Article

pubs.acs.org/JAFC

Formation of 3Monochloro-1,2-propanediol (3-MCPD) Di- and

Monoesters from Tristearoylglycerol (TSG) and the Potential
Catalytic Eect of Fe2+ and Fe3+
Zhongfei Zhang,, Boyan Gao,,, Xiaowei Zhang, Yuanrong Jiang,# Xuebing Xu,#
and Liangli (Lucy) Yu*,,

Institute of Food and Nutraceutical Science, School of Agriculture and Biology, Shanghai Jiao Tong University, Shanghai 200240,
China

Department of Nutrition and Food Science, University of Maryland, College Park, Maryland 20742, United States
#
Wilmar Biotechnology Research and Development Center (Shanghai) Company, Ltd., 118 Gaodong Road, Pudong New District,
Shanghai 200137, China
*
S Supporting Information

ABSTRACT: This study investigated whether and how triacylglycerol (TAG) may serve as a precursor for 3-monochloro-1,2-
propanediol (3-MCPD) fatty acid ester formation using tristearoylglycerol (TSG). TSG was reacted with inorganic chloride
compounds including NaCl, KCl, FeCl2, CuCl2, ZnCl2, FeCl3 and dry HCl, or organic chlorine compound lindane at dierent
temperatures. Only FeCl2 and FeCl3 were able to form 3-MCPD esters from TSG. Further electron spin resonance (ESR)
determination of TSG, Fe2(SO4)3 and 5,5-dimethylpyrroline-N-oxide (DMPO) reactions revealed potential of Fe ion in
promoting free radical generations under the experimental conditions. To further conrm the eect of Fe ion, chelating agent
(EDTA-2Na) was added to the model reactions. The results showed for the rst time that EDTA-2Na was able to reduce the
generation of 3-MCPD esters. In addition, FT-IR examination indicated a possible involvement of a carbonyl group during the
reaction. Taking all the observations together, the possible mechanisms, involving the formation of either a cyclic acyloxonium or
a glycidol ester radical intermediate, were proposed for generating 3-MCPD fatty acid di- and mono- esters from TAG under a
high temperature and low moisture condition, as well as the coformation of glycidol esters. The results from this study may be
useful for reducing the level of 3-MCPD esters and related toxicants in the rened edible oils and food products.
KEYWORDS: 3-MCPD esters, formation mechanism, iron ion, ESR, lipid peroxidation, EDTA-2Na

INTRODUCTION
3-Monochloro-1,2-propanediol (3-MCPD) fatty acid esters,
rst identied in 1980 by Velis ek and Davidek
in hydrolyzed
vegetable proteins, are a group of processing-induced food
contaminants with nephrotoxicity.1,2 3-MCPD esters have been
detected in many foods, including rened edible oils, cakes,
cereals, cheeses37 and infant and baby foods,8 as well as
human breast milk.9 The deodorization procedure during the
edible oil rening process has been proven as a critical step for
3-MCPD ester formation.10,11 Five possible chemical mecha-
nisms for the formation of 3-MCPD esters have been proposed
to better develop eective and practical approaches to mitigate
their levels in food and edible oil products.1215 Two of them
involved direct substitution by a chlorine anion at a glycerol
carbon atom replacing either a hydroxyl group or a fatty acid
ester group.13 Another two mechanisms consist of the
formation of an intermediate epoxide ring or acyloxonium
cation before their nucleophilic attack by a chlorine
anion.13,1618 In 2013, our group reported a novel free radical
mechanism, including the formation of a cyclic acyloxonium
free radical intermediate, followed by its reaction with a
chlorine radical or a chlorine compound.14 This mechanism
was supported by the detection of the cyclic acyloxonium free
radical and 5,5-dimethylpyrroline-N-oxide (DMPO) adducts
through quadrupoletime-of-ight mass spectrometer (Q-TOF
MS) analysis, as well as the generation of 3-MCPD diesters
from sn-1,2-stearoylglycerol reacting with lindane or chlorine
gas.14 However, this study revealed only the possible formation
mechanism of 3-MCPD diesters using a diacylglycerol (DAG)
precursor. Whether and how 3-MCPD esters could be formed
from triacylglycerol (TAG) through reacting with organic and
inorganic chlorine-containing compounds remains unclear.
In 2012, Destaillats and others compared the ability of TAG
and DAG to form 3-MCPD diesters upon heating with
chlorinated compounds.15 The results indicated that TAG
yielded 3-MCPD esters in a greater abundance than DAG.
TAG was the main acylglycerol class undergoing elimination of
fatty acids and the primary acylglycerol class involved in 3-
MCPD diester formation,15 suggesting the importance of
understanding the formation mechanism of 3-MCPD esters
from TAG for developing possible solutions to mitigate the 3-
MCPD ester level in edible oils and food products. It was
hypothesized that TAG would undergo rearrangement by the
Received: December 17, 2014
Revised: January 27, 2015
Accepted: January 29, 2015
Published: January 29, 2015

