Isotope-Labeling Studies on the Formation Pathway of Acrolein during

Article pubs.acs.org/JAFC

Isotope-Labeling Studies on the Formation Pathway of Acrolein during Heat Processing of Oils Alice Ewert,† Michael Granvogl,‡ and Peter Schieberle*,†,‡ †

Deutsche Forschungsanstalt für Lebensmittelchemie and Lehrstuhl für Lebensmittelchemie, Technische Universität München, Lise-Meitner-Straße 34, D-85354 Freising, Germany

‡

ABSTRACT: Acrolein (2-propenal) is classified as a foodborne toxicant and was shown to be present in significant amounts in heated edible oils. Up to now, its formation was mainly suggested to be from the glycerol part of triacylglycerides, although a clear influence of the unsaturation of the fatty acid moiety was also obvious in previous studies. To unequivocally clarify the role of the glycerol and the fatty acid parts in acrolein formation, two series of labeled triacylglycerides were synthesized: [13C3]triacylglycerides of stearic, oleic, linoleic, and linolenic acid and [13C54]-triacylglycerides with labeled stearic, oleic, and linoleic acid, but with unlabeled glycerol. Heating of each of the seven intermediates singly in silicon oil and measurement of the formed amounts of labeled and unlabeled acrolein clearly proved the fatty acid backbone as the key precursor structure. Enzymatically synthesized pure linoleic acid and linolenic acid hydroperoxides were shown to be the key intermediates in acrolein formation, thus allowing the discussion of a radical-induced reaction pathway leading to the formation of the aldehyde. Surprisingly, although several oils contained high amounts of acrolein after heating, deep-fried foods themselves, such as donuts or French fries, were low in the aldehyde. KEYWORDS: Acrolein, foodborne toxicant, frying, fatty acid hydroperoxides, labeling studies

■

Umano and Shibamoto7 assumed the necessity of free fatty acids for acrolein formation, and acrolein generation was suggested from linolenic acid via its 13-hydroperoxide as key intermediate.10 Esterbauer et al.32 proposed that acrolein originates from the center of the fatty acid chain involving βcleavage of the 9-hydroperoxide. However, this formation mechanism needs ω-3 or ω-6 fatty acids, respectively, as precursors, but it remains unclear how acrolein can also be generated from, for example, oleic acid. The ability of free fatty acids as well as their esters to generate acrolein was also observed by Tamura et al.33 as well as by Pan et al.34 Recently, it was shown by our group that the amount of acrolein generated during heat processing of different edible fats and oils was well correlated with the content of unsaturated fatty acids in the oil sample.6 However, the reaction pathway leading to acrolein from triacylglycerides is still unclear. Thus, the objectives of this study were, first, to identify the most efficient precursor structures using several 13 C-labeled triacylglycerides as well as unlabeled fatty acid hydroperoxides in a model system with silicon oil, and, second, to quantitate acrolein in different types of fried foods and to correlate the amounts with the concentrations present in the frying oil.

INTRODUCTION Acrolein (2-propenal) is a highly reactive aldehyde commonly formed by incomplete combustion of organic material. Thus, the aldehyde has been found in cigarette smoke, fireplaces, incense, candles, and fuels.1−4 Furthermore, acrolein has been detected in the exhaust of kitchens and thermally processed cooking oils5−10 as well as in biological samples.11,12 Acrolein has been classified by the U.S. Environmental Protection Agency (EPA) as a pulmonary toxicant and is among the air toxicants posing the greatest health risk.13−18 Its toxicity is related to a reaction with, for example, the sulfhydryl group of cysteine in proteins and glutathione, leading to interferences in the intermediary cell metabolism and to modified proteins by introducing inter- and intramolecular cross-links.8,19,20 Furthermore, administration of acrolein to rats decreased lung and liver levels of glutathione, ascorbic acid, and α-tocopherol.21 In addition, acrolein has been reported to induce the generation of oxygen radicals22,23 and has been associated with several illnesses, such as atherosclerosis,24 Alzheimer’s disease,25 and cancer.26,27 Despite extensive toxicological studies, data on the formation pathway of acrolein from food constituents are scarcely available in the literature. Presently, two published major hypotheses can be found: acrolein may be generated by either partial hydrolysis of triacylglycerides, followed by dehydration of liberated glycerol to acrolein,28,29 or, in particular, a degradation of the unsaturated fatty acid backbone.7,30 Fujisaki et al.26 showed that the amount of acrolein in processed oil was well correlated with the experimental oxygen concentration (2−20%) in the air, and Bastos and Pereira31 observed a clear increase of acrolein during oil processing in the presence of metal ions, suggesting that hydroperoxide formation is involved. © 2014 American Chemical Society

■

MATERIALS AND METHODS

CAUTION: Acrolein, as well as [13C3]-acrolein, is classified as toxic and may cause cancer. Be careful when handling these chemicals. Food Samples. Vegetable oils, potato chips, precooked French fries, and donuts were purchased in local supermarkets. Received: Revised: Accepted: Published: 8524

