Combination of saponification and dispersive liquid–liquid microextraction for the determination of tocopherols and tocotrienols in cereals by reversed-phase high-performance liquid chromatography

Journal of Chromatography A, 1300 (2013) 31–37

Contents lists available at SciVerse ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Combination of saponification and dispersive liquid–liquid microextraction for the determination of tocopherols and tocotrienols in cereals by reversed-phase high-performance liquid chromatography Balakrishnan Shammugasamy, Yogeshini Ramakrishnan, Hasanah M. Ghazali, Kharidah Muhammad ∗ Faculty of Food Science and Technology, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor Darul Ehsan, Malaysia

a r t i c l e

i n f o

Article history: Available online 25 March 2013 Keywords: Tocopherol Tocotrienol Analysis HPLC Microextraction Cereal

a b s t r a c t A simple sample preparation technique coupled with reversed-phase high-performance liquid chromatography was developed for the determination of tocopherols and tocotrienols in cereals. The sample preparation procedure involved a small-scale hydrolysis of 0.5 g cereal sample by saponification, followed by the extraction and concentration of tocopherols and tocotrienols from saponified extract using dispersive liquid–liquid microextraction (DLLME). Parameters affecting the DLLME performance were optimized to achieve the highest extraction efficiency and the performance of the developed DLLME method was evaluated. Good linearity was observed over the range assayed (0.031–4.0 ␮g/mL) with regression coefficients greater than 0.9989 for all tocopherols and tocotrienols. Limits of detection and enrichment factors ranged from 0.01 to 0.11 ␮g/mL and 50 to 73, respectively. Intra- and inter-day precision were lower than 8.9% and the recoveries were around 85.5–116.6% for all tocopherols and tocotrienols. The developed DLLME method was successfully applied to cereals: rice, barley, oat, wheat, corn and millet. This new sample preparation approach represents an inexpensive, rapid, simple and precise sample cleanup and concentration method for the determination of tocopherols and tocotrienols in cereals. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Tocols, commonly known as vitamin E, is a collective term for a group of eight naturally occurring chemical isomers, tocopherols (␣-, ␤-, ␥-, ␦-T) and their four respective unsaturated tocotrienols (␣-, ␤-, ␥-, ␦-T3). Their presence in food samples have been reported in cereals, palm oil, vegetable oils, nuts and seeds [1–3]. A high number of tocol analysis procedures using highperformance liquid chromatography (HPLC) for food samples have been published [1,2,4–6]. However, sample preparation remains as the most demanding and challenging part in these procedures. Saponification assisted extraction [7–9] and direct solvent extraction [7,9,10] are the two popular sample preparation techniques used for the analysis of tocols, besides the Soxhlet [11], solid-phase [12], supercritical-fluid [13] and pressurized liquid [14] extractions. Saponification is employed to liberate the tocols from the sample matrix, to convert the esterified derivatives of tocols to

∗ Corresponding author. Tel.: +60 389468394; fax: +60 389423552. E-mail address: [email protected] (K. Muhammad). 0021-9673/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.chroma.2013.03.036

their free forms and to reduce the load of material extracted into the organic phase [2,4,6,7,13]. This extraction results in a much cleaner extract [7] with good tocol separation and detection selectivity [4,13]. The use of large amount of organic solvents, long operating time and tedious extraction procedure, and the tendency of emulsion layer formation in liquid–liquid extraction step are the drawbacks of this extraction method. On the other hand, direct solvent extraction overcomes these drawbacks with the advantages of rapidity and simplicity of method. This extraction method was reported to yield lower amount of tocols in cereals than saponification assisted extraction [5,7,9]. High load of the extracted material, however, could decrease the selectivity of tocol detection [5] and life-time of the column. Nevertheless, these two extraction methods require considerable amount of extraction solvents to extract the tocols and frequently a pre-concentration step to increase the tocol detection sensitivity. These methods also generate a large amount of laboratory waste, which is an added cost for its disposal. Moreover, multiple steps in the sample preparation and the solvent evaporation step that is time consuming, limit the possibility to analyse more samples and could introduce error in the results.

