Académique Documents
Professionnel Documents
Culture Documents
Pavel Jandera1 ker kov1 Veronika S Lucie Rehov2 Hjek1 Toms Lucie Baldrinov1 kopov1 Gabriela S Vladimr Kellner3 Horna4 Ales
1
RP-HPLC analysis of phenolic compounds and flavonoids in beverages and plant extracts using a CoulArray detector
Methods were developed for the analysis of natural antioxidants including phenolic compounds and flavonoids in beverages and plant extracts using gradient HPLC with multi-channel electrochemical coulometric detection. Suitability of various reversedphase columns for this purpose was compared; pH and mobile phase gradients were optimized with respect to the separation selectivity and sensitivity of detection. Because of different target compounds in various sample types, the overlapping resolution maps and the normalized resolution product approaches described earlier were used to select optimum columns and gradients to suit the analysis of the individual sample types. The methods were applied to the analysis of phenolic compounds and flavonoids in beer, wine, tea, and yacon extracts. 32 phenolic compounds were identified and determined, including derivatives of benzoic and cinnamic acids, flavones, and a few related glycosides. Eight-channel CoulArray detection offers high selectivity and sensitivity with limits of detection in the low lg L 1 range, at least an order of magnitude lower than single-channel coulometric detection using the Coulochem detector. No special sample pretreatment is necessary and, because of the compatibility of the CoulArray detector with gradient elution, phenolic antioxidants of different polarities can be determined in a single run. In addition to the retention times, the ratios of the areas of the pre-dominant and post-dominant peaks to the area of the dominant peak can be used for improved identification of natural antioxidants.
Key Words: Phenolic compounds; Flavonoids; Natural antioxidants; Coulometric detection; Beer; Wine; Tea; Yacon; Received: January 4, 2005; revised: February 9, 2005; accepted: February 21, 2005 DOI 10.1002/jssc.200500003
Department of Analytical Chemistry, Faculty of Chemical Technology, University of s. Legi 565, Pardubice, nm. C CZ-532 10 Pardubice, Czech Republic 2 Instutute af Macromolecular Chemistry, Czech Academy of Sciences, Heyrovskho nm. 2, CZ-162 06 Prague 6, Czech Republic 3 Research Institute of Brewing and Malting Plc, Lpov 15, CZ-12044 Prague 4 Department of Food Engineering and Chemistry, Faculty of Technology, T. Bata University, CZ-76212 Zlin, Czech Republic
1 Introduction
Various phenolic and flavonoid natural antioxidants can influence the taste, flavour, color, and durability of beer, wine, tea, fruit juice, and other beverages. Furthermore, these compounds have beneficial effects on human health by decreasing the level of free radicals in the organism. Their presence in beverages arises from the plants used in the production process, such as barley, hops, and hop products used in beer brewing technology. Analysis of phenolic compounds occurring in fruit, vegetables, tea, and other beverages has attracted considerable attention due to the recent discovery of beneficial physiological and anti-carcinogenic properties for human health of many of these compounds [1 4]. Various techniques have been used for the determination of natural antioxidants, including high performance liquid chromatography (HPLC) with electrochemical [5 15], HPLC/MS [16 24],
Correspondence: Pavel Jandera, Department of Analytical Chemistry, Faculty of Chemical Technology, University of Pardu s. Legi 565, CZ-532 10 Pardubice, Czech Republic. bice, nm. C Phone: +420 466 037023. Fax: +420 466 037068. E-mail: Pavel.Jandera@upce.cz.
Because of a high selectivity and sensitivity, HPLC with electrochemical detection has become increasingly popular for the analysis of natural antioxidants in food and beverages. Both amperometric and coulometric detectors were used for reversed-phase HPLC, usually with an octadecyl silica column and aqueous-organic mobile phases containing acidic buffers to suppress the dissociation of weakly acidic phenolic compounds. Amperometric detection was used in isocratic HPLC of phenolic acids and flavonoids in beer and other beverages with mobile phases containing ammonium phosphate [5] or acetate buffer and citric acid [13] in aqueous acetonitrile, or in mixed ternary mobile phases containing methanol, 1-propanol, and water [14]. Coulometric detection usually provides improved sensitivity and signal stability in comparison to amperometric detection. The Coulochem II electrochemical detector with two coulometric detection cells in series was used for
www.jss-journal.de
Original Paper
UV spectrophotometric [5, 8, 9, 25 36], or fluorimetric [28, 36] detection, capillary zone electrophoresis [37], thin-layer chromatography [21], or nuclear magnetic resonance spectrometry [22].
