MS/MS were carried out to determine the DMPO radical adducts. The mixtures of DPG and DMPO heated at 200 °C for 15 min were injected into QTOF MS/MS system. The MS and MS/MS spectra were recorded in the range m/z 100−800 and are shown in Figure 7 and Figure 8, respectively. As exhibited in Figure 7A, m/z 314.5035 and 553.9145, attributed to [CAI′]+ (calcd 314.5017)

Figure 7. Q-TOF MS spectra of 1.0 mg/mL DPG heated without (A) and with (B) DMPO at 200 °C for 15 min. CAI, cyclic acyloxonium intermediate. PGE, glycidyl palmitate. 4173


Article


Figure 8. Q-TOF MS/MS spectra of parent ion at m/z 706.0681.

and [CAI″]+ (calcd 553.9172), respectively, shown in Figure 9A, were observed in the heated DPG system without DMPO, suggesting the formation of two different CAI. Glycidyl palmitate (PGE) (m/z 312.4838, calcd 312.4870) was also detected, which was consistent with FTIR measurement.9 Analysis of the mixture of DPG and DMPO under the same conditions as above found two pairs of sodiated adduct ions, that is, at m/z 466.6428 most likely attributed to [CAI′ + DMPO + O + Na]+ (calcd 466.6489) and m/z 706.0681 for [CAI” + DMPO + O + Na]+ (calcd 706.0644), as well as their corresponding hydrogenated ions (Figure 7B). However, neither [CAI′]+ nor [CAI″]+ was observed in Figure 7B due to their complete reaction with DMPO, suggesting that the CAI existed in a free radical form, readily captured by DMPO. In addition, [PGE]+ was detected in the heated mixture of DPG and DMPO, which indicated the absence of GE radical intermediate formation at high temperature. This observation was not consistent with the speculative description of Zhang et al.13 that GEs might be formed through epoxide ring free radical intermediate. These findings provided support for the DMPO radical adducts being involved in the formation of GEs derived from DPG in this model system at high temperature. To further confirm the formation of DMPO radical adducts, Q-TOF MS/MS of the selected parent ion at m/z 706.0681 was also conducted. From Figure 8, the fragment ions at m/z 553.9152 corresponding to [CAI″]+, as well as m/z 685.0949 and 686.0984 consistent with [CAI″ + DMPO + O + H]+ (calcd 685.0904) and [CAI″ + DMPO + OH + H]+ (calcd 685.0983), respectively, were detected upon collisionally activated dissociation of parent oil. Furthermore, [DMPO + Na]+ and [DMPO + O + Na]+ were also found in the fragment ions, which suggested that the parent ion at m/z 706.0681 was indeed the sodiated adduct of CAI and DMPO with one

oxygen atom bound. Hence, these above data adequately demonstrated that GEs were formed through the free radical intermediate at high temperature. Proposed Influence Mechanism of Lipid Oxidation on the Formation of GEs at High Temperature. According to the above data, a free radical mediated mechanism for the formation of GEs from DAGs at high temperature was demonstrated in the present study (Figure 9A). The reaction process involved in the formation of cyclic acyloxonium free radical intermediates (CAI′ and CAI″) might be initiated by the removal of either fatty acid radical (R•) at the sn-2 position or hydroxyl radical at the sn-3 position combining with the thermally resulting radical (such as ROO•) that most likely originated from oil peroxidation, or the alternatively direct fracture of covalent bond under high-temperature conditions. The ester carbonyl group at the sn-1 position might be immediately attacked by the sn-2 carbon-centered radical to form CAI′ free radical. Also, the ester carbonyl group at the sn2 position might also be attacked by the sn-3 carbon-centered radical to form CAI″ free radical. As shown in Figure 9A, subsequently, CAI free radical could perform a further intramolecular rearrangement to form GEs in which the free radical might participate in the termination stage of oil autoxidation.36 As a consequence, the generated free radical during lipid oxidation could be in favor of the formation of cyclic acyloxonium free radical intermediates, which explained why lipid oxidation promoted GE formation (Figure 9B). Thus, the effects of Fe3+, TBHQ, and VE on the formation of GEs, as shown above, were well understood. More importantly, Fe3+ facilitated the formation of free radical, and therefore also positively (+) affected CAI formation. On the contrary, antioxidant, such as TBHQ, VE, et al., could negatively (−) affect CAI formation by scavenging free radical. In brief, the 4174


Article


Figure 9. Free radical mediated mechanisms for GE formation from DAGs (A), and proposed influencing mechanisms of lipid oxidation on GE formation (B).

