Quantitation of hydroperoxy-, keto- and hydroxy-dienes during oxidation of FAMEs from high-linoleic and high-oleic sunflower oils

Article abstract of DOI:10.1007/s11746-010-1624-8

The objective of this work was to study the quantitative formation of hydroperoxydienes, ketodienes and hydroxydienes during autoxidation at 40 °C of fatty acid methyl esters derived from two sunflower oils with different degree of unsaturation, high-linoleic sunflower oil and high-oleic sunflower oil. The analysis of the oxidation compounds was carried out by NP-HPLC-UV and results were compared to the specific extinction at 232 nm (K ₂₃₂) and the peroxide value (PV). Analysis of FAME polymers by HPSEC was also performed to discard samples of advanced oxidation. Results showed that the contents of hydroperoxydienes with respect to the PV were higher for the high linoleic (HL) sample. At the end of the period of slow polymerization (ΔPol ≤ 1 wt%), the content of hydroperoxydienes was found to be 86.0 and 30.7 μg/mg for the HL and high oleic (HO) samples, respectively. Throughout this period, hydroperoxydienes constituted around 90 and 50 wt% of the total hydroperoxides in the HL and HO samples, respectively, suggesting that a significant oxidation of oleic acid also occurred in both samples. The contents of ketodienes and hydroxydienes as a whole constituted 2-3 wt% of the diene compounds analyzed at the end of the period of slow polymerization. Higher contents of ketodienes than of hydroxydienes were found throughout the oxidation time, and the ratio between the contents of ketodienes and hydroxydienes increased with a factor that changed from 1 to 2 throughout the period of slow polymerization.

Full text of DOI:10.1007/s11746-010-1624-8

J Am Oil Chem Soc (2010) 87:1271–1279
DOI 10.1007/s11746-010-1624-8
ORIGINAL PAPER
Quantitation of Hydroperoxy-, Keto- and Hydroxy-Dienes During
Oxidation of FAMEs from High-Linoleic and High-Oleic
Sunflower Oils
•
Arturo Morales Carmen Dobarganes
•
•
Gloria Marquez-Ruiz Joaquın Velasco
´
´
Received: 4 March 2010 / Revised: 16 April 2010 / Accepted: 7 June 2010 / Published online: 23 June 2010
Ó AOCS 2010
Abstract The objective of this work was to study the
quantitative formation of hydroperoxydienes, ketodienes
and hydroxydienes during autoxidation at 40 °C of fatty
acid methyl esters derived from two sunflower oils with
different degree of unsaturation, high-linoleic sunflower oil
and high-oleic sunflower oil. The analysis of the oxidation
compounds was carried out by NP-HPLC–UV and results
the contents of ketodienes and hydroxydienes increased
with a factor that changed from 1 to 2 throughout the
period of slow polymerization.
Keywords Hydroperoxides Á Hydroperoxydienes Á
Ketodienes Á Hydroxydienes Á Autoxidation Á FAMEs
were compared to the specific extinction at 232 nm (K232
)
and the peroxide value (PV). Analysis of FAME polymers
by HPSEC was also performed to discard samples of
advanced oxidation. Results showed that the contents of
hydroperoxydienes with respect to the PV were higher for
the high linoleic (HL) sample. At the end of the period of
slow polymerization (DPol B 1 wt%), the content of
hydroperoxydienes was found to be 86.0 and 30.7 lg/mg
for the HL and high oleic (HO) samples, respectively.
Throughout this period, hydroperoxydienes constituted
around 90 and 50 wt% of the total hydroperoxides in the
HL and HO samples, respectively, suggesting that a sig-
nificant oxidation of oleic acid also occurred in both
samples. The contents of ketodienes and hydroxydienes as
a whole constituted 2–3 wt% of the diene compounds
analyzed at the end of the period of slow polymerization.
Higher contents of ketodienes than of hydroxydienes were
found throughout the oxidation time, and the ratio between
Introduction
Lipid oxidation is one of the main causes of quality dete-
rioration in foods because it is responsible for the devel-
opment of products that impair the food flavor. In recent
years there has been special concern about the potentially
detrimental nutritional effects of oxidized lipids in foods as
a consequence of implications of oxidized lipids and other
reactive oxygen species in a variety of chronic diseases
[1–8]. In order to evaluate their actual contribution to the
promotion and/or development of diseases, it is first nec-
essary to know the nature and quantities of lipid oxidation
products in foods [4].
The qualitative analysis of the main oxidation products
of the major oxidizing fatty acids in foods has been well
established through model lipid systems [9]. Nevertheless,
the quantitative analysis of lipid oxidation products in real
foods finds serious limitations because of the large number
of products involved, their low contents as individual
species and their relatively high instability.
A. Morales Á C. Dobarganes Á J. Velasco (&)
Instituto de la Grasa, Consejo Superior de Investigaciones
In previous studies carried out in our lab, the quantita-
tive analysis of lipid oxidation products in thermoxidized
and frying oils was tackled through a derivatization step of
the triacylglycerol molecules into fatty acid methyl esters
(FAMEs) followed by gas–liquid chromatography analysis
with flame ionization detection (GC-FID) [10–12]. While
the compounds accumulated in thermoxidized oils are
´
Cientıficas (CSIC), Avda. Padre Garcıa Tejero, 4,
41012 Sevilla, Spain
e-mail: jvelasco@cica.es
´
´
G. Marquez-Ruiz
Instituto del Frıo, Consejo Superior de Investigaciones
´
´
Cientıficas (CSIC), c/Jose Antonio Novais, 10,
28040 Madrid, Spain
´
123

1272
J Am Oil Chem Soc (2010) 87:1271–1279
relatively stable secondary oxidation products, the more
abundant compounds during autoxidation are hydroperox-
ides, whose relatively low stability makes their analysis
through derivatization steps a challenging task.
peroxide value. In addition, the analysis of polymers by
high-performance size-exclusion chromatography with
refraction-index detection (HPSEC-RFI) was performed as
a rapid control measure in order to discard samples within
advanced oxidation [15].