2015 American Chemical Society 1839 DOI: 10.1021/jf5061216

J. Agric. Food Chem. 2015, 63, 18391848
Journal of Agricultural and Food Chemistry
elimination of a fatty acid to form a cyclic acyloxonium cation
reactive intermediate, followed with a nucleophilic substitution
with a chlorine anion to form 3-MCPD diesters.15 However, no
direct evidence was provided to support the hypothesis.
Whether TAG could be the precursor for 3-MCPD monoester
formation also remained unknown.
The present study aimed to investigate whether TAG may
serve as precursors for 3-MCPD di- and monoesters and to
explore possible chemical mechanisms for their formation.
TAG and selected chlorine-containing compounds were reacted
under a high-temperature and low-moisture condition for 3-
MCPD ester formation. An electron spin resonance (ESR)
spectrometer was utilized to detect free radicals, whereas
Fourier transform infrared spectroscopy (FT-IR) was used to
examine the possible involvement of carbonyl group(s). The
results from this study may serve as the scientic foundation for
reducing the level of 3-MCPD esters in rened edible oils and
other food products.
MATERIALS AND METHODS
Materials. Tristearoylglycerol (TSG) and lindane were purchased
from J&K Scientic (Beijing, China). Sodium acetate (MS grade), iron
chloride, DMPO, ferric sulfate, and ferrous sulfate were purchased
from Sigma-Aldrich (St. Louis, MO, USA). Hexadecane and iron
dichloride were purchased from Aladdin Reagent (Shanghai, China).
Sodium hydroxide, toluene, sodium chloride, potassium chloride,
copper chloride, zinc chloride, and concentrated HCl were bought
from Ambrosia Pharmaceuticals (Shanghai, China). High-purity
nitrogen (99.999%) was purchased from Ruili Chemical Gas Co.,
Ltd. (Shanghai, China). 2-Propanol and methanol (LC-MS grade)
were obtained from Merck (Darmstadt, Germany). GHP membrane
(0.2 m) was purchased from Pall Corp. (New York, NY, USA). The
rened palm oil was a gift from the Yihai Kerry/Wilmar Biotechnology
Research and Development Center (Shanghai, China).
Model Reaction between TSG and Selected Chlorine-
Containing Compounds. For the reaction between solid chloride
agents and TSG, 2 mg of NaCl, KCl, FeCl2, CuCl2, ZnCl2, FeCl3, or
lindane was added into 10 mL of 0.5 mg/mL TSG solution in
hexadecane, respectively, according to a previous study.14 HCl gas was
prepared by heating concentrated HCl, followed by drying through
98% H2SO4, and introduced into 10 mL of a 0.5 mg/mL TSG
solution. To examine whether Fe2+ or Fe3+ could catalyze the
formation of 3-MCPD esters, 2 mg of FeSO4 or Fe2(SO4)3 was added
into the mixture of 2 mg of KCl, CuCl2, or lindane and 10 mL of 0.5
mg/mL TSG solution, respectively. Each reaction was reuxed with
stirring for 20 min under ambient temperature or 120 or 240 C,
respectively. The nonpolar reaction system was low in moisture
content. After cooling to ambient temperature, the mixture was diluted
to 1/100 (v/v) with methanol, ltered through a GHP 0.2 m
membrane, and subjected to UPLC-Q-TOF-MS analysis (Waters
Xevo G2 Q-TOF-MS, Milford, MA, USA).14
ESR Determination of Radical Signals. To determine whether
free radical was generated in the formation of 3-MCPD esters from
TSG at dierent temperatures, TSG solution, TSG solution containing
DMPO, TSG solution containing DMPO and FeCl3, and TSG
solution containing DMPO, KCl and Fe2(SO4)3 were kept at ambient
temperature or 60, 80, or 120 C, respectively, for 20 min and
measured on a Bruker EMX-8/2.7C spectrometer equipped with a
variable-temperature system (Bruker Optics, Karlsruhe, Germany) in a
0.1 mL quartz tube for their ESR signals.14 The concentration of TSG
solution was 0.5 mg/mL in toluene. The nal DMPO concentration
was 20 mmol/L in the reaction mixture, whereas the concentration of
FeCl3 or Fe2(SO4)3 was 0.2 mg/mL. The ESR conditions were the
same as described before.14
FT-IR Measurements. TSG was directly applied onto the
attenuated total reectance (ATR) crystal and scanned immediately
at 25 C and then heated to 120 C and scanned immediately.14 FT-IR