32

B. Shammugasamy et al. / J. Chromatogr. A 1300 (2013) 31–37

Current research in analytical chemistry has focused on smallscale, simplified, efficient and particularly environmentally friendly extraction technique that inspires toward the development of microextraction method and its application in different samples. To the best of our knowledge, no microextraction method has been developed for the extraction of tocols in food samples. Dispersive liquid–liquid microextraction (DLLME), one variant of the microextraction method developed by Assadi and co-workers [15], is possible to be used for tocol extraction. DLLME is mainly used on clean samples such as water samples, and the extend of its application on complex matrix samples are limited [16]. Since the matrices of food samples are often complex while the tocols are usually associated with lipid components of the food matrix, the conventional sample pretreatment steps prior to microextraction, such as homogenization, direct extraction, dilution or filtration of sample, for the extraction of tocols are insufficient. Saponification can be employed as a sample pretreatment step; however, saponified extract contains saponication chemicals and extraneous matrix component that could interfere with the successful application of the subsequent microextraction step. Thus, we evaluated the applicability of DLLME after saponification for the extraction of tocols in food samples using cereals, which are regarded as the richest source of tocols [7,9,10,13,17]. In this paper, a combination of saponification and DLLME with HPLC provides a rapid and sensitive method for the determination of tocols in cereal samples, with minimal sample treatment and chemical solvent usage that can be used in routine analysis. 2. Materials and methods 2.1. Chemicals Tocopherols (purity ≥ 95.0%) and tocotrienols (purity ≥ 99.5%) were purchased as isomer kits from Calbiochem (San Diego, CA, USA) and Chromadex (Santa Ana, CA, USA), respectively. HPLCgrade ethanol, methanol and acetonitrile were obtained from Merck (Darmstadt, Germany), and pyrogallol (purity > 98.0%) was purchased from Fluka (Seelze, Germany). All other reagents including tetrachloromethane (CCl4 ) used were of analytical grade. 2.2. Standard solutions and cereal samples All standard stock solutions of tocols were prepared in absolute ethanol with an approximate concentration of 50 ␮g/mL and stored at −20 ◦ C. The concentration of each tocol standard stock solution was confirmed spectrophotometrically using the known 1 cm absorption coefficient of each isomer in ethanol (E1% ) [2]. From these stock solutions, working standards mixtures consisting of all eight isomers in the concentration range of 0.156–40 ␮g/mL of each isomer were prepared. Rice samples were obtained from PadiBeras Nasional Berhad (Malaysia) and cereals (barley, oat, wheat, corn, and millet) were obtained from local stores. All samples were ground, sieved through a 500-␮m sieve and stored in refrigerator at 4 ◦ C prior to analysis.

was set at Ex/Em = 295/330 nm. Initially, the column was conditioned with the mobile phase for at least 30 min until a linear baseline was obtained before chromatographic runs were made. 2.4. Saponification Cereal sample was saponified according to the method reported by Fratianni et al. [13] with minor modifications. Briefly, a mixture, consisting of 0.5 g of sample, 0.5 mL of potassium hydroxide (600 g/L), 0.5 mL of ethanol (95%), 0.5 mL of aqueous sodium chloride (10 g/L), 1.25 mL of ethanolic pyrogallol (60 g/L) and an additional 2.25 mL of ethanol (95%), in a screw cap glass test tube was flushed with nitrogen gas and incubated in a shaking water bath at 70 ◦ C for 45 min. After that, the tube was cooled to room temperature, vortexed for 10 s, and centrifuged at 3000 rpm for 5 min to sediment the floating particles. A 1 mL of clear saponified extract containing tocols was then subjected to the DLLME. 2.5. Dispersive liquid–liquid microextraction A 1.1 mL of a mixture of CCl4 :acetonitrile (1:10, v/v), followed by 1 mL of saponified extract were placed in a 10 mL glass tube with conical bottom. CCl4 was the extraction solvent and acetonitrile was the dispersive solvent selected. Subsequently, 3 mL of deionised water was injected rapidly into the glass tube using 5 mL autopipette. The formed cloudy solution was shaken gently and allowed to stand for 5 min. Then, the tube was vortexed for 10 s and centrifuged at 3000 rpm for 5 min. The centrifugation time was extended when white precipitate appeared at the bottom of the glass tube after centrifugation. The sediment of CCl4 (about 75 ± 3 ␮L) at the bottom of the glass tube was collected using a 50-␮L micro-syringe and transferred into a vial-insert for the HPLC analysis. A thin layer of deionised water was placed on top of the CCl4 extract in the vial-insert to prevent the CCl4 from evaporating. Besides evaluation of the type and volume of extraction and dispersive solvents, the extraction and centrifugation times were also optimized. All experiments in the DLLME optimization stage were performed in duplicates. 2.6. Enrichment factor and recovery Enrichment factor (EF) and recovery (R%) are the two important parameters for the DLLME optimization. The EF was defined as the ratio between the analyte concentration in the sedimented phase (Csed ) and the initial concentration of analyte (Csam ) within the sample. Meanwhile, R% was defined as the percentage of the total analyte amount that was extracted into the sedimented phase. The following equations EF = Csed /Csam and R% = 100 × (Csed × Vsed )/(Csam × Vsam ) were used to calculate EF and R%, respectively. Vsed and Vsam are the volume of sedimented phase and volume of aqueous sample, respectively. Csed was calculated from calibration curves obtained by direct injection of working standards mixtures in the range of 0.5–40 ␮g/mL.

