Nova Biotechnol Chim (2018) 17(1): 95-102 DOI: 10.2478/nbec-2018-0010  Corresponding author: katarina.hrobonova@stuba.sk Nova Biotechnologica et Chimica HPLC separation and determination of dicoumarol and other simple coumarins in sweet clover Katarína Hroboňová1,, Jana Sádecká1 and Jozef Čižmárik2 1Institute of Analytical Chemistry, Faculty of Chemical and Food Technology, Slovak University of Technology in Bratislava, Radlinského 9, Bratislava, SK-812 37, Slovak Republic 2Comenius University, Faculty of Pharmacy, Department of Pharmaceutical Chemistry, Odbojárov 10, Bratislava, SK-812 37, Slovak Republic Article info Article history: Received: 15th March 2018 Accepted: 16th May 2018 Keywords: HPLC-DAD Coumarins Sweet clover Synchronous fluorescence Abstract Dicoumarol is a mycotoxin, that acts as a blood anticoagulant, is formed during the microbial action of molds and fungi in spoiled hay or silage containing high- coumarin plant. A HPLC-DAD method for determination of coumarins, including dicoumarol, coumarin, and 4-hydroxycoumarin was developed. Methanol and acetic acid were used as mobile phase with gradient elution. The simultaneous separation was performed using C18 type of stationary phase. The recoveries were 88.6 – 92.6 %, 91.8 – 95.0 %, and 89.7 – 94.1 % (evaluated for three concentration levels) for dicoumarol, coumarin, and 4-hydroxycoumarin respectively. The parameters of system suitability (repeatability of retention times and peak areas) were determined for evaluation of the method. The method showed a good linearity in the concentration range 0.7 – 100 µg.mL-1 for dicoumarol, 0.05 – 100 µg.mL-1 for coumarin and 4-hydroxycoumarin with correlation coefficients higher than 0.9885. Extracts of sweet clover herb, hay, and spoiled hay were subjected to HPLC- DAD analysis. The most abundant compound in sweet clover herb and hay extracts was coumarin. In spoiled sweet clover hay extract the 4-hydroxycoumarin was detected in addition. The formation of 4-hydroxycoumarin was also observed in the synchronous fluorescence spectra recorded at the wavelength difference of 90 nm (difference between emission and excitation wavelength).  University of SS. Cyril and Methodius in Trnava Introduction Sweet clover (Melilotus albus or Melilotus officinalis L.) is widespread plant in central Europe. Besides of vitamin C, allantoin, tannins, mineral salts and flavonoids it contains coumarin compounds. During the microbial action of molds and fungi (Penicillium jenseni, Penicillium nigricans) in spoiled hay or silage containing high- coumarin plant, the dicoumarol is formed (4-hydroxycoumarin is the intermediate of dicoumarol production from coumarin (Poulton et al. 1980). Dicoumarol, a symmetrical biscoumarin (3,3´-methylenebis(4- hydroxycoumarin; Fig. 1), is potentially toxic and as potent vitamin K antagonist and anticoagulant may cause animal death due to internal or external hemorrhaging (Stahmann et al. 1941; Smith et al. 1965; Casper et al. 1983). Its discovery led to the development of another modern anticoagulant drug, Warfarin (4-hydroxy-3-(3-oxo-1- phenylbutyl)-coumarin). The European Committee prohibited the presence of dicoumarol in cosmetic products and food (Regulation (EC) Bereitgestellt von Slovenská poľnohospodárska knižnica | Heruntergeladen 28.02.20 10:01 UTC Nova Biotechnol Chim (2018) 17(1): 95-102 96 No 1334/2008); EU Regulation (EC) No 1223/2009). Fig. 1. Chemical structures of coumarin, 4-hydroxycoumarin, and dicoumarol (respectively). Several analytical techniques are available for determination of dicoumarol, coumarin, and 4-hydroxycoumarin such as colorimetry (Emery et al. 1970), fluorescence spectrometry (Marcolan et al. 2011; Poláček et al. 2015), and liquid chromatography including thin layer chromatography (Bellis et al. 1967). The literature provides several studies involving the use of HPLC with UV, MS and fluorescence detectors for determination of individual coumarin, 4-hydroxycoumarin, and dicoumarol or in group with other phenolic compounds. Only a few papers deal with simultaneous separation and determination of selected compounds in sweet clover. Table 1 summarizes some studies using HPLC for the determination of investigated coumarins in different matrices. Techniques, such as ultrasonic liquid extraction for plant and cosmetics samples (Muir et al. 1992; Ikeda et al. 2009; Xi et al. 2010) or accelerated solvent extraction for food sample (Vieriková et al. 2009) were applied for sample pretreatment. The aim of the present work was to develop a rapid and sensitive HPLC method for simultaneous determination