Journal of Chromatography A, 1218 (2011) 84968502

Contents lists available at SciVerse ScienceDirect

Journal of Chromatography A
j our nal home page: www. el sevi er . com/ l ocat e/ chr oma
Method for simultaneous analysis of eight vitamin E isomers in various foods by
high performance liquid chromatography and uorescence detection
Helena Maria Pinheiro-SantAna
a,
, Michele Guinazi
a
, Daniela da Silva Oliveira
b
,
Ceres Mattos Della Lucia
a
, Brbara de Lazzari Reis
a
, Sebastio Csar Cardoso Brando
b
a
Laboratory of Vitamins Analysis, Department of Nutrition and Health, Federal University of Vic osa, Avenida Purdue, s/n, Campus Universitrio, Vic osa, MG CEP 36561-000, Brazil
b
Department of Food Technology, Federal University of Vic osa, Avenida Purdue, s/n, Campus Universitrio, Vic osa, MG CEP 36561-000, Brazil
a r t i c l e i n f o
Article history:
Received 4 May 2011
Received in revised form
16 September 2011
Accepted 20 September 2011
Available online 1 October 2011
Keywords:
Tocopherols
Tocotrienols
Vegetables
Vegetable oils
Eggs
Seeds
a b s t r a c t
The objective of this study was to optimize a method to investigate the occurrence and to quantify the full
isomeric composition of vitamin E (-, -, - and -tocopherols and tocotrienols) in 6 vegetables (raw
and cooked), 3 herbs/spices, raw and cooked eggs, vegetable oils (canola, olive and soybean), axseed
and sorghum (our and seeds) and soy (our) by HPLC with uorescence detection. Different conditions
of extraction and analysis were tested. The optimized method consisted of direct extraction with sol-
vent (hexane:ethyl acetate, 85:15, v/v). For analysis normal phase column was used with mobile phase
consisting of hexane:isopropanol:acetic acid (98.9:0.6:0.5) with isocratic elution and uorescence detec-
tion. Excellent separation of all isomers was obtained along with adequate quantication in the foods
analyzed. Recovery rates of standards ranged from 91.3 to 99.4%. The linearity range for each isomer
varied from 2.5 to 137.5 ng/mL (R
2
greater than 0.995 in all cases). Detection limits ranged from 21.0
to 48.0 ng/mL for tocopherols and from 56.0 to 67.0 ng/mL for tocotrienols, while quantication limits
ranged from 105.0 to 240.0 ng/mL for tocopherols and from 280.0 to 335.0 ng/mL for tocotrienols. The
optimized method was considered simple, fast and reliable, and also preserved vitamin E isomers when
compared to validated methods involving saponication.
2011 Elsevier B.V. All rights reserved.
1. Introduction
VitaminE is composedby a groupof eight isomers, -, -, - and
-tocopherols and tocotrienols, which have biological activity. This
vitamin is considered one of the major antioxidants, since it pro-
tects cell membranes fromoxidation and lowdensity lipoproteins
against lipidperoxidation. Inadditiontoits antioxidant activity and
its ability to scavenge free radicals, it may reduce the risk of cancer
and prevent progression of precancerous lesions. Thus, knowledge
of the vitamin E content in foods is of great importance to ensure
the optimal daily intake as an essential factor in human health [1].
Vitamin E is widespread in plant tissues and to a lesser extent in
animal foods such as eggs and liver. Vegetable oils are considered
the most important sources, followed by oilseeds and dark green
leafy vegetables [2]. Among these foods, oils are the focus of many
studies and thus are better characterized regarding their vitamin E
content. Moreover, although both vegetables and oilseeds consti-
tute important sources of vitamin E, there is little data available in
literature on its contents in these foods [3].

Corresponding author. Tel.: +55 31 3899 1684.