March July 2, July 4, July 4,

31, 2014 2014 2014 2014

dx.doi.org/10.1021/jf501527u | J. Agric. Food Chem. 2014, 62, 8524−8529

Journal of Agricultural and Food Chemistry

Article

[13C54]-Glyceryl trioleate: MS-APCI+, m/z (%) 940.6 ([M + H]+; 100), 639.5 (35), 663.2 (30), 439.2 (25), 383.2 (15), 523.3 (10). [13C3]-Glyceryl trilinoleate: MS-APCI+, m/z (%) 883.4 ([M + H]+; 100), 620.2 (25), 602.0 (20), 390.9 (15), 278.9 (10). [13C54]-Glyceryl trilinoleate. MS-APCI+, m/z (%) 934.6 ([M + H]+; 100), 635.5 (35), 600.3 (10). [13C3]-Glyceryl trilinolenate: MS-APCI+, m/z (%) 877.3 ([M + H]+; 100), 739.3 (50), 598.3 (20), 321.4 (10), 472.5 (10). Preparation of Hydroperoxides. Fatty acid hydroperoxides were prepared as previously described.36 Either linoleic acid or linolenic acid (200 mg each) was treated with soybean lipoxygenase type I (4 mg) for 60 min under an atmosphere of pure oxygen at room temperature. Separation of the hydroperoxides from the residual fatty acids was performed by means of preparative flash-column chromatography using a LiChroprep Diol stationary phase and n-hexane/2-propanol/ methanol/water (900:50:2:1; v/v/v/v) as the mobile phase. Due to the high regio- and stereospecifity of LOX I (94−97% of 13hydroperoxy-(Z,E)-9,11-octadienoic acid), no further separation of the isomers was needed.37 The 13-hydroperoxides were identified by LCMS (APCI+). Quantitation of the fatty acid hydroperoxides was performed photometrically.38 13-Hydroperoxy-(Z,E)-9,11-octadecadienoic acid: MS-APCI+, m/ z (%) 279.9 (100), 294.9 (50), 207.0 (40), 313.0 ([M + H]+; 35), 608.2 (30), 625.3 (25), 592.2 (20), 349.1 (20). 13-Hydroperoxy-(Z,E,Z)-9,11,15-octadecatrienoic acid: MSAPCI+, m/z (%) 311.1 ([M + H]+; 100), 277.8 (60), 293.1 (60), 603.5 (40), 587.6 (20), 621.4 (10). Heating of Oils. Rapeseed oil (100 g) or a commercial frying fat (100 g) was mixed either with D,L-all-rac-α-tocopherol (100 mg) or with L-ascorbic acid (100 mg; finely powdered, forming a suspension), respectively, before heat treatment. For comparison, rapeseed oil or the frying fat was analyzed without antioxidants. All samples were heated at 180 °C for 4 h. Degradation of Labeled Triacylglycerides or Monohydroperoxides. Silicon oil (5 mL) was mixed singly either with each of the isotopically labeled triacylglycerides or with the hydroperoxides as well as with the respective free fatty acids (20 mg each). A control experiment was performed by heating silicon oil without any additives to prove the absence of signals and background interferences within the mass traces of acrolein (m/z 57 in chemical ionization (CI) mode) or [13C3]-acrolein (m/z 60), respectively. Samples were heated at 140 °C for 4 h under continuous stirring in closed glass vessels in a thermocontrolled metal block. Preparation of Food and Frying. Donuts. Yeast (42 g) was suspended in warm water, and wheat flour (500 g), whole milk (150 mL), sugar (50 g), eggs (100 g), and water (25 mL) were added to obtain a homogeneous dough, which was divided into pieces of about 50 g and stored at 30 °C for 2 h before frying. Dif ferent foods (potato chips, precooked French f ries, and donuts) were fried at 180 °C either in rapeseed oil, in safflower oil, or in a commercial frying fat: potato chips (peeled, 1.5 mm of cross section) for 2 min, French fries for 3.5 min, and donuts for 6 min. To provide similar reaction conditions, the first five batches of the respective fried foods were discarded, and only the sixth batch was used for analysis. The frying process (100 g of each food) was carried out in a DF320 Kenwood deep-fryer (Heusenstamm, Germany) containing 2.3 L of oil, which was preheated for 10 min prior to frying. Quantitation of Acrolein. Model System with Antioxidants, Free Fatty Acids, and Hydroperoxides. Immediately after cooling to room temperature, the internal standard [13C3]-acrolein (dissolved in dichloromethane; 1−214 μg, depending on the amounts determined in preliminary experiments) was added to an aliquot of each sample (5 g). The sample was vortexed (1 min) and ultrasonified (5 min), and aliquots (1 g) were weighed into gastight headspace vials (20 mL), which were immediately capped and stirred for 20 min at room temperature prior to headspace GC-MS measurement (CI mode) as recently described.6 Model System with Labeled Triacylglycerides. Quantitation was performed by preparing acrolein standards in dichloromethane from a stock solution (1.372 mg/mL) in the concentration range from 1 to