2.3. High-performance liquid chromatography 3. Results and discussion An Agilent Technologies 1200 system chromatograph consisting of a degasser, a quaternary pump, an autosampler and a fluorescence detector was used. Separation of tocols at temperature of 25 ◦ C was accomplished using a COSMOSIL ␲-NAP column (250 mm × 46 mm; 5 ␮m) (Nacalai Tesque, Kyoto, Japan) with a 20-␮L injection volume. The mobile phase consisting of water: methanol: acetonitrile (13:80:7) was delivered isocratically at a flow rate of 1 mL/min and the fluorescence detection wavelength

3.1. Optimization of dispersive liquid–liquid microextraction The type and volume of extraction and dispersive solvents, salt addition, extraction time and centrifugation time are the DLLME parameters that would affect the EF and R% of tocols. A series of experiments changing one variable at one time was designed to achieve the highest R%.

B. Shammugasamy et al. / J. Chromatogr. A 1300 (2013) 31–37

33

Fig. 1. Recovery and enrichment factor of tocols obtained using different dispersive solvents. Extraction conditions: sample volume, 1 mL; aqueous volume, 3 mL; dispersive solvent volume, 1 mL; extraction solvent, 100 ␮L CCl4 .

3.2. Extraction solvent selection

3.3. Dispersive solvent selection

Dichloromethane, chloroform and tetrachloromethane (CCl4 ) were evaluated to select a suitable extraction solvent for the DLLME system based on these properties: (a) higher density than water, (b) low solubility in water, (c) strong capability to extract analyte of interest, and (d) good chromatographic compatibility [15]. Meanwhile, ethanol, acetonitrile and acetone were chosen as dispersive solvents. All possible combinations of selected extraction and dispersive solvents in the presence of saponified extract were tested for the formation of cloudy solution after addition of purified water and formation of sediment in the DLLME system after centrifugation. As the saponified extract contained mainly ethanol, it also acted as dispersive solvent that influence the dispersive process in the DLLME system. A cloudy solution was formed only with CCl4 as extraction solvent in all tested dispersive solvents, meanwhile a two-phase system was observed in all combinations after the mixtures were centrifuged. Cloudy solution was formed because of the fine particles of the extraction solvent, which was dispersed entirely in the aqueous phase. Formation of fine particles is essential for extraction in DLLME, thus CCl4 was selected as an extraction solvent.