1006
the HPLC analysis of phenolic compounds and flavonoids in beer with mobile phases containing formic acid [8, 9] or acetic acid [7, 15] in aqueous methanol or acetonitrile as the mobile phase. Many natural antioxidants are strongly polar compounds, requiring highly aqueous mobile phases for successful separation. On the other hand, some flavonoids are rather non-polar and mobile phases with a higher concentration of organic modifier are necessary to accomplish their elution. To enable HPLC separation of compounds strongly differing in polarities in a single run, gradient elution with increasing concentration of an organic modifier has to be used. However, amperometric and conventional coulometric electrochemical detection are generally not compatible with the gradient elution mode. This problem has been solved by the development of a multi-channel CoulArray detector with arrays of 4, 8, 12, or 16 three-electrode electrochemical cells connected in series, controlled by software compensating for the baseline drift during gradient runs. Gradient elution with increasing concentration of methanol or acetonitrile in aqueous mobile phases buffered with sodium dihydrogenphosphate and phosphoric acid and eight-channel [6] or twelve-channel [12] CoulArray detectors were used for the analysis of phenolic compounds and flavonoids in wort and beer [6, 38] and in fruit and vegetables [12]. Achilli et al. [10] and Gamache et al. [11] reported applications of a 16-channel CoulArray detector for the analysis of more than 30 phenolic compounds and flavonoids in juice and other beverages. They used gradient elution with increasing concentrations of methanol or acetonitrile in aqueous-organic mobile phases with sodium dodecylsulphate, sodium phosphate, and phosphoric or nitriloacetic acid additives. The electrochemical CoulArray detection is not only very sensitive and highly selective, but also provides different and reproducible signal responses to the sample compounds at the reduction or oxidation potentials applied across the individual flow-through cells connected in series (in 4, 8, 12, or 16 channel arrays). Hence, the ratios of the areas of the dominant peak (recorded in the channel providing the highest response) to the areas of so-called pre-dominant and the post-dominant peaks (in the channels just before and after the dominant peak channel, respectively) can be used to confirm the identification of sample compounds. Recently, we presented an optimised method for the analysis of 27 phenolic and flavonoid antioxidants in beer using gradient HPLC with CoulArray detection [39]. In the present work, we adapted this method for the analysis of 32 natural antioxidants in various beverages, including beer, wine, tea, and plant extracts.
J. Sep. Sci. 2005, 28, 1005 1022 www.jss-journal.de
1007
Table 1. Phenolic and flavonoid antioxidants (structures in Figure 1) and CoulArray detection conditions. Dom dominant peaks, Pre pre-dominant peaks, Post post-dominant peaks, pH = 3.14. No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
a) b)
Compound Gallic acid, 3,4,5-trihydroxybenzoic acid Protocatechuic acid, 3,4-dihydroxybenzoic acid p-Hydroxybenzoic acid
Potential of the dominant peak 600 mV 600 mV 900 mV 900 mV 900 mV 900 mV 800 mV 600 mV 600 mV 600 mV 800 mV 900 mV 900 mV 900 mV 600 mV 900 mV 600 mV 600 mV 900 mV 600 mV 600 mV 900 mV 300 mV 600 mV 800 mV 600 mV 900 mV 600 mV 800 mV 900 mV 900 mV 800 mV
Ratio of peak areas Pre/Dom.a) 0.03 0.34 0.08 0.67 0.57 0.54 0.34 0.15 0.19 0.11 0.06 0.76 0.42 0.19 0.52 0.43 0.10 0.18 0.09 0.12 0.04 0.48 0.10 0.17 0.07 0.62 0.41 0.74 0.42 0.84 Post/Dom.b) 0.62 0.38 0.71 0.18 0.20 0.34 0.90 0.51 0.52 0.12 0.12 0.19 0.32 0.71 0.24 0.13 0.61 0.70
Esculin, 6,7-dihydroxycoumarin-6-b-D-glucopyrano- Sigma Aldrich side 4-Hydroxyphenylacetic acid Vanillic acid, 4-hydroxy-3-methoxybenzoic acid (+)-Catechin, trans-3,3949,5,7-pentahydroxyflavane Chlorogenic acid, 5-O-(3,4-dihydroxycinnamoyl)D-quinic acid Caffeic acid, 3,4-dihydroxycinnamic acid Syringic acid, 3,5-dimethoxy-4-hydroxybenzoic acid Vanilline, 4-hydroxy-3-methoxybenzaldehyde Salicylic acid, 2-hydroxybenzoic acid p-Coumaric acid, trans-4-hydroxycinnamic acid Umbelliferone, 7-hydroxycoumarin ( )-Epicatechin, cis 3,39,49,5,7-pentahydroxyflavane Scopoletin, 7-hydroxy-5-methoxycoumarin Ferulic acid, 3-(4-hydroxy-3-methoxyphenyl)propenoic acid Sinapic acid, 3,5-dimethoxy-4-hydroxycinnamic acid 4-Hydroxycoumarin, 4-hydroxy-1-benzopyran-2one Rutin, quercetin-3-rutinoside Quercetin-3-arabinoside, quercetin-3-O-b-D-arabinofuranoside Naringin, naringenin-7-rhamnosidoglucoside Myricetin, 3,39495,597-hexahydroxyflavone Quercetin, 3,39495,7-pentahydroxyflavone Apigenin, 495,7-trihydroxyflavone Quercetindimethylether, 5,3949-trihydroxy-3,7dimethoxyflavone Biochanin A, 5,7-dihydroxy-49-methoxyisoflavone Esculetin, 6,7-dihydroxy-2H-1-benzopyran-2-one trans-Resveratrol, 3,49,5-trihydroxystilbene Naringenin Hesperetin, 5,7,39-Trihydroxy-49-methoxyflavanone Pinosylvin, 3,5-dihydroxystilbene Pre-dominant to dominant peak area ratio. Post-dominant to dominant peak area ratio. Sigma Aldrich Fluka Fluka Fluka Fluka Sigma Aldrich Fluka Fluka Sigma Aldrich Sigma Aldrich Fluka Sigma Aldrich Sigma Aldrich Sigma Aldrich Sigma Aldrich Fluka Fluka Fluka Sigma Aldrich Fluka Fluka Fluka Sigma Aldrich Sigma Aldrich VPS Sigma Aldrich Sigma Aldrich VPS
www.jss-journal.de
1008
Table 2A. Beer samples. No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Beer Heineken Stella Artois premium Amstel Bitburger Premium Pils Kronenbourg 1664 Bud Budweiser ZIPFER Original, Premium Veltins Budweiser Brgerbru Pilsner Urquell Pilot sweet wort Pilot hopped wort Pilot fresh beer Pilot aged beer Manufacturer Brewery Heineken, Amsterdam, The Netherlands Brewery Interbrew Belgium, Brussels, Belgium Brewery Heinecken, Amsterdam, The Netherlands Bitburger Brewery, Bitburg/Eifel, Germany Kronenbourg Brewery, Strasbourg, France Birra Peroni, Roma, Italy Brewery Zipf, Austria Brewery C and A Veltins, Germany esk Bude jovice, Czech Republic Town Brewery Samson C Brewery Pilsner Urquell, Czech Republic Research Inst. of Brewing and Malting, Praha, Czech Rep. Research Inst. of Brewing and Malting, Praha, Czech Rep. Research Inst. of Brewing and Malting, Praha, Czech Rep. Research Inst. of Brewing and Malting, Praha, Czech Rep.