Chemicals. [13C3]-Acrolein was obtained from Isotec Stable Isotopes (St. Louis, MO), and acrolein and dicyclohexyl carbodiimide were from Fluka (Steinheim, Germany). Glycerol, stearic acid, oleic acid, α-linoleic acid, α-linolenic acid, [13C3]-glycerol, [13C18]-stearic acid, [13C18]-oleic acid, [13C18]-linoleic acid, 4-(dimethylamino)pyridine, soybean lipoxygenase type I, and xylenol orange were obtained from Sigma-Aldrich (Taufkirchen, Germany). L-Ascorbic acid, D,L-all-rac-α-tocopherol, silicon oil, sulfuric acid (96%), and ferrous ammonium sulfate hexahydrate were from VWR International (Darmstadt, Germany). Silica gel (230−400 mesh, grade 60) and LiChroprep Diol (40−63 μm) were from Merck (Darmstadt). Argon, oxygen, and liquid nitrogen were obtained from Linde (Munich, Germany). All other reagents were of analytical grade. Synthesis of 13C-Labeled Triacylglycerides. The glycerides were prepared according to a general procedure published previously.35 Either unlabeled glycerol (0.15 mmol) was reacted with each of the carbon-13 labeled fatty acids (0.5 mmol), or [13C3]glycerol was reacted with each of the unlabeled fatty acids dissolved in dichloromethane (2.5 mL) under an argon atmosphere. The solution was stirred and cooled to 0 °C while dicyclohexyl carbodiimide (0.75 mmol) and 4-(dimethylamino)pyridine (0.05 mmol) were added within 5 min. The reaction mixture was stirred for 3 h at room temperature, and the dicyclohexyl urea formed was removed by filtration. After extraction with hydrochloric acid (0.5 mol/L; total volume = 40 mL) and an aqueous saturated sodium bicarbonate solution (total volume = 40 mL), the organic phase was dried over anhydrous sodium sulfate and concentrated to a final volume of ∼1 mL using a rotary evaporator. The crude product obtained was purified by column chromatography (220 mm × 22 mm i.d.) on silica gel (230−400 mesh, grade 60). Elution was performed with dichloromethane/n-pentane (8:2; v/v). Structures of the seven target compounds were confirmed by LC-MS in positive atmospheric pressure chemical ionization (APCI+) mode as well as by 1H and 13C NMR spectroscopy ([13C3]-labeled compounds). Yields of synthesized 13 C-labeled triacylglycerides were determined by LC-MS (APCI+) and calculated from external five-point calibration curves using defined triacylglyceride solutions. Reference solutions were prepared in dichloromethane from a stock solution (5 mg/mL) in the concentration range between 10 and 200 μg/mL. Analytical Data of the Synthesized Triacylglycerides. [13C3]Glyceryl tristearate: MS-APCI+, m/z (%) 611.3 (100), 551.3 (30), 648.9 (15), 895.4 ([M + H]+; 10). 1 H NMR (400 MHz; CDCl3; 298 K): δ 0.90 (t, 7.0 Hz, H-C18′/HC18′′), 1.21−1.39 (m, H-C4′−H-C17′/H-C4′′−H-C17′′), 1.55−1.70 (m, H-C3′/H-C3′′), 2.33 (t, 7.6 Hz, H-C2′), 2.34 (t, 7.5 Hz, H-C2′′), 4.17 (dm, 145.6 Hz, Ha-C1), 4.23 (dm, 148.3 Hz, Hb-C1), 5.29 (dm, 149.2 Hz, H-C2). Numbering of carbon atoms refers to Figure 1.

Figure 1. Structure of [13C3]-labeled glyceryl tristearate. (■) labeling.

13

C-

13 C NMR (100 MHz; CDCl3; 298 K): δ 14.1 (C18′/C18′′), 22.7 (C17′/C17′′), 24.88 (C3′), 24.92 (C-3′′), 29.0-29.9 (C4′-C15′/C4′′C15′′), 31.9 (C16′/C16′′), 34.1 (C2′), 34.2 (C2′′), 62.1 (d, 43.6 Hz, C1), 68.9 (t, 43.3 Hz, C2), 172.9 (C1′′), 173.3 (C1′). [13C54]-Glyceryl tristearate: MS-APCI+, m/z (%) 643.3 (100), 579.3 (30), 880.3 (15), 946.3 ([M + H]+; 10). [13C3]-Glyceryl trioleate: MS-APCI+, m/z (%) 888.9 ([M + H]+; 100), 606.8 (30), 270.0 (30), 551.2 (30), 458.0 (25), 523.2 (20), 663.3 (20), 796.3 (10).

