Analysis of isoflavones in soy drink by capillary zone electrophoresis coupled with electrospray ionization mass spectrometry

Analytica Chimica Acta 709 (2012) 113–119

Contents lists available at SciVerse ScienceDirect

Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca

Analysis of isoflavones in soy drink by capillary zone electrophoresis coupled with electrospray ionization mass spectrometry ˜ ∗ , R. Carabias-Martínez, J. Domínguez-Álvarez M. Bustamante-Rangel, M.M. Delgado-Zamarreno Departamento de Química Analítica, Nutrición y Bromatología Universidad de Salamanca 37008 Salamanca, Spain

a r t i c l e

i n f o

Article history: Received 18 July 2011 Received in revised form 7 October 2011 Accepted 8 October 2011 Available online 17 October 2011 Keywords: Capillary electrophoresis Electrospray mass spectrometry Isoflavones Soy drink

a b s t r a c t Capillary zone electrophoresis coupled with electrospray ionization mass spectrometry (CZE–ESI-MS) has been applied for the first time for the separation and quantification of isoflavones in soy products. The proposed method was successfully applied to the determination of seven isoflavones, including aglycones and glucosides, in soy drink. The target compounds were the glucosides daidzin and genistin, and the aglycones daidzein, genistein, formononetin, biochanin A and glycitein. During CE separation in positive mode, the analytes were present as anions, and MS detection was carried out in ESI positive-ion mode. To prevent the frequent drops in current and to improve the resolution in the separation of analytes in anionic form, a programmed nebulizing gas pressure (PNP) was applied along the analysis. Extraction of isoflavones from soy drinks was carried out by liquid–liquid extraction using ethanol. The proposed extraction procedure is simple, efficient, and affords reproducible results. Quantification of the isoflavones in soy drinks using the external standard method did not produce good results; therefore, both internal standard and standard addition quantification methods were used, obtaining significantly similar results. The detection limits found were lower than 3.2 ␮g L−1 . © 2011 Elsevier B.V. All rights reserved.

1. Introduction Phytoestrogens are secondary plant metabolites that have estrogen-like properties and that have been associated with a lower incidence of steroid-hormone-dependent cancers. Dietary phytoestrogens are thought to protect against chronic diseases such as cancer, osteoporosis, cardiovascular disease and menopausal symptoms. The major classes of phytoestrogens are natural phenolic products of isoflavones and lignans, as well as coumestanes, flavonoids and stilbenes. Phytoestrogens are found in plants and in many food products as glycosidic conjugates. In fermented foods, they are deconjugated to their aglycones [1–3]. Isoflavonoids are a characteristic and very important subclass of flavonoids. Their structures are based on the 3-phenylchromen skeleton. Isoflavones have been classified in four major subgroups according to their functional groups: aglycones, glucosides, malonyl glucosides and acetyl glucosides. There are many positive biological activities associated with isoflavones, including a reduction in osteoporosis and the prevention of cancer and cardiovascular disease, and they can also be used for the treatment of the symptoms of menopause. There is growing evidence that isoflavones may play a role in the treatment of metabolic disorders

∗ Corresponding author. Tel.: +34 923294483. ˜ E-mail address: [email protected] (M.M. Delgado-Zamarreno). 0003-2670/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2011.10.015

[4–8]. As one of the major classes of phytoestrogens, isoflavones are widely distributed throughout the plant kingdom, but accumulate predominantly in plants of the Leguminosae family. The best known sources of isoflavones are soy and soy derived products, as well as red clover [1]. The analytical methods for the determination of phytoestrogens can be categorized, depending on the separation methodology used, into chromatographic (HPLC, GC) and non-chromatographic methods (CE). GC–MS has been the technique most commonly used due to its marked potential of high resolution, selectivity and sensitivity. HPLC separation is generally carried out on reversed-phase columns with a mobile phase of methanol or acetonitrile and water containing small amounts of acid as a modifier [1,3,9–11]. Several CE techniques have been developed, including the widely used capillary zone electrophoresis (CZE) [12,13], micellar electrokinetic capillary chromatography (MEKC) [14–16] and capillary electrochromatography (CEC) [17]. Since most phytoestrogens and their metabolites contain phenolic hydroxyl groups and have a weak acidic nature, CZE methods are generally performed based on a borate buffer run at alkaline pH to ensure the analytes will be present as anions for electrophoretic separation [1]. Detection is usually performed with UV [18–20] and electrochemical detection (ED) [21–24]. There are few references in which CE–MS coupling has been used in the analysis of isoflavones. Aramendia et al. [25] separated and identified several isoflavones using CE–ESI-MS in negative ion mode from standard samples. García-Villalba et al. [26] used