dicoumarol, coumarin, and 4-hydroxycoumarin in sweet clover herb and hay. Experimental Chemicals and sample The standards of dicoumarol (98 %), 4-hydroxycoumarin (98 %), and coumarin (99 %), were purchased from Sigma-Aldrich (St. Louis, USA). Methanol (HPLC grade) and acetic acid (99 %, an analytical grade) were purchased from Merck (Darmstadt, Germany). The deionized water (resistivity of 18.2 MΩ/cm) was obtained from a AquaMax ultra (series 370) water purification system. The sample of sweet clover (Melilotus officinalis L.; aerial parts) herb was purchased from Slovakia. The herb was dried to constant weight during 4 hour at 30 °C and powdered. The fresh herb was incubated without access to light and air at 23 °C for 14 and 21 days. Preparation of standard solutions The stock solutions of dicoumarol, coumarin, and 4-hydroxycoumarin were prepared by dissolving of standards in methanol (concentration of analyte 1 mg.mL-1) and stored at 8 °C. The working standard solutions and mixed standard working solution (concentration of each analyte 100 µg.mL-1) were prepared by appropriate dilution of stock solutions with methanol. The solutions were filtered through a 0.45 µm nylon membrane filter. Solutions of dicoumarol were prepared daily. Solutions of 4-hydroxycoumarin and coumarin were stable during a week. Sample preparation An accurately weighted portion of sample (approximately 0.1 g) was mixed with 20 mL of methanol and extracted under stirring on mechanical shaker for 60 min at 23 °C. The mixture was centrifuged for 10 min (4,000 rpm). The supernatant was used for HPLC analysis. HPLC analysis HPLC system (Agilent Technologies, series 1200) consisted of a binary pump, an autosampler, a column oven, a diode array detector, and a fluorescence detector. Data were acquired and controlled by the Agilent ChemStation chromatographic data system software. Separations of analytes were performed in reverse phase mode using a Symmetry C18 (150 mm x 3.9 mm I.D., 5 µm particle size; column I), Symmetry C18 (75 mm x 4.6 mm I.D., 3.5 µm particle size; column II), XBridge C18 (50 mm x 4.6 mm I.D., 3.5 µm particle size; column III) (Waters, USA) column. The isocratic and gradient elutions were performed using the mobile phases consisted of methanol with the addition 0.3 % acetic acid (A) Bereitgestellt von Slovenská poľnohospodárska knižnica | Heruntergeladen 28.02.20 10:01 UTC Nova Biotechnol Chim (2018) 17(1): 95-102 97 Table 1. HPLC methods for separation and determination of coumarin, 4-hydroxycoumarin, and dicoumarol. Stationary phase Mobile phase Detection LOD Linear range Sample Reference Coumarin C18 acetonitrile/0.1 % formic acid gradient elution MS – 0.026-13.5 g.mL-1 Green tea with sweet-herbaceous odour Yang et al. (2009) C18 acetonitrile/0.1% formic acid (50/50 v/v) DAD 0.02 g.kg-1 0.05-500 g.kg-1 Food, biscuits Vieriková et al., (2009) 4-hydroxycoumarin C18 methanol/phosphate buffer pH 3 (65/35 v/v) UV (286 nm) FL (372/290 nm) 0.0121 g.mL-1 0.0125 g.mL-1 – Plant sample (Morinda citrifolia) Ikeda et al. (2009) C18 methanol/acetonitrile/ phosphate buffer pH 4 gradient elution UV (280 nm) 0.0200 g.mL-1 0.07-4 g.mL-1 Pharmareutical preparative Vijayakumar et al. (2009) Dicoumarol phenyl methanol/acetonitrile/ water gradient elution DAD 0.2 g.mL-1 0.5-60 g.mL-1 Cosmetics Xi et al. (2010) C18 methanol/acetate buffer pH 6 (25/75 v/v) UV (303 nm) 2 g.mL-1 – Sweet clover hay, silage Muir et al. (1992) DAD – diode array detection; FL – fluorescence detection; LOD – limit of detection; MS – mass spectrometry; UV – ultraviolet detection. and 0.3 % aqueous solution of acetic acid (B). The gradient profiles and flow rates used are summarised in Table 2. The injection volume was 10 µL, and column temperature was maintained Table 2. The gradient profiles and flow rates for HPLC separation of target coumarins on tested columns. Column Flow rate [mL.min-1] Time [min] A [%] B [%] Column I 0.8 0 7 10 11 65 40 0 0 35 60 100 100 Column II 1.6 0 2.1 2.7 4.0 65 40 0 0 35 60 100 100 Column III 1.6 0 2.1 2.7 4.0 65 40 0 0 35 60 100 100 A – 0.3 % acetic acid; B – methanol containing 0.3 % acetic acid at 23 °C. The diode array detector was operated in the wavelength range of 190 – 400 nm and chromatograms were obtained at 300 nm for dicoumarol and 280 nm for other compounds under study. The fluorescence detector was operated in the emission wavelength range of 270 – 500 nm and chromatograms were obtained at excitation wavelength 300 nm and emission wavelength 390 nm for detection of dicoumarol. Synchronous fluorescence analysis Lumina Fluorescence Spectrometer (Thermo Scientific) equipped with