8525



Article

500 μg acrolein/mL containing a fixed amount of internal standard (10 μg/mL). Calibration curves were newly prepared prior to each measurement and showed good-quality fit (r2 > 0.993). For acrolein quantitation in foods, the material was frozen in liquid nitrogen and ground with a laboratory mill. Aliquots (5 g) were weighed into gastight headspace vials (100 mL) containing silicon oil (30 mL). After addition of [13C3]-acrolein (100 ng), the vials were immediately capped, and the suspensions were stirred for 20 min at 30 °C prior to headspace GC-MS measurement (CI mode) as recently described.6 High-Performance Liquid Chromatography−Mass Spectrometry (HPLC-MS). Mass spectrometry was performed by means of an LCQ ion trap mass spectrometer connected to an SCM 1000 high-performance liquid chromatography system (both Finnigan MAT, Bremen, Germany) equipped with a 150 mm × 2.0 mm i.d., 4 μm Synergi Polar RP 80 Å column kept at 30 °C and connected to a 4 mm × 2.0 mm i.d. polar RP precolumn (both Phenomenex, Aschaffenburg, Germany). Elution was done isocratically using a mixture of n-hexane/2-propanol/methanol/water (900:50:2:1; v/v/v/ v). Flow rate was set at 0.8 mL/min, and injection volume was 50 μL. Analytes were detected in full scan using APCI+ mode. The sheath and auxiliary gas nitrogen nebulized the effluent with flows of 80 and 20 arb, respectively. Further parameters were the following: discharge current, 5 μA; vaporizer temperature, 400 °C; heated capillary temperature, 200 °C; capillary voltage, 4 V; and tube lens offset, 30 V. The ion trap operated at a helium pressure of 1.3 × 10−4 kPa. Nuclear Magnetic Resonance (NMR) Spectroscopy. 1H and 13 C NMR spectra were recorded using a 400 MHz DRX spectrometer (Bruker, Rheinstetten, Germany). Sample was dissolved in deuterated chloroform (CDCl3; 0.7 mL), and chemical shifts (1H NMR) were determined using either tetramethylsilane (0.03%) as the internal standard or the carbon signal of CDCl3 at 77.0 ppm (13C NMR), respectively.

Figure 2. Mass spectrum (APCI+) of [13C3]-glyceryl tristearate.

reduced when the antioxidant BHA was added to the cooking oil. The more effective reduction of acrolein found for αtocopherol in comparison to L-ascorbic acid might be explained by the better solubility of α-tocopherol in the fatty matrix. On the basis of these results, it can be assumed that acrolein generation in oils involves a radical-induced degradation of the triacylglycerides. Stable Isotope Labeling Studies with Triacylglycerides. To get further insight into the formation pathway of acrolein from triacylglycerides, experiments using isotopically labeled precursors were undertaken. Recently, it was shown that acrolein formation from oils containing linoleic or linolenic acid is significantly higher than from oils mainly consisting of saturated or monounsaturated fatty acids.6 However, up to now, no systematic investigations on the formation pathway of acrolein were reported in the literature. Thus, two kinds of labeled triacylglycerides were synthesized: in one group of four triacylglycerides only the glycerol backbone was carbon-13 labeled, and in the other group of three labeled triacylglycerides the entire fatty acid moiety of either stearic, oleic, and linoleic acid was labeled, whereas glycerol was left unlabeled. All compounds were characterized by LC-MS analysis in APCI+ mode and by NMR spectroscopy ([13C3]-labeled compounds). As an example, the mass spectrum of [13C3]-tristearate (Figure 2) and the respective 13 C NMR spectrum (Figure 3) are shown. In the next step, the four triacylglycerides labeled in the glycerol backbone were individually heated at 140 °C for 4 h in silicon oil, and the formation of acrolein was measured. In a control experiment, the oil did not show any mass traces at m/z 57 (acrolein) or m/z 60 ([13C3]-acrolein) when heated for 4 h. When [13C3]-trioleate was heated (expt 2) (Table 2), the amount of unlabeled acrolein was higher by a factor of >8 compared to the carbon-13 labeled acrolein, clearly indicating the fatty acid moiety as the main precursor structure. The labeled acrolein formed from [13C3]-trilinoleate and [13C3]trilinolenate (expts 3 and 4) (Table 2) was also as low as found for [13C3]-trioleate, but the amounts of unlabeled acrolein increased with the number of double bonds in the fatty acid moiety, amounting to 6240 μmol/mol from the linolenic acid residues. By contrast, from the saturated [13C3]-tristearate no acrolein was formed (expt. 1) (Table 2). Altogether, these results proved that the acrolein originates from the fatty acid