China (Project No. 31471677), the National Hi-tech Research and 620 Development Project of China (Project No. 2013AA102103), the Public Welfare (Agriculture) Research Project (Project No. 201303072), and the Fundamental Research Funds for the Central Universities of China.

relationship between GE formation and lipid oxidation is exhibited in Figure 9B. In summary, the significant effect of Fe3+, TBHQ, and VE on GE formation at high temperature was demonstrated in both model oil system and chemical model system. It was confirmed that the formation of GEs involved the free radical mediated mechanisms and the formation of cyclic acyloxonium free radical intermediates, which explained why the radicalgenerated lipid oxidation was in favor of GE formation. Therefore, it was well understood that Fe3+ promoted the formation of GEs, which was also inhibited by TBHQ and VE, not only by indirectly regulating lipid oxidation/free radical, but also by directly affecting CAI formation. These results will enlighten the development of effective elimination and inhibition methods for GE formation in the production of refined edible oil.

■

Notes

The authors declare no competing financial interest.

■

ABBREVIATIONS USED GEs, glycidyl esters; 3-MCPD, 3-chloropropane-1,2-diol; MPO, model palm oil; MCO, model camellia oil; MSO, model soybean oil; MLO, model linseed oil; DAGs, diacylglycerols; MAGs, monoacylglycerols; CAI, cyclic acyloxonium intermediate; TPC, total polar component; FFAs, free fatty acids; PGE, glycidyl palmitate; PBA, phenylboronic acid; TBHQ, tertbutylhydroquinone; VE, α-tocopherol; DMPO, 5,5-dimethylpyrroline N-oxide; oxTAGs, oxidized triacylglycerols; TGPs, triacylglycerol polymers; TGDs, triacylglycerol dimers; POV, peroxide value; p-AV, p-anisidine value; TOTOX, total oxidation value; SFA, saturated fatty acid; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid; ATRFTIR, attenuated total reflectance Fourier transform infrared; GC-MS, gas chromatography−mass spectrometry; HPSEC, high performance size exclusion chromatography; Q-TOF, quadrupole time-of-flight; ESI, electrospray ionization; DGF, German Society for Fat Science; AOCS, American Oil Chemists’ Society

AUTHOR INFORMATION

Corresponding Author

*Tel: 8620-8711-4262. Fax: 8620-8711-3875. E-mail: guoqin@. scut.edu.cn. ORCID

Weiwei Cheng: 0000-0002-4907-7872 Funding

The work was supported by the National Key Research and Development Program of China (Project No. 2016YFD0400401-5), the National Natural Science Fund of 4175


