The Formation of 3-Monochloropropanediol Esters and Glycidyl Esters during Heat-Induced Processing Using an Olive-Based Edible Oil

With the prevalence of edible diacylglycerol (DAG) oil, which is beneficial to human, the generation of 3-monochloropropanediol esters (3-MCPDE) and glycidyl esters (GE) as well as the stability of physical properties during heat-induced processing still need to be explored. In this study, the experiment used olive-based edible oil with different contents of DAG (40, 60, and 80%) to make crackers and fry chicken. They were heated at 160 and 180 °C to determine the changes in 3-MCPDE and GE, the crackers’ hardness and gumminess, and the physical properties of the oil. During baking and frying, 3-MCPDE decreased, while the content of GE slightly increased with the prolonged heating duration. Finally, 3-MCPDE and GE were lower than 1.25 mg/kg and 1.00 mg/kg, respectively. The AV increased proportionally as duration increased and POV was below 0.30 g/100 g. In general, the changes in 3-MCPDE and GE were related to the heating temperature and duration, and not significantly (p > 0.05) related to the content of DAG.


3-monochloropropanediol ester
and glycidyl esters (GE) are heat-induced processed contaminants found in refined vegetable oil [1]. GE can be metabolized into free glycidol, which is classified as "probably carcinogenic to human" (Group 2A) by the International Agency for Research on Cancer (IARC) [2]. In February 2018, the European Commission (EC) set the maximum GE limit at 1.00 mg/kg in edible oils. A stringent GE limit was 0.50 mg/kg for oils used to formulate food. Meanwhile, 3-MCPD is assessed as a non-genotoxic carcinogen at a tolerable daily intake (TDI) of 2.00 µg/kg body weight in human [3]. According to Regulation (EU) 2020/1322, the EC proposed two maximum levels for 3-MCPDE in refined oils, which were 1.25 mg/kg and 2.50 mg/kg, in September 2020.
The natural and unrefined oils do not contain 3-MCPDE or GE, and they are formed due to the high temperature during the refining deodorization step of vegetable oil [1]. Therefore, applications of edible oils, such as baking and frying, can also result in the production of 3-MCPDE and GE because of high temperature [4,5]. In addition, the content of NaCl [6], type of frying media [7], frying methods [8], frying temperature [9], antioxidants [10] and additives [11] can affect the formation of 3-MCPDE and GE. The monoacylglycerol (MAG), diacylglycerol (DAG) and triacylglycerol (TAG) also contribute to the formation of 3-MCPDE and GE [12]. Additionally, 3-MCPDE was found in savoury foods, such as salty crackers and biscuits [13,14].
Natural edible oils mainly contain TAG, and the DAG is generally less than 10% [15]. Several studies have proved that DAG has potentially beneficial properties, such as reducing body fat, having hypolipidemic activity and regulating blood sugar level [16,17]. Therefore, DAG edible oils are considered to be functional lipids. In September 2009, it was found that the GE content was above the limit of the European Union (EU) in Econa Healthy Cooking Oil (a DAG edible oil product). With a high content of DAG (6-10%), palm oil contained the highest GE among vegetable oils [15].
Generally, olive oils with high oleic acids demonstrate good frying performance because they are stable at high temperature [18]. However, the stability of an olive-based DAG was not reported in the literature. In this study, the stability of olive-based DAG oil was investigated in both baking and frying processes. The effects of DAG content, heating temperature and duration on the formation of 3-MCPDE, GE and oil qualities were evaluated.

Materials and Chemicals
The olive-based TAG oil was obtained from Wilmar Oleo Co., Ltd. (Dongguan, China), and the olive-based DAG oil was obtained using laboratory methods [19]. Olive-based DAG oil was prepared at 40, 60 and 80% by mixing olive-based DAG oil and olive-based TAG oil at ratios of 4:6, 6:4 and 8:2. Materials for baking and chicken for frying were obtained from a local market.