www.jss-journal.de
1009
Vrbovec winery, Znojmo region, Znovn Znojmo, atov, Czech Republic S Vrbovec winery, Znojmo region, Agrovno Vrbovec, Czech Republic Vrbovec winery, Znojmo region, Agrovno Vrbovec, Czech Republic Velk Blovice winery, Velk Popovice region, Vladimr Tetur, Velk Blovice, Czech Republic
Table 2C. Tea samples. No. 1 2 3 4 5 6 7 8 9 10 11 Type Ceylon OPA Tea Ceylon OPI Pettigalia Darjeeling Leaf Tea Gundpowder Keemun Congoun Sikkim FTGFOP 1 Pickwick Morning Tea Assam GFBOP Kenya GFBOP Yunnan FOP Napal FTGFOP Maloom
system (Millipore, Bedford, MA, USA). Mobile phases were prepared by dissolving ammonium acetate in water or in acetonitrile-water and adjusted to appropriate pH by adding a few drops of formic acid. In gradient elution, the concentration ratio of the component B to the component A continuously increased. 0.3868 g of ammonium acetate was dissolved in 1 L of water and pH was adjusted with formic acid to the appropriate value to prepare the mobile phase A (weak eluent). The mobile phase B (strong eluent) was prepared by diluting the mobile phase A with acetonitrile in the volume ratio 1:1.
Table 2D. Yacon extract samples (in ethanol). No. 1 2 3 4 Yacon plant part Caudexes Leaves Stalks Tubers
from Mr. O. Bujnoch, Golden Pike Hotel, Pardubice, Czech Republic. The tea samples were donated by the Friends of Tea Society, Prague, Czech Republic (Table 2.C). Yacon (Smallanthus sonchifolius) ethanolic extracts from different plant parts tubers, leaves, stalks, and caudexes were obtained from Prof. Lachman, Agricultural University, Prague, Czech Republic Table 2.D.
1010
Table 3. Column types, dimensions and gradient programs (Fm flow-rate, mL min 1). Column, Manufacturer A B Aqua - C18, 15063.0 mm, 5 lm, Phenomenex, Torrance, CA, USA LiChrospher STAR 60, RP SelectB, 15063.0 mm, 5 lm Merck, Darmstadt, Germany Purospher STAR, RP - 18e, 15063.0 mm, 5 lm Merck, Darmstadt, Germany Zorbax SB - C18, 15063.0 mm, 5 lm, Agilent, Palo Alto, CA, USA A1: B1: Gradient program 0 min: 5% ACN; 25 min: 5%ACN; 60 min: 14% ACN; 70 min: 21% ACN; 100 min: 50% ACN 0 min: 7.7% ACN; 12 min: 7.7% ACN; 35 min: 21.8% ACN 0 min: 2% ACN; 20 min: 2% ACN; 50 min: 9% ACN; 65 min: 19% ACN; 90 min: 50% ACN pH 3.14 3.00 Fm 0.6 1.0
C1:
3.41
0.40
Step, 0 min: 5% ACN; 15 min: 5% ACN; 15 min: 20% D1: ACN, 70 min: 20% ACN D2: 0 min: 3% ACN; 30 min: 5% ACN; 35 min: 8% ACN; 50 min: 10% ACN; 65 min: 16% ACN; 75 min: 27% ACN; 85 min: 32% ACN
3.14 3.14
0.23 0.23
cratic mode. A lower concentration of acetonitrile in the mobile phase had to be used for elution of polar analytes 1 9 (5 mmol L 1 ammonium acetate in 5:95 acetonitrile water) than for the elution of less polar more strongly retained antioxidants 10 20 (5 mmol L 1 ammonium acetate in 20:80 acetonitrile water). Because the CoulArray detector allows the use of gradient elution, all the antioxidants could be analyzed in a single run. The stronger solvent B (5 mmol L 1 ammonium acetate in water acetonitrile, 1:1), was mixed with the solvent A (5 mmol L 1 ammonium acetate in water) in increasing proportions according to pre-set linear or multi-segmented programs (Table 3). The temperature of the column and the detector array was maintained at 36 l 0.18C.
the suitability of 11 different stationary phases for simultaneous analysis of 32 flavonoids and phenolic antioxidants [39], we selected four columns providing best separation for the analysis of various types of samples in the present work. We used the normalized resolution product, r, and the overlapping resolution maps to optimize the pH (in the range from 2.74 to 4.4), the flow rate, and the gradient program. The optimized pH (3.0 to 3.41) and gradient conditions for the separation of natural antioxidants in different sample matrices on the columns A D are given in Table 3.