E-mail address: helena.santana@ufv.br (H.M. Pinheiro-SantAna).
Historically, -tocopherol has been reported as the isomer with
greatest biological activity. Therefore, most methods for determin-
ing vitamin E available in the literature were developed exclusively
for determination of -tocopherol. However, many studies have
shown that the antioxidant activity of other isomers also has great
importance [4].
The structural complexity and the different antioxidant poten-
tial of the group of compounds with vitamin E activity require that
reliable analytical techniques are applied for extraction, separa-
tion, identication and quantication of individual components in
various types of food matrices [5]. Among these techniques, high
performance liquidchromatography(HPLC) has beenusedsuccess-
fully in the analysis of compounds present in small amounts in
complex matrices [6].
Fluorescence is the chromatographic detection technique used
for the analysis of tocopherols and tocotrienols in most studies per-
formed with HPLC. This preference for detection by uorescence is
attributed to its higher selectivity, sensitivity and specicity com-
paredtoabsorbancedetection[7]. LCMSmethods for thevitaminE
analysis of foods, both with normal and reversed chromatographic
phases have been found in the literature [8].
The full isomeric composition of vitamin E in Brazilian
food has rarely been studied. In Brazil, there is great need
for studies in this area because there is no information
0021-9673/$ see front matter 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.chroma.2011.09.067
H.M. Pinheiro-SantAna et al. / J. Chromatogr. A 1218 (2011) 84968502 8497
available regarding the vitamin E content on food composition
tables.
Thus, the objective of this study was to optimize a method to
investigate the occurrence and content of eight different vitamin E
isomers by HPLC with uorescence detection in vegetables, eggs,
vegetable oils, ours and seeds.
2. Materials and methods
2.1. Raw material
Foods (vegetables, vegetable oils, eggs, seeds and ours) were
selected based on those indicated in literature as major sources of
vitamin E, and for most of these samples no specic information
regarding the content of tocopherols and tocotrienols was encoun-
tered, especially in Brazilian foods. The samples were:
(a) Vegetables andherbs/spices: watercress (Nastrirtiumofcinale),
chicory (Hieracium commursonil), broccoli (Brassica oleracea,
var. italica), arugula (Eruca sativa), kale (Brassica oleracea, var.
acephala), spinach (Spinacia oleracea), green peppers (Cap-
sicum annuum), chives (Allium schonoeprasum) and parsley
(Petroseliumcrispum);
(b) Egg yolk (rawand cooked);
(c) Vegetable oils: soybean, canola and olive;
(d) Whole seeds: sorghum (Sorghum bicolor) and axseed (Linum
usiatissimimL.);
(e) Whole ours: sorghum, axseed and soybean (Glycine max L.).
2.2. Reagents and other materials
For sample preparation, the following analytical grade reagents
were used: ethyl acetate, hexane and isopropanol (Synth, Brazil),
sodium sulfate (Vetec, Brazil), and butylhydroxytoluene (BHT)
(Synth, Brazil).
The vitamin standards used (-, -, - and -tocopherols and
tocotrienols) were purchased fromCalbiochem

, EMD Biosciences,
Inc. (USA) and stored at 18

C until use.
The mobile phases were prepared using HPLC grade reagents:
hexaneandisopropanol (Mallinckrodt, USA) andacetic acid(Merck,
Brazil).
Filtration of the samples included use of lter paper n

JP41
(J. Prolab, Brazil), 3mL sterile syringes (Rymco, Colombia), and
polyethylene lter units Millex HV (Millipore, 0.45m, Brazil).
2.3. Equipment
The following equipment were used for extraction of vitamin
E in the samples: microgrinder, model MA 102 (Marconi, Brazil),
vacuum pump model CA (FANEN, Brazil) and rotative evaporator,
Q-344.1 (Quimis, Brazil). The mobile phases were degassed in an
ultrasonic vibrator T-14 (Odontobras, Brazil). A spectrophotome-
ter UV 1601 (Shimadzu, Japan) was used to scan the absorption
spectra of vitamin standards. Analyses were performed by a high
performance liquid chromatography system (Shimadzu, Japan)
equipped with a high-pressure pump, model LC-10AD VP; auto
injector SIL-10AF withloopof 50L; columnLiChrosorb 5m, Si60
Phenomenex, 250mm4mm; and uorescence detector model
RF-10A XL. The systemwas controlled by the software Multi Sys-
tem Class VP 6.1.
2.4. Methods
2.4.1. Collection and sample preparation
The vegetable and egg samples, purchased locally, were brought
to the laboratory where they were stored in refrigerator for a
maximumperiod of 5 days at temperatures between 0