114

M. Bustamante-Rangel et al. / Analytica Chimica Acta 709 (2012) 113–119

capillary electrophoresis–time-of-flight-mass spectrometry to compare the metabolic profiles of conventional and genetically modified soybeans, including isoflavones, aminoacids, carboxylic acids and peptides. In both references, separation was carried out in positive mode at pH 9–9.5 and detection was achieved in negative ionization mode, but quantification was not reported. In CE–MS separations of anions in alkaline media with uncoated capillaries and positive separation mode – that is, the anode at the inlet and the cathode at the ESI interface – it is frequently observed that when the electroosmotic flow arrives at the tip of the ESI needle frequent drops in current occur. Furthermore, a lack of resolution in electrophoretic separation has been found. The application of a programmed nebulizing gas pressure (PNP) along the analysis [27] has the advantage of providing high resolution and drops in the current are avoided. The aim of this work was to develop a simple, efficient and sensitive method for the quantitative analysis of isoflavones, including aglycones and glucosides, using capillary zone electrophoresis coupled with electrospray ionization mass spectrometry (CZE–ESI-MS) as a good alternative to traditional techniques such as LC and GC. The compounds analyzed were the glucosides daidzin and genistin, and the aglycones daidzein, genistein, formononetin, biochanin A and glycitein. The most abundant isoflavones in soy drinks are daidzein, genistein and glycitein and their glucosides, although some references report very low levels of biochanin A and formononetin [28]. Accordingly, we tested these seven isoflavones. The proposed method was successfully applied to the determination of seven isoflavones in commercially available soy drink samples. To our knowledge, there are no methods that describe the quantification of isoflavones in food samples using CE–ESI-MS. 2. Experimental 2.1. Chemicals The isoflavones studied – daidzin, genistin, daidzein, genistein, formononetin, biochanin A and glycitein, with apigenin as an internal standard – were obtained from Aldrich (Alcobendas, Madrid, Spain). Stock solutions of each standard were prepared in methanol at concentrations ranging from 500 to 1000 mg L−1 and stored at −18 ◦ C. Standard solutions were prepared by dilution of stock solutions in water with a final methanol/water ratio of 1:10 (v/v). Table 1 shows the chemical structure and some properties of these compounds. Acetonitrile, methanol, ethanol and isopropanol were from Merck (Darmstadt, Germany). All were of special HPLC quality. All chemicals used for the preparation of the background electrolytes (BGE) and all other chemicals were of analytical reagent grade. The soy drink samples analyzed were from commercial sources, with contents ranging from 6% to 14% of soy beans. 2.2. CE system The experiments were carried out with a Hewlett-Packard HP3D Capillary Electrophoresis instrument (Agilent Technologies, Waldbronn, Germany) equipped with a UV–vis DAD device working at 214 nm with a bandwidth of 16 nm. Fused-silica capillaries were from Polymicro Technologies (Phoenix, AZ, USA), supplied by Composite Metal Services Ltd. (West Yorkshire, UK). Samples were injected hydrodynamically, using a pressure of 50 mbar for 5 s. The temperature of the capillary was kept at 25 ◦ C. The BGE was an aqueous solution of ammonium acetate adjusted to pH 11.0 with a concentrated aqueous solution of ammonia. Electrophoretic separation was achieved in positive polarity mode by applying a voltage of 25 kV during analysis.