1011
Figure 2. pH effects on the peak areas at various potentials; CoulArray detector. 2A: Epicatechin: dominant potential = 600 mV at pH = 2.75 3.91 and 300 mV at 4.35. 2B: Sinapic acid: dominant potential = 600 mV at pH = 2.75 3.91. 2C: Syringic acid: dominant potential = 800 mV at pH = 2.75, 700 mV at pH = 3.14, 900 mV at pH = 3.55, 600 mV at pH = 3.91, and 800 mV at pH = 4.35.
Figure 3. Chromatograms of a beer sample. 3A: CoulArray detector, column D (Zorbax SB-C18), gradient program D2 (Table 3). 3B: COULOCHEM II detector, column D (Zorbax SB-C18), Step gradient D1 (Table 3). Numbers of peaks are as in Table 1.
sent work yielded a sufficient signal intensity for quantitative analysis at the dominant potentials given in Table 1 and the baseline noise interferences were low enough to achieve the limits of detection in the low lg L 1 range. The optimized methods were used to analyze beer and wine samples and tea and yacon extracts. The individual antioxidants were identified on the basis of their retention times and according to the ratios of the areas of the predominant and the post-dominant peaks to the area of the dominant peak (Table 1).
J. Sep. Sci. 2005, 28, 1005 1022 www.jss-journal.de
1012
Table 4A. Parameters of calibration equation: A = q + k.c. A = peak area (mC), c = concentration (mg L 1), q = intercept, k = slope, rk = correlation coefficient, sx = standard deviation of determination in the concentration range 0.005 1 mg L 1, tr = retention times, LOQ = limits of quantitation at the signal ten times the baseline noise. Commercial beer samples, column D, gradient D2 for CoulArray a), isocratic (5 mM ammonium acetate in 15% acetonitrile) for Coulochem II b, pH = 3.14. Solute (Table 1) Coul Array a) tr, min 3.0 5.7 10.5 13.7 16.7 15.6 17.9 21.3 25.0 28.2 43.6 37.2 40.2 53.3 56.3 64.5
a)
Coulochem IIb)
a)
LOQ lg L 1 1.86 0.98 7.89 2.12 5.55 0.42 1.90 118.70 3.63 5.13 23.31 3.61 17.51 3.64 2.63 6.30
Parameters of calibration curves q, mC 0.27 l 0.25 0.11 l 0.13 ( )0.11 l 0.20 0.03 l 0.15 ( )0.42 l 0.27 ( )0.06 l 0.12 ( )0.67 l 0.36 ( )0.01 l 0.11 0.03 l 0.31 0.10 l 0.19 ( )0.14 l 0.08 ( )0.43 l 0.09 ( )0.31 l 0.23 ( )0.39 l 0.11 ( )0.11 l 0.27 ( )0.16 l 0.07 k, mg L 1.mC 1 22.39 l 0.05 25.77 l 0.26 6.30 l 0.62 19.14 l 0.60 22.11 l 0.54 10.77 l 0.22 13.84 l 0.62 18.67 l 0.24 15.61 l 0.61 16.69 l 0.33 6.27 l 0.27 13.30 l 0.17 14.17 l 0.44 10.60 l 0.20 16.95 l 0.52 3.43 l 0.13 rk
sx,% 1.41 1.81 1.44 1.46 1.36 1.43 1.36 1.60 1.23 1.66 1.52 1.81 1.70 1.77 1.65 1.68
tr, min b) min 2.8 5.1 10.3 14.5 17.1 16.3 18.0 22.0 25.3 28.7 44.2 37.9 40.9 50.1 53.6 62.6
LOQ b) lg L 1 23.28 12.20 98.56 26.44 69.43 5.27 23.70 1483.68 45.41 64.07 291.38 45.06 218.82 45.53 32.85 78.78
1 2 3 5 6 7 8 9 10 11 12 13 15 17 18 20
0.9988 0.9998 0.9952 0.9985 0.9991 0.9996 0.998 0.9997 0.9977 0.9996 0.9981 0.9998 0.9986 0.9997 0.9986 0.9979
Figure 4. Chromatogram of a mixture of phenolic and flavonoid antioxidant standards. Column C (Purospher STAR, RP-18e, 2.06150 mm, 5 lm), gradient C1 (Table 3), pH = 3.14, flow rate 0.4 mL min 1. Numbers of peaks are as in Table 1. J. Sep. Sci. 2005, 28, 1005 1022 www.jss-journal.de
1013
LOQ, lg L 1 1.89 1.88 55.25 17.41 5.28 5.20 0.29 1.90 8.65 11.27 8.28 9.68 10.59 6.69 9.84 7.43 79.18 6.21 94.56 63.02 26.33 14.98 143.37 66.01 139.86 14.31 42.46
the separation conditions optimized for the analysis of other sample types (Table 4.B D). The quantitation limits calculated as the concentrations yielding a signal ten times the baseline noise are significantly lower, by at least one order of magnitude, with the CoulArray detector than those obtained than with the Coulochem II for most natural antioxidants. Figure 3.A and Figure 3.B illustrate the application of the two detection techniques for the analysis of a beer sample on a Zorbax SB-C18 column (D). A smooth baseline was recorded using a multi-segment linear gradient D2 with Coul-Array detection (Figure 3.A), whereas a significant sudden baseline shift corresponding to the step change in mobile phase composition is apparent on the Coulochem II detector record when a two-step gradient D1 was used (Figure 3.B).