spectra were recorded on a Spectrum 100 FT-IR spectrometer
1840
Article
(PerkinElmer, Waltham, MA, USA) equipped with a pressure
application device, combined with a deuterated triglycine sulfate
(DTGS) detector and a single bounce diamond temperature
controller.
Eect of Chelating Agent. Two milligrams of FeCl3 was added
into 5 mL of 0.5 mg/mL TSG solution, followed by ultrasonic
oscillation for 30 min after the addition of a series of dierent amount
of EDTA-2Na (1.2, 2.3, 4.6, 9.2, and 23.0 mg). The control reaction
contained no EDTA-2Na. The reaction was carried out at 120 C for
20 min with continuous stirring. When the mixtures were cooled to
ambient temperature, they were diluted with 50-fold of methanol/2-
propanol (1:1, v/v) and examined by UPLC-Q-TOF MS after
ltration.14
The experiment was also conducted using rened palm oil under a
high-temperature and low-moisture condition. FeCl3 (400 mg) was
added into 80 g of rened palm oil and mixed well, then 0, 0.938, or
4.655 g of EDTA-2Na was added, respectively, for the control reaction
and low and high levels of EDTA-2Na. The mixture was kept at 240
C for 1 h with a nitrogen ow rate of 2 L/min at a pressure of about 5
mbar. After the reaction, inlet N2 was stopped, and vacuum was
released when the temperature was below 100 C. The mixture was
diluted with 2-propanol, ltered through a GHP 0.2 m membrane,
and analyzed by UPLC-Q-TOF MS.
Q-TOF MS Examination of DMPO Radical Adducts. TSG
solutions (0.5 mg/mL in toluene) with Fe2(SO4)3 and DMPO were
heated to 120 C and kept for 20 min. The nal concentrations of
Fe2(SO4)3 and DMPO in the TSG solutions were 0.2 mg/mL and 20
mmol/L, respectively. After cooling to ambient temperature, the TSG
solution was further diluted with methanol for 100-fold and injected
directly into the Q-TOF mass spectrometer for analysis according to a
method described previously.14 Briey, an ESI positive mode was
selected with lockspray as an internal standard; capillary voltage for
positive ion mode was 3.0 kV; sampling cone voltage was 30.0 kV;
source temperature was 120 C with a desolvation temperature at 450
C; cone gas ow and desolvation gas ow were 50 and 800 L/h,
respectively; scan range was m/z 501000 Da; scan time was 0.1 s.
Data were collected and analyzed using Masslynx 4.1 software, and all
peak areas were determined using the total ion counts in the extracted
ion chromatogram.
Statistical Analysis. Data were expressed as the mean standard
deviation (SD) for each experiment. Dierences between means were
determined by the analysis of one-way ANOVA and Tukeys test,
using SPSS software (SPSS for Windows, version rel. 16.0, SPSS Inc.,
Chicago, IL, USA). Statistical signicance was declared at P < 0.05.
RESULTS AND DISCUSSION
Potential of Forming 3-MCPD Esters from TSG. NaCl,
KCl, FeCl2, CuCl2, ZnCl2, FeCl3, and dry HCl gas were
examined as the inorganic chlorine donors, whereas lindane was
selected as an organic chloride-containing compound to react
with TSG because they were able to react with sn-1,2-
stearoylglycerol (DSG) to form 3-MCPD fatty acid diester.14
UPLC-Q-TOF-MS analysis of the reaction products showed
that only FeCl2 and FeCl3 at 120 and 240 C and HCl gas at
240 C formed 3-MCPD diesters. It was reported that NaCl,
KCl, FeCl2, CuCl2, ZnCl2, FeCl3, HCl gas, and lindane were
able to react with DSG at 240 C to form 3-MCPD diester.14 It
was also reported that only HCl gas, FeCl2, and FeCl3 were able
to react with DSG and form detectable amounts of 3-MCPD
diester at 120 C.14 Taken together, these results suggested that
TSG might be less active in reacting with chlorine-containing
compounds to form 3-MCPD diesters compared to DSG. Also,
FeCl2 and FeCl3 might have greater reactivity with TAG/TSG
and DAG/DSG than HCl gas and other inorganic and organic
chlorine compounds to generate 3-MCPD esters.
In addition, 3-MCPD fatty acid monoester was also
generated in the reaction of FeCl3 and TSG at both 120 and
DOI: 10.1021/jf5061216
J. Agric. Food Chem. 2015, 63, 18391848
Journal of Agricultural and Food Chemistry Article