Acetone, acetonitrile and ethanol were selected as dispersive solvents because of their good miscibility in both water and extraction solvent. As tocols have different degree of solubility in each dispersive solvent, the effect of different dispersive solvents on EF and R% of tocols was studied by using 1 mL of each dispersive solvent. The results in Fig. 1 indicated that the maximum R% and EF were achieved using acetonitrile–CCl4 as dispersive and extraction solvents pair compared to acetone and ethanol. Tocols have lower solubility in acetonitrile than in acetone and ethanol, which resulted in the highest R%. Therefore, acetonitrile was selected as the dispersive solvent. 3.4. Extraction solvent volume Different volumes of CCl4 (60, 80, 100, 120 and 140 ␮L) were subjected to the DLLME procedure to evaluate its effect on EF and R% of tocols. The results as shown in Fig. 2 demonstrated that an increase of CCl4 volume from 60 to 100 ␮L increased slightly the R% and further increase of CCl4 volume decreased slightly the R%. The sediment volume of CCl4 at 60 ␮L and 140 ␮L were 33 ␮L and

Fig. 2. Recovery and enrichment factor of tocols obtained using different volumes of extraction solvent. Extraction conditions: sample volume, 1 mL; aqueous volume, 3 mL; dispersive solvent volume, 1 mL; extraction solvent, 60–140 ␮L CCl4 .

34

B. Shammugasamy et al. / J. Chromatogr. A 1300 (2013) 31–37

Fig. 3. Recovery and enrichment factor of tocols obtained using different volumes of dispersive solvent. Extraction conditions: sample volume, 1 mL; aqueous volume, 3 mL; dispersive solvent volume, 0.85–1.75 mL; extraction solvent, 100 ␮L CCl4 .

132 ␮L, respectively. It is clear that an increase in the volume of CCl4 increased the sediment phase volume and subsequently reduced the EF. Thus, 100 ␮L of CCl4 was selected in order to obtain good R% with suitable EF. 3.5. Dispersive solvent volume Different volumes of acetonitrile, ranging from 0.7 to 1.75 mL in 0.15-mL intervals, were used to evaluate its effect on EF and R% of tocols. Fig. 3 shows that no obvious variation in R% was observed with increasing volume of acetonitrile from 1 to 1.45 mL. However, lower R% was obtained at 0.7, 0.85 and 1.75 mL of acetonitrile. It seems that at lower volumes of acetonitrile, the formation of tiny extraction droplet was not effective which resulted in the low R%. Moreover, using 0.7 mL of acetonitrile, a white precipitation was formed at the bottom of the centrifuge tube that caused a poor separation between the sediment and aqueous phases. Meanwhile, the solubility of tocols in aqueous phase increased at higher volumes of acetonitrile. It reduced the tocols partition into the extracting droplets and consequently the R%. An increase in the volume of acetonitrile from 1 to 1.75 mL reduced the volume of the sedimented phase from 82 to 39 ␮L, thereby increased the EF (Fig. 3). As the R% of tocols was similar in the range of 1–1.45 mL of acetonitrile, a 1 mL was selected to achieve suitable sediment volume for HPLC analysis with good EF. 3.6. Salt addition, extraction and centrifugation times An addition of salt (NaCl) in the range of 0–150 mg in the aqueous phase decreased the EF and R% of tocols. The addition of salt also resulted in a precipitation of hydrolysed cereal sample matrix under salting effect at the bottom of the aqueous phase in the centrifuge tube that made it difficult to collect the extraction solvent after centrifugation. Therefore, DLLME was carried out without the addition of salt. Meanwhile, the effects of extraction time (0, 2.5, 5 and 10 min) and centrifugation time (2.5, 5, 7.5 and 10 min) on the peak area of each tocol isomer were evaluated. Extraction time was defined as the interval time between the injection of purified water into the mixture of dispersive solvent, extraction solvent and sample, and starting of the centrifugation. No obvious variations of peak areas of tocols over tested extraction and centrifugation times were observed, which indicated that the DLLME was independent of extraction and centrifugation times. However, 5 min of extraction time and 5 min of centrifugation time were chosen to achieve complete extraction in order to obtain maximum EF and R%.