C and +5

C
until extraction and analysis. Vegetable oils were purchased from
local markets. For a better representation of the analysis, care was
taken to purchase oil products that were of the same brand, but
fromdifferent lots.
Flaxseed was bought locally. Sorghumwas developed and sup-
plied by Embrapa Maize and Sorghum (Sete Lagoas, Minas Gerais,
Brazil). The sorghum seeds were manually selected and subjected
to sieving to remove dirt and impurities.
To obtain axseed our, the seeds were crushed in a blender at
mediumspeed for 5min and sieved to obtain a maximumparticle
size of 20 mesh.
For the production of our from sorghum and soybean, the
seeds were ground with the pericarp in a knife grinder (Braben-
der, Germany), with sieve-limit number 0 (zero) for 10min or until
all material passed through the sieve.
2.4.2. Preparation of vitamin standards and analytical curves
Stock solutions were prepared with 5mg of each standard dis-
solved in hexane containing BHT 0.01% and the volume completed
to 100mL, resulting in concentrations of 50ng/mL.
Real concentrations of the vitamin E isomers were veried by
spectrophotometry and corrected using the LambertBeer equa-
tion. The coefcients of molar absorptivity (in E
1%
) were: 70.8 for
-tocopherol, 86.4for -tocopherol, 92.8for -tocopherol and91.2
for -tocopherol in a 96% ethanol solution. Since the specic values
of the molar absorptivity coefcients of tocotrienols are not found
in literature, we used in this study the maximum wavelength of
each tocopherol isomer corresponding to the real concentration of
-, -, - and -tocotrienols, as indicated by Piironen et al. [9]. In
additionto spectrophotometric analysis, standardsolutions of each
of the isomers of vitamin E were individually analyzed by HPLC in
order to conrmtheir purity. As we detected only one peak for each
of the standards, we concluded that they had a high level of purity
and did not co-elute with any other component on the same wave-
length. The percentage of purity of the standards can be seen in
Table 1. Although it was observed that all eight standards showed
high levels of purity, no correction of the gravimetric concentra-
tions could result in errors in the quantication of compounds in
samples.
Analytical curves of the vitamin E isomers were constructed
using ve increasing concentrations of each isomer, with injections
of each concentration in triplicate. A linear correlation between
peak areas and concentrations of each injected isomer was deter-
mined. The linear regressionequationobtainedfor eachisomer was
used to calculate the content of isomers in food samples.
2.5. Optimization of the extraction method
All tests were performed in the Laboratory of Vitamin Analysis
of the Department of Nutrition and Health, Federal University of
Vic osa, Minas Gerais, Brazil.
For the extraction of vitamin E isomers, different methods
described in literature were tested and evaluated in accordance
with the response in resolution and quantication of compounds
by HPLC.
Methods involving saponication using different combinations
of time, temperature and potassium hydroxide (KOH) concentra-
tion, as well as a direct extraction technique with solvents were
evaluated. The procedures tested for extraction prior to chromato-
graphic analysis are presented in Table 2.
Because the method involving direct extraction with solvents
presented the best results in this study, it was used for the extrac-
tion of vitamin E isomers, as described below.
The extraction procedure (based on [11]) involved the extrac-
tion of the isomers, evaporation of solvents and dissolution in a
8498 H.M. Pinheiro-SantAna et al. / J. Chromatogr. A 1218 (2011) 84968502
Table 1
Maximumwavelength, absorbance and purity of standards of tocopherols and tocotrienols in solution of ethanol 96%.
Standard max (nm) Absorbance (A) Gravimetric concentration
a
(g/mL) Spectrophotometric concentration
b
(g/mL) Purity (%)
-T 292 0.380 55.00 53.67 97.58
-T 295.5296.5 0.477 57.00 55.21 96.86
-T 297298 0.475 57.00 52.08 91.37
-T 297298 0.468 56.00 50.23 90.05
-T3 289.5291 0.343 54.00 48.45 89.72
-T3 296 0.447 56.00 51.74 92.39
-T3 298.6299.5 0.497 57.00 53.56 93.86
-T3 297298.5 0.484 56.00 53.07 94.77
a
Uncorrected for reference standard purity.
b
Purity corrections made through HPLC with uorescence detection (excitation at 290nmand emission at 330nm) and spectrophotometer analysis.
known volume. For this, the following steps were taken. Roughly
10g of vegetable and egg and 4g of samples of sorghum, axseed
(our andseeds) andsoy(our) were weightedina tube. Next, 4mL
of heated water (about 80

C) was added with 5g of anhydrous

sodium sulfate and mixed with a spatula. Subsequently, 10mL
of isopropanol and 1mL of hexane containing BHT 0.05% were
added. After adding 25mL of the solvent mixture (hexane:ethyl
acetate, 85:15, v/v), the samples were ground using microgrinder
on mediumspeed for 1min. Once crushed, the samples were vac-
uum ltered using quick ltering paper. The residue retained on
the lter was washed with 15mL of the solvent mixture. The
residue was then transferred to the tube and the extraction step
was repeated by adding 5mL of isopropanol and 30mL of the sol-
vent mixture, followedbyhomogenizationandsubsequent vacuum
ltration. The residue was again washed with 5mL of the solvent
mixture, then the ltrate was transferred to a ask and the solvents
were evaporatedina rotative evaporator at 70

Ctoconcentrate the
volume of the extract and re-dissolve it later in a known volume.
Subsequently, theextracts werequantitativelytransferredto25mL
volumetric asks and the volume completed witha mixture of hex-
ane and ethyl acetate (85:15, v/v). After complete homogenization
in the ask, an aliquot of 5mL was removed and evaporated under
nitrogen gas, stored in hermetically sealed amber glass asks at
temperatures of 5

Cto0

Cuntil chromatographic analysis, which

occurred in less than 4h.
Fluorescent andnatural light was avoidedinall operations using
amber glassware or protective aluminumfoil.
Analysis of oils did not require the extraction stage. They were
simply diluted in hexane (0.1g in 10mL of hexane) and directly
injected into the column for analysis (adapted from[14]).
Table 2
Extraction conditions evaluated for determination of tocopherols and tocotrienols
in foods by HPLC.
Extraction conditions Original analysis Reference
Saponication under heating
Use of KOH 60%, at 70