Figure 1. UPLC-QTOF MS extracted-ion chromatograms for the reaction products of TSG and lindane. The extracted ions were for 3-MCPD (A)
diester ion, (B) monoester, (C) diester with Fe2(SO4)3, (D) monoester with Fe2(SO4)3, (E) diester with FeSO4, and (F) monoester with FeSO4.
The concentration was 0.5 mg/mL for TSG and the reaction temperature was 120 C. The concentration was 0.2 mg/mL for Fe2(SO4)3 and FeSO4.
indicates the peak area for a 3-MCPD di- or monoester ion.

240 C, suggesting that TSG or other TAGs might also serve as
precursors for 3-MCPD fatty acid monoester formation under
the same high-temperature and low-moisture condition.
Interestingly, no detectable amount of 3-MCPD monoester
was formed in the reactions of DSG with all of the tested
chlorine-containing compounds in an earlier study.14 To our
knowledge, this was the rst time that TAG was found to form
3-MCPD monoester under a high-temperature and low-
moisture condition. The fact that only FeCl2 or FeCl3 could
react with TAG to form 3-MCPD esters under a high-
temperature and low-moisture condition, similar to that during
1841
oil deodorization, also raised the question of whether Fe3+ or
Fe2+ could play a role in 3-MCPD ester formation.
Possible Role of Fe3+ or Fe2+ in 3-MCPD Ester
Formation. To examine the possible role of Fe3+ or Fe2+ in
the formation of 3-MCPD fatty acid esters from TAG,
Fe2(SO4)3 and FeSO4 were tested as the Fe3+ and Fe2+ donors,
respectively. TSG was reacted with both inorganic chlorides
including KCl and CuCl2 and an organic chlorine compound
lindane in the presence of Fe3+ or Fe2+ at room temperature
and 120 and 240 C, respectively. Figure 1 reports the 3-
MCPD ester formation in the reaction mixture of TSG and
lindane at 120 C. No detectable amount of either 3-MCPD
DOI: 10.1021/jf5061216
J. Agric. Food Chem. 2015, 63, 18391848
Journal of Agricultural and Food Chemistry Article

Figure 2. UPLC-QTOF MS extracted-ion chromatograms for the reaction products of TSG and KCl. The extracted ions were for 3-MCPD (A)
diester ion, (B) monoester, (C) diester with Fe2(SO4)3, (D) monoester with Fe2(SO4)3, (E) diester with FeSO4, and (F) monoester with FeSO4.
The concentration was 0.5 mg/mL for TSG and the reaction temperature was 120 C. The concentration was 0.2 mg/mL for Fe2(SO4)3 and FeSO4.
indicates the peak area for a 3-MCPD di- or monoester ion.

diester (Figure 1A) or monoester (Figure 1B) was observed
under the experimental conditions. Signicant amounts of 3-
MCPD diesters at the retention time of 7.71 min with an m/z
value of 665.5248 (calculated as 665.5252) were detected with
the presence of Fe2(SO4)3 or FeSO4 (Figure 1, panels C and E,
respectively), with signicant levels of the monoester at 3.45
min with m/z 399.2640 (calculated as 399.2642) (Figure
1D,F). The structures of the 3-MCPD fatty acid mono- and
diesters were conrmed with high-resolution MS analysis (data
not shown). The possible catalytic eect of Fe3+ or Fe2+ in 3-
1842
MCPD ester formation was also conrmed with an inorganic
chlorine compound KCl (Figure 2). No 3-MCPD esters were
detected in the reaction mixture of TSG and KCl (Figure
2A,B), whereas signicant contents of 3-MCPD diester at 7.69
min and monoester at 3.42 min were detected when Fe2(SO4)3
was added (Figure 2C,D) or FeSO4 was added in the reaction
(Figure 2E,F). Furthermore, no dierence in catalytic activity
was observed between Fe2(SO4)3 and FeSO4. Interestingly, the
ratio of 3-MCPD monoester to diester was approximately 10:1
when TSG reacted with KCl in the presence of Fe2(SO4)3 or
DOI: 10.1021/jf5061216
J. Agric. Food Chem. 2015, 63, 18391848
Journal of Agricultural and Food Chemistry Article

Figure 3. UPLC-QTOF MS extracted-ion chromatograms for the reaction products of TSG and dierent chlorine sources: (A) 3-MCPD diester ion
with KCl and Fe2(SO4)3; (B) 3-MCPD monoester ion with KCl and Fe2(SO4)3; (C) 3-MCPD diester ion with FeCl3; (D) monoester ion with
FeCl3. The retention times of both diester and monoester are dierent between injections because the UPLC columns were dierent. The
concentration was 0.5 mg/mL for TSG the and reaction temperature was 120 C. The concentration was 0.2 mg/mL for Fe2(SO4)3 and FeSO4.
indicates the peak area for a 3-MCPD di- or monoester ion.

FeSO4 (Figure 2DF, respectively), whereas the ratio was 1:10
when TSG reacted with FeCl3 (Figure 3). The results suggested
that both Fe3+ and Fe2+ were able to catalyze the formation of
3-MCPD fatty acid di- and monoesters from TAG and might
favor the formation of 3-MCPD monoesters under a high-
temperature and low-moisture condition, indicating their
important role in forming 3-MCPD esters during the oil
deodorization process. The results also suggested that the
overall amount and prole of 3-MCPD esters also depended on
the type of chlorine sources. This was the rst report about the
possible catalytic eect of Fe3+ or Fe2+ in 3-MCPD ester
formation from TAG, suggesting a possible approach to
mitigate 3-MCPD esters in rened oils through reducing the
Fe availability. This was also the rst report that chlorine source
(FeCl3 vs KCl with Fe2(SO4)3 or FeSO4) may alter the nal 3-
MCPD ester prole.
In addition, the two facts that 3-MCPD esters were formed
through the reaction of lindane and TSG in the presence of
Fe3+ or Fe2+ and that lindane is an organic chlorine compound
1843
also suggested a possible free radical mediated mechanism for
3-MCPD ester formation from TAG. It was interesting why or
how Fe3+ or Fe2+ was able to facilitate the 3-MCPD ester
formation under a low-moisture and high-temperature
condition.
Electron Spin Resonance Examination of Possible
Free Radicals. ESR examination was performed to examine
and conrm the presence of free radicals under the
experimental conditions, because our previous study showed
a free radical mediated mechanism for 3-MCPD fatty acid
diester formation from diacylglycerol.14 As shown in Figure 4A,
heating TSG with DMPO, a spin trapping agent, at 60 C for
20 min did not result in any detectable level of free radicals,
whereas signicant ESR signal was detected in TSG with
DMPO kept at 120 C for 20 min (Figure 4B), indicating the
generation of free radicals due to the thermal treatment. To test
the role of Fe3+, the mixture of TSG and DMPO with
Fe2(SO4)3 or FeCl3 was heated for 20 min under the same
conditions and examined for possible ESR signals. Weak ESR
DOI: 10.1021/jf5061216
J. Agric. Food Chem. 2015, 63, 18391848
Journal of Agricultural and Food Chemistry Article