All new capillaries were conditioned sequentially by flushing with 1.0 M sodium hydroxide, UHQ water, and BGE for 10 min each. Before each analysis, the capillary was rinsed with BGE for 2 min. At the end of each session, the capillary was washed and stored with water. The BGE was refreshed every four runs. Optimization of electrophoretic separation was carried out with a 75 ␮m capillary, with total length of 57 cm and 50 cm to the UV detector. When the MS detector was used, the experiments were carried out with a 50 ␮m capillary, with a total length 87.5 cm and 21 cm to the UV detector. 2.3. ESI-MS system MS was performed using an Agilent LC/MSD SL mass spectrometer (Agilent technologies) equipped with a single quadrupole analyzer. The MS device was controlled by Agilent HP ChemStation software, version B.02.01 SR1. An Agilent coaxial sheath-liquid sprayer was used for CZE–MS coupling (Agilent technologies). The fused-silica capillary was mounted in such a way that the tip just protruded from the surrounding steel needle, ∼1/2 of the capillary o.d. The sheath liquid was an isopropanol/water mixture (1:1, v/v) containing 0.05% (w/v) acetic acid and was delivered at a flow rate of 1.0 mL min−1 by an Agilent 1100 series pump, equipped with a 1:100 flow-splitter. The ESI capillary voltage was set at 3500 V; the drying gas flow rate and drying gas temperature were set at 6 L min−1 and 350 ◦ C, respectively. The nebulizing gas pressure was switched off during injection, the minimum permitted pressure (1 psi) was applied during separation, and the optimum pressure for the ionization of analytes (5 psi) was applied 1.0 min before the analytes arrived at the detector. The mass spectrometer was operated in positive-ion mode. Analyte quantification was carried out under the SIM acquisition mode using protonated molecules. The ions monitored are shown in Table 1. 2.4. Sample preparation Extraction of the analytes from the soy drink samples was carried out as follows: 1.0000 g of sample mixed with the internal standard (apigenin) was extracted using an automatic inversion mixer SBS (ABT-2) at room temperature. The extractions were performed using ethanol at a sample: solvent ratio of 1:2 for 30 min. The extract obtained was centrifuged at 5000 rpm and filtered through a 0.45 ␮m nylon syringe filter. 100 ␮L of the filtered extract was diluted in UHQ water to a final volume of 1.0 mL. The samples were analyzed immediately or stored at −18 ◦ C for later analysis. 3. Results and discussion 3.1. Optimization of CE conditions Optimization of electrophoretic separation was carried out using a UV–vis DAD device, with a 75 ␮m capillary with a total length of 57 cm and 50 cm to the UV detector. A 5 mg L−1 standard solution of six isoflavones (daidzin, genistin, biochanin A, formononetin, daidzein and genistein) was used in these studies. Among the electrolytes suitable for CE–MS analysis, an aqueous ammonium acetate buffer was used. The effect of the buffer concentration on the separation of six isoflavones was investigated in a range from 10 to 50 mM. When the acetate concentration was increased, the migration time of six isoflavones was longer; however, their resolution was not significantly improved. The buffer concentration affects not only the resolution and migration time of the analytes, but also the peak current [21,22]. Thus, an electrolyte concentration of 15 mM was chosen, taking into account resolution, analysis time, and electrical current intensity.

M. Bustamante-Rangel et al. / Analytica Chimica Acta 709 (2012) 113–119

115

Table 1 Chemical structure and properties of the target compounds .

OH O

OH

O

R1

R2 R4

OH

O

O

R3

apigenin

Isoflavone Isoflavone

R1

R2

R3

R4

pKa

m/z [M–H]+

Daidzin (Di) Genistin (Gi) Daidzein (De) Genistein (Ge) Formononetin (For) Biochanin A (Bio) Glycitein (Gly) Apigenin (IS)

O–C6 O5 H11 O–C6 O5 H11 OH OH OH OH OH

H H H H H H OCH3

H OH H OH H OH H

OH OH OH OH OCH3 OCH3 OH

– – 9.74a 9.81a – 9.9b 9.98a –

417 433 255 271 269 285 285 271

a b

Calculated by McLeod and Shepherd [37] using capillary electrophoresis. Obtained by us using potentiometric analysis.

The pH of the run buffer affected the separation of compounds, such that high values (above the pKa of analytes, see Table 1) improved the separation and allowed the analytes to reach the ESI device as anions. pH was studied in a range from 9.5 to 11.5. The pH of the buffer solution was modified by adding ammonia. When the standard solution was injected in pure methanol, the peaks of the isoflavones, mainly daidzin and genistin, were broad and asymmetric. Dilution with the buffer electrolyte produced a similar effect, whereas when the standard solution was prepared using water the peak shape improved when the concentration of methanol was lower than 10–15%. Therefore, standard solutions were prepared by the dilution of stock solutions in water with a final methanol/water ratio of 1:10 (v/v). The separation voltage and injection time were also studied. The separation voltage was set at 25 kV; higher values afforded a higher current intensity, whereas lower voltages elicited a longer migration time. Injection was carried out hydrodynamically at a pressure of 50 mbar for 5 s, and the temperature of the capillary was kept at 25 ◦ C. 3.2. Optimization of CE–ESI-MS conditions To achieve coupling between the CE and ESI-MS devices, the previously optimized parameters for the electrophoretic separation were applied. Working in the PNP mode described above in a 50 ␮m i.d. capillary improved the stability and resolution of the signals; therefore, for the rest of the experiments a 50 ␮m capillary, with a total length of 87.5 cm and 21 cm to the UV detector, was used. Standard solutions of six isoflavones (daidzin, genistin, biochanin A, formononetin, daidzein and genistein) at 1 mg L−1 in methanol/water (1:10, v/v) were used for the optimization of the parameters affecting the CE–ESI-MS coupling. The direct way of introducing species in MS that have previously been separated as anions in CZE should be in ESI negative-ion mode. However, the ESI source may prove to be much more effective in positive mode instead of in negative mode for the analysis of anions [29]. In our case, single MS spectra of the isoflavones in both positive and negative acquisition modes were tested to check which of them produced the best signals. Acetic acid and ammonia, respectively, at concentrations ranging from 0.05% to 1% were added to the sheath liquid, composed of mixtures of isopropanol/water (1:1, v/v). An increase in the concentration of the ammonia or acid did not produce any significant improvement. Using the negative-ion