■

RESULTS AND DISCUSSION Model Studies. Effect of Antioxidants on Acrolein Formation. To check if a radical mechanism is involved in Table 1. Amounts of Acrolein Formed by Heating of Rapeseed Oil and a Common Frying Fat at 180 °C for 4 h in the Presence or Absence of Antioxidants concn (mg/kg)a in expt

additive

rapeseed oil

frying fat

1 2 3

α-tocopherolb ascorbic acidb withoutc

9.8 (7.6−12.1) 16.9 (15.8−17.1) 67.5 (63.2−69.7)

19.6 (17.8−21.4) 35.4 (33.4−37.6) 51.8 (48.9−53.1)

a

Mean values of three independent experiments. Concentration ranges in parentheses. b100 mg per 100 g of oil and fat. cControl experiment without antioxidant addition.

acrolein generation, in a first series of experiments, the antioxidant α-tocopherol or L-ascorbic acid was singly added to rapeseed oil as well as to a commercial frying fat before heating. The oils were selected on the basis of their different fatty acid compositions: the frying fat consisted of >50% of saturated fatty acids, whereas rapeseed oil mainly consisted of unsaturated fatty acids. Compared to the samples heated without antioxidant addition (expt 3) (Table 1), in the presence of both antioxidants, acrolein formation was strongly reduced during heating, although to a lesser extent in the frying fat (expts 1 and 2) (Table 1). In the presence of α-tocopherol, which was the more effective antioxidant, acrolein concentration in rapeseed oil was reduced to ∼15% compared to the oil heated without the antioxidant. The results supported data on wok-cooking19 indicating that acrolein emissions were 8526



Article

Figure 3. 13C NMR spectrum of [13C3]-glyceryl tristearate. (■) 13C-labeling.

Table 2. Amounts of Unlabeled Acrolein and [13C]-Acrolein Formed by Heating of Triacylglycerides Labeled either in the Glycerol Backbone ([13C3]-Compounds) or in the Fatty Acid Moieties ([13C54]-Compounds) at 140 °C for 4 h

Table 3. Amounts of Acrolein Formed either from Free Linoleic Acid and Free Linolenic Acid or from 13Hydroperoxy-(Z,E)-9,11-octadecadienoic Acid (13-HPOD) and 13-Hydroperoxy-(Z,E,Z)-9,11,15-octadecatrienoic Acid (13-HPOT) during Heating at 140 °C for 4 h

concna (μmol/mol precursor) expt

precursor

[13C3]-acrolein

acrolein

expt

precursor

1 2 3 4

[13C3]-tristearate [13C3]-trioleate [13C3]-trilinoleate [13C3]-trilinolenate

ndb 30 (10−40) 30 (20−40) 40 (30−60)

nd 260 (240−280) 710 (680−740) 6240 (5980−6310)

1 2 3 4

linoleic acid linolenic acid 13-HPOD 13-HPOT

5 6 7

13

[ C54]-tristearate [13C54]-trioleate [13C54]-trilinoleate

nd 240 (220−260) 780 (760−810)

a

nd 20 (10−40) 20 (10−30)

concna (μmol/mol precursor) 350 2210 560 3100

(260−390) (2050−2390) (210−840) (2930−3310)

Mean values of triplicates. Concentration ranges in parentheses.

acrolein were formed as compared to unlabeled acrolein, with trilinoleate being more effective than trioleate (expts 6 and 7) (Table 2). Again, the labeled tristearate did not form acrolein, which leads to the suggestion that at least one double bond in the fatty acid moiety is needed to initiate the formation of acrolein from the fatty acid moiety or the glycerol backbone. Thus, these results are additional proof that the fatty acid moiety plays the major role in acrolein formation. For the first time, the high precursor potential of unsaturated fatty acid moieties compared to the glycerol backbone, which was considered as a main acrolein precursor until now,28 was proven. Hence, this model system is well in line with former findings that linseed oil containing high amounts of linolenic acid showed the highest precursor potential.6 Stable Isotope Labeling Studies with Fatty Acid Hydroperoxides. To get more information about the transient reaction steps of acrolein formation from fatty acids, the precursor potential of fatty acid hydroperoxides was studied. The high regio- and stereospecifity of soybean lipoxygenase type I was used to synthesize the 13-hydroperoxides of linoleic as well as of linolenic acid. The mass spectrum of the 13hydroperoxide from linolenic acid showed high fragmentation (APCI+) (Figure 4), which was advantageous for the identification of the target molecule. Single-model experiments with defined amounts of either free linoleic acid or linolenic acid as well as the respective 13hydroperoxides were performed in silicon oil at 140 °C for 4 h. Also, the free linoleic acid and the free linolenic acid generated acrolein, with linolenic acid being the better precursor (expts 1

a

Mean values of three independent heat processings. Concentration ranges in parentheses. bNot detected (limit of detection = 0.25 μmol).

Figure 4. Mass spectrum (APCI+) of 13-hydroperoxy-(E,Z)-9,11octadecadienoic acid.

moiety of triacylglycerides containing at least one double bond. Further experiments with the triglycerides containing the labeled fatty acid moieties corroborated this finding. From these precursor triacylglycerides, higher amounts of [13C3]8527



Article

Figure 5. Reaction pathway leading to acrolein via 9-hydroperoxy-(E,Z)-10,12-octadecadienoic acid (following the idea of ref 32).