Figure 4. ESR examination of TSG mixtures. The TSG reaction mixtures were kept at (A) 60 C without Fe, (B) 120 C without Fe, (C) 60 C with
FeCl3, (D) 120 C with FeCl3, (E) 60 C with Fe2(SO4)3, and (F) 120 C with Fe2(SO4)3. All TSG reaction mixtures contained 0.5 mg/mL TSG
and 20 mmol/L DMPO, as well as dierent chlorine compounds. The concentration was 0.2 mg/mL for FeCl3 and Fe2(SO4)3. All reactions were
carried out for 20 min in toluene at the temperature before ESR analysis.

signal was observed when the mixture of FeCl3, TSG, and
DMPO was heated at 25 C (Supporting Information, Figure
S1) and 60 C (Figure 4C), but no signicant ESR signal was
detected in the mixture kept at 80 C (Figure S1) and 120 C
(Figure 4D). Interestingly, the mixture of Fe2(SO4)3, TSG, and
DMPO kept at 60 C (Figure 4E), 80 C (Figure S2), and 120
C (Figure 4F) had signicant levels of ESR signals. These
results suggested that Fe2(SO4)3 might promote radical
generation at a high-temperature and low-moisture condition
and consequently stimulate 3-MCPD ester formation from
TSG or other TAGs when a chlorine source is also available.
The absence of ESR signals in the mixture of TSG, DMPO, and
FeCl3 at a relatively higher temperature might be explained by
the availability of chlorine in FeCl3, which immediately reacted
with the radical intermediate(s) to form the 3-MCPD ester nal
product(s) under the experimental conditions. The catalytic
activity of Fe3+ or Fe2+ in generating free radicals might depend
on their electronic structures and electron availability. The
weaker ESR signals detected in the FeCl3 and TSG reactions
might also be partially explained by the lesser electron
availability of FeCl3 compared to that of Fe2(SO4)3.
Eect of Chelating Agent. To further examine the role of
Fe3+ or Fe2+ in forming 3-MCPD esters, a chelating agent,
EDTA-2Na, was selected and investigated for its possible
inuence on 3-MCPD ester formation in the presence of Fe3+.
Figure 5A shows that EDTA-2Na at a molar ratio to FeCl3 of
1:1 or greater signicantly decreased 3-MCPD diester
1844
formation from TSG in a dose-dependent manner, but had
little eect on the monoester amount (data not shown).
The eect of EDTA on 3-MCPD ester formation was also
investigated using rened palm oil. As shown in Figure 5B, the
addition of EDTA in the reaction mixture was able to suppress
the formation of 1-palmitoyl, 1-oleoyl, and 1-linoleoyl 3-MCPD
esters in a dose-dependent matter. The levels of other 3-MCPD
esters were not estimated as their levels were low in the
reaction mixture and could not be accurately quantied. The
result further conrmed the catalytic role of Fe3+ or Fe2+ in
forming 3-MCPD fatty acid esters from TAG under a high-
temperature and low-moisture condition and suggested that
chelating agents might suppress 3-MCPD fatty acid formation
from TAG through reducing Fe2+ or Fe3+ availability.
Taken together, the data in Figure 5 indicate that the oil
composition might alter the 3-MCPD ester formation and
modulate the catalytic eect of Fe3+ or Fe2+, suggesting that 3-
MCPD ester formation during oil rening could be very
complicated and warrant additional research. This nding was
partially supported by the earlier observation that the oil
composition or type of oil might alter the generation of 3-
MCPD esters.19
FT-IR Determination of TSG at 25 and 120 C. To
further examine the proposed free radical mechanism(s) for 3-
MCPD di- and monoester formation, TSG kept at 25 and 120
C was measured for FT-IT spectra. As shown in Figure 6A,
because the carbonyl ester groups at C-1 and C-3 positions
have the same chemical environment, TSG exhibited two
DOI: 10.1021/jf5061216
J. Agric. Food Chem. 2015, 63, 18391848
Journal of Agricultural and Food Chemistry
Figure 5. Eect of EDTA on 3-MCPD ester formation: (A) 3-MCPD
diester formation from TSG; (B) formation of the three 3-MCPD
monoesters from a rened palm oil. The relative amounts of 3-MCPD
esters are reported. Dierent letters (ac) in panel A reect signicant
dierences in the relative amount of 3-MCPD diester in TSG (P <
0.05), whereas dierent letters (ac, ac, and ac) in panel B
represent signicant dierence in relative amount of 1-oleoyl, 1-
linoleoyl, and 1-palmitoyl 3-MCPD esters in the palm oil, respectively
(P < 0.05).
carbonyl absorbance bands at 25 C, one centered at 1737
cm1 and the other centered at 1729 cm1. However, when
heated to 120 C, only one carbonyl absorbance band centered
at 1748 cm1 was detected, reecting a hypochromic shift of