a 150 Watt Ozone-free Xenon lamp was were used. Excitation and emission slits were both set at 5.0 nm. Scan speed was 200 nm/min. PMT Voltage was set at 500 V. The samples were placed in a 1 cm quartz cell. Method validation For the validation of the HPLC-DAD method, parameters such as linearity, limits of detection (LOD), limits of quantification (LOQ), precision, and accuracy were evaluated. The calibration curves of analytes were constructed after injection of standard solutions and obtained by plotting a graph of mean peak area versus corresponding concentration of analyte (seven concentration Bereitgestellt von Slovenská poľnohospodárska knižnica | Heruntergeladen 28.02.20 10:01 UTC Nova Biotechnol Chim (2018) 17(1): 95-102 98 levels, three replicate injections of each solution). The LODs and LOQs values were determined empirically by the injection of the series of diluted solutions of the analytes and were calculated as a signal to noise ratio of 3 and 10, respectively. The recovery tests were performed in sample of dried sweet clover spiked at three concentration levels of dicoumarol, coumarin, and 4-hydroxycoumarin. Spiked sample was held for 1 hour before extraction. Spiked and unspiked samples were treated by the same procedure. Seven independent extractions of each sample were carried out. Intra-day and inter-day precisions of the method were evaluated for seven preparations of sample within one day and three days, making triplicate injections under the working conditions and (expressed as RSD %). Results and Discussion HPLC separation Reversed phase separation mode is most commonly used for HPLC separation of mixtures of coumarin derivatives including of dicoumarol as documented Table 1. The work was focused on development of HPLC method for separation and determination of three of comarins, coumarin, 4-hydroxycoumarin, and dicoumarol. Coumarin is precursor of dicoumarol and 4-hydroxycoumarin is a probable intermediate compound. (Smith et al. 1938) The 4-hydroxycoumarin is converted into dicoumarol in the presence of formaldehyde (Stahmann et al. 1941). At the beginning of the study, a suitable analytical column, which would offer sufficient separation for quantification of analytes was estimated. Three reversed-phase C18 type analytical columns, Symmetry C18, and XBridge C18 with different particle size and column length, were tested. The mobile phases consist of methanol and acetic acid at different ratios (Fig. 2) were investigated for isocratic elution. The effectiveness of the separation was evaluated using standard solution containing coumarin, 4-hydroxycoumarin, and dicoumarol. The baseline separations of analytes were achieved, but the retention factors of dicoumarol were significantly higher in comparison to retention factors of coumarin and 4-hydroxycoumarin (higher than 7 for all tested stationary and mobile phases), which lead to longer analysis times. Gradient elution profiles shown in Table 2 were investigated, which resulted in effective separation of target compounds in analysis time lower than 15 min. 0 2 4 6 8 10 12 14 16 A/B (5 0/50) Column I A/B (6 0/40) A/B (5 0/50) Column II A/B (6 0/40) A/B (5 0/50) Column III A/B (6 0/40) k coumarin 4-hydroxycoumarin dicoumarol Fig. 2. Effect of mobile phase composition on retention factor of coumarins obtained for three analytical columns and isocratic elution with mobile phase A/B. A – 0.3 % acetic acid; B – methanol; flow rate 0.8 mL.min-1; DAD detection. The complete separation of analytes was achieved (resolution values higher than 2) using all tested columns as is shown in Fig. 3. The chromatographic characteristics for tested columns are summarised in Table 3. The repeatability of the retention times and areas (determined for standard solutions at concentration level of analytes 10 µg.mL-1) were lower than 1.1 % for all analytes and columns under study. All tested stationary phases were suitable for separation of coumarin, 4-hydroxycoumarin, and dicoumarol. The satisfactory resolution and symmetrical peaks of analytes were obtained. While the Symmetry C18 type of column is stable under pH range 2 – 8, the XBridge C18 stationary phase is very resistant to harsh conditions (pH 1 – 12) resulted from bridged ethylene hybrid technology. The results indicated, that fast separation using XBridge C18 was achieved, but this column was characterized by a lower column efficiency (greater HETP values) in comparison to other columns under study. Therefore, the Symmetry C18 type column was chosen as suitable column for next analyses of real samples. The decrease of particle size and column