Hopia and coworkers [13] have applied normal-phase
high-performance liquid chromatography (NP-HPLC)
with UV light absorption detection for the direct analysis
of the main oxidation products of linoleic acid, i.e.
hydroperoxydienes, ketodienes and hydroxydienes, in
oxidized samples of methyl linoleate. Due to the diene
structure, the hydroperoxydienes and hydroxydienes were
detected at 233 nm, while the ketodienes were recorded at
268 nm. Relative contents of the secondary oxidation
products, i.e. ketodienes and hydroxydienes, were repor-
ted by referring their areas to the area of hydroperoxides
and considering their molar extinction coefficients.
However, the absolute contents of the oxidation products
were not given.
Materials and Methods
Chemicals
Linoleic acid (C99%) and soybean lipoxygenase (lipoxi-
dase preparation Type I-B) were purchased from Sigma-
´
Aldrich Quımica (Madrid, Spain). Methyl linoleate
(C99%) was acquired from Nu-Chek-Prep (Elysian, MN,
USA). Diethyl ether stabilized with 1% v/v ethanol (Super
purity solvent, HPLC grade) was purchased from Romil
LTD (Cambridge, UK) and n-heptane (99% purity, HPLC
grade) was purchased from Carlo Erba Reactifs-SDS (Val
de Reuil, France). Both HPLC solvents were used as
received.
The determination of hydroperoxydienes, ketodienes
and hydroxydienes has only been carried out in a pure lipid
model system as is methyl linoleate. However, the analysis
of these compounds has not yet been performed in more
complex lipid systems such as mixtures of FAMEs with
different oxidizing fatty acids. FAMEs derived from oils
may be used as model systems to study oxidation in oils.
They can be analyzed directly by NP-HPLC–UV for the
separation and determination of oxidation products con-
taining conjugated double bonds. Nevertheless, the analysis
of these analytes in oxidized samples of oils would require
the derivatization of triacylglycerols into FAMEs. There-
fore, the analysis of oxidized samples of FAMEs derived
from oils can be considered as a necessary step previous to
the development of the analysis of real samples of oils. As
in previous studies, oxidized samples of FAMEs would
allow us to investigate how derivatization procedures affect
the compounds of interest [14].
Samples
High linoleic sunflower oil (HLSO) and high oleic sun-
´
flower oil (HOSO) were supplied by Koipe S.A. (Andujar,
´
Jaen, Spain). The fatty acid composition was determined to
be 6.7% C16:0, 0.2% C16:1, 3.6% C18:0, 33.0% C18:1,
55.2% C18:2 and 1.3% others in HLSO and 4.1% C16:0,
4.6% C18:0, 77.6% C18:1, 11.7% C18:2 and 2.0% others
in HOSO.
Sample Treatments
The oils were purified with aluminum oxide according to
Yoshida et al. [16]. Then the purified oils were subjected to
a transmethylation procedure to obtain FAMEs. This pro-
cedure consisted of the reaction of the oils with 2 M KOH
in methanol at room temperature. First, an aliquot of 30 g
of purified oil was dissolved in 100 mL n-hexane. A vol-
ume of 10 mL of the KOH solution was added and stirring
was applied for 30 min at room temperature and in the
dark. The medium was neutralized with 30 mL of 0.5 M
H2SO4 in MeOH. The organic phase was separated, filtered
through filter paper containing anhydrous Na2SO4, and the
solvent was evaporated in a rotary evaporator at 40 °C.
Finally, the lipid sample was bubbled with N2 to remove
the residual solvent. The transmethylation reaction was
found to be complete by thin layer chromatography (TLC).
Silica plates and hexane:diethyl ether:acetic acid (80:20:1,
v/v) as the mobile phase were used. The spots on the plate
were revealed by iodine vapor.
The objective of this work was to study the quantitative
formation of hydroperoxydienes, ketodienes and hydrox-
ydienes during autoxidation of FAMEs derived from
sunflower oils with different degree of unsaturation, high-
linoleic sunflower oil (HLSO) and high-oleic sunflower oil
(HOSO). The analysis of the oxidation compounds was
carried out by NP-HPLC with a UV-light detector fol-
lowing the studies of Hopia and coworkers in methyl
linoleate [13]. Unlike the studies of Hopia et al., in this
study we tackle the quantitative analysis of the compounds
of interest. To this end, compounds representative of each
group of products were synthesized and the corresponding
HPLC response factors were determined. In order to relate
the contents of the oxidation products under study to the
oxidation state of the samples, oxidation was characterized
by the specific extinction at 232 nm (K232) and the
123

J Am Oil Chem Soc (2010) 87:1271–1279
Oxidation Conditions
1273
the procedure described by Hidalgo et al. [17]. Thus, oxi-
dation was performed in dry pyridine under stirring for
24 h at room temperature. The reaction was stopped with
water and the ketodiene was extracted with diethyl ether.