3.7. Analytical performance A series of experiments were performed under the optimized DLLME conditions to obtain method linearity, limit of detection (LOD), limit of quantitation (LOQ), enrichment factor (EF), precision and accuracy. For linearity, LOD and LOQ determinations, 1.0 mL of saponified extract (without sample) containing different concentrations of standards mixture was subjected to the DLLME. The results for linearity, LOD, LOQ and EF are shown in Table 1. The developed DLLME method demonstrated good linearity in the tested concentration range at six different levels with the regression coefficient (R2 ) ≥ 0.9989 for all tocol isomers. LOD and LOQ for each tocol, based on signal-to-noise of 3 and 10, were low and ranged from 0.01 to 0.11 ␮g/mL and 0.05 to 0.30 ␮g/mL, respectively. EFs were high because of high tocol pre-concentration in small volume of CCl4 . Fig. 4a illustrates chromatograms obtained for a standards mixture after 1 mL of the standards mixture at 0.6 ␮g/mL was subjected to DLLME and for direct injection of standards mixture at 6 ␮g/mL. The DLLME gave higher detection sensitivity (Fig. 4a-i) than direct injection (Fig. 4a-ii); although, the concentration of standards mixture used for the DLLME was 10 fold lower than in the direct injection. Intraand inter-day precision of the method were evaluated by successive triplicate analyses of a sample at three spiked concentration levels on the same day and on three different days, respectively. The RSDs for intra- and inter-day precision were less than 6.9 and 8.9%, respectively, indicating that this method has an acceptable repeatability (Table 2). Meanwhile, the accuracy of the DLLME method was evaluated in terms of percentage of recovery, which was calculated from the difference between the concentration of tocols in the spiked sample and concentration of tocols in the unspiked sample compared to the concentration of tocols added. A sample in triplicate was spiked at three concentration levels with known amount of tocol standards (1, 3 and 6 ␮g/sample) prior to saponification and subjected to the entire extraction method. The proposed method gave satisfactory recoveries of the added tocols being in the range of 85.5 and 116.6% (Table 2). The strong ␲–␲ interactions between the napthylethyl group bonded silica stationary phase and tocols resulted in a better separation of ␤ and ␥ isomers than that achievable in conventional reverse-phase column. The selectivity of the method was demonstrated by good separation of tocols with no interfering peaks in the elution region of the analytes for the cereal matrices (Fig. 4a–h). Furthermore, the purity of peaks was evaluated by measuring peak height of each tocol at different excitation wavelengths (270, 280 and 290 nm)

B. Shammugasamy et al. / J. Chromatogr. A 1300 (2013) 31–37

35

Table 1 Analytical performance of proposed DDLME method for the determination of tocols. Tocol ␦-T3 ␥-T3 ␤-T3 ␦-T ␣-T3 ␥-T ␤-T ␣-T a b c

Calibration rangea (␮g/mL) 0.031–4.0 0.031–4.0 0.031–4.0 0.031–4.0 0.063–4.0 0.031–4.0 0.031–4.0 0.125–4.0

Regression equation (y = mx + c)

Regression coefficient (R2 )

LODb (␮g/mL)

LOQc (␮g/mL)

Enrichment factor

y = 2055.3x − 14.4 y = 1443.1x − 15.7 y = 1244.5x − 15.8 y = 2120.4x − 21.4 y = 423.6x − 10.2 y = 1491.6x − 20.7 y = 1289.1x − 19.7 y = 364.0x − 11.7

0.9993 0.9992 0.9993 0.9995 0.9994 0.9993 0.9993 0.9989

0.01 0.01 0.02 0.02 0.06 0.02 0.03 0.11

0.05 0.05 0.06 0.08 0.17 0.07 0.09 0.30

73 71 70 63 63 55 58 50

Concentration of tocols in 1 mL of saponified extract subjected to DLLME. LOD, Limit of detection. LOQ, Limit of quantitation.

with a constant emission wavelength (330 nm) [18]. Except for ␤-T, an identical peak height ratios of 280/290, 270/290 and 280/270 for sample and tocols standards solutions were obtained, which indicate that the peaks were the tocols. The peak height ratio of ␤-T in sample deviated slightly from that of the standard, which may be due to interference from unknown compound in this peak. 3.8. Matrix effect As the blank calibration curve was prepared without the addition of sample, the possibility of matrix effect on calibration curve was investigated by comparing the slope of matrix-matched

calibration curve with the slope of the blank calibration curve [19]. Since it was not possible to find a blank cereal, matrix-matched calibration curve was constructed by analysing cereal samples (milled rice and brown rice) which were spiked with standards solutions in the range of 0.75–6 ␮g/mL. Table SI (supplement data) shows regression equation and the slope ratio of the matrix to the blank for each tocol. The slope ratios were in the range of 0.9–1.1, which indicated that there was mild matrix effect on the calibration curve. Although a mild matrix effect on the calibration curve was observed, quantitation of analytes in the cereal samples using matrix-matched calibration is recommended to obtain results that are more accurate.