mode, the signals obtained had a poorer peak shape, mainly for daidzin and genistin, which appeared as a double peak. The best results were obtained in positive mode using an isopropanol/water mixture (1:1, v/v) containing 0.05% (w/v) acetic acid. The sheath liquid flow rate was varied from 0.7 to 1 mL min−1 (before 1:100 split) and, since the most satisfactory results were obtained with the latter rate, 1 mL min−1 was set at optimum for the rest of the studies. It was observed that the capillary voltage, the drying gas flow rate, and the drying gas temperature did not have a significant influence on either the migration time or the peak shape. These parameters ranged from 2500 to 4000 V; 6 to 12 L min−1 , and 200 to 350 ◦ C, respectively. An MS capillary voltage of 3500 V, a flow rate of 6 L min−1 , and a temperature of 350 ◦ C were chosen because the analytical signal was slightly higher. However, the nebulizing gas pressure had a strong influence on both the analysis time and on resolution. The addition of a nebulizing gas, combined with a high voltage, is generally necessary to assist solvent evaporation. However, gas has been reported to have an aspirating effect that affects the quality of separation because of a pressure-induced flow [27,30]. To investigate the effect of the nebulizing gas pressure on separation performance, pressure was varied between 3 and 25 psi (3, 5, 10, 15, 20 and 25 psi were tested). As shown in Fig. 1 – where only 5, 15 and 25 psi are shown to simplify the figure, both the migration time and resolution were significantly affected by the pressure of the nebulizing gas. As expected, the migration time of the analytes decreased when the pressure was increased from 5 to 25 psi. This decrease in the migration time can be attributed to aspiration into the capillary. As a result, a laminar flow was formed inside the capillary, and at pressures above 15 psi the resolution between analytes was hindered to a considerable extent. It was also observed that the nebulizing gas pressure not only affected the CZE separation, but also the sensitivity of detection. As the nebulizing gas pressure increased from 5 to 25 psi, sensitivity increased. The optimal nebulizing gas pressure was chosen as a compromise between separation and sensitivity. As pressure, 5 psi was chosen instead of 15 psi owing to the fact that the separation was better, mainly taking into account that real samples could present interferences. To minimize the entry of air into the capillary when working in positive separation mode, some authors have proposed switching off the nebulizing gas during sample injection [31]. Domínguez-Álvarez et al. [27] have reported that the

116

M. Bustamante-Rangel et al. / Analytica Chimica Acta 709 (2012) 113–119

3,4

separation of anions in positive mode using ESI in positive-ion mode the application of an optimized PNP provided high resolution and prevented drops in the current.

5 psi 15 psi

3.3. Optimization of the extraction of isoflavones from soy drink samples

25 psi

3,4 3,4 67

12

12 1,2

10000 0

0

2

4

67

6,7

6

8

10

12

14

min

Fig. 1. Total ion electropherograms for CE–ESI-MS analysis of a solution of isoflavones at different nebulizing gas pressures. Conditions as in text. Analyte identification: daidzin (1), genistin (2), formononetin (3), biochanin A (4), daidzein (6), genistein (7).

application of a programmed nebulizing gas pressure during injection and electrophoretic separation facilitates the analysis of anions by CE–ESI-MS in positive separation mode. In order to improve the resolution in the separation of analytes and to prevent drops in the current, a study was carried out applying PNP along the analysis. PNP involves switching off the nebulizing gas pressure during sample injection; the application of a lower value of 1 psi (the minimum value permitted during the analysis) during electrophoretic separation, and the application of the optimum pressure 1.0 min before the analytes arrive at the ESI interface. This effect was studied at two final pressure values: 5 and 15 psi. As can be seen in Fig. 2, the application of PNP improved the resolution between daidzein and genistein, whereas sensitivity was not affected. The optimum pressure chosen for the ionization of the analytes was 5 psi. The results obtained revealed that in CE 3