Article


■

(20) Cao, W. M.; Zhang, K. Y.; Xue, B.; Chen, F. X.; Jin, Q. Z.; Wang, X. G. Determination of oxidized triacylglycerol polymers by preparative flash chromatography and high-performance size-exclusion Chromatography. Asian J. Chem. 2013, 25, 9189−9194. (21) AOCS/JOCS. Joint AOCS/JOCS Official Method Cd 29c-13: Fatty-acid-bound 3-chloropropane-1,2,diol (3-MCPD) and 2,3-epoxipropane-1-ol (glycidol), Determination in Oils and Fats by GC/MS (Differential Measurement). Official Methods and Recommended Practices of the AOCS, 3rd printing, 2013−2014. Additions and Revisions, 6th ed.; AOCS: Urbana, IL 2013. (22) Deutsche Gesellschaft für Fettwissenschaft: DGF Standard Method C III 18b (2009): Determination of ester-bound 3chloropropane-1,2-diol (3-MCPD esters) and 3-MCPD forming substances in fats and oils by means of GC-MS. Deutsche Einheitsmethoden zur Untersuchung von Fetten, Fettprodukten, Ed., Tensiden und verwandten Stoffen: Wissenschaftliche Verlagsgesellschaft, Stuttgart, Germany, 2009. (23) Shimizu, M.; Weitkamp, P.; Vosmann, K.; Matthäus, B. Temperature dependency when generating glycidyl and 3-MCPD esters from diolein. J. Am. Oil Chem. Soc. 2013, 90, 1449−1454. (24) Destaillats, F.; Craft, B. D.; Dubois, M.; Nagy, K. Glycidyl esters in refined palm (Elaeis guineensis) oil and related fractions. Part I: Formation mechanism. Food Chem. 2012, 131, 1391−1398. (25) Becalski, A.; Feng, S.; Lau, B. P.; Zhao, T. A pilot survey of 2and 3-monochloropropanediol and glycidol fatty acid esters in foods on the Canadian market 2011−2013. J. Food Compos. Anal. 2015, 37, 58−66. (26) Samaras, V. G.; Giri, A.; Zelinkova, Z.; Karasek, L.; Buttinger, G.; Wenzl, T. Analytical method for the trace determination of esterified 3- and 2-monochloropropanediol and glycidyl fatty acid esters in various food matrices. J. Chromatogr. A 2016, 1466, 136−147. (27) Marc, C.; Drouard-Pascarel, V.; Retho, C.; Janvion, P.; Saltron, F. Determination of MCPD Esters and Glycidyl Esters by Microwave Extraction in different Foodstuffs. J. Agric. Food Chem. 2016, 64, 4353−4361. (28) Lucas-Torres, C.; Pérez, Á .; Cabañas, B.; Moreno, A. Study by 31 P NMR spectroscopy of the triacylglycerol degradation processes in olive oil with different heat-transfer mechanisms. Food Chem. 2014, 165, 21−28. (29) Correia, A. C.; Dubreucq, E.; Ferreira-Dias, S.; Lecomte, J. Rapid quantification of polar compounds in thermo-oxidized oils by HPTLC-densitometry. Eur. J. Lipid Sci. Technol. 2015, 117, 311−319. (30) Choe, E.; Min, D. Chemistry of deep-fat frying oils. J. Food Sci. 2007, 72, R77−R86. (31) Craft, B. D.; Nagy, K.; Seefelder, W.; Dubois, M.; Destaillats, F. Glycidyl esters in refined palm (Elaeis guineensis) oil and related fractions. Part II: Practical recommendations for effective mitigation. Food Chem. 2012, 132, 73−79. (32) Schafer, F. Q.; Qian, S. Y.; Buettner, G. R. Iron and free radical oxidations in cell membranes. Free Radical Biol. Med. 2000, 46, 657− 662. (33) Chen, B.; Mcclements, D. J.; Decker, E. A. Minor components in food oils: a critical review of their roles on lipid oxidation chemistry in bulk oils and emulsions. Crit. Rev. Food Sci. Nutr. 2011, 51, 901− 916. (34) Berdeaux, O.; Fontagné, S.; Sémon, E.; Velasco, J.; Sébédio, J. L.; Dobarganes, C. A detailed identification study on high-temperature degradation products of oleic and linoleic acid methyl esters by GC− MS and GC−FTIR. Chem. Phys. Lipids 2012, 165, 338−347. (35) Alamed, J.; Chaiyasit, W.; Mcclements, D. J.; Decker, E. A. Relationships between free radical scavenging and antioxidant activity in foods. J. Agric. Food Chem. 2009, 57, 2969−2976. (36) St. Angelo, A. J.; Vercellotti, J.; Jacks, T.; Legendre, M. Lipid oxidation in foods. Crit. Rev. Food Sci. Nutr. 1996, 36, 175−224.