Fig. 4. HPLC chromatograms of (a-i) tocols standards mixture after 1 mL of the standards mixture at 0.6 ␮g/mL was extracted by DLLME; (a-ii) standards mixture at 6 ␮g/mL by direct injection; (b-i) brown rice by direct solvent extraction; (b-ii) brown rice by saponification assisted extraction, and (c) brown rice; (d) barley; (e) oat; (f) wheat; (g) corn and (h) millet by DLLME. A ␲-NAP stationary phase with chromatographic conditions, mobile phase, water:methanol:acetonitrile (13:80:7); flow rate, 1 mL/min; temperature, 25 ◦ C, was used for the tocol separation. The peaks are labelled as follows: ␦-tocotrienol (␦-T3), ␥-tocotrienol (␥-T3), ␤-tocotrienol (␤-T3), ␦-tocopherol (␦-T), ␣-tocotrienol (␣-T3), ␥-tocopherol (␥-T), ␤-tocopherol (␤-T), and ␣-tocopherol (␣-T).

36

B. Shammugasamy et al. / J. Chromatogr. A 1300 (2013) 31–37

Table 2 Intra- and inter-day precision (RSD, %) and recoveries (%) of tocols in spiked brown rice sample. Tocol

␦-T3

␥-T3

␤-T3

␦-T

␣-T3

␥-T

␤-T

␣-T

Added (␮g)

Inter-day (n = 3 × 3)

Intra-day (n = 3)

0 1 3 6 0 1 3 6 0 1 3 6 0 1 3 6 0 1 3 6 0 1 3 6 0 1 3 6 0 1 3 6

RDS (%)

Recovery (%)

RDS (%)

Recovery (%)

1.4 6.2 2.1 3.4 4.3 4.5 2.7 3.7 ND 3.9 3.9 4.0 1.1 6.2 2.2 3.2 1.8 6.0 4.6 3.1 6.9 6.0 2.8 3.6 3.9 5.3 2.8 4.0 6.3 4.2 2.5 4.7

– 96.2 89.0 87.3 – 109.9 102.6 96.4 – 116.6 104.9 100.3 – 95.4 90.1 96.3 – 94.3 89.5 93.9 – 114.5 96.7 93.9 – 97.0 92.5 91.8 – 105.3 91.1 94.2

3.6 3.9 5.1 6.0 8.2 5.1 4.9 4.4 ND 8.9 5.5 4.1 4.2 8.6 4.7 6.8 8.2 8.0 6.0 4.9 8.4 4.8 5.5 5.6 6.0 5.1 6.2 3.5 6.1 7.0 6.8 4.5

– 103.4 97.4 94.7 – 85.5 91.9 97.2 – 107.1 101.1 102.0 – 97.4 97.7 95.2 – 100.4 99.3 93.5 – 111.8 97.6 94.6 – 102.8 93.6 96.0 – 101.2 101.7 89.7

ND, not detected.

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chroma. 2013.03.036.

used in the conventional extraction methods. Significant reduction in the volume of extraction solvent not only saves cost, but also reduces the risk on human health and the environment. It is also possible to scale-down the saponification sample size as low as 0.2 g because only 1 mL of saponified extract is required for the DLLME. The DLLME procedure after saponification is simple, rapid and the handling of the sample is reduced with the elimination of double extraction and solvent evaporation step. The elimination of these two steps reduces the time needed for each analysis and more samples can be analysed in a short time. Although the described method in this study is not as simple as the direct extraction method due to the presence of the saponification step, the latter step is recommended for cereal samples to release the tocols from their matrices and hydrolyse the lipid compounds [6]. It can be seen that a large extraneous matrix component peak observed in the direct extraction method chromatogram (Fig. 4b-i) was not found in the chromatograms of the other two methods that included the saponification step (Fig. 4b-ii and Fig. 4c).