25000

0

7

1

8

9

10

11

3

12

13

8

7 2

7

8

11

12

13

14 min

1psi 0

4

7 1

10

15psi

B.2 15 min

7

0

7

8

11 min

P

5psi

0

10

3

2

9

9

1psi

6

7

0

15 min

6

40000

A.2

1

14 min

P

4

0

4

7

2

25000

0

B.1 15 min

6

15psi

P

5psi

4

0

3

40000

P

A.1

1

Extraction of the analytes from the samples was carried out using an inversion mixer at ambient temperature. The variables affecting the extraction process were studied: namely, the solvent used, the sample: solvent ratio, and the extraction time. In order to determine the best solvent to achieve the extraction of the seven isoflavones, methanol and ethanol were tested. Similar results were obtained using both solvents, but cleaner extracts and a higher degree of reproducibility were obtained when the solvent used was ethanol. It is known that the extraction of isoflavones from soy samples with pure solvents affords lower yields than extraction with a certain amount of water in the extraction solvent [32]. In this case, the sample itself contained a large amount of water; accordingly, no additional water was added. The optimization of the sample: solvent ratio was performed by adding different amounts of ethanol to the sample at proportions ranging from 1:2 to 1:5. Higher sample proportions were not employed because cloudy extracts were obtained. No significant differences were observed when the percentage of ethanol was increased. Taking into account that a higher amount of solvent produces greater dilution, a sample: solvent ratio of 1:2 was selected. In order to obtain higher yields, the extraction time was studied. Several portions of soy drink samples were extracted, using ethanol as solvent at the chosen sample: solvent ratio for times ranging from 10 to 60 min. Although in all cases similar results were obtained, 30 min was chosen as the extraction time for ensuing experiments because a slight increase in the analytical signal was observed and the RSD decreased when the extraction time was increased. Before CE–ESI-MS analysis, the extracts were centrifuged at 5000 rpm for 10 min, filtered through a 0.45 ␮m nylon syringe filter, and diluted. As mentioned above, the injection of samples diluted in pure organic solvents produced broader and asymmetric peaks. The same effect was observed when the extracts were diluted in

15 min

7

6 7

2

9

10

11

12 min

Fig. 2. Electropherograms (selected m/z values) for CE–ESI-MS analysis of a solution of isoflavones at two nebulizing gas pressures: 5 psi (A) and 15 psi (B), without (A.1 and B.1) and with (A.2 and B.2) the optimal programmed nebulizing gas pressure. Conditions as in text. Analyte identification as in Fig. 1.

M. Bustamante-Rangel et al. / Analytica Chimica Acta 709 (2012) 113–119

117

Analyte area IS area

Analyte area

0.18

30000 25000

without IS

with IS

without IS after 4 days

with IS after 4 days

0.16 0.14 0.12

20000

0.1 15000

0.08 0.06

10000

0.04 5000

0.02 0

0

Di

Gi

De

Ge

For

Bio

Gly

Fig. 3. Correction of the analytical signal in the presence of internal standard.

the running buffer. Dilution of the extracts in water produced good analytical signals. Taking into account that the extracts contained aqueous sample:ethanol at a ratio of 1:2, a study was carried out to find the maximum amount of organic solvent – ethanol – that could be present in the injected extracts. Several ethanolic extracts were diluted in water at concentrations ranging from 5% to 30%. As in the standard solutions, ethanol percentages higher than 15% afforded broader peaks for daidzin and genistin, and in some cases even double peaks. Accordingly, the ethanolic extracts were diluted in water at a ratio of 1:9. Finally, the injection time was studied. The best results were obtained when injection was carried out hydrodynamically at a pressure of 50 mbar for 50 s.

3.4. Performance of the analytical procedure Isoflavone quantification cannot be achieved using the external standard method because, using CE–MS, the analytical signals are not reproducible between days. In order to obtain more reproducible results, the use of an internal standard (IS) was tested. Some substances used as internal standards in the analysis of isoflavones are 2-methoxyflavone [33], dihydroxybenzaldehyde [34], apigenin [35,36]. We tested whether apigenin, a flavone analogue of the isoflavone genistein, could be used as an IS because it was separated from the rest of analytes under the working conditions and it was ionized in the MS system. In order to check that the addition of apigenin did improve the results obtained, several extracts of soy drink samples were analyzed with and without the addition of the IS before extraction of the analytes. These extracts were analyzed immediately and after 4 days. Fig. 3 shows that the addition of apigenin corrected the imprecision of the analytical signals. The linearity ranges using apigenin as an IS were tested in a concentration range between 10 and 5000 ␮g L−1 (n = 11) for daidzein, genistein, formononetin, biochanin A and glycitein, and from 25 to 5000 ␮g L−1 (n = 10) for daidzin and genistin. Triplicate determinations of each calibration standard were made, and good correlation coefficients (>0.998) were obtained for all compounds (Table 2). The detection limits, calculated on the basis of a signal to noise ratio (S/N) of 3, were between 0.65 ␮g L−1 for glycitein and 4.8 ␮g L−1 for genistin. The intra-day precision was evaluated by analyzing a standard solution containing 100 ␮g L−1 of each isoflavone six times on the same day. The RSD ranged from 1.3% to 5.8%. The RSD values for the inter-day precision, evaluated on four different days at the same concentration level, ranged from 4.4% to 11% (Table 2).