the carbonyl group (Figure 6B). The results indicated the
involvement of the carbonyl ester group at the C-2 position or
two of the three carbonyl groups at a high temperature in
forming 3-MCPD esters.14
Proposed Mechanism for 3-MCPD Ester Formation
from TAG. On the basis of the data, a new mechanism was
proposed for 3-MCPD ester formation from a triacylglycerol
under a high-temperature and low-moisture condition. As
shown in Figure 7, the reaction might be initiated by the
formation of a carbon-centered radical (I) by eliminating a
RCOO from C-1 or C-3 of TSG (TAG) in step a due to the
thermal treatment and/or the catalytic eect(s) of Fe3+ or Fe2+.
The carbon-centered radical was not stable and might
immediately attack the ester carbonyl group located at the C-
2 and form a cyclic acyloxonium free radical (II, CAFR)
through step b or form a epoxide ring free radical (III, ERFR)
through step c. CAFR and ERFR could be relatively stable
because of the dislocation of the unpaired electron among
multiple atoms within the ring structures. CAFR could either
react with a chlorine radical or extract a chlorine radical from a
chlorine-containing compound, such as HCl, Cl2, NaCl, KCl,
FeCl2, FeCl3, CuCl2, or ZnCl2, to form 3-MCPD diesters
Article
Figure 6. FT-IR spectra of TSG kept at (A) 25 C and (B) 120 C.
through step d or e, respectively, whereas EAFR could undergo
a similar reaction to form 3-MCPD monoesters (Figure 7). The
proposed mechanism(s) could explain the formation of both 3-
MCPD fatty acid diesters and monoesters under a high-
temperature and low-moisture condition, such as that during
the deodorization procedure in the oil-rening process. The
mechanism(s) could also explain the copresence of 3-MCPD
and glycidol fatty acid esters in the rened edible oils.20
It was reported that both Fe2+ and Fe3+ were associated with
free radical generation21 and lipid peroxidation.22 Fe2+ and Fe3+
may serve as a detrimental catalyst in 3-MCPD and glycidol
fatty acid ester formation through promoting radical generation
and forming complexes with the radical intermediates to
enhance their stability. These might be explained by their
special electron structures. For instance, Fe2+ and Fe3+ have
been proposed to form ironoxygen complexes such as
preferred ion (Fe2+O2 or Fe3+O2) or ferryl ion (FeO2+) to
facilitate lipid peroxidation.22
To further examine the proposed radical mechanism(s), TSG
was heated with Fe2(SO4)3 and DMPO to 120 C for 20 min,
and the reaction mixtures were examined using Q-TOF MS.
However, neither the intermediate free radicalDMPO adducts
nor other possible product ions of the intermediates were
detected.
Possible Role of Thermal Treatment on 3-MCPD Ester
Formation. The eect of thermal treatment degree on 3-
MCPD ester formation was also investigated by reacting TSG
with FeCl3 at 120 and 240 C, and the concentrations of
products were semiquantied on the basis of the peak areas of
1845 DOI: 10.1021/jf5061216
J. Agric. Food Chem. 2015, 63, 18391848
Journal of Agricultural and Food Chemistry
Figure 7. Proposed chemical mechanisms for 3-MCPD ester formation from triacylglycerol (TAG).
their extracted ion chromatogram. At 120 C, the total ion
count for 3-MCPD di- and monoesters was 8969 and was 5588
when the reaction was carried out at 240 C, suggesting that a
higher degree of thermal treatment might result in a lower
overall 3-MCPD ester amount in the oil products (Figure 8).
The nding was further conrmed with the reactions of TSG
and KCl in the presence of either Fe3+ or Fe2+. Compared to
the reactions carried out at 240 C, greater levels of total 3-
MCPD esters were detected at 120 C, regardless of the form
of Fe (Supporting Information, Figures S3 and S4). This
nding was in contrast to the observation by Svejkovska and
others that MCPD ester concentration was proportional to the
reaction temperature,23 but supported an earlier observation
that 3-MCPD esters undergo thermal degradation with the rate
proportional to the thermal treatment degree.24
In addition, the ratio of 3-MCPD diester to monoester was
more than 5:1 at 120 C, whereas the ratio became about 1.3
when reacted at 240 C (Figure 8), suggesting the possible
inuence of thermal treatment degree on the ratio of 3-MCPD
di- and monoesters. Also noted was that TSG with KCl and
Fe3+ also resulted in an ion count of 943 for 3-MCPD
monoester and 82 for diester at 120 C, respectively, whereas
the reaction generated an ion count of 788 for 3-MCPD
monoester without detectable diester (Supporting Information,