The corresponding methyl ester was obtained and purified
following the same procedure as described above. The
ketodiene fraction was shown to be pure by TLC. Confir-
mation of identity and purity was also performed by 1H and
13C NMR.
Aliquots of 20 g of FAME samples were oxidized in Petri
dishes (14.5 cm i.d.) at 40 °C in the dark by using an oven
with continuous air circulation. The surface-to-volume
ratio of the FAME samples was as high as 10 cm-1, so
oxidation was developed under non-limited oxygen
conditions.
Synthesis of Oxidation Compounds
Mixture of Hydroperoxydienes from Methyl Linoleate
Methyl 13-Hydroperoxy- (Z)-9, (E)-11-Octadecadienoate
Methyl linoleate was oxidized in the dark at room tem-
perature. An aliquot of 5 g of methyl linoleate was
weighed in a 250-mL beaker and oxidized for 3 d. For-
mation of hydroperoxides was checked by TLC using a
blend of hexane, diethyl ether and acetic acid (80:20:1,
v/v) as mobile phase. The spots were revealed by iodine
vapor and the formation of hydroperoxydienes was pre-
viously confirmed by UV light. Hydroperoxides were
isolated by fractionation in a glass chromatographic col-
umn filled with silica gel following the same procedure as
described above. Unreacted methyl linoleate was oxidized
again under the same oxidation conditions. Hydroperox-
ides were isolated and added to the first batch. This
procedure was repeated three times in total. The fraction
obtained was found to be pure by TLC and confirmation
Methyl 13-hydroperoxy- (Z)-9, (E)-11-octadecadienoate
was obtained through the enzymatic oxidation of linoleic
acid with soybean lipoxygenase following the procedure
described by Hidalgo et al. [17]. First, linoleic acid was
oxidized by lipoxygenase in aqueous solution under basic
conditions (pH 9), oxygen saturation and at a low tem-
perature (0 °C). The reaction was stopped by adding 2 M
HCl up to pH 3 and the hydroperoxide formed was
extracted with diethyl ether. Then, the corresponding
methyl ester was obtained by the reaction of the free fatty
acid with diazomethane according to the procedure
described by Cohen [18]. Finally, methyl 13-hydroperoxy-
(Z)-9, (E)-11-octadecadienoate was purified by fraction-
ation in a glass chromatographic column filled with silica
gel. A blend of hexane and diethyl ether (7:3, v/v) was used
as the elution solvent. The hydroperoxydiene fraction
obtained was shown to be pure by TLC. Confirmation of
1
of identity and purity was also performed by H and 13C
NMR. The hydroperoxides were kept under nitrogen at
-32 °C until analysis.
1
identity and purity was performed by H and 13C NMR.
Analysis of FAME Polymers
Methyl 13-Hydroxy- (Z)-9, (E)-11-Octadecadienoate
Analysis of polymers was performed according to IUPAC
standard method 2.508 [19]. An HPSEC chromatograph
equipped with a Rheodyne 7725i injector with a 10-lL
sample loop, a Knauer 120 HPLC pump (Knauer, Berlin,
Germany) and a Merck L-7490 refractive index detector
(Merck, Darmstadt, Germany) was used. The separation
Methyl 13-hydroxy- (Z)-9, (E)-11-octadecadienoate was
obtained from the reduction of 13-hydroperoxy- (Z)-9,
(E)-11-octadecadienoic acid with NaBH4 following the
procedure described by Hidalgo et al. [17]. Briefly, before
stopping the lipoxygenase reaction, NaBH4 was added to
the reaction medium and stirring was applied for 1 h under
a nitrogen atmosphere. Then, 2 M HCl was added up to an
acidic pH and methyl 13-hydroxy- (Z)-9, (E)-11-octade-
cadienoate was obtained applying the same procedure as
the one described above for the hydroperoxide. Likewise,
the hydroxydiene fraction was shown to be pure by TLC.
˚
was performed on two 100- and 500-A Ultrastyragel col-
umns (25 cm 9 0.77 cm i.d.) packed with porous, highly
cross-linked styrene-divinylbenzene copolymers (10 lm)
(Agilent Technologies, Palo Alto, CA) connected in series
by using tetrahydrofuran as the mobile phase at a flow rate
of 1 mL/min.
1
Confirmation of identity and purity was performed by H
and 13C NMR.
Peroxide Value
Methyl 13-keto- (Z)-9, (E)-11-Octadecadienoate
The peroxide value was determined by the iodometric
assay according to IUPAC standard method 2.501 [19].
Samples of 500 mg FAMEs and 1 9 10-2 M Na2S2O3
solution were used.
Methyl 13-keto- (Z)-9, (E)-11-octadecadienoate was
obtained from oxidation of 13-hydroxy- (Z)-9, (E)-11-oc-
tadecadienoic acid with chromium (VI) oxide according to
123

1274
J Am Oil Chem Soc (2010) 87:1271–1279
Ultraviolet Light Absorption at 232 nm (K232
)
methyl linoleate, respectively, for the determination of
HPLC response factors. The analysis of these compounds
by HPLC showed the presence of minor contents of posi-
tion and geometrical isomers due to the formation of the
9-hydroperoxide and the double-bound isomerization to the
trans,trans forms, respectively (Fig. 1). The presence of
isomers at variable quantities would not affect significantly
the value of a global response factor for each group of
compounds because the absorption spectra of the different
isomers are quite similar. As an example, the spectra of the
cis, trans and the trans, trans isomers of the hydro-
peroxydienes have been reported to be broad bands with
very close absorption maxima, namely, at 235 and 230 nm,
respectively [13]. Therefore, the total area of each group of
compounds was used for the determination of the response
factors.