3.5. Analysis of isoflavones in soy drinks In order to evaluate the applicability of the proposed method for the determination of isoflavones in soy drink samples, four commercially available samples were analyzed following the proposed procedure (see Section 2). The electropherograms of a standard mixture of seven isoflavones and apigenin (A) and a soy drink sample (B) are depicted in Fig. 4. Quantification of the isoflavones was performed using the internal standard and the standard addition methods. To achieve quantification using the internal standard method, three replicates of each soy drink sample were analyzed. The results obtained from the analysis of the four kinds of soy drink samples are given in Table 3. The concentration of the isoflavones and the standard deviation for each sample are shown. To carry out quantification using the standard addition method, the samples were spiked with five standard solutions of seven isoflavones at concentration levels ranging from 10 to 500 ␮g L−1 for daidzein, genistein, formononetin, biochanin A and glycitein, and at a concentration range

IS

30000

A 3,4

5 6

1 2

0

7

8

9

10

11

12

7

13

14

15

16 min

2

35000

B

1

IS

3,4 5 67

0

7

8

9

10

11

12

13

14

15

16 min

Fig. 4. Total ion electropherograms for CE–ESI-MS analysis of (A) a standard solution of the isoflavones studied (1 mg L−1 ) and the internal standard; (B) an extract of a soy drink sample. Conditions as in text. Analyte identification: daidzin (1), genistin (2), formononetin (3), biochanin A (4), glycitein (5), daidzein (6), genistein (7), apigenin (IS).

118

M. Bustamante-Rangel et al. / Analytica Chimica Acta 709 (2012) 113–119

Table 2 Analytical characteristics using the internal standard method. Isoflavone d

Di Gid De Ge For Bio Gly a b c d

Intercept (−2.5 (−1.9 (−4.1 (−2.9 (−6.2 (−2.8 (−7.4

± ± ± ± ± ± ±

Slope −2

0.9) × 10 1.0) × 10−2 1.7) × 10−2 1.2) × 10−2 2.2) × 10−2 0.9) × 10−2 10) × 10−3

(4.70 (5.36 (1.48 (1.013 (2.59 (1.072 (2.34

± ± ± ± ± ± ±

−

0.05) × 10 3 0.05) × 10−3 0.01) × 10−2 0.007) × 10−2 0.01) × 10−2 0.006) × 10−2 0.01) × 10−2

R2

LOD (␮g L−1 )a

RSDb

RSDc

0.9985 0.9981 0.9983 0.9983 0.9991 0.9991 0.9993

3.2 2.3 1.0 1.7 0.67 1.8 0.52

3.4% 1.3% 2.5% 5.8% 2.2% 3.2% 3.7%

11% 4.4% 9.5% 5.1% 8.6% 8.9% 6.8%

S/N = 3. Intra-day precision, calculated by analyzing a standard solution containing 100 ␮g L−1 of each isoflavone six times on the same day. Inter-day precision, calculated by analyzing a standard solution containing 100 ␮g L−1 of each isoflavone on four different days. Concentration range 25–5000 ␮g L−1 ; the rest 10–5000 ␮g L−1 .

Table 3 Isoflavone contents (mg L−1 ± RSD) of soy drink samples using the internal standard method (IS method) and the standard addition quantification method (SA method). Contents (mg L−1 ± RSD) Soy drink samples

1

2

Isoflavone

IS method

Di Gi De Ge For Bio Gly Total isoflavones

45 49 2.2 1.6 0.9 1.7 0.4 101

± ± ± ± ± ± ± ±

3 2 0.2 0.2 0.1 0.1 0.1 4

SA method 47 47 2.2 1.5 0.83 1.5 0.62 101

± ± ± ± ± ± ± ±

1 1 0.1 0.1 0.08 0.1 0.09 1

3

IS method 81 83 2.9 2.0 0.24 1.4 0.26 171

± ± ± ± ± ± ± ±

4 3 0.1 0.1 0.04 0.2 0.02 5

from 100 to 5000 ␮g L−1 for daidzin and genistin. All samples analyzed were injected in triplicate (Table 3). As can be seen, the different soy drink samples had very different concentrations of each isoflavone, but in all cases the highest content corresponded to the glycosides, daidzin and genistin, and the lowest concentration corresponded to formononetin or glycitein. The total concentration of isoflavones also differed among the different samples. It can be observed that the highest total isoflavone concentration corresponded to the sample with the highest amount of soy in its composition (sample 4, with a content of 14% soy seeds), and the lowest total isoflavone content corresponded to sample 3, with a concentration in soy seed of only 6%. In order to compare the results obtained with the internal standard and the standard addition methods, a statistical test was carried out. The test included a variance comparison (F-test) and means comparison based on Student’s t-test. There were no significant differences between the results obtained with either quantification method (at a level of significance of 0.05), except in the case of biochanin A in sample 3. The slopes corresponding to the calibration curves obtained using the standard addition method were compared with those obtained with external standard calibration method; i.e., prepared by the dilution of stock solutions in water. No significant differences were seen between the slopes obtained for any of the isoflavones, and hence it may be concluded that there was no matrix effect. Thus, the quantification of isoflavones in soy drink samples can be performed using internal standard calibration. The repeatability and reproducibility of the method were evaluated. Intra-day precision was determined by extracting six replicates of a soy drink sample under the previously optimized conditions. Inter-day precision was evaluated by analyzing four replicates of a soy drink sample extracted on four non-consecutive days. The RSD values ranged between 3.8% and 13% for the intra-day precision, and between 8.5% and 21% for the inter-day precision; the highest RSDs were seen for the lowest isoflavone contents.