3.9. Comparison of tocol extraction methods The performance of the developed DLLME method to extract tocols from brown rice sample was compared with that of saponification assisted extraction [13] and direct solvent extraction [9]. The amount of tocols obtained using this developed DLLME method was statistically not different (ANOVA, p > 0.05) from the results of the two latter extraction methods (Table 3). The results demonstrated that the developed DLLME method has an equivalent extraction performance with the conventional extraction methods for the extraction of tocols in rice samples. In addition, the characteristics of these three extraction methods were also compared. In this developed DLLME method, only 1 mL of acetonitrile and 0.1 mL of CCl4 were required for the extraction of tocols after saponification instead of approximately 10 mL of extraction solvent

Table 3 Concentrations of tocols (␮g/mL) in brown rice sample measured using three different extraction methods. Method

Saponification assisted extraction [4] Developed method (present work) Direct solvent extraction [7]

Tocol (␮g/g) ␦-T3

␥-T3

␦-T

␣-T3

␥-T

␤-T

␣-T

0.79 ± 0.01a 0.90 ± 0.08a 0.93 ± 0.00a

22.45 ± 0.00ab 21.60 ± 0.25b 23.95 ± 0.69a

0.17 ± 0.03a 0.16 ± 0.03a 0.09 ± 0.01b

0.81 ± 0.20a 0.55 ± 0.14a 0.86 ± 0.01a

3.28 ± 0.10a 3.20 ± 0.16a 3.44 ± 0.14a

0.67 ± 0.15a 0.93 ± 0.18a 0.60 ± 0.10a

3.05 ± 0.19a 3.22 ± 0.22a 2.94 ± 0.15a

Each value is the mean ± standard deviation of triplicate extractions. Values that are followed by different letters within columns are significantly different (p < 0.05).

B. Shammugasamy et al. / J. Chromatogr. A 1300 (2013) 31–37

37

Table 4 Tocopherol and tocotrienol contents of cereals (␮g/g). Cereal

␦-T3

␥-T3

␤-T3

␦-T

␣-T3

␥-T

␤-T

␣-T

Total tocols

Brown rice Red rice Black rice White rice White rice (basmati) Husked barley Oat Wheat Corn Millet

0.86 ± 0.02 0.27 ± 0.01 1.97 ± 0.15 0.28 ± 0.01 Tr Tr Tr 0.54 ± 0.04 Tr 3.82 ± 0.01

25.01 ± 0.52 8.47 ± 0.02 21.75 ± 0.47 6.00 ± 0.04 2.58 ± 0.08 0.92 ± 0.02 Tr Tr 7.66 ± 0.08 4.94 ± 0.05

Nd Nd Tr Nd Nd 1.07 ± 0.01 1.62 ± 0.08 19.99 ± 0.59 Nd Nd

Tr Tr 0.78 ± 0.25 Tr Tr Nd Nd Nd 0.46 ± 0.01 1.68 ± 0.00

Tr Tr 1.65 ± 0.13 Tr Tr 3.06 ± 0.03 12.81 ± 0.06 12.02 ± 0.15 3.15 ± 0.05 2.36 ± 0.04

5.28 ± 0.13 2.51 ± 0.02 6.32 ± 0.31 Tr Tr Tr Nd Nd 15.63 ± 0.19 69.09 ± 0.21

Tr Tr 1.25 ± 0.24 Tr Nd Tr Tr 4.83 ± 0.10 Tr Nd

4.43 ± 0.04 1.15 ± 0.06 7.01 ± 0.43 Tr Tr Tr 4.92 ± 0.50 14.42 ± 0.38 2.94 ± 0.03 4.53 ± 0.04

35.58 12.40 40.73 6.28 2.58 5.06 19.35 51.80 29.38 86.41

Each value is the mean ± standard deviation of three triplicate extractions. Nd, not detected. Tr, below limit of quantitation.