C
for 50min
Extraction with ethanol
and hexane
Determination of - and
-tocopherol and tocotrienol in
tomato and broccoli
[6]
Saponication at room
temperature
Use of KOH 50%, at room
temperature, overnight
Extraction with ethanol
and hexane
Determination of tocopherols and
tocotrienols in vegetables, fruits
and diets
[9,10]
Direct extraction with
solvents
Use of isopropanol and
mixture of solvents
compound by hexane and
ethyl acetate (85:15, v/v)
Determination of tocopherols in
infant cereals extrudates
[11]
2.5.1. Optimization of HPLC conditions for analysis
The efciency of chromatographic separation of eight vitamin
E isomers was evaluated from tests with different proportions of
components inthemobilephase, owrateandwavelengthof great-
est sensitivity of the detector, according to conditions suggested in
theliteratureandreportedinTable3. For this, weusedtheHPLCand
a normal phase system, with greater efciency for the separation
of isomers, especially -tocopherol and -tocotrienol.
The choice of mobile phase components and the excitation and
emission wavelengths was made considering the use of a normal
phase systemand the conditions described in literature for studies
with similar analyses [7,12]. In most cases, the separation of -
tocopherol and -tocotrienol isomers was the decisive factor in
assessing the conditions tested.
Identication of compounds was performed by comparison
of retention times obtained for standards and samples, analyzed
under the same conditions, by co-chromatography andcomparison
of quantitative results observed in literature to verify consistency
of data.
Analyses of the vitamin E isomers were conducted using
the chromatographic condition that presented the best results.
Thus, the selected conditions included: a normal phase column
LiChrosorb Si60, 5m; 250mm 4mm, uorescence detec-
tion (Ex 290nm and Em 330nm), mobile phase consisting of
hexane:isopropanol:acetic acid (98.9:0.6:0.5), and ow rate of
1mL/min.
After extraction, aliquots (5mL) were dried in nitrogen, redis-
solved in hexane (2mL) and ltered through 0.45m lter units.
Analyses were performed by injecting two different volumes (5
and 50L) for each sample in order to obtain the detection of all
compounds in quantities appropriate for identication and quan-
tication, since some vitamin E isomers can only be adequately
quantied when injected in volumes of 50L, due to their lowcon-
centrations in the samples. Fromthe start of sample processing to
extractiontoinjection, the total time required, for eachsample, was
one and a half hour.
2.5.2. Determination of linearity range
The linearity range was obtainedusing the previously optimized
chromatographic conditions fromthe injection, in triplicate, of six
increasing concentrations of the standard solution of each isomer.
The data obtainedfor the peak areas were usedfor linear regression
analysis. The correlation coefcient (R
2
) obtained in each case was
used to assess linearity [16].
2.5.3. Analysis of the standards recovery
Standards recovery tests were performed from the addition of
knownconcentrations of the different isomers insamples of water-
cress, peppers, parsley, arugula and axseed.
To evaluate recovery, samples were carefully weighed and
homogenized in two tubes in order to obtain two very similar
aliquots. In one sample, standards in concentrations sufcient to
H.M. Pinheiro-SantAna et al. / J. Chromatogr. A 1218 (2011) 84968502 8499
Table 3
Chromatographic conditions tested for determination of tocopherols and tocotrienols in foods by HPLC.
Mobile phase Flowrate (mL/min) excitation and emission (nm) Reference
HX 1.0 290 and 330 [7,12]
HX:IP (99:1) 1.0 285 and 325 [12,13]
HX:IP (99.1:0.9) 1.0 285 and 325 [11,12]
HX:IP (99.4:0.6) 1.0 290 and 330 [12]
HX:IP (99.4:0.6) 1.5 285 and 325 [7,12]
HX:IP (99.5:0.5) 1.0 290 and 330 [12,15]
HX:IP (99.7:0.3) 1.0 290 and 330 [12,14]
HX:IP (99.7:0.3) 1.5 285 and 325 [7,12]
HX:IP:HAC (99.35:0.6:0.05) 1.0 290 and 330 [12]
HX:IP:HAC (99.33:0.6:0.07) 1.0 290 and 330 [12]
HX:IP:HAC (99.33:0.6:0.07) 1.0 285 and 325 [12]
HX:IP:HAC (99.33:0.6:0.07) 1.5 290 and 330 [12]
HX:IP:HAC (99.33:0.6:0.07) 2.0 290 and 330 [12]
HX:IP (99.6:0.6), pH adjusted with HAC 2.0 290 and 330 [12]
HX, hexane; IP, isopropanol; HAC, acetic acid.
Table 4
Evaluation of extraction conditions for determination of tocopherols and
tocotrienols in foods by HPLC.
Extraction conditions Resolution of
compounds
a
Analytical
values
b
Saponication under heating +
Saponication at roomtemperature +
Direct extraction with solvents +++ +++
a
Rs: , Rs <0.6; , Rs =0.6; +, Rs =0.7 a 0.8; ++, Rs =0.9; +++, Rs =1.0 a 1.25.
Rs: 0.6, bad; 0.8, weak; 1.0, enough; 1.25, very good.
b
Values of the peak area observed: , lower values; , intermediate values;
+++, higher values.
represent about 50% of the original vitamin content of each sample
were added, and the other tube was used as control. The two tubes
of each sample were then subjected to the processes of extraction
and analysis. All procedures were performed in triplicate and the
samples injected in duplicate. The recovery values were obtained
from the percent difference between the contents analyzed and
added.
2.5.4. Determination of limits of detection and quantication
The limit of detection (LOD) was determined as the concentra-
tion where the ratio between the peak height of interest and the
baseline noise was 3. The limit of quantication (LOQ) was consid-
ered to be 5 times the detection limit [17].
2.5.5. Experimental design and statistical analysis of data
A randomized design was used with 3 repetitions for vegetable
oils, ours and axseed and sorghum; 5 repetitions for soybean
our, and 6 repetitions for vegetables and egg yolk. The contents
of tocopherols and tocotrienols were reported as meanstandard
deviation for each sample. The analysis of variance and Duncan test
at 5% probability were applied to detect differences between the
mean content of isomers of different food groups (vegetables and
herbs/spices, sorghum seedsorghum our, axseedaxseed
our; soybean oil olive oil canola oil). For the soybean our no
statistical analysis were used because there were no similar foods
to be compared with the specic sample. Statistical analysis was
conducted using SAS