J. Sep. Sci. 2005, 28, 1005 1022 www.jss-journal.de
Figure 4 shows the chromatogram of a mixture of standard compounds, 0.25 mg L 1 each, at the optimized HPLC separation selectivity and CoulArray sensitivity under gradient conditions on a Purospher Star column (C, Table 3).
1014
Table 4C. Wine samples, CoulArray detector, column C, gradient C1. Solute (Table 1) 1 2 3 5 6 7 8 9 10 12 13 14 15 17 18 19 20 21 22 26 27 29 30 31 32 Parameters of calibration curves tr, min 2.5 5.7 10.6 15.4 19.0 29.3 21.9 24.3 31.7 17.0 39.0 41.7 45.4 48.8 54.0 56.5 62.9 65.2 66.3 82.7 85.5 68.2 75.5 77.1 81.1 q, mC 0.10 l 0.11 0.26 l 0.29 ( )0.02 l 0.05 0.15l 0.21 0.55 l 0.26 ( )0.02 l 0.07 ( )0.02 l 0.03 0.15 l 0.15 0.06 l 0.11 0.13 l 0.04 0.22 l 0.10 0.32 l 0.09 0.18 l 0.09 0.03 l 0.07 0.07 l 0.06 0.37 l 0.23 ( )0.05 l 0.03 ( )0.01 l 0.006 0.03 l 0.05 ( )0.02 l 0.04 0.33 l 0.22 ( )0.38 l 0.30 0.06 l 0.07 0.12 l 0.05 ( )0.03 l 0.09 k, mg L 1.mC 1 19.07 l 0.23 30.53 l 0.64 15.57 l 0.11 29.49 l 0.45 21.24 l 0.56 12.51 l 0.15 14.59 l 0.06 29.29 l 0.33 10.41 l 0.13 7.83 l 0.09 10.93 l 0.11 11.78 l 0.19 6.20 l 0.11 19.71 l 0.08 10.67 l 0.07 3.89 l 0.30 4.54 l 0.03 2.53 l 0.006 4.44 l 0.05 9.28 l 0.04 13.83 l 0.44 12.22 l 0.33 20.12 l 0.14 8.33 l 0.12 14.73 l 0.18 rk 0.9997 0.9991 0.9999 0.9995 0.9986 0.9997 0.9999 0.9998 0.9996 0.9997 0.9998 0.9995 0.9992 0.9999 0.9999 0.9857 0.9999 0.9999 0.9997 0.9999 0.9984 0.9986 0.9999 0.9996 0.9998 sx ,% 1.97 1.89 1.01 1.41 1.5 1.13 1.05 1.29 1.23 1.08 1.19 1.17 1.2 1.15 1.13 1.54 1.05 1.01 1.09 1.07 1.37 1.57 1.12 1.11 1.15 LOQ, lg L 1 1.87 1.84 4.35 3.75 2.15 14.14 6.99 2.90 5.64 9.86 7.33 6.66 4.97 6.36 6.19 27.57 33.11 113.99 19.59 10.87 14.79 6.65 6.52 2.04 7.33
Table 4D. Tea and yacon extract samples, CoulArray detector, column B, gradient B1. Solute (Table 1) 4 7 9 10 14 16 17 19 28 Parameters of calibration curves tr, min 5.9 7.6 11.6 13.5 20.4 23.2 25.0 29.4 10.5 q, mC ( )0.86l 0.26 0.36 l 0.34 ( )0.39 l 0.50 0.85 l 0.47 ( )0.81 l 0.08 ( )0.33 l 0.57 ( )1.66 l 0.76 0.45 l 0.62 ( )0.68 l 0.49 k, mg L 1.mC 1 14.28 l 0.25 44.07 l 0.33 54.87 l 0.48 44.88 l 0.44 5.53 l 0.06 36.16 l 0.55 41.60 l 0.73 4.89 l 0.48 26.14 l 0.46 rk 0.9992 0.9999 0.9998 0.9998 0.9998 0.9994 0.9992 0.9994 0.9992 sx ,% 1.46 1.6 1.89 1.7 1.09 2.01 2.33 1.87 1.84 LOQ, lg L 1 7.70 2.50 2.01 2.45 19.89 9.57 2.63 32.07 4.21
Table 5.A shows the results of the analysis of 10 commercial lager beer samples (see Table 2.A). The concentrations of 15 natural antioxidants were determined from the
J. Sep. Sci. 2005, 28, 1005 1022 www.jss-journal.de
dominant peak signals using the characteristics of the calibration curves listed inTable 4.A. Significant differences were found between the concentrations of the indi-
1015
Figure 5. Chromatograms of two commercial beer samples. Column D (Zorbax SB C18, 2.16150 mm, 5 lm), gradient program D2 (Table 3). Numbers of peaks are as in Table 1. 8A: Pilsner Urquell, Brewery Pilsner Urquell, Czech Republic. 8B: Zipfer Original, Premium lager beer, Brewery Zipf, Austria.