3.10. Application to cereals

Acknowledgement

Tocol content of cereals (rice, barley, oat, wheat, corn and millet) was determined using the developed DLLME method. As shown in Table 4, the amount and composition of tocols varied from cereal to cereal. Among the cereal samples analysed, millet contained the highest amount of total tocols followed by wheat, black rice, brown rice, corn and oat. The composition of tocols in cereals found in this study is similar to that reported in previous studies [7,12,14]. However, differences in the amount of tocols in cereals might be due to factors such as genotype and growth location. Moreover, these cereal samples were spiked with standards mixtureto investigate their recoveries. Good recoveries in the range of 99.2–124.6% were obtained and it demonstrated the ability of this DLLME method in extracting tocols from cereal samples. Fig. 4c–h show chromatograms of cereals analysed using the developed DLLME method. As can be seen, good separation of tocols with no interference of impurity peaks was observed.

The authors would like to thank PadiBeras Nasional Berhad for funding this study.

4. Conclusion This paper outlined the successful development and application of dispersive liquid–liquid microextraction (DLLME) coupled with RP-HPLC for the analysis of tocopherols and tocotrienols in cereals. The parameters that affect the DLLME efficiency in extracting tocols from saponified extract were optimized. Results of the validation study indicated that the developed DLLME method is precise and accurate in determining the tocol content of cereals. Simplicity, short time, low cost and minimum extraction solvent consumption are the advantages of the method. Thus, the method is suitable to be used for routine tocol analysis in cereals. It is also possible to use this DLLME procedure as an alternative extraction method to extract tocols after saponification for samples besides cereals.

References [1] R.A. Moreau, A.-M. Lampi, in: Z. Xu, L.R. Howard (Eds.), Analysis of Antioxidant-Rich Phytochemicals, Wiley-Blackwell, United Kingdom, 2012, p. 353. [2] R.R. Eitenmiller, L. Ye, W.O. Landen, Vitamin Analysis for the Health and Food Sciences, 2nd ed., CRC Press, New York, 2008, p. 119. [3] L. Packer, J. Fuchs (Eds.), Vitamin E in Health and Disease, Marcel-Dekker, New York, 1993, p. 3. [4] F.J. Rupérez, D. Martín, E. Herrera, C. Barbas, J. Chromatogr. A 935 (2001) 45. [5] J.K. Lang, M. Schillaci, B. Irvin, in: A.P. DeLeenheer, W.E. Lambert, H.J. Nelis (Eds.), Modern Chromatographic Analysis of Vitamins, Marcel Dekker, New York, 1992, p. 153. [6] G.F.M. Ball, Fat-Soluble Vitamin Assays in Food Analysis: A Comprehensive Review, Elsevier, New York, 1988. [7] G. Panfili, A. Fratianni, M. Irano, J. Agric. Food Chem. 51 (2003) 3940. [8] M. Ryynänen, A.M. Lampi, P. Salo-Väänänen, V. Ollilainene, V. Piironene, J. Food Comp. Anal. 17 (2004) 2749. [9] D.M. Peterson, C.M. Jensen, D.L. Hoffman, B. Mannerstedt-Fogelfors, Cereal Chem. 84 (2007) 56. [10] M.M. Nielsen, A. Hansen, Cereal Chem. 85 (2008) 248. [11] A.G. González, F. Pablos, M.J. Martın, M. León-Camacho, M.S. Valdenebro, Food Chem. 73 (2001) 93. [12] M.N. Irakli, V.F. Samanidou, I.N. Papadoyannis, J. Agric. Food Chem. 60 (2012) 2076. [13] A. Fratianni, M. Caboni, M. Irano, G. Panfili, Eur. Food Res. Technol. 215 (2002) 353. [14] M.M. Delgado-Zamarreno, M. Bustamante-Rangel, S. Sierra-Manzano, M. Verdugo-Jara, R. Carabias-Martinez, J. Sep. Sci. 32 (2009) 1430. [15] M. Rezaee, Y. Assadi, M.R.M. Hosseini, E. Aghaee, F. Ahmadi, S. Berijani, J. Chromatogr. A 1116 (2006) 1. [16] D. Han, K.H. Row, Microchimica Acta 176 (2012) 1. [17] M.N. Irakli, V.F. Samanidou, I.N. Papadoyannis, J. Sep. Sci. 34 (2011) 1375. [18] Y. Haroon, D.S. Bacon, J.A. Sadowski, Clin. Chem. 32 (1986) 1925. [19] F. Gosetti, E. Mazzucco, D. Zampieri, M.C. Gennaro, J. Chromatogr. A 1217 (2010) 3929.