Specific extinction at 232 nm was determined in cyclo-
hexane as a measure of the total conjugated dienes
according to AOCS standard method Ch 5-91 [20].
Analysis of Oxidation Compounds by HPLC
A Waters 600 HPLC chromatograph (Waters Corp., Mil-
ford, MA, USA) equipped with a 600 Waters pump, a
Rheodyne injector valve (20-lL sample loop), a silica
HPLC column (Si 60, 5 lm particle size) (Merck,
Darmstadt, Germany), a 486 Waters tunable absorbance
detector (10 mm path length) and a 600 Waters controller
was used. The mobile phase was n-heptane:diethyl ether
(82:18, v/v) with a flow rate of 1 mL/min. Ethanol present
in diethyl ether as a stabilizer was not removed. Hydro-
peroxydienes and hydroxydienes were monitored at
234 nm, while ketodienes were recorded at 268 nm. The
FAME samples were dissolved in n-hexane in the range
1–50 mg/mL prior to analysis. The scale of absorbance
units was set at 2,000 units/volt.
Table 1 lists the purity of each compound and the
content of the different isomers, the response factors and
the limit of detection (LOD) and quantification (LOQ)
defined at a signal-to-noise ratio of 3 and 10, respectively.
For comparative purposes, the extinction coefficients were
also determined in cyclohexane and included in Table 1.
1H and 13C NMR
1
1H- and 13C-NMR spectra were recorded at 300 and
75.4 MHz, respectively, in a Bruker AC-300P (Karlsruhe,
Germany). Tetramethylsilane was used as an internal
standard. Two-dimensional NMR was used to assign the
13C-NMR spectra.
a
3
2
4
Statistical Analysis
5
Analytical determinations were carried out in triplicate and
results were expressed as mean values. Linear regression
analysis was performed in Microsoft Excel 2000 (Micro-
soft Corporation, Redmond, WA, USA). Comparisons
between the slopes obtained in the linear regression anal-
ysis were made by the Student’s t test in Microsoft
Excel 2000 (Microsoft Corporation, Redmond, WA, USA).
Significance was defined at p \ 0.05.
b
6
7
8
Results and Discussion
c
Determination of HPLC Response Factors
9
Methyl 13-hydroperoxy-(Z)-9,(E)-11-octadecadienoate
(13-cis,trans-MeLOOH), methyl 13-keto-(Z)-9,(E)-11-oc-
tadecadienoate (13-cis,trans-MeLO) and methyl 13-hydro-
xy-(Z)-9,(E)-11-octadecadienoate (13-cis,trans-MeLOH)
were synthesized and used as compounds representative of
the hydroperoxides, ketodienes and hydroxydienes of
0
5
10
15
20
25
min
Fig. 1 HPLC chromatograms of the synthesized compounds.
a Methyl 13-keto-(Z)-9,(E)-11-octadecadienoate recorded at 268 nm.
b Methyl 13-hydroperoxy-(Z)-9,(E)-11-octadecadienoate recorded at
234 nm. c Methyl 13-hydroxy-(Z)-9,(E)-11-octadecadienoate recorded
at 234 nm. For chromatographic peak numbers see Fig. 2
123

J Am Oil Chem Soc (2010) 87:1271–1279
1275
Table 1 HPLC response factors, limits of detection and quantification, and molar extinction coefficients of the synthesized compounds
Ketodienes
1
Hydroperoxydienes
Hydroxydienes
2
3
4
5
6
7
8
9
10
% AreaHPLC
Conc. range (lg mL-1
FR (lg-1 mL)
r2
71.0
11.2
12.5
5.3
82.8
13.9
3.4
96.7
2.1
ND
)
10–100
10–300
219.2 1.0
0.9998
0.1
10–300
226.9 2.5
0.9992
0.2
223.6 1.3
0.9997
0.1
LOD (lg mL-1
LOQ (lg mL-1
)
)
0.3
0.3
0.6
e (mol-1 L cm-1
)
25,700
25,300
26,200
FR response factor determined at 268 nm for ketodienes and at 234 nm for hydroperoxydienes and hydroxydienes, r2 linear regression coef-
ficient, LOD limit of detection, LOQ limit of quantification, e molar extinction coefficient in cyclohexane at 268 nm for ketodienes and at 233 nm
for hydroperoxydienes and hydroxydienes, ND not detected. For chromatographic peak numbers see Fig. 2
a
6
7
8
8
9
9
10
10
5
3 4
12
0
5
10
15
20
25
min
b
3 4
6
12
7
5
0
5
10
15
20
25
min
Fig. 2 HPLC chromatograms of a FAME sample derived from
HLSO and oxidized at 40 °C for 68 h. a Chromatogram recorded at
234 nm. b Chromatogram recorded at 268 nm. Peak assignation: 1
13-hydroperoxy- (E)-9,(E)-11-octadecadienoate and methyl 9-hydro-
peroxy-(E)-10,(Z)-12-octadecadienoate, methyl 9-hydroperoxy-
(E)-10,(E)-12-octadecadienoate, methyl 13-hydroxy-(Z)-9,(E)-
7
8
methyl 13-keto-(Z)-9,(E)-11-octadecadienoate,
3
octadecadienoate, 4 methyl 9-keto-(E)-10,(E)-12-octadecadienoate,
5 methyl 13-hydroperoxy-(Z)-9,(E)-11-octadecadienoate, 6 methyl
2
methyl 9-keto-
(E)-10,(Z)-12-octadecadienoate, methyl 13-keto-(E)-9,(E)-11-
11-octadecadienoate, 9 methyl 13-hydroxy-(E)-9,(E)-11-octadecadie-
noate and methyl 9-hydroxy-(E)-10,(Z)-12-octadecadienoate,
10 methyl 9-hydroxy-(E)-10,(E)-12-octadecadienoate
HPLC Chromatograms of an Oxidized FAME Sample
Linearity was found in the whole range of concentrations
assayed. The response factors for hydroperoxydienes and
hydroxydienes were very similar to each other, as well as
between these determined at 234 nm and the one for ke-
todienes determined at 268 nm.