Figure S3). For the reaction of TSG with KCl and Fe3+, the
ratios between diester and monoester were 1:10.6 and 1:14.5 at
120 and 240 C, respectively (Figure S4). These data suggested
the thermal degradation of 3-MCPD esters might include both
1846
Article
dechlorination and deacylation, and deacylation may be more
preferred under the experimental conditions. This nding was
in contrast to the conclusion by Ermacora and Hrncirik that
dechlorination was preferred over deacylation for 3-MCPD
dipalmitate and dilaurate.24 The degradation experiment was
conducted with the residues of two 3-MCPD diesters in a
temperature-controlled oven and kept at 180, 240, and 260 C
for 20 min. The conclusion was based on the analysis of 3-
MCPD diesters and their degradation products. In the present
study, the overall amounts for 3-MCPD di- and monoesters
from TSG reacting with a chlorine compound were the results
of both formation and degradation reactions. Taken together,
these data suggested a possibility to reduce nal 3-MCPD ester
concentration in the rened oils through an extended thermal
treatment degree. It needs to be pointed out that the thermal
treatment may promote lipid oxidation and other chemical
changes in the edible oil products through reducing their 3-
MCPD ester contents.
In summary, the possible mechanisms behind the formation
of 3-MCPD mono- and diesters from triglycerides might
involve either a cyclic acyloxonium or a glycidol ester radical
intermediate under a high-temperature and low-moisture
condition. This study also reported that glycidol esters might
be a coproduct from TAG during 3-MCPD ester formation
under the experimental conditions. In addition, Fe2+ and Fe3+
might catalyze both 3-MCPD and glycidol esters formation
from TAG. These results suggested that reducing the
availability of Fe2+ and Fe3+ through controlling ingredients,
DOI: 10.1021/jf5061216
J. Agric. Food Chem. 2015, 63, 18391848
Journal of Agricultural and Food Chemistry
Figure 8. UPLC-QTOF MS extracted-ion chromatograms for the reaction products of TSG and FeCl3. The reactions were carried out for 20 min at
120 or 240 C. The extracted ions were for 3-MCPD (A) diester ion at 120 C, (B) monoester ion at 120 C, (C) diester ion at 120 C, and (D)
monoester ion at 120 C. The concentration was 0.5 mg/mL for TSG, and the concentration was 0.2 mg/mL for FeCl3. indicates the peak area for
a 3-MCPD di- or monoester ion.
equipment, and containers and processing agents including
water, using chelating agents in selected processing steps, as
well as improvement of the thermal processing degree may lead
to reduced levels of 3-MCPD esters and related toxicants such
as glycidol esters in the rened edible oils and food products.
*
ASSOCIATED CONTENT
S Supporting Information
Figures S1S4. This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*(L.Y.) Phone: (301) 405-0761. Fax: (301) 314-3313. E-mail:
lyu5@umd.edu.
Author Contributions
These authors made equal contributions to the manuscript.
Funding
This research was supported by a special fund for Argo-
scientic Research in the Public Interest (Grant 201203069),
two grants from the National High Technology Research and
Development Program of China (Grants 2013AA102202 and
2013AA102207), a grant from the Wilmar (Shanghai)
Biotechnology Research and Development Center Co., Ltd.,
1847
Article
and grants from SJTU 985-III disciplines platform and talent
funds (Grants TS0414115001 and TS0320215001).
Notes
The authors declare no competing nancial interest.
ACKNOWLEDGMENTS
We thank Yihai Kerry/Wilmar (Shanghai) Biotechnology
Research and Development Center for providing rened
palm oil and Dr. Hai Zhang for his editorial assistance and
scientic comments during manuscript revision.
ABBREVIATIONS USED
3-MCPD, 3-monochloro-1,2-propanediol; TAG, triacylglycerol;
ESR, electron spin resonance; DMPO, 5,5-dimethylpyrroline-
N-oxide; Q-TOF MS, quadrupoletime-of-ight mass spec-
trometer; DAG, diacylglycerol; MAG, monoacylglycerol; FT-
IR, Fourier transform infrared
REFERENCES
(1) Velisek, J.; Davidek, J.; Kubelka, V.; Janicek, G.; Svobodova, Z.;
Simicova, Z. New chlorine-containing organic compounds in protein
hydrolysates. J. Agric. Food Chem. 1980, 28, 11421144.
(2) Liu, M.; Gao, B.; Qin, F.; Wu, P.; Shi, H.; Luo, W.; Ma, A.; Jiang,
Y.; Xu, X.; Yu, L. Acute oral toxicity of 3-MCPD mono- and di-
DOI: 10.1021/jf5061216
J. Agric. Food Chem. 2015, 63, 18391848
Journal of Agricultural and Food Chemistry Article