Preparation of Crackers
The crackers were prepared with olive-based DAG oil (40, 60 and 80%) according to the American Association of Cereal Chemists Method 10-54 with minor modifications [20]. Flour (100 g), yeast (1 g), baking soda (0.5 g), skim milk (60 g), oil (20 g) and salt (1 g) were combined to form a smooth dough. Then the dough was pressed to 3 mm thickness and cut into 40 mm × 40 mm pieces. Finally, the crackers were baked at 160 and 180 • C for 12 min. In addition, the oil without materials (as control) was heated in the oven under the same conditions. The oil in the crackers was extracted with n-hexane (ultrasonic bath) to analyze the acid value (AV), peroxide value (POV), fatty acid composition, 3-MCPDE and GE.

Intermittent Frying Condition
A 1.5 L volume of olive-based DAG oil (40, 60 and 80% DAG) was added to the fryer and heated at 160 and 180 • C for 10 min. One frying cycle consisted of 3 min of frying and 27 min of heating without chicken. There was a total of six cycles (180 min) of frying per day continuous for 3 days without refuelling. Meanwhile, another six systems combining DAG content (40, 60 and 80%) and heating temperature (160 and 180 • C) were established without chicken as controls. Samples of 50 mL of oil were collected after frying and stored at 4 • C every day.

Physico-Chemical Characteristics
The hardness and gumminess of crackers were tested using a texture analyzer (TA-XT plus, Stable Micro System, Godalming, UK) equipped with a P/36 R probe. Each sample was compressed twice to 60% of its original height at 1 mm/s [21]. Each group had 10 parallel samples.
The AV (an important parameter indicating the content of free fatty acids) and POV (an indicator of the degree of oxidation of oils) were examined according to AOCS Official Method Ca 5a-40 [22] and AOCS Official Method Ja 8-87 [23], respectively.
The fatty acid composition was determined with a gas chromatography-flame ionization detector (GC-FID) (Agilent Technologies, Mississauga, ON, Canada) with CP-Sil 88 (60 m × 0.25 mm × 0.2 mm, Dikma Technologies, Beijing, China). The oil was pretreated with KOH-methanol, and then detected with GC-FID. The percentage of fatty acid was calculated using the area normalization method [19].

Determination of 3-MCPDE and GE
According to AOCS Official Method Cd 29a-13 [24], 100-110 mg of oil was accurately weighed in a screw cap tube. Acidified sodium bromide solution was used to convert GE into bromopropanol ester. After that, the solution was reacted with a H 2 SO 4 -methanol solution at 40 • C for 16 h, and then was derivatized using PBA. It was analyzed by gas chromatography-mass spectrometry (SIM) (TQ8050, Shimadzu, Japan) with a capillary column Equity-1 (30 m × 0.25 mm × 1 µm, Supelco, Shanghai, China).

Statistical Analyses
Samples were analyzed in triplicate, and standard deviation (±SD) was reported. Oneway analysis of variance and the Pearson correlation test were performed using the SPSS software (version 14.0 demo, SPSS Inc., Chicago, IL, USA). The partial least squares (PLS) analysis was performed using Minitab 16 software (Minitab Inc., State College, PA, USA).

Physico-Chemical Properties of Baking Trial
The hardness and gumminess of three groups of crackers (40, 60 and 80% olive-based DAG oil) were analyzed ( Figure 1A).
The AV and POV were not significantly different between oil extracted from crackers and oil heated in the oven. However, the AV and POV of both oil extracted from crackers and oil heated in the oven were significantly higher than fresh oil, as shown in Figure 1B,C. Moreover, the AV and POV were higher at 180 • C compared to 160 • C. Overall, the AV was below 0.8 mg/g and the POV did not exceed 0.3 g/100 g in all oil samples.
In terms of fatty acid composition, there was no significant change in oleic acid (C18:1). Oleic acid, the main monounsaturated fatty acid found in olive oil, was increased by 0.14-0.23% after baking (Table 1). In addition, no harmful trans-fatty acids (TFA) were formed during baking.

Intermittent Frying of Chicken
Frying is a repetitive process and the heating duration is relatively longer than baking. The physico-chemical characteristics of the olive-based DAG oil were measured after 3 days of frying at 160 and 180 • C.
As expected, the AV increased proportionally as temperature increased, until it reached 3.30 mg/g ( Figure 2A). With the extension of frying duration, the POV increased slowly and then decreased. Interestingly, the POV reached its peak value (0.23 g/100 g) on the first day and then decreased to 0.19 g/100 g ( Figure 2B). In addition, there was no significant difference (p > 0.05) between frying chicken and control (without chicken).