Figure 2 illustrates HPLC chromatograms recorded at 234
and 268 nm of an oxidized FAME sample. Because of the
specificity of the light absorption of conjugated diene
structures, the chromatograms of mixtures of oxidized
FAMEs were not complex. At the chromatographic con-
ditions applied, ketodienes and hydroxydienes were com-
pletely resolved or separated between each other.
Ketodienes eluted as four chromatographic peaks, while
hydroperoxydienes were only resolved as three peaks.
The values found for LOD and LOQ were as low as
0.1–0.2 and 0.3–0.6 lg/mL, respectively, or 0.3–0.6 and
1.0–2.0 expressed in lM. The molar extinction coefficients
were coherent with those reported in the literature [21, 22],
as well as their relative values with the HPLC response
factors obtained in this study.
123

1276
J Am Oil Chem Soc (2010) 87:1271–1279
Formation of Hydroperoxydienes, Ketodienes
Likewise, hydroxydienes eluted as three peaks, the second
of which was split depending on the concentration.
Although the determination of the different isomers was
out of the scope of this study, we attempted to assign the
different chromatographic peaks to oxidation compounds.
Thus, assignation was performed according to the work of
and Hydroxydienes: Influence of the Degree
of Unsaturation
Results for FAMEs derived from HLSO and HOSO during
autoxidation at 40 °C are listed in Tables 2 and 3,
respectively. The analysis of polymers by HPSEC, which
was used to discard samples with advanced oxidation,
showed two well-defined oxidation periods. A first stage
was characterized by very low contents of polymers,
reaching values up to around 1.0% with respect to the
initial content. Then, polymerization was accelerated in the
second period. The end of the first period was detected at
76 h in the high linoleic (HL) sample (Table 2) and at
144 h in the high oleic sample (HO) (Table 3).
¨
Hopia et al. [13] and Makinen and Hopia [23]. Because of
its changing profile along oxidation, the second peak of
hydroxydienes can be assigned to the coelution of the
9-cis,trans and 13-trans,trans isomers, and the same was
adopted for the hydroperoxydienes. Regarding ketodienes,
¨
Makinen and Hopia [23] obtained two pairs of peaks that
were assigned in order of elution to the cis,trans and
trans,trans isomers, so a specific assignation to the two
regioisomers was not given. In this study, the analysis of
the 13-cis,trans-MeLO that was synthesized showed that
the major peak corresponded to the peak of the shortest
retention time (Fig. 1), so the first and the second peaks
were assigned to the 13-cis,trans and the 9-cis,trans-
MeLO, respectively, and the same was tentatively adopted
for the trans,trans-MeLO isomers.
Hydroperoxydienes were already detected in the starting
samples at very low concentrations (0.10 lg/mg HL sam-
ple and 0.02 lg/mg HO sample), even when the PV was
zero. Ketodienes and hydroxydienes were also detected
in the starting samples, however, their contents were
below the LOQ, i.e. lower than 0.02 and 0.05 lg/mg,
Table 2 Autoxidation of purified FAMEs derived from high linoleic sunflower oil at 40 °C in the dark
Time (h)
Pol (%)
PV (meq/kg)
K232
Hydroperoxydienes (lg/mg)
Ketodienes (lg/mg)
Hydroxydienes (lg/mg)
0
1.0
1.0
1.3
1.3
2.0
2.0
6.5
8.6
19.3
0.0 0.0
11.3 1.2
128 4.7
184 1.4
485 1.4
809 7.8
1453 17.4
–
2.23 0.09
0.10 0.00
1.04 0.01
12.74 0.20
19.01 0.90
55.65 0.31
86.04 0.95
131.33 1.73
–
\0.02
\0.05
20
44
51
68
76
91
100
115
2.73 0.09
11.60 0.11
17.96 0.89
44.84 3.90
73.77 3.20
107.69 8.58
–
\0.02
\0.05
0.082 0.005
0.112 0.007
0.662 0.026
1.107 0.004
3.094 0.099
–
0.060 0.005
0.066 0.005
0.212 0.011
0.550 0.030
1.338 0.080
–
–
–
–
–
–
Pol FAME polymers, PV peroxide value, K232 specific extinction coefficient at 232 nm
Table 3 Autoxidation of FAMEs derived from purified high oleic sunflower oil at 40 °C in the dark
Time (h)
Pol (%)
PV (meq/kg)
K232
Hydroperoxydienes (lg/mg)
Ketodienes (lg/mg)
Hydroxydienes (lg/mg)
0
0.9
1.1
1.1
1.2
1.3
1.5
1.8
2.3
4.0
5.3
6.7
0.0 0.0
20.0 0.7
79.0 4.1
163 1.4
275 0.0
384 7.8
558 2.1
818 0.7
1053 16.3
–
2.19 0.03
0.02 0.00
0.90 0.01
4.15 0.16
8.66 0.06
15.16 0.49
21.36 0.06
30.71 0.07
36.45 0.33
42.29 0.08
–
\0.02
\0.05
25
2.83 0.08
4.45 0.07
8.32 0.00
13.37 0.00
18.27 0.38
26.12 0.03
39.82 0.99
45.13 0.85
–
\0.02
\0.05
48
0.031 0.003
0.084 0.006
0.231 0.030
0.397 0.001
0.664 0.010
1.239 0.027
1.726 0.030
–
0.056 0.004
0.096 0.005
0.162 0.003
0.198 0.002
0.291 0.005
0.459 0.004
0.590 0.035
–
72
96
120
144
168
192
216
240
–
–
–
–
–
Pol FAME polymers, PV peroxide value, K232 specific extinction coefficient at 232 nm
123

J Am Oil Chem Soc (2010) 87:1271–1279
1277
100.00
80.00
60.00
40.00
20.00
0.00
respectively, at the analytical conditions applied. The
coefficients of variation found for the three groups of
compounds were satisfactory, being B5% for hydro-
peroxydienes and B7% for ketodienes and hydroxydienes,
respectively.