Table 4. Influence of the Type of Fat Used in the Frying Process of Several Foods on Acrolein Formation acrolein concna (μg/kg) in food after frying using

a b

sample

safflower oil

rapeseed oil

frying fat

commercial samplesb

potato chips French fries donuts

17.0 (15.3−17.9) 18.1 (15.5−19.8) 14.4 (11.3−15.5)

23.3 (22.0−24.9) 14.8 (12.6−17.1) 15.0 (14.1−15.9)

16.3 (15.1−17.1) 19.9 (17.4−22.5) 14.1 (11.7−16.3)

18.3 (16.9−19.7) 18.5 (16.4−19.8) 16.9 (15.3−18.0)

Mean values from three independent frying processes. Concentration ranges in parentheses. LoD = 3 μg/kg, LoQ = 9 μg/kg according to ref 39. For comparison.

conditions, 180 °C/24 h),6 the fried food itself does not seem to be a significant source of free acrolein in the human diet. In conclusion, the formation pathways of the toxicologically relevant acrolein were proven on the basis of labeling experiments. A formation pathway from the carbon backbone of unsaturated fatty acids, in particular, linolenic acid via the 13hydroperoxide, seems very likely, whereas the glycerol backbone does not play a key role. Surprisingly, however, acrolein was low in different types of foods, although the investigated oils and fat were high in acrolein. This result suggests that acrolein might be bound to, for example, macromolecular structures in foods, thus no longer existing in the free state. Experiments to study this assumption are currently underway due to the fact that Watzek et al.40 were able to show that the amounts of acrolein-related mercapturic acids excreted in urine after potato crisp intake were much higher than calculated from the previously analyzed amounts of free acrolein in the crisps.

and 2) (Table 3). Heating of the respective hydroperoxides led to higher amounts of acrolein with the 13-linolenic acid hydroperoxide generating higher amounts (expts 3 and 4) (Table 3). Thus, these results suggest that the formation of hydroperoxides is an important step in acrolein formation, leading to the introduction of the oxygen moiety. In the literature, it was proposed that acrolein may originate from the unsaturated part of the aliphatic chain, that is, by βcleavage of 9-hydroperoxides resulting in the respective aldehyde and an alkyl fragment. By Hock fragmentation, the released carbonyls lead to two aldehyde fragments (Figure 5). Contrary to the suggested mechanism via linolenic acid 13hydroperoxide,10 this pathway may also explain acrolein generation from the monounsaturated oleic acid. To the best of our knowledge, these results are the first proof for acrolein generation from fatty acid hydroperoxides. Identification and Quantitation of Acrolein in Fried Foods. In the next experiments, it was clarified whether a high concentration of acrolein in the frying oil must consequently result in high amounts in the fried food. Thus, French fries, donuts, and potato chips were fried either in oils high in unsaturated fatty acids, such as safflower oil and rapeseed oil, or in a commercial frying fat. As previously shown,6 rapeseed oil in particular generated the highest amounts of acrolein simply by heating without foods. Acrolein quantitation in the selfprepared as well as in the commercial samples revealed comparable amounts of acrolein, nearly independently from the type of frying oil used (Table 4). Acrolein concentrations in fried potato chips ranged from 16.3 μg/kg prepared with a common frying fat to 23.3 μg/kg using rapeseed oil, and the analysis of commercial potato chips revealed an acrolein concentration of 18.3 μg/kg. The average acrolein content in French fries ranged from 14.8 μg/kg prepared in rapeseed oil to 19.9 μg/kg in frying fat. Analysis of fried donuts showed slightly lower acrolein concentrations compared to the other food samples with 14.1 μg/kg in donuts fried in frying fat, 15.0 μg/ kg in donuts prepared in rapeseed oil, and 16.9 μg/kg in donuts from a local bakery (Table 4). Compared to the acrolein amounts measured in heated fats and oils from previous studies (46.8 mg/kg of safflower oil, 55.8 mg/kg of frying fat, 157 mg/kg of rapeseed oil; heating

■

AUTHOR INFORMATION

Corresponding Author

*(P.S.) Phone: +49 8161 71 2932. Fax: +49 8161 71 2970. Email: [email protected]. Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS The authors thank Dr. Johanna Kreiβl and Dr. Johannes Polster for measurement and interpretation of the NMR spectra. REFERENCES