SA method 86 84 2.8 1.9 0.28 1.2 0.25 176

± ± ± ± ± ± ± ±

1 1 0.1 0.1 0.01 0.1 0.01 1

4

IS method

SA method

IS method

29 ± 3 35 ± 2 1.7 ± 0.1 1.6 ± 0.1 0.3 ± 0.1 1.1 ± 0.1 n.d. 69 ± 4

25 ± 1 36 ± 1 1.6 ± 0.1 1.4 ± 0.1 0.22 ± 0.03 0.87 ± 0.04 n.d. 65 ± 1

85 81 7.6 4.4 0.26 1.1 0.40 180

± ± ± ± ± ± ± ±

3 3 0.1 0.2 0.09 0.2 0.08 4

SA method 83 83 7.5 4.4 0.24 0.9 0.27 179

± ± ± ± ± ± ± ±

1 1 0.1 0.1 0.03 0.1 0.03 1

Table 4 Accuracy of the proposed method. Data obtained in the analysis of a soy drink (sample 2). Isoflavone

RSDa

RSDb

Recoveryc

Di Gi De Ge For Bio Gly

3.2% 4.4% 3.8% 6.8% 5.8% 8.6% 13%

9.5% 8.5% 8.5% 7.1% 8.8% 6.2% 21%

96% 95% 118% 114% 112% 105% 117%

a Intra-day precision, calculated by extracting six replicates of a soy drink sample on the same day. b Inter-day precision, calculated by analizing four replicates of a soy drink sample extracted of four non-consecutive days. c Recovery studies were carried out by spiking 1 mL of soy drink sample with the seven isoflavones studied at concentration levels close to those present in the original samples.

In the absence of certified or standard materials, the method was validated by measuring the percentage of recovery after the addition of known amounts of standard to the samples. Recovery studies were carried out by spiking 1 mL of soy drink sample with the seven isoflavones studied at concentration levels close to those present in the original samples. The signal obtained for each of the analytes was introduced into the corresponding internal standard calibration line. In all cases, satisfactory recovery values were obtained (Table 4). 4. Conclusions A new method for the determination of isoflavones in soy drink samples has been developed using capillary zone electrophoresis–electrospray ionization–mass spectrometry. The proposed method was successfully applied to the analysis of the seven isoflavones (daidzin, genistin, daidzein, genistein,

M. Bustamante-Rangel et al. / Analytica Chimica Acta 709 (2012) 113–119

formononetin, biochanin A, and glycitein) in four soy drink samples. The analytes were separated as anions in positive separation mode, and ionization was carried out in ESI positive-ion mode. The proposed extraction procedure is very simple and efficient and affords reproducible results. It should be noted that as far as we are aware the literature does not contain any references concerning the quantification of isoflavones in food samples using capillary electrophoresis coupled with mass spectrometry. The use of a programmed nebulizing gas pressure along analysis prevents the frequent drops in current and improves the resolution in the separation of analytes in anionic form. Acknowledgements This work was supported by the Ministerio de Ciencia e Innovación, Spain (Project CTQ 2008-02200/BQU) and the Junta de Castilla-León (Grupo de Excelencia GR-65). References [1] Q. Wu, M. Wang, J.E. Simon, J. Chromatogr. B 812 (2004) 325–355. [2] J. Liggins, L.J.C. Bluck, S. Runswick, C. Atkinson, W.A. Coward, S.A. Bingham, J. Nutr. Biochem. 11 (2000) 326–331. [3] C.C. Wang, J.K. Prasain, S. Barnes, J. Chromatogr. B 777 (2002) 3–28. [4] F.R. Marin, J.A. Perez-Alvarez, C. Soler-Rivas, Stud. Nat. Prod. Chem. 32 (2005) 1177–1207. [5] P.Y. Lin, H.M. Lai, J. Agric. Food Chem. 54 (2006) 3807–3814. [6] S. Rochfort, J. Panozzo, J. Agric. Food Chem. 55 (2007) 7981–7994. [7] W.M. Mazur, J.A. Duke, K. Wähälä, S. Rasku, H. Adlercreutz, J. Nutr. Biochem. 9 (1998) 193–200. [8] W. Mazur, Baillière’s Clin. Endocrinol. Metab. 12 (1998) 729–742. [9] E. Rijke, P. Out, W.M.A. Niessen, F. Ariese, C. Gooijer, U.A.Th. Brinkman, J. Chromatogr. A 1112 (2006) 31–63. [10] J. Valls, S. Millán, M.P. Martí, E. Borràs, L. Arola, J. Chromatogr. A 1216 (2009) 7143–7172. [11] J. Vacek, B. Klejdus, L. Lojková, V. Kubán, J. Sep. Sci. 31 (2008) 2054–2067.