a
Hydroperoxydienes increased exponentially over oxi-
dation time, reaching different values for the HL and HO
samples at the end of the period of slow polymerization.
Thus, while 86.0 lg/mg was found for the HL sample, the
content of hydroperoxydienes was as low as 30.7 lg/mg
for the HO sample. Likewise, ketodienes and hydroxydi-
enes increased exponentially throughout the oxidation
time, but, as expected, in a much lower range of concen-
trations. In fact, the content of ketodienes and hydroxydi-
enes as a whole constituted 2–3 wt% of the analyzed
compounds at the end of the period of slow polymerization.
Higher contents of ketodienes than of hydroxydienes were
found throughout oxidation time, and the ratio between the
contents of ketodienes and hydroxydienes increased with a
factor that changed from 1 to 2 throughout the period of
slow polymerization. A faster formation of ketodienes, as
compared to hydroxydienes, has also been found by
0.00
20.00
40.00
60.00
80.00
232
K
100.00
80.00
60.00
40.00
20.00
0.00
b
0
200
400
600
800
1000
PV (meq/kg)
¨
Makinen et al. [24] during oxidation of methyl linoleate at
Fig. 3 a Linear correlation between the content of hydroperoxydienes
and DK232 in FAMEs derived from HLSO (solid) and HOSO (hollow)
during autoxidation at 40 °C (HL: [LOOH] (lg/mg) = 1.24( 0.02)
40 °C. Although the exact mechanism responsible for the
formation of hydroxy- and ketodiene-products from lipid
hydroperoxides has not yet been established, a few mech-
anisms have been proposed [25, 26].
DK232, r = 0.999; HO: [LOOH] (lg/mg) = 1.28( 0.02) DK232
,
r = 0.999). b Linear correlation between the content of hydro-
peroxydienes and the PV in samples of FAMEs derived from HLSO
(solid circles) and HOSO (hollow circles) during autoxidation at 40 °C,
and in solutions of pure hydroperoxydienes (squares) in hexane
(HL: [LOOH] (lg/mg) = 0.110( 0.002) PV (meq/kg), r = 0.999;
HO: [LOOH] (lg/mg) = 0.0555( 0.0003) PV (meq/kg), r = 0.999;
Pure hydroperoxydienes: [LOOH] (lg/mg) = 0.119( 0.004) PV
(meq/kg), r = 0.998)
The content of hydroperoxydienes showed, as expected,
a significant linear correlation with the extinction coeffi-
cient at 232 nm (K232) during the period of slow poly-
merization for the two FAME samples. For comparative
purposes, the increment of the extinction coefficient with
respect to the starting sample (DK232) was examined
(Fig. 3a). Results showed no significant differences in the
linear correlation analysis between the HL and HO sam-
ples, indicating that the compounds analyzed, i.e. hydro-
peroxydienes, were to a great extent the most abundant
compounds absorbing light at 232 nm.
Although oxidation of oleic acid is known to be much
slower than that of linoleic acid when both substrates are
tested independently [27], the results showed that oxidation
of oleic acid was favored in mixtures of oleic and linoleic
acids and that the participation of oleic acid in the oxida-
tion chain was a function of its concentration.
The content of hydroperoxydienes also showed a sig-
nificant linear correlation with the PV during the period of
slow polymerization, but clear differences were found
between the HL and HO samples (Fig. 3b). As an example,
when the PV was 100 meq/kg the content of hydro-
peroxydienes was determined to be 11.5 and 5.6 lg/mg for
the HL and HO samples, respectively. These differences
indicate that a significant formation of hydroperoxides
other than hydroperoxydienes occurred in the HO sample.
As the only oxidizing substrates were oleic and linoleic
acids and the formation of polymerization compounds was
very low, the differences between the HL and HO samples
can be mainly attributed to the oxidation of oleic acid,
whose hydroperoxides do not possess conjugated double
bonds and are not detected in the HPLC–UV analysis.