palmitic esters in Swiss mice and their cytotoxicity in NRK-52E rat (24) Ermacora, A.; Hrncirik, K. Study on the thermal degradation of
kidney cells. Food Chem. Toxicol. 2012, 50, 37853791. 3-MCPD esters in model systems simulating deodorization of
(3) Hamlet, C. G.; Sadd, P. A.; Gray, D. A. Chloropropanols and vegetable oils. Food Chem. 2014, 150, 158163.
their esters in baked cereal products. In Abstracts, 224th National
Meeting of the American Chemical Society; ACS: Washington, DC, USA,
2004.
(4) Hamlet, C. G.; Asuncion, L.; Velisek, J.; Dolezal, M.; Zelinkova,
Z.; Crews, C. Formation and occurrence of esters of 3-chloropropane-
1,2-diol (3-CPD) in foods: what we know and what we assume. Eur. J.
Lipid Sci. Technol. 2011, 113, 279303.
(5) Pudel, F.; Benecke, P.; Fehling, P.; Freudenstein, A.; Matthaeus,
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, 368373.
(6) Svejkovska, B.; Novotny, O.; Divinova, V.; Reblova, Z.; Dolezal,
M.; Velisek, J. Esters of 3-chloropropane-1,2-diol in foodstuffs. Czech J.
Food Sci. 2004, 22, 190196.
(7) Zelinkova, Z.; Svejkovska, B.; Velisek, J.; Dolezal, M. Fatty acid
esters of 3-chloropropane-1,2-diol in edible oils. Food Addit. Contam.
2006, 23, 12901298.
(8) Zelinkova, Z.; Dolezal, M.; Velsek, J. Occurrence of 3-
chloropropane-1,2-diol fatty acid esters in infant and baby foods.
Eur. Food Res. Technol. 2009, 228, 571578.
(9) Zelinkova, Z.; Novotny, O.; Schurek, J.; Velisek, J.; Hajslova, J.;
Dolezal, M. Occurrence of 3-MCPD fatty acid esters in human breast
milk. Food Addit. Contam. 2008, 25, 669676.
(10) Weisshaar, R. 3-MCPD-esters in edible fats and oils a new
and worldwide problem. Eur. J. Lipid Sci. Technol. 2008, 110, 671672.
(11) Franke, K.; Strijowski, U.; Fleck, G.; Pudel, F. Influence of
chemical refining process and oil type on bound 3-chloro-1,2-
propanediol contents in palm oil and rapeseed oil. LWTFood Sci.
Technol. 2009, 42, 17511754.
(12) 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.
(13) Rahn, A. K. K.; Yaylayan, V. A. What do we know about the
molecular mechanism of 3-MCPD ester formation? Eur. J. Lipid Sci.
Technol. 2011, 113, 323329.
(14) Zhang, X.; Gao, B.; Qin, F.; Shi, H.; Jiang, Y.; Xu, X.; Yu, L. Free
radical mediated formation of 3-monochloropropanediol (3-MCPD)
fatty acid diesters. J. Agric. Food Chem. 2013, 61, 25482555.
(15) Destaillats, F.; Craft, B. D.; Sandoz, L.; Nagy, K. Formation
mechanisms of monochloropropanediol (MCPD) fatty acid diesters in
refined palm (Elaeis guineensis) oil and related fractions. Food Addit.
Contam. 2012, 29, 2937.
(16) Collier, P. D.; Cromie, D. D. O.; Davies, A. P. Mechanism of
formation of chloropropanols present in protein hydrolysates. J. Am.
Oil Chem. Soc. 1991, 68, 785790.
(17) Vel se k, J.; Dolezal, M.; Crews, C.; Dvorak , T. Optical isomers
of chloropropanediols: mechanisms of their formation and decom-
position in protein hydrolysates. Czech J. Food Sci. 2002, 20, 161170.
(18) Sonnet, P. E. A short highly regio- and stereoselective synthesis
of triacylglycerols. Chem. Phys. Lipids 1991, 58, 3539.
(19) Ermacora, A.; Hrncirik, K. Influence of oil composition on the
formation of fatty acid esters of 2-chloropropane-1, 3-diol (2-MCPD)
and 3-chloropropane-1, 2-diol (3-MCPD) under conditions simulating
oil refining. Food Chem. 2014, 161, 383389.
(20) Weihaar, R.; Perz, R. Fatty acid esters of glycidol in refined fats
and oils. Eur. J. Lipid Sci. Technol. 2010, 112, 158165.
(21) Welch, K. D.; Davis, T. Z.; Aust, S. D. Iron autoxidation and free
radical generation: effects of buffers, ligands, and chelators. Arch.
Biochem. Biophys. 2002, 397, 360369.
(22) Schafer, F. Q.; Qian, S. Y.; Buettner, G. R. Iron and free radical
oxidations in cell membranes. Cell. Mol. Biol. 2000, 46, 657662.
(23) Svejkovska, B.; Dolezal, M.; Velisek, J. Formation and
decomposition of 3-chloropropane-1,2-diol esters in models simulating
processed foods. Czech J. Food Sci. 2006, 24, 172179.

1848 DOI: 10.1021/jf5061216

J. Agric. Food Chem. 2015, 63, 18391848