119

[12] Z.K. Shihabi, T. Kute, L.L. Garcia, M. Hinsdale, J. Chromatogr. A 680 (1994) 181–185. [13] O. Mellenthin, R. Galensa, J. Agric. Food Chem. 47 (1999) 594–602. [14] G.A. Micke, N.M. Fujiya, F.G. Tonin, A.C.O. Costa, M.F.M. Tavares, J. Pharm. Biomed. Anal. 41 (2006) 1625–1632. [15] Y. Zhang, J. Chen, L. Zhao, Y.P. Shi, Biomed. Chromatogr. 21 (2007) 987–992. [16] M. Xiao, J. Ye, X. Tang, Y. Huang, Food Chem. 126 (2011) 1488–1492. [17] J.A. Starkey, Y. Mechref, C.K. Byun, R. Steinmetz, J.S. Fuqua, O.H. Pescovitz, M.V. Novotny, Anal. Chem. 74 (2002) 5998–6005. [18] G. Dinelli, I. Aloisio, A. Bonetti, I. Marotti, A. Cifuentes, J. Sep. Sci. 30 (2007) 604–611. [19] C. Fang, X. Wan, H. Tan, C. Jiang, Ann. Chim. 96 (2006) 117–124. [20] C. Li, J. Zhang, A. Chen, J. Xie, X. Chen, Z. Hu, Biomed. Chromatogr. 19 (2005) 369–374. [21] Y. Peng, Q. Chu, F. Liu, J. Ye, Food Chem. 87 (2004) 135–139. [22] Y.Y. Peng, J.N. Ye, Fitoterapia 77 (2006) 171–178. [23] Z. Chen, J. Zhang, G. Chen, J. Chromatogr. B 863 (2008) 101–106. [24] Q. Chu, L. Fu, T. Wu, J. Ye, Biomed. Chromatogr. 19 (2005) 149–154. [25] M.A. Aramendia, I. García, F. Lafont, J.M. Marinas, J. Chromatogr. A 707 (1995) 327–333. [26] R. García-Villalba, C. León, G. Dinelli, A. Segura-Carretero, A. Fernández˜ A. Cifuentes, J. Chromatogr. A 1195 (2008) Gutiérrez, V. García-Canas, 164–173. [27] J. Domínguez-Álvarez, E. Rodríguez-Gonzalo, J. Hernández-Méndez, R. Carabias-Martínez, Anal. Chem. 83 (2011) 2834–2839. [28] J.T. Dwyer, B.R. Goldin, N. Saul, L. Gualtieri, S. Barakat, H. Adlercreutz, J. Am. Diet. Assoc. 94 (1994) 739–743. [29] R.J. Soukup-Hein, J.W. Remsburg, P.K. Dasgupta, D.W. Armstrong, Anal. Chem. 79 (2007) 7346–7352. [30] J.C. Severs, R.D. Smith, in: R.B. Cole (Ed.), Electrospray Ionization Mass Spectrometry; Fundamentals, Instrumentation and Applications, Wiley, New York, 1997, p. 343. [31] T. Soga, T. Ishikawa, S. Igarashi, K. Sugawara, Y. Kakazu, M. Tomita, J. Chromatogr. A 1159 (2007) 125–133. [32] M.A. Rostagno, A. Villares, E. Guillamón, A. García-Lafuente, J.A. Martínez, J. Chromatogr. A 1216 (2009) 2–29. [33] R.C. Alves, I.M.C. Almeida, S. Casal, M.B.P.P. Oliveira, Food Chem. 122 (2010) 914–919. [34] M.A. Rostagno, M. Palma, C.G. Barroso, Anal. Chim. Acta 597 (2007) 265–271. [35] D.L. Luthria, R. Biswas, S. Natarajan, Food Chem. 105 (2007) 325–333. [36] M.W. Collison, J. AOAC Int. 91 (2008) 489–500. [37] G.S. McLeod, M.J. Shepherd, Phytochem. Anal. 11 (2000) 322–326.