[18], licensed to UFV, 2008.

3. Results and discussion
3.1. Optimization of extraction conditions
Results of the HPLC analysis showed that the use of about 10g of
each sample of vegetables, herbs/spices and eggs and 4g for our
and seed was sufcient to obtain suitable chromatograms for the
identication and quantication of the eight vitamin E isomers.
Among the three tested extraction methods, direct extraction
of the vitamin with organic solvents showed the best results, both
in relation to the separation of eight isomers, especially the com-
pounds -tocopherol and -tocotrienol, and the analytical values
(Table 4). Optimization of the method for analysis of tocopherols
and tocotrienols in vegetables, eggs, oils, seeds and our was
considered an improvement, since it was based on an original
Table 5
Inuence of different chromatographic conditions in the resolution of -tocopherol and -tocotrienol in foods by HPLC.
Chromatographic conditions Resolution -T and -T3
a
HX; 1mL/min; Ex 290 and Em 330nm
HX:IP (99:1); 1mL/min; Ex 285 and Em 325nm
HX:IP (99.1:0.9); 1mL/min; Ex 290 and Em 330nm
HX:IP (99.4:0.6); 1mL/min; Ex 290 and Em 330nm
HX:IP (99.4:0.6); 1.5mL/min; Ex 285 and Em 325nm
HX:IP (99.5:0.5); 1mL/min; Ex 290 and Em 330nm
HX:IP (99.7:0.3); 1mL/min; Ex 290 and Em 330nm
HX:IP (99.7:0.3); 1.5mL/min; Ex 285 and Em 325nm
HX:IP: HAC (99.35:0.6:0.05); 1mL/min; Ex 290 and Em 330nm +
HX:IP: HAC (99.33:0.6:0.07); 1mL/min; Ex 290 and Em 330nm ++
HX:IP: HAC (99.33:0.6:0.07); 1mL/min; Ex 285 and Em 325nm +
HX:IP: HAC (99.33:0.6:0.07); 1.5mL/min; Ex 290 and Em 330nm
HX:IP: HAC (99.33:0.6:0.07); 2.0mL/min; Ex 290 and Em 330nm
HX:IP: HAC (98.9:0.6:0.5); 1mL/min; Ex 290 and Em 330nm +++
-T, -tocoferol; -T3, -tocotrienol.
a
Rs: , Rs <0.6; , Rs =0.6; +, Rs =0.7 a 0.8; ++, Rs =0.9; +++, Rs =1.0 a 1.25.
Rs: 0.6, bad; 0.8, weak; 1.0, enough; 1.25, very good.
8500 H.M. Pinheiro-SantAna et al. / J. Chromatogr. A 1218 (2011) 84968502
Fig. 1. HPLC analysis of tocopherols and tocotrienols in foods. (A) Mixture of standards, (B) raw cabbage (C) raw arugula, (D) cooked egg yolk, (E) sorghum our, (F)
soybean our, (G) gold axseed our and (H) soybean oil. Chromatographic conditions: mobile phase: hexane:isopropanol:acetic acid (98.9:0.6:0.5), column LiChrosorb Si60,
uorescence detection (Ex 290 and Em 330nm), owrate 1mL/min.
method for determination of tocopherols only in infant cereal
extracts.
It is observed that although the extraction methods involving
saponication did not interfere signicantly in the chromato-
graphic separation, they were determinant in preservation
of the isomers. This signies that for the samples ana-
lyzed here, the saponication step, independent of temperature
and time applied, exerted a destructive effect on the vita-
min E content, similar to observations made in previous
studies [6,7].
3.2. Optimization of chromatographic conditions
Qualitative assessment of the interference of the chromato-
graphic conditions for separation of the vitamin E isomers,
especially -tocopherol and -tocotrienol, in the studied samples
are presented in Table 5. According to the results, it is observed that
mobile phases commonly used insimilar studies, consisting mainly
of hexane and/or isopropanol were not efcient (resolutions 0.6)
to promote a satisfactory separation of these two compounds. Sep-
aration was obtained only in the column used in this study when
H.M. Pinheiro-SantAna et al. / J. Chromatogr. A 1218 (2011) 84968502 8501
Table 6
Content
a
of tocopherols and tocotrienols in rawvegetables (mg/100g fresh weight).
Food -T -T -T -T -T3 -T3 -T3 -T3
Vegetables