In order to determine the relationship between pure
hydroperoxydienes and PV, methyl linoleate was oxidized
and hydroperoxides were isolated and purified. Then
solutions of known concentrations of the pure hydro-
peroxydienes in hexane were titrated with sodium thio-
sulfate applying the same conditions as those of the PV
test. For comparative purposes, the results of the titration
were expressed supposing the same amount of lipid sample
that was used in the PV test for the HO and HL samples,
i.e. 0.5 g. Likewise, the hydroperoxydiene concentrations
(lg/mL) were also referred to as a hypothetical amount of
sample of 0.5 g. As both parameters were modified by a
same factor, the relationship between each other remained
unaffected. Results showed an excellent linear correlation
123

1278
J Am Oil Chem Soc (2010) 87:1271–1279
2.00
1.00
0.00
between the content of pure hydroperoxydienes and the PV
(Fig. 3b). The slope of the curve was quite close to that of
the HL sample, but statistically different. From the slopes
of the curves we can estimate the content of hydro-
peroxydienes relative to that of the compounds giving
response in the PV test, i.e. total hydroperoxides. Thus, the
hydroperoxydienes analyzed in this study constituted 92%
( 3) of total hydroperoxides in the HL sample and 47%
( 3) in the HO sample. Therefore, this approach suggests
that formation of hydroperoxides from oleic acid was
around 10 and 50% in the HL and HO samples, respec-
tively. These figures should be considered as estimations,
since accurate results would also require the direct deter-
mination of the hydroperoxides from oleic acid. The results
obtained in this study are in agreement with those of a
previous report, in which oxidation of purified HLSO,
HOSO and methyl linoleate was studied at 40 °C in the
dark [28]. Results for the relationship between K232 and
the PV for the three lipid substrates were similar to those of
the present study (data not shown). In addition, the loss of
fatty acids in HLSO and HOSO exhibited that while the
oxidation of oleic acid was very low in HLSO, the loss of
oleic acid was similar to that of linoleic acid in HOSO.
The contents of both ketodienes and hydroxydienes with
respect to hydroperoxydienes were significantly different for
the HO and HL samples. Surprisingly, for a given content of
their precursor hydroperoxydienes, the contents of ketodi-
enes and hydroxydienes were higher for the sample with a
lower content of linoleic acid, i.e. the HO sample (Fig. 4).
Therefore, factors controlling the formation of these sec-
ondary oxidation products other than the content of hydro-
peroxydienes were involved. Among them, a lower tendency
to polymerization of monounsaturated oils as compared to
polyunsaturated oils has been clearly observed at high tem-
perature [29, 30]. Under these circumstances, the monomers-
to-polymers ratio would be favored for the less unsaturated
substrate in the complex set of oxidation reactions.
a
0.00
20.00
40.00
60.00
80.00
100.00
Hydroperoxydienes ( g/mg)
2.00
1.00
0.00
b
0.00
20.00
40.00
60.00
80.00
100.00
Hydroperoxydienes ( g/mg)
Fig. 4 Contents of ketodienes (a) and hydroxydienes (b) as a
function of the content of hydroperoxydienes in FAMEs derived
from HLSO (solid) and HOSO (hollow) during autoxidation at 40 °C
total hydroperoxides in the HL and HO samples, respec-
tively, suggesting that a significant oxidation of oleic acid
also occurred in both lipid substrates.
Formation of the secondary oxidation products, i.e.
ketodienes and hydroxydienes, was already detected in the
non-heated samples, when the content of hydroperoxydienes
was as low as 0.02 lg/mg and the PV was zero. The content
of ketodienes and hydroxydienes as a whole constituted 2–3
wt% of the diene compounds analyzed at the end of the
period of slow polymerization. Higher contents of ketodi-
enes than of hydroxydienes were found throughout the
oxidation time, and the ratio between the contents of keto-
dienes and hydroxydienes increased by a factor that changed
from 1 to 2 throughout the period of slow polymerization.
Conclusions
The quantitative analysis of hydroperoxydienes during
autoxidation of FAMEs derived from HLSO and HOSO has
showed that with respect to the oxidation extent, as detected
by the PV, the formation of hydroperoxydienes depends on
the relative composition of the oxidizing fatty acids, i.e. oleic
and linoleic acids. Thus, for a given PV, higher contents of
hydroperoxydienes were found in the sample with a higher
content of linoleic acid. At the end of the period of slow
polymerization (DPol B 1 wt%), the content of hydro-
peroxydienes was found to be 86.0 and 30.7 lg/mg for the
HL and HO samples, respectively. Throughout this period
hydroperoxydienes constituted around 90 and 50 wt% of
Acknowledgments This work was funded by ‘‘Ministerio de
´
Ciencia e Innovacion’’ through project AGL 2007-62922/ALI and by
´
‘‘Junta de Andalucıa’’ through project AGR-4622. The authors thank
Prof. Javier Hidalgo and Prof. Rosario Zamora for assistance in the
synthesis and identification of compounds.
References
1. Cohn J (2002) Oxidized fat in the diet, postprandial lipaemia and
cardiovascular disease. Curr Opin Lipidol 13:19–24
2. Riemersma RA (2002) Analysis and possible significance of
oxidised lipids in foods. Eur J Lipid Sci Technol 104:419–420
123

J Am Oil Chem Soc (2010) 87:1271–1279
1279
3. Kanazawa A, Sawa T, Akaike T, Maeda H (2002) Dietary lipid
peroxidation products and DNA damage in colon carcinogenesis.