vidual natural antioxidants in various beer samples. Some antioxidants were not detected in concentrations above the detection limits (1 5 lg/L). In most beer samples, ferulic acid (17), rutin (20) vanillic acid (6), and epicatechin (15) were found at concentrations higher than 0.5 lg/L. We also analyzed experimental samples of sweet wort, hopped wort, fresh beer, and aged beer, prepared at the
J. Sep. Sci. 2005, 28, 1005 1022 www.jss-journal.de
Research Institute of Brewing and Malting in Prague to monitor the changes in antioxidant concentration during beer brewing and storage. In this case, analysis of some less polar and more strongly retained flavones (compounds 21 27, Table 1) was required, in addition to the more polar phenolic antioxidants. We used a Phenomenex Aqua C18, 5 lm, 3.06150 mm column and a gradient program with extended acetonitrile concentration
1016
Table 5A. Concentrations (c) of antioxidants identified in analyzed samples. Commercial beer samples (Table 2.A), CoulArray detector, column D, gradient D2, c in mg L-1. Solute (Table 1) 1 2 3 5 6 7 9 8 10 11 12 13 15 17 18 20
a)
5 a) 1.25
a)
9 0.35 0.38
a)
10 0.36 0.54 1.47 0.65 a) 0.19 2.25 0.26 a) a) 2.08 1.38 6.70 1.39 0.43
0.42 a) a) 0.55
a)
range, up to 50% (column A, gradient A1 in Table 3), which provide improved separation for these sample types over the Zorbax SB-C18 and gradient D2 conditions. From this series of experiments, only the results for the consequent stages of one experimental technological process (sweet wort (11), hopped wort (12), fresh beer (13), and aged beer (14) samples 11 14 in Table 2.A) are shown (Figure 6.A D). Table 4.B lists the parameters of calibration curves used for the determination of the antioxidants in the pilot beer samples under optimized separation conditions. The results are shown in Table 5.B. The concentrations of most antioxidants were approximately constant or slightly increased during the brewing process (sweet wort hopped wort fresh beer), but decreased more or less during storage (fresh beer aged beer). Of the antioxidants analyzed, gallic acid, 4-hydroxybenzoic acid, 4-coumaric acid, 4-hydroxycoumarine, salicylic acid, ferulic acid, catechin, epicatechin, rutin, sinapic acid, and quercetin were found in all types of samples at concentrations 0.2 23 mg/L; other antioxidants occurred either in some samples only, or at lower concentrations.
range should be used (ending at 50% acetonitrile, like in the analysis of beer samples, but starting at 2% acetonitrile gradient program C1). Column C (Purospher STAR-RP-18e, 12563 mm ID) provided optimum separation for wine samples and was used for their analysis. Table 5.C shows the concentrations of 25 natural antioxidants in the wine samples, calculated using the parameters of the calibration curves under optimized gradient conditions inTable 4.C. Higher concentrations of (+)-catechin and ( )-epicatechin were found in red wines than in white wine samples. Sinapic acid, syringic acid, and hesperetin were contained in red wines only, whereas 4-hydroxycoumarin was found only in white wines. Gallic acid could not be identified in red wine samples, because of high interference concentrations (probably red colorants). The sample 1, Chardonnay white wine, contained a lower concentration of resveratrol than other samples, possibly because of different climatic conditions in the year 2003 (the other samples were from the year 2002) and of different kind of grapes. Generally, significantly higher concentrations of natural phenolic compounds were found in wine samples than in beer samples. Figure 7.A D show the chromatograms of the four wine samples analyzed.