Watercress 0.5677c 0.022 0.0119b 0.004 0.1431b 0.036 0.0163a 0.003 0.0088b0.003 0.0706a 0.064 0.0355a 0.013 nf
Spinach 0.1881d 0.015 0.0004b 0.013 0.0104c 0.004 0.0064a 0.001 0.00550.002 nf 0.0203a 0011 nf
Arugula 0.8825b 0.187 0.0180b 0.006 0.2868a 0.090 0.0077a 0.006 0.0111b0.003 0.0064a 0.014 0.2739a 0.422 nf
Chicory 0.4137c 0.041 0.0059b 0.003 0.2392ab 0.036 0.0127a 0.010 0.0061b0.002 nf 0.0312a 0.016 nf
Broccoli 0.4156c 0.058 0.0096b 0.002 0.2286ab 0.101 0.0058a 0.001 0.0136b0.004 0.0562a 0.030 0.0084b0.003 nf
Kale 1.1462a 0.282 0.0430a 0.031 0.3218a 0.161 0.0143a 0.012 0.0472a 0.035 0.0704a 0.102 0.0613a 0.045 nf
Herbs/spices
Chive 0.3688b 0.275 0.0411a 0.069 0.2099a 0.116 0.1059a 0.125 0.0167a 0.010 0.0231a 0.019 0.0730a 0.109 nf
Parsley 0.8383a 0.327 0.0078a 0.035 0.1915b 0.050 0.0078a 0.004 0.0214a 0.008 nf 0.0755a 0.057 nf
Green Pepper 0.2628b 0.144 0.0050a 0.009 0.0188b 0.016 0.0079a 0.045 nf 0.0089a 0.076 0.2739a 0.422 nf
nf, not found.
Means between the same types of foods in the column (vegetables and herbs/spices: n=6) followed by at least one same letter in column do not differ (=5%) by Duncan
test.
a
Meanstandard deviation.
low concentrations of acetic acid were added to mobile phase, as
indicated by Shin and Godber [19].
Theseparationof -tocopherol and-tocotrienol was onlysatis-
factory with the addition of acetic acid to the mobile phase in small
concentrations. Shin and Godber [19] reported in their study that
acetic acid favors the stability of the silica column and therefore
can assist in the resolution.
The silica column used in this study provided a good separation
of the -tocopherol and -tocotrienol isomers, which is consistent
with that reported in the literature for the same type of stationary
phase [5,12,20]. The resolution obtained (between 1.0 and 1.25)
was consideredsufcient for the quanticationof bothcompounds,
with good reliability [21].
However, there was great variability in run times during the
optimization of chromatographic conditions, causing this stage to
be the slowest. This variation, although at lowlevels, continued in
the samples analysis step, which resulted in differences in reten-
tion times. Dionisi et al. [14] raised this question in their work,
afrming that the greatest disadvantage of the normal phase sys-
temis the lowreproducibility of retention times, as well as column
equilibration and analyses being more time consuming.
3.3. Qualitative analysis of vitamin E
Fig. 1A illustrates the typical chromatographic prole obtained
for the mixture of standardsolutions of the eight vitaminE isomers.
Elutionof thecompounds for thespecic mobilephasefollowedthe
order already well established in the literature for analysis using
the normal phase system, namely: -T-T3-T-T-
T3-T3-T-T3 [10].
Fig. 1BD presents typical chromatograms of the vegetable and
egg yolk samples analyzed in this work (B, raw cabbage; C, raw
arugula; D, cooked egg yolk). It is veried that the extraction
method of the samples allowed for perfect identication and quan-
tication of the compounds of interest.
From the chromatograms of the vegetable and egg yolk sam-
ples analyzed, it is observed that all vitamin E compounds occurred
in the samples and the - and -tocopherol isomers were pre-
dominant. The other compounds occurred in small quantities, and
the -tocotrienol isomer was the rarest in the samples. Qualitative
data are coincident with the information found in the literature
[10], indicating that -tocopherol is the predominant compound
in dark green leafy vegetables and eggs, whereas -tocopherol,
-tocotrienol and -tocotrienol isomers occur in small amounts.