Eur J Lipid Sci Technol 104:439–447
17. Hidalgo FJ, Zamora R, Vioque E (1992) Syntheses and reactions
of methyl (Z)-9, 10-epoxy-13-oxo-(E)-11-octadecenoate and
methyl (E)-9, 10-epoxy-13-oxo-(E)-11-octadecenoate. Chem
Phys Lipids 60:225–233
18. Cohen JD (1984) Convenient apparatus for the generation of
small amounts of diazomethane. J Chromatogr 303:193–196
19. IUPAC (1992) Standard methods for the analysis of oils, fats and
derivatives, 7th edn. International Union of Pure and Applied
Chemistry. Blackwell Scientific, Oxford
´
4. Dobarganes MC, Marquez Ruiz G (2003) Oxidized fats in foods.
Curr Opin Clin Nutr Metab Care 6:157–163
5. Staprans I, Pan XH, Rapp J, Feingold KR (2005) The role of
oxidized cholesterol and oxidized fatty acids in the development
of atherosclerosis. Mol Nutr Food Res 49:1075–1082
´
6. Marquez-Ruiz G, Dobarganes MC (2007) Nutritional and phys-
iological effects of used frying fats. In: Perkins EG, Erickson MD
(eds) Deep frying: chemistry nutrition and practical applications.
AOCS Press, Champaign, pp 173–203
20. AOCS (1998) Official methods and recommended practices of
the American Oil Chemists’ Society, 5th edn. AOCS Press,
Champaign
7. Kanner J (2007) Dietary advanced lipid oxidation end products are
risk factors to human health. Mol Nutr Food Res 51:1094–1101
21. Chan HWS, Levett G (1977) Autoxidation of methyl linleate.
Separation and analysis of isomeric mixtures of methyl linoleate
hydroperoxides and methyl hydroxylinoleates. Lipids 12:99104
22. Fishwick MJ, Swoboda PA (1977) Measurement of oxidation of
polyunsaturated fatty acids by spectrophotometric assay of con-
jugated derivatives. J Sci Food Agric 28:387–393
´
8. Guillen MD, Goicoechea E (2008) Toxic oxygenated a, b-
unsaturated aldehydes and their study in foods: a review. Crit Rev
Food Sci Nutr 48:119–136
9. Frankel EN (2005) Lipid oxidation, 2nd edn. In: Frankel EN (ed).
The Oily Press, Dundee
¨
23. Makinen EM, Hopia AI (2000) Effects of a-tocopherol and
ascorbyl palmitate on the isomerization and decomposition of
´
10. Velasco J, Berdeaux O, Marquez-Ruiz G, Dobarganes MC (2002)
Sensitive and accurate quantitation of monoepoxy fatty acids in
thermoxidized oils by gas-liquid chromatography. J Chromatogr
A 982:145–152
methyl linoleate hydroperoxides. Lipids 35:1215–1223
¨
24. Makinen EM, Kamal-Eldin A, Lampi A-M, Hopia A (2000)
Effects of a- and c-tocopherol on formation of hydroperoxides
and two decomposition products from methyl linoleate. J Am Oil
Chem Soc 77:802–806
´
11. Velasco J, Marmesat S, Berdeaux O, Marquez-Ruiz G,
Dobarganes C (2005) Quantitation of short-chain glycerol-bound
compounds in thermoxidized and used frying oils. A monitoring
study during thermoxidation of olive and sunflower oils. J Agric
Food Chem 53:4006–4011
25. Russell GA (1957) Deuterium-isotope effects in the autoxidation
of aralkyl hydrocarbons: mechanism of the interaction of peroxy
radicals. J Am Chem Soc 79:3871–3877
12. Marmesat S, Velasco J, Dobarganes MC (2008) Quantitative
determination of epoxy acids, keto acids and hydroxy acids
formed in fats and oils at frying temperatures. J Chromatogr A
1211:129–134
13. Hopia A, Huang SH, Frankel EN (1996) Effect of a-tocopherol
and Trolox on the decomposition of methyl linoleate hydroper-
oxides. Lipids 31:357–365
26. Labuza TP (1971) Kinetics of lipid oxidation in foods. CRC Crit
Rev Food Technol 2:355–405
27. Holman RT, Elmer OC (1947) The rates of oxidation of unsat-
urated fatty acids and esters. J Am Oil Chem Soc 24:127–129
28. Marmesat S, Morales A, Velasco J, Ruiz-Mendez MV,
Dobarganes MC (2009) Relationship between changes in perox-
ide value and conjugated dienes during oxidation of sunflower
oils with different degree of unsaturation. Grasas y Aceites
60:155–160
´
14. Berdeaux O, Marquez-Ruiz G, Dobarganes C (1999) Selection of
methylation procedures for quantitation of short-chain glycerol-
bound compounds formed during thermoxidation. J Chromatogr
A 863:171–181
´
29. Marquez-Ruiz G, Tasioula-Margari M, Dobarganes MC (1995)
Quantitation and distribution of altered fatty acids in frying fats.
J Am Oil Chem Soc 72:1171–1176
´
´
15. Marquez-Ruiz G, Martın-Polvillo M, Dobarganes MC (1996)
Quantitation of oxidized triglyceride monomers and dimers as a
useful measurement for early and advanced stages of oxidation.
Grasas y Aceites 47:48–53
´
30. Velasco J, Marmesat S, Berdeaux O, Marquez-Ruiz G, Dobar-
ganes C (2004) Formation and evolution of monoepoxy fatty
acids in thermoxidized olive and sunflower oils and quantitation
in used frying oils from restaurant and fried-food outlets. J Agric
Food Chem 52:4438–4443
16. Yoshida H, Kondo I, Kajimoto G (1992) Participation of free
fatty acids in the oxidation of purified soybean oil during
microwave heating. J Am Oil Chem Soc 11:1136–1140
123