1017
Table 5C. Concentrations (c) of antioxidants identified in analyzed samples. Wine samples (Table 2.B), CoulArray detector, column C, gradient C1, c in mg L-1. Solute (Table 1) c (mg L 1) Sample (Table 2.B) 1 1 2 3 5 6 7 8 9 10 12 13 14 15 17 18 19 20 21 22 26 27 29 30 31 32
a b
13 5.66 0.31 2.05 3.63 0.40 0.59 1.80 0.86 0.20 0.19 a 5.12 0.37
a
14
a)
3
b
4 b 3.36 0.58 2.16 3.20 24.30 a 2.12 5.25 a 7.97 a 122.9 0.46 3.49 b 7.58 4.47 6.93 a a 1.58 0.44 0.68 a
0.21 2.69 1.81 0.36 0.57 1.46 1.02 0.10 0.11 0.22 3.78 0.35
a
3.09 0.61 1.15 3.59 23.86 a 2.79 5.25 a 8.27 a 68.20 0.38 5.45
b
0.20 0.09 0.27 0.80 0.33 0.27 0.34 1.31 0.45 8.99 4.92 0.04 0.09 0.08 0.44 0.53 0.01 0.12
0.46 4.95 a 23.21 0.55 b 4.84 5.97 1.34 1.73 0.12 a 0.56
b b
0.47 1.41 0.53 13.92 3.01 0.03 a a 0.55 0.53 0.01 0.04
0.02 0.08
dants found in the tea and yacon extract samples: esculin, catechin, caffeic acid, syringic acid, umbelliferone, ferulic acid, scopoletin, and esculetin. The parameters of the calibration curves of standards under optimized conditions in Table 4.D were used for quantitative analysis of tea and yacon extract samples. Figure 8.A,B show two examples of the chromatograms of tea samples. The results of the determination of phenolic compounds in 11 tea samples (in mg per 1 g of dry tea) are shown in Table 5.D. 4-Hydroxycoumarin (19) was found at highest concentrations in all tea samples. On the other hand, syringic and ferulic acids were found at low concentrations only. However, significant differences in
J. Sep. Sci. 2005, 28, 1005 1022 www.jss-journal.de
the concentrations of the individual antioxidants were found between various tea samples. In yacon extract samples, we found similar antioxidants as in tea samples, except for umbelliferone, hence we used the parameters of the calibration curves in Table 4.D for the quantitative analysis of yacon extracts. The concentrations of antioxidants (in mg per 1 L of liquid extract) are given in Table 5.E. Different concentrations of the individual antioxidants were found in the extracts from different parts of the yacon plant. In agreement with earlier reported results [43, 44], we found high concentrations of caffeic and ferulic acids in caudex extracts and somewhat lower concentrations in the leaves and stalks extracts, whereas these phenolic acids were not detected in the
1018
Figure 6. Chromatograms of pilot beer samples (Table 2.A). Column A (Aqua C18, 3.06150 mm, 5 lm), gradient program A1 (Table 3). Numbers of compounds are as in Table 1. 6A: hopped wort sample No. 11; 6B: sweet wort sample No. 12; 6C: fresh beer samples No. 13; 6D: aged beer sample No.14.
Figure 7. Chromatograms of Moravian wine samples (Table 2.B). Column C (Purospher STAR, RP C18, 2.06125 mm, 5 lm), gradient program C1 (Table 3); 7A: Chardonnay 2003 (sample 1), 7B: Riesling 2002 (sample 2), 7C: St. Lawrence 2002 (sample 3), 7D: Ruland blue 2002 (sample 4).
tuber extracts. We found relatively high concentrations of esculetin in the leaf and stalk extracts and of 4-hydroxycoumarin in stalk, tuber, and caudex extracts. Tubers conJ. Sep. Sci. 2005, 28, 1005 1022 www.jss-journal.de
tained catechin, which was not detected in other plant parts. Figure 9.A D shows chromatograms of yacon extracts from the different parts of the plant.
1019
Figure 8. Chromatograms of tea samples (Table 2.C). Column B (LiChrospher STAR 60, RP Select B, 4.06125 mm, 5 lm), gradient program B1 (Table 5); 8A: Ceylon OPA (sample 1), 8B: Assam GFOP (sample 8).
Table 5D. Concentrations (c) of antioxidants identified in analyzed samples. Tea samples (Table 2.C), CoulArray detector, column B, gradient B1, c in mg.g-1. Sample (Table 2C) 4 1 2 3 4 5 6 7 8 9 10 11
a
c (mg L 1) Solute (Table 1) 7 0.60 0.41 0.48 0.48 0.01 0.21 2.00 0.62 0.59 0.90 0.43 9 1.10 1.80
a
10
a
14 4.20 a
a a
16 0.10 0.16
a
17
a
19 5.80 6.80
a
a
a a
a
a a
a
a a
2.10 a
a
5.40 a
a
0.21 2.12
a
a
a
a
a
0.01
a a
0.05
a
7.80
a
0.85 0.27 a
a
0.20 0.27 a
a
a a
a
a a
a
a a
a
www.jss-journal.de
1020
Figure 9. Chromatograms of yacon extract samples (Table 2.D). Column B (LiChrospher STAR 60, RP Select B, 4.06125 mm, 5 lm), gradient program B1 (Table 5); 9A: caudexes (sample 1), 9B: leaves (sample 2), 9C: stalks (sample 3), 9D: tubers (sample 4).
www.jss-journal.de
1021
Table 5E. Concentrations (c) of antioxidants identified in analyzed samples. Yacon extract samples (Table 2.D), CoulArray detector, column B, gradient B1, c in mg L1. Sample (Table 2D) 4 1 2 3 4
a b
10 a
a a
16 a
a a
19 25.0
a
72.0 49.0
7.9
4 Concluding remarks
CoulArray multichannel electrochemical detection of flavonoids and phenolic antioxidants is very sensitive and selective and is compatible with gradient elution HPLC. The results are little affected by matrix interferences, so that the samples of the beverages do not require any special pre-treatment, just dilution with the mobile phase and filtration before injection. In addition to the retention times, the ratios of the areas of the pre-dominant and the postdominant peaks to the area of the dominant (most intense) peaks, recorded at different applied potentials can be used to improve the identification of sample compounds. The separation conditions (pH, flow-rate, and gradient profile) for natural phenolic and flavonoid antioxidants should be optimized with respect to the sample types and target analytes, such as 32 antioxidants occurring in beer, wort, wine, tea, and yacon extract samples analyzed in the present work. Phenomenex Aqua C18, Purospher Star RP-18 e, Zorbax SB C18, and LiChrospher STAR 60 RP Select B columns provided the best resolution. Gradient elution with increasing concentration of acetonitrile in mobile phases containing 0.005 M ammonium acetate acidified by addition of formic acid to pH = 3.14 3.41 provided excellent sensitivity in connection with multi-channel coulometric detection. The optimum separation and detection conditions were applied to the analysis of real samples. Significant differences were found between the occurrence and average concentrations of the identified antioxidants in the individual samples. Tens of peaks were detected in the chromatograms, from among which we identified 32 compounds using comparison with authentic standards. In addition to the retention times, the ratios of the areas of the pre-dominant, the post-dominant, and the dominant (most intense) peaks, recorded at different applied potentials are very helpful for peak identity confirmation. Unfortunately, standards were not available for all peaks in the chromatograms. We are investigating possibilities of connecting CoulArray detecJ. Sep. Sci. 2005, 28, 1005 1022 www.jss-journal.de
tion with HPLC/MS for improving the identification of unknown sample compounds.