Thetypical chromatographic proleof theour of sorghum, soy-
bean and axseed samples is shown in Fig. 1EG. The isomers -
and -tocopherol were identied in all samples, both in seeds and
in our, -tocopherol being the predominant in all samples.
Insoybeanoil (Fig. 1H), -tocopherol was thepredominant com-
pound; -tocopherol occurredinlower concentrations. Soybeanoil
presented -tocopherol, a compound that occurs less frequently in
other foods.
3.4. Tocopherol and tocotrienol contents in food
The contents of tocopherols and tocotrienols in the samples of
vegetables, herbs/spices, egg yolks, ours, seeds and vegetable oils
are presented in Tables 6 and 7.
It is observed that the tocopherol and tocotrienol composi-
tions varied widely according to the type of food evaluated, but
Table 7
Content
a
of tocopherols and tocotrienols in eggs, whole seeds, whole ours and vegetable oils (mg/100g fresh matter).
Food -T -T -T -T -T3 -T3 -T3 -T3
Rawegg yolk 0.9766a 0.137 0.00430.001 tr 0.00210.001 0.0873a 0.005 tr 0.0084a 0.006 nf
Cooked egg yolk 0.8653a 0.056 tr 0.07590.0445 tr 0.0776a 0.010 0.17170.222 0.0147a 0.008 tr
Sorghumour 0.0846b 0.021 nf 0.2008a 0.045 nf nf nf nf nf
Sorghumseed 0.1247a 0.019 nf 0.2244a 0.048 nf nf nf nf nf
Gold axseed our 0.2677a 0.026 nf 5.5418a 0.494 0.1056a 0.037 nf nf nf nf
Brown axseed our 0.0942b 0.009 nf 5.4254a 0.589 0.917a 0.022 nf nf nf nf
Raw soybean our 0.5217 0.303 nf 2.14230.247 0.76060.044 nf nf nf nf
Soybean oil 12.1401b 1.624 2.8146b0.441 64.2734a 3.982 22.572a 1.220 nf nf 0.8104a 0.141
Canola oil 18.3931a 0.964 5.1578a 1.429 39.0238b8.929 1.1900b0.031 nf nf 1.1416a 0.218 nf
Olive oil 14.0436b 1.681 0.4741c 0.087 1.5150c 0.157 0.3691b0.190 nf nf 1.2283a 0.088
nf, not found; tr, trace =detected but not quantied.
Means between the same types of foods in the column followed by at least one same letter in column do not differ (=5%) by Duncan test.
a
Meanstandard deviation.
8502 H.M. Pinheiro-SantAna et al. / J. Chromatogr. A 1218 (2011) 84968502
in general, tocopherol isomers were found in quantities exceed-
ing tocotrienols and in a larger number of samples. Besides
-tocopherol, samples showedmoderate amounts of -tocopherol,
and reduced amounts of - and -tocopherols, -, - and -
tocotrienols. The -tocotrienol isomer was present only in trace
amounts in the samples.
It was alsonotedthat themethodusedinthis studywas effective
in extracting all isomers present in the samples, including those in
lower concentrations, which may be suitable for analysis of vita-
min E in various food matrices. The method was also suitable for
qualitative and quantitative analysis of the vitamin E isomers in
cooked foods (boiled broccoli and stir fried endive and kale) and
also toasted soybean our (data not shown).
3.5. Linearity range
The linearity range varied between 2.65 and 92.75ng for
-tocopherol, 2.70 and 135.0ng for -tocopherol, 2.65 and
132.5ng for -tocopherol, 2.65 and 132.5ng for -tocopherol,
2.50 and 130.0ng for -tocotrienol, 2.75 and 137.5ng for -
tocotrienol, 2.75 and 137.5ng for -tocotrienol and 2.70 and
135.0ng for -tocotrienol, using the conditions optimized in this
study.
It can be observed that the linearity range for each compound
was wide, which ensures acquisition of reliable data for foods with
lowand high contents of vitamin E isomers. The correlation coef-
cient (R
2
) was greater than 0.995 in all cases.
3.6. Recovery of standards
The percent recoveries ranged from 83.7 (arugula) to 101.4%
(watercress) for the tocopherol isomers and 87.5 (arugula) to
100.1%(salsa) for tocotrienol isomers. Inthe sample of rawaxseed