The Effects of Salt, Duration, Temperature and DAG Content on 3-MCPDE and GE Content in Baking and Frying
When baking crackers with vegetable oil, the addition of salt affected the production of 3-MCPDE [13]. It can be observed in Table 3 that the content of 3-MCPDE was relatively Foods 2022, 11, 4073 6 of 10 low in oil heated in the oven (without salt and other materials) compared to the oil extracted from crackers. In other words, 3-MCPDE was 1.08 mg/kg for fresh oil, 0.97 mg/kg for oil extracted from crackers and 0.84 mg/kg for oil heated in the oven (40% DAG at 160 • C). 1.08 ± 0.01a 0.62 ± 0.01b 0.33 ± 0.01c ND ND ND 60% 1.23 ± 0.02a 1.15 ± 0.05a 1.04 ± 0.02b ND <LOQ ND 80% 1.12 ± 0.07a 0.69 ± 0.03b 0.63 ± 0.01b ND <LOQ <LOQ One-way analysis of variance was used to indicate significant differences between each oil sample (p < 0.05) as shown by different lowercase letters. LOQ, limit of quantification. ND, not detected (below the limit of quantification: GE LOQ = 0.60 mg/kg, GE LOD = 0.24 mg/kg). Table 4 clearly shows that 3-MCPDE decreased to 0.30 mg/kg, while the content of GE slightly increased with the prolonging of heating duration. The maximum content of GE was still below 1 mg/kg. However, a higher level of GE was found in the control after 3 days of frying (1.29 mg/kg). Similarly, under the condition of intermittent deep-frying, 3-MCPDE showed a significant decline, while GE showed an increasing trend of [9]. It was found that the formation of 3-MCPDE during long-term heating was lower than the rate of its decomposition, while GE increased and was relatively stable [7]. One-way analysis of variance was used to indicate significant differences between each frying interval (p < 0.05) as shown by different lowercase letters. LOQ, limit of quantification. ND, not detected. There was no correlation between the production of the contaminants and DAG content. The GE was below the limit of quantification with the increase in DAG content at 160 • C, even after long-term heating (Table 4). After 3 days of heating, the 3-MCPDE was the lowest in the 40% olive-based DAG oil (1.07 mg/kg).

Effect of Temperature, Duration and DAG Content on 3-MCPDE and GE
The influence of temperature, heating duration and DAG content on 3-MCPDE and GE were analyzed using the Pearson correlation test and PLS during baking and frying.
During baking, temperature showed a very strong correlation with 3-MCPDE (r > 0.9) and GE (r > 0.9), while the correlations were not significant between DAG content and both 3-MCPDE (P = 0.365) and GE (p = 0.563). In frying oil, temperature displayed a negative and very strong correlation with 3-MCPDE (r > 0.9) and a strong correlation with GE (0.7 < r < 0.9). Moreover, changes in 3-MCPDE and GE were correlated with the frying duration (r > 0.9). Interestingly, DAG showed moderate correlation with 3-MCPDE (0.5 < r < 0.7). On the other hand, the correlations were not significant between DAG content and GE (p = 0.754).
The PLS model employed had two components with an R 2 value of 0.7775 and a p-value less than 0.05, thus authenticating its validity. The load diagram compares the relative influence of each prediction variable on the response. In this study, the line corresponding to DAG content was short, indicating the weak correlation with 3-MCPDE and GE (Figure 3). The lines of heating temperature and duration were very long, indicating that they were more relevant to 3-MCPDE and GE.