(1) Ho, S. S. H.; Yu, J. Z. Feasibility of collection and analysis of airborne carbonyls by on-sorbent derivatization and thermal desorption. Anal. Chem. 2002, 74, 1232−1240. (2) Kataoka, H.; Kondo, T.; Sumida, A. Gas chromatographic determination of aldehydes in combustion smoke samples. Anal. Chim. Acta 1998, 358, 269−275. (3) Singer, B. C.; Hodgson, A. T. Gas-phase organics in environmental tobacco smoke. 1. Effects of smoking rate, ventilation, and furnishing level on emission factors. Environ. Sci. Technol. 2002, 5, 846−853. (4) Gilbert, N. L.; Guay, M.; Miller, J. D.; Judek, S.; Chan, C. C.; Dales, R. E. Levels and determinants of formaldehyde, acetaldehyde,

8528



Article

and acrolein in residential indoor air in Prince Edward Island, Canada. Environ. Res. 2005, 99, 11−17. (5) Seaman, V. Y.; Bennett, D. H.; Cahill, T. M. Indoor acrolein emission and decay rates resulting from domestic cooking events. Atmos. Environ. 2009, 43, 6199−6204. (6) Ewert, A.; Granvogl, M.; Schieberle, P. Development of two stable isotope dilution assays for the quantitation of acrolein in heatprocessed fats. J. Agric. Food Chem. 2011, 59, 3582−3589. (7) Umano, K.; Shibamoto, T. Analysis of acrolein from heated cooking oils and beef fat. J. Agric. Food Chem. 1987, 35, 14−18. (8) Casella, I. G.; Contursi, M. Quantitative analysis of acrolein in heated vegetable oils by liquid chromatography with pulsed electrochemical detection. J. Agric. Food Chem. 2004, 52, 5816−5821. (9) Fullana, A.; Carbonell-Barrachina, A. A.; Sidhu, S. Comparison of volatile aldehydes present in the cooking fumes of extra virgin olive, olive, and canola oils. J. Agric. Food Chem. 2004, 52, 5207−5214. (10) Da Silva, T. O.; de Paula Pereira, P. A. Influence of time, surfaceto-volume ratio, and heating process (continuous or intermittent) on the emission rates of selected carbonyl compounds during thermal oxidation of palm and soybean oils. J. Agric. Food Chem. 2008, 56, 3129−3135. (11) Takamoto, S.; Sakura, N.; Namera, A.; Yashiki, M. Monitoring of urinary acrolein concentration in patients receiving cyclophosphamide and ifosphamide. J. Chromatogr., B 2004, 806, 59−63. (12) Paci, A.; Rieutord, A.; Guillaume, D.; Traoré, F.; Ropenga, J.; Husson, H.-P.; Brion, F. Quantitative high-performance liquid chromatographic determination of acrolein in plasma after derivatization with Luminarin® 3. J. Chromatogr., B 2000, 739, 239−246. (13) International Agency for Research on Cancer. Acrolein. IARC Monogr. Eval. Carcinog. Risks Hum. 1995, 63, 337−372. (14) Yoshida, M.; Tomitori, H.; Machi, Y.; Hagihara, M.; Higashi, K.; Goda, H.; Ohya, T.; Niitsu, M.; Kashiwagi, K.; Igarashi, K. Acrolein toxicity: comparison with reactive oxygen species. Biochem. Biophys. Res. Commun. 2009, 378, 313−318. (15) LoPachin, R. M.; Gavin, T.; Petersen, D. R.; Barber, D. S. Molecular mechanism of 4-hydroxy-2-nonenal and acrolein toxicity: nucleophilic targets and adduct formation. Chem. Res. Toxicol. 2009, 22, 1499−1508. (16) Horton, N. D.; Mamiya, B. M.; Kehrer, J. P. Relationships between cell density, glutathione and proliferation of A549 human lung adenocarcinoma cells treated with acrolein. Toxicology 1997, 122, 111−122. (17) Rudra, P. K.; Krokan, H. E. Acrolein cytotoxicity and glutathione depletion in n-3 fatty acid-sensitive and -resistant human tumor cells. Anticancer Res. 1999, 19, 461−469. (18) U.S. Environmental Protection Agency.. Control of hazardous air pollutants from mobile sources: final rule. 40 CFR Parts 59, 80, 85, and 86. Fed. Regist. 2007, 72, 8428−8570. (19) Shields, P. G.; Xu, G. X.; Blot, W. J.; Fraumeni, J. F., Jr.; Trivers, G. E.; Pellizzari, E. D.; Qu, Y. H.; Gao, Y. T.; Harris, C. C. Mutagens from heated Chinese and U.S. cooking oils. J. Natl. Cancer Inst. 1995, 87, 836−841. (20) Beauchamp, R. O., Jr.; Andjelkovich, D. A.; Kligerman, A. D.; Morgan, K. T.; Heck, H. D. A critical review of the literature on acrolein toxicity. Crit. Rev. Toxicol. 1985, 14, 309−380. (21) Arumugam, N.; Sivakumar, V.; Thanislass, J.; Sadasivan Pillai, K.; Niranjali Devaraj, S.; Devaraj, H. Acute pulmonary toxicity of acrolein in rats − underlying mechanism. Toxicol. Lett. 1999, 104, 189−194. (22) Adams, J. D., Jr.; Klaidman, L. K. Acrolein-induces oxygen radical formation. Free Radical Biol. Med. 1993, 15, 187−193. (23) Patel, J. M. Stimulation of cyclophosphamide-induced pulmonary microsomal lipid peroxidation by oxygen. Toxicology 1987, 45, 79−91. (24) Srivastava, S.; Sithu, S. D.; Vladykovskaya, E.; Haberzettl, P.; Hoetker, D. J.; Siddiqui, M. A.; Conklin, D. J.; D’Souza, S. E.; Bhatnagar, A. Oral exposure to acrolein exacerbates atherosclerosis in apoE-null mice. Atherosclerosis 2011, 215, 301−308.