our, recovery ranged from85 to 94% for tocopherols.
The values foundhere were similar tothose reportedbyLee et al.
[11] whoextracting the compounds directly withsolvents obtained
an average recovery of 84%, 96.4%, 101% and 100.4% for -, -, -
and -tocopherols, respectively.
In general, excellent recovery percentages were found for the
analyzed components, which decreases the chances of losses
during extraction and analysis, and guarantees reliability of the
optimized method.
3.7. Limits of detection and quantication
The LOD obtained ranged from21.0 to 48.0ng/mL for the toco-
pherols and was between 56.0 and 67.0ng/mL for tocotrienols,
while the LOQ ranged from 105.0 to 240.0ng/mL for the toco-
pherols and between 280.0 and 335.0ng/mL for tocotrienols. It
can be noticed from the values obtained that LOD and LOQ were
low, suggesting that the standard method allows detection of small
quantities of these compounds.
4. Conclusions
The optimized methods for extraction and quantication of the
eight vitamin E isomers in the foods analyzed were efcient and
allowed the achievement of reliable results, which was considered
an improvement because it was based on an original method to
determine only tocopherols in infant cereals extracts. Detection
and quantication of the eight vitamin E isomers by HPLC was
performed in only two runs for each food, using shorter run times
(approximately 24min).
The method of direct extraction with solvents, optimized for
extraction of tocopherols and tocotrienols in the foods analyzed
was considered simple and fast, and also preserved vitamin E iso-
mers when compared to methods involving saponication that
were evaluated in this study.
Acknowledgements
We are grateful to FAPEMIG for funding the research project, to
CAPES and CNPq for granting masters and undergraduate scholar-
ships for research and to Leandro de Morais Cardoso, for editing of
the gures.
References
[1] A. Sanchz-Prez, M.M. Delgado-Zamarreno, M. Bustamante-Rangel, J.
Hernandez-Mendez, J. Chromatogr. A 881 (2000) 229.
[2] N.I. Krinsky, Mol. Aspects Med. 24 (2003) 317.
[3] A. Podsedek, LWT 40 (2007) 1.
[4] A.K. Hewavitharana, M.C. Lanari, C. Becu, J. Chromatogr. A. 1025 (2004) 313.
[5] A. Kamal-Eldin, S. Grgen, J. Petterson, A.M. Lampi, J. Chromatogr. A 881 (2000)
217.
[6] J. Lee, Y. Kin, W.O. Landen, R.R. Eitenmiller Jr., J. Food Comp. Anal. 3 (2000) 45.
[7] F.J. Ruprez, D. Martn, E. Herrera, C. Barbas, J. Chromatogr. A 935 (2001) 45.
[8] O. Heudi, M.J. Trisconi, C.J. Blake, J. Chromatogr. A. 1022 (2004) 115.
[9] V. Piironen, P. Varo, E.L. Syvoja, K. Salminen, P. Koivistoinen, Int. J. Vitam. Nutr.
Res. 53 (1984) 35.
[10] V. Piironen, E.L. Syvoja, P. Varo, K. Salminen, P. Koivistoinen, J. Agric. Food
Chem. 34 (1986) 742.
[11] J. Lee, K. Suknark, Y. Kluvitse, R.D. Phillips, R.R. Eitenmiller, J. Food Sci. 64 (1999)
968.
[12] S.L. Abidi, J. Chromatogr. A 881 (2000) 197.
[13] H.E. Indyk, Analyst 113 (1988) 1217.
[14] F. Dionisi, J. Prodolliet, E. Tagliaferri, J. Am. Oil Chem. Soc. 72 (1995) 1505.
[15] J.Y. Chen, J.D. Latshaw, H.O. Lee, D.B. Min, J. Food Sci. 63 (1998) 919.
[16] S. Albal-Hurtado, S. Novella-Rodrguez, M.T. Veciana-Nogus, A. Marin-Font,
J. Chromatogr. A 778 (1997) 243.
[17] D.B. Rodriguez-Amaya, Arch. Latinoam. Nutr. 49 (1999) 74S.
[18] SAS Institute Inc., SAS/STAT Users Guide. Version 6, 2, Fourth Edition, SAS
Institute Inc., Cary, NC, 2008, p. 846.
[19] T.S. Shin, J.S. Godber, J. Am. Oil Chem. Soc. 70 (1993) 1289.
[20] A. Pyka, J. Sliwiok, J. Chromatogr. A 935 (2001) 71.
[21] L.R. Snyder, J.J. Kirkland, Introduction to Modern Liquid Chromatography,
Wiley, NewYork, 1979.