Discussion
The crackers were relatively lower in hardness and gumminess when baked at 160 °C, and there was no significant difference (p > 0.05) among the three groups of crackers. This might be caused by the emulsifying properties of DAG [16]. The physico-chemical characteristics in the baking trial were not significantly different (p > 0.05) between oil extracted from crackers and oil heated. This observation was related to the good thermal stability of high-oleic-acid oil at 180 °C [18].
During frying, the increase in AV can be explained by the increase in free fatty acids due to the hydrolysis reaction [25]. The reduced POV was caused by the decomposition and transformation of the primary oxidation product [26,27]. During frying, the fatty acid composition changed significantly ( Table 2). There is a higher tendency of unsaturated oils to oxidation and polymerization reactions during frying or heating compared to oils with less unsaturation such as olive oil, which has more oleic fatty acid [28]. Based on previously reported data, heating could result in the isomerization of unsaturated fatty acids' double bonds from cis to trans forms [29].
Salt, as a chlorine source, can provide a precursor for the formation of 3-MCPDE [30]. The results indicated that 3-MCPDE was generally lower in oil extracted from crackers and oil heated in the oven than in the fresh oil. In addition, the content of 3-MCPDE was higher in the presence of salt (oil extracted from crackers) under the same baking conditions. In the case of GE, the maximum value of GE was still below the limit of quantification (0.60 mg/kg) among the three groups, indicating that the presence of salt might not be related to the formation of GE in this study [11]. From the findings, 3-MCPDE decreased with the increase in temperature, while GE was lower at 160 °C (Table 4). During

Discussion
The crackers were relatively lower in hardness and gumminess when baked at 160 • C, and there was no significant difference (p > 0.05) among the three groups of crackers. This might be caused by the emulsifying properties of DAG [16]. The physico-chemical characteristics in the baking trial were not significantly different (p > 0.05) between oil extracted from crackers and oil heated. This observation was related to the good thermal stability of high-oleic-acid oil at 180 • C [18].
During frying, the increase in AV can be explained by the increase in free fatty acids due to the hydrolysis reaction [25]. The reduced POV was caused by the decomposition and transformation of the primary oxidation product [26,27]. During frying, the fatty acid composition changed significantly ( Table 2). There is a higher tendency of unsaturated oils to oxidation and polymerization reactions during frying or heating compared to oils with less unsaturation such as olive oil, which has more oleic fatty acid [28]. Based on previously reported data, heating could result in the isomerization of unsaturated fatty acids' double bonds from cis to trans forms [29].
Salt, as a chlorine source, can provide a precursor for the formation of 3-MCPDE [30]. The results indicated that 3-MCPDE was generally lower in oil extracted from crackers and oil heated in the oven than in the fresh oil. In addition, the content of 3-MCPDE was higher in the presence of salt (oil extracted from crackers) under the same baking conditions. In the case of GE, the maximum value of GE was still below the limit of quantification (0.60 mg/kg) among the three groups, indicating that the presence of salt might not be related to the formation of GE in this study [11]. From the findings, 3-MCPDE decreased with the increase in temperature, while GE was lower at 160 • C (Table 4). During heating, GE increased to 1.29 mg/kg with the increase in temperature. Similarly, during deep-frying, 3-MCPDE declined, while GE showed an increasing trend [31]. Researchers have opined that the effect of heating duration is more relevant than temperature [4]. Merkle et al. found that there was no significant difference in the contaminants when frying fish with a mixture of MAG/DAG [32]. In addition, the production of 3-MCPD and free glycidol was confirmed in heated MAG, but not from DAG [12]. DAG could be a precursor in the heating process depending on the conditions. The results of Pearson and PLS analyses suggest that DAG led to the formation of 3-MCPDE and GE in heat-induced processing and was closely related to the conditions, mainly heating temperature and duration. The order of the effects was duration > temperature > DAG content. The previous studies also found that the role of DAG in the formation of 3-MCPDE and GE was not clear, but high temperature was indeed the key factor [33].

Conclusions
It was concluded that olive-based DAG oil is safe and stable to be used in both baking and frying. The temperature and duration of heating treatment affect the physico-chemical properties of the olive-based DAG oil. As the temperature and duration increased, the AV continued to increase, the primary oxidation products were produced and rapidly transformed, and the decomposition of fatty acids was more likely to occur. The 3-MCPDE and GE in crackers and frying oil were lower than the EU limit (1.25 mg/kg and 1 mg/kg, respectively). Under the heating conditions of 160 and 180 • C, 3-MCPDE and GE showed a correlation with duration (r > 0.9). However, the content of DAG had no significant effect on 3-MCPDE and GE (r < 0.5). In conclusion, the factors in forming 3-MCPDE and GE were mostly associated with heating duration, followed by temperature and lastly DAG content.