(25) Calingasan, N. Y.; Uchida, K.; Gibson, G. E. Protein-bound acrolein: a novel marker of oxidative stress in Alzheimer’s disease. J. Neurochem. 1999, 72, 751−756. (26) Fujisaki, M.; Endo, Y.; Fujimoto, K. Retardation of volatile aldehyde formation in the exhaust of frying oil by heating under low oxygen atmospheres. J. Am. Oil Chem. Soc. 2002, 79, 909−914. (27) Crook, T. R.; Souhami, R. L.; McLean, A. E. Cytotoxicity, DNA cross-linking, and single strand breaks induced by activated cyclophosphamide and acrolein in human leukemia cells. Cancer Res. 1986, 46, 5029−5034. (28) Nawar, W. Thermal and radiolytic decomposition of lipids. In Chemical Changes in Food during Processing; Richardson, T., Finley, J., Eds.; AVI: Westport, CT, 1985; pp 79−105. (29) Zhu, X.; Wang, K.; Zhu, J.; Koga, M. Analysis of cooking oil fumes by ultraviolet spectrometry and gas chromatography-mass spectrometry. J. Agric. Food Chem. 2001, 49, 4790−4794. (30) Schauer, J. J.; Kleeman, M. J.; Cass, G. R.; Simoneit, B. R. Measurement of emissions from air pollution sources. 4. C1-C27 organic compounds from cooking with seed oils. Environ. Sci. Technol. 2002, 36, 567−575. (31) Bastos, L. C. S.; de Paula Pereira, P. A. Influence of heating time and metal ions on the amount of free fatty acids and formation rates of selected carbonyl compounds during the thermal oxidation of canola oil. J. Agric. Food Chem. 2010, 58, 12777−12783. (32) Esterbauer, H.; Schaur, R. J.; Zollner, H. Chemistry and biochemistry of 4-hydroxynonenal, malonaldehyde and related aldehydes. Free Radical Biol. Med. 1991, 11, 81−128. (33) Tamura, H.; Kitta, K.; Shibamoto, T. Formation of reactive aldehydes from fatty acids in a Fe2+/H2O2 oxidation system. J. Agric. Food Chem. 1991, 39, 439−442. (34) Pan, X.; Kaneko, H.; Ushio, H.; Ohshima, T. Oxidation of all-cis7,10,13,16,19-docosapentaenoic acid ethyl ester. Hydroperoxide distribution and volatile characterization. Eur. J. Lipid Sci. Technol. 2005, 107, 228−238. (35) Neises, B.; Steglich, W. Simple method for the esterification of carboxylic acids. Angew. Chem., Int. Ed. Engl. 1978, 17, 522−524. (36) Schieberle, P.; Grosch, W.; Kexel, H.; Schmidt, H.-L. A study of oxygen isotope scrambling in the enzymic and non-enzymic oxidation of linoleic acid. Biochim. Biophys. Acta 1981, 666, 322−326. (37) Browne, R. W.; Armstrong, D. HPLC analysis of lipid-derived polyunsaturated fatty acid peroxidation products in oxidatively modified human plasma. Clin. Chem. 2000, 46, 829−836. (38) Jiang, Z.-Y.; Woollard, A. C. S.; Wolff, S. P. Lipid hydroperoxide measurement by oxidation of Fe2+ in the presence of xylenol orange. Comparison with the TBA assay and an iodometric method. Lipids 1991, 26, 853−856. (39) Granvogl, M. Development of three stable isotope dilution assays for the quantitation of (E)-2-butenal (crotonaldehyde) in heatprocessed edible fats and oils as well as in food. J. Agric. Food Chem. 2014, 62, 1272−1282. (40) Watzek, N.; Scherbl, D.; Feld, J.; Berger, F.; Doroshyenko, O.; Fuhr, U.; Tomalik-Scharte, D.; Baum, M.; Eisenbrand, G.; Richling, E. Profiling of mercapturic acids of acrolein and acrylamide in human urine after consumption of potato crisps. Mol. Nutr. Food Res. 2012, 56, 1825−1837.

8529


Isotope-Labeling Studies on the Formation Pathway of Acrolein during

Recommend Documents