Acknowledgements
Financial support of this research by the Ministry of Education, Youth and Sports of Czech Republic under research project No. MSM 0021627502 and by the Grant Agency of Czech Republic under Project No. 203/04/0917 is gratefully acknowledged.
References
[1] M.G.L. Hertog, P.C.H. Hollman, M.B. Katan, J. Agric. Food Chem. 1992, 40, 2379 2386. [2] M.G.L. Hertog, P.C.H. Hollman, Van de Putte, J. Agric. Food Chem. 1993, 41, 1242 50. [3] H. Tsuchiya, M. Sato, K. Hirotsugu, T. Okubo, L.R. Juneja, M. Kim, J. Chromatogr. B 1997, 703, 253 258. [4] J.J. Dalluge, B.C. Nelson, J.B. Thomas, L.C. Sander, J. Chromatogr. A 1998, 793, 265 274. [5] S.M. Lunte, J. Chromatogr. 1987, 384, 371 382. [6] T.M. Kenyhercz, P.T. Kissinger, J. Agric. Food Chem. 1977, 25, 959 961. [7] D.A. Roston, P.T. Kissiger, Anal. Chem. 1981, 53, 1695 1699. [8] P.J. Hayes, M.R. Smyth, I. McMurrough, Analyst 1987, 112, 1197 1204. [9] P.J. Hayes, M.R. Smyth, I. McMurrough, Analyst 1987, 112, 1205 1207. [10] M. Nardini, A. Ghiselli, Food Chem. 2004, 84, 137 143. [11] M.N. Peyrat-Maillard, S. Bonnely, C. Berset, Talanta 2000, 51, 709 716. [12] Ch. Guo, G. Cao, E. Sofic, R.L. Prior, J. Agric. Food Chem. 1997, 45, 1787 1796. [13] S. Floridi, O. Marconi, L. Montanari, Brewing Science Group, Bulletin 2002, December 2002, 49 56. [14] G. Achilli, G.P. Cellerino, P.H. Gamache, G.V. Melzi Eril, J. Chromatogr. 1993, 632, 111 117. [15] P. Gamache, E. Ryan, I.N. Acworth, J. Chromatogr. 1993, 635, 143 150. [16] U. Justesen, J. Chromatogr. A 2000, 902, 369 379. [17] M. Careri, A. Mangia, M. Musci, J. Chromatogr. A 1998, 794, 263 297.
1022
[18] F. Snchez-Rabaneda, O. Juregui, R.M. Lamuela-Ravents, J. Bastida, F. Viladomat, C. Codina, J. Chromatogr. A 2003, 1008, 57 72. [19] K. Robards, J. Chromatogr. A 2003, 1000, 657 691. [20] G. Shui, L.P. Leong, J. Chromatogr. A 2004, 1022, 67 75. ek, K. Valentov, J. Ulri[21] B. Simonovska, I. Vovk, S. Andrens chov, J. Chromatogr. A 2003, 1016, 89 98. [22] A.M. Gil, I.F. Duarte, M. Godejohann, U. Braumann, M. Maraschin, M. Spraul, Anal. Chim. Acta 2003, 488, 35 51. [23] S. McDonald, P.D. Prenzler, M. Antolovich, K. Robards, Food Chem. 2001, 73, 73 84. [24] G.K. Poon, J. Chromatogr. A 1998, 794, 63 74. [25] A. Escarpa, M.D. Morales, M.C. Gonzles, Anal. Chim. Acta 2002, 460, 61 72. [26] I. McMurrough, G.P. Hennigan, M.J. Loughrey, J. Inst. Brew. 1983, 89, 15 23. [27] Y. Amakura, M. Okada, S. Tsuji, Y. Tonogai, J. Chromatogr. A 2000, 891, 183 188. [28] M.A. Rodrguez-Delgado, S. Malovan, J.P. Prez, T. Borges, F.J. Garca Montelongo, J. Chromatogr. A 2001, 912, 249 257. [29] G. Shui, Lai Peng Leong, J. Chromatogr. A 2002, 977, 89 96. [30] L. Yao, Y. Jiang, N. Datta, R. Singanusong, X. Liu, J. Duan, K. Raymont, A. Lisle, Y. Xu, Food Chem. 2004, 84, 253 263. [31] I. McMurrough, G.P. Roche, K.G. Cleary, J. Inst. Brew. 1984, 90, 181 187.
www.jss-journal.de