REFERENCES

(1) Weißhaar, R.; Perz, R. Fatty acid esters of glycidol in refined fats and oils. Eur. J. Lipid Sci. Technol. 2010, 112, 158−165. (2) Wong, Y. H.; Muhamad, H.; Abas, F.; Lai, O. M.; Nyam, K. L.; Tan, C. P. Effects of temperature and NaCl on the formation of 3MCPD esters and glycidyl esters in refined, bleached and deodorized palm olein during deep-fat frying of potato chips. Food Chem. 2017, 219, 126−130. (3) MacMahon, S.; Begley, T. H.; Diachenko, G. W. Occurrence of 3MCPD and glycidyl esters in edible oils in the United States. Food Addit. Contam., Part A 2013, 30, 2081−2092. (4) Pudel, F.; Benecke, P.; Fehling, P.; Freudenstein, A.; Matthäus, B.; Schwaf, A. On the necessity of edible oil refining and possible sources of 3-MCPD and glycidyl esters. Eur. J. Lipid Sci. Technol. 2011, 113, 368−373. (5) Inagaki, R.; Hirai, C. Formation of Glycidol Fatty Acid Esters in Meat Samples Cooked by Various Methods. J. Food Process. Technol. 2016, 7, 557−562. (6) Wöhrlin, F.; Fry, H.; Lahrssen-Wiederholt, M.; Preiß-Weigert, A. Occurrence of fatty acid esters of 3-MCPD, 2-MCPD and glycidol in infant formula. Food Addit. Contam., Part A 2015, 32, 1810−1822. (7) Irwin, R. NTP Toxicology and Carcinogenesis Studies of Glycidol (CAS No. 556-52-5) in F344/N Rats and B6C3F1 Mice (Gavage Studies). Natl. Toxicol. Program Tech. Rep. Ser. 1990, 374, 1− 229. (8) IARC. Some industrial chemicals; IARC monographs on the evaluation of carcinogenic risk of chemicals to humans; International Agency for Research on Cancer: Lyon, France, 2000; Vol. 77. (9) Cheng, W.; Liu, G.; Liu, X. Formation of Glycidyl Fatty Acid Esters Both in Real Edible Oils during Laboratory-Scale Refining and in Chemical Model during High Temperature Exposure. J. Agric. Food Chem. 2016, 64, 5919−5927. (10) Cheng, W.; Liu, G.; Wang, L.; Liu, Z. Glycidyl Fatty Acid Esters in Refined Edible Oils: A Review on Formation, Occurrence, Analysis, and Elimination Methods. Compr. Rev. Food Sci. Food Saf. 2017, 16, 263−281. (11) Rahn, A. K. K.; Yaylayan, V. A. Monitoring cyclic acyloxonium ion formation in palmitin systems using infrared spectroscopy and isotope labelling technique. Eur. J. Lipid Sci. Technol. 2011, 113, 330− 334. (12) Zhang, X.; Gao, B.; Qin, F.; Shi, H.; Jiang, Y.; Xu, X.; Yu, L. L. Free radical mediated formation of 3-monochloropropanediol (3MCPD) fatty acid diesters. J. Agric. Food Chem. 2013, 61, 2548−2555. (13) Zhang, Z.; Gao, B.; Zhang, X.; Jiang, Y.; Xu, X.; Yu, L. L. Formation of 3-monochloro-1,2-propanediol (3-MCPD) di- and monoesters from tristearoylglycerol (TSG) and the potential catalytic effect of Fe2+ and Fe3+. J. Agric. Food Chem. 2015, 63, 1839−1848. (14) Zhao, Y.; Zhang, Y.; Zhang, Z.; Liu, J.; Wang, Y. L.; Gao, B.; Niu, Y.; Sun, X.; Yu, L. Formation of 3-MCPD Fatty Acid Esters from Monostearoyl Glycerol and the Thermal Stability of 3-MCPD Monoesters. J. Agric. Food Chem. 2016, 64, 8918−8926. (15) Li, C.; Jia, H.; Shen, M.; Wang, Y.; Nie, S.; Chen, Y.; Zhou, Y.; Wang, Y.; Xie, M. Antioxidants Inhibit Formation of 3-Monochloropropane-1,2-diol Esters in Model Reactions. J. Agric. Food Chem. 2015, 63, 9850−9854. (16) Zhang, H.; Jin, P.; Zhang, M.; Cheong, L. Z.; Hu, P.; Zhao, Y.; Yu, L.; Wang, Y.; Jiang, Y.; Xu, X. Mitigation of 3-Monochloro-1,2propanediol Ester Formation by Radical Scavengers. J. Agric. Food Chem. 2016, 64, 5887−5892. (17) AOCS. AOCS Official Method Cd 8b-90. Peroxide Value, Acetic Acid-Isooctane Method. In Official Methods and Recommended Practices of the American Oil Chemists’ Society; AOCS Press: Champaign, IL, 1996. (18) AOCS. AOCS Official Method Cd 18-90. p-Anisidine Value. In Official Methods and Recommended Practices of the American Oil Chemists’ Society; AOCS Press: Champaign, IL, 1996. (19) Shahidi, F.; Wanasundara, U. Methods for measuring oxidative stability in edible oils. Food Lipids: Chemistry, Nutrition and Biotechnology; CRC Press: New York, 2008; pp 387−388. 4176