length and use of higher flow rate (Column II) resulted in decrease of analysis time (~2.5 times in comparison with Column I) Bereitgestellt von Slovenská poľnohospodárska knižnica | Heruntergeladen 28.02.20 10:01 UTC Nova Biotechnol Chim (2018) 17(1): 95-102 99 A Column I Coumarin B Column II 4-hydroxycoumarin C Column III Dicoumarol Fig. 3. Chromatographic separations of coumarin (1), 4-hydroxycoumarin (2), and dicoumarol (3) on three analytical columns and UV spectra of compounds under study. Mobile phase composition and flow rate as Table 2, DAD detection (chromatograms at 280 nm). and finally decrease of solvent consumption. Coumarins allow UV absorbance and some of them fluorescence too (Hroboňová et al. 2013). Spectrophotometric detection of coumarins was realised by DAD detector operated in the interval from 190 to 400 nm where the absorption maxima of target compounds are presented (280 nm for coumarin and 4-hydroxycoumarin, 300 nm for dicoumarol; Fig. 2). The maximum excitation and emission wavelength for fluorescence detection of dicoumarol were selected from excitation and emission spectra. As a result, the 300/390 nm (excitation/emission) wavelength pair was chosen for purpose of this study. Validation of the method The analytical figures of merit were determined under optimized separation and detection conditions. Parameters including LOD, LOQ, linear range, precision and accuracy were calculated to demonstrated the validation of the developed HPLC-DAD method (listed in Table 4). The calibration curves of the three analytes were Table 3. Retention times (tr), resolution (Rs), symmetry factor (As), and height equivalent of theoretical plate (HEPT) for HPLC separation of coumarins on tested analytical columns. a mobile phase and gradient profiles as in Table 2. tr [min] Rs As HEPT [µm] Column I coumarin 6.47 7.5 0.81 11.0 4-hydroxycoumarin 8.7 10.6 0.81 6.5 dicoumarol 9.86 0.87 1.7 Column II coumarin 2.36 5.9 0.88 11.1 4-hydroxycoumarin 3.1 6.1 0.85 5.4 dicoumarol 3.53 0.87 2.2 Column III coumarin 1.59 4.7 1.7 25.0 4-hydroxycoumarin 2.22 10.3 1.8 9.3 dicoumarol 3.29 0.92 2.3 Bereitgestellt von Slovenská poľnohospodárska knižnica | Heruntergeladen 28.02.20 10:01 UTC Nova Biotechnol Chim (2018) 17(1): 95-102 100 Table 4. Linearity and sensitivity of the HPLC-DAD method. Parameter Coumarin 4-hydroxycoumarin Dicoumarol Linear range [µg.mL-1] 0.05-100 0.05-100 0.7-100 Regression equation y = 70165c + 3.9 y = 139610c + 11.4 y = 22050c + 38.8 R2 0.9983 0.9896 0.9885 LOD [µg.mL-1] 0.01 0.02 0.2 LOQ [µg.mL-1] 0.05 0.05 0.7 created after the injection of a mixed standard solution. Results showed good linear relationship for each analyte (coefficient of determination in the range 0.9885 – 0.9983). The LODs of the method were calculated as a signal to noise ratio of 3 using the injection of series of more diluted standard solutions of analytes. The LOQs were evaluated using signal to noise ratio of 10. The LODs and LOQs of coumarin and its derivatives were in the range 0.01 – 0.2 µg.mL-1 and 0.05 – 0.7 µg.mL-1, respectively. Since the LODs and LOQs of dicoumarol obtained for fluorescence detection were higher (2.1 and 6.5 µg.mL-1, respectively) DAD detector was used in plant samples analyses. Recovery tests of method were performed for dried sample spiked with dicoumarol levels at three concentration levels. The recovery values of dicoumarol, coumarin, and 4-hydroxycoumarin were higher than 85 %. As documented Table 5, good repeatability of the recovery was achieved (RSDs less than 4 % for all spiked concentration levels). The intra-day and inter-day precisions of developed method, determined for spiked sample were lower than 5 % and 9 %, respectively. Analysis of real plant samples In order to show reliability and applicability of the developed method, the extracts of sweet clover samples were analysed. Samples of herb, dried plant (aerial parts), and herb incubated for 14 and 21 days, were used in this study. Sample extracts were prepared in methanol. Comparison of retention times from HPLC and UV spectra of sample peaks and standards were used for characterization of coumarins in extracts. Analyses of dried plant and herb extracts confirmed the absence of dicoumarol and 4-hydroxycoumarin in tested samples (Fig 4A, B). The content of coumarin in dried plant sample was 3.3 – 0.1 mg.g-1. The 4-hydroxycoumarin was formed in the fresh herb incubated for 14 and 21 days (Fig. 4C). The formation of 4-hydroxycoumarin was also observed in the synchronous fluorescence spectra recorded at the wavelength difference (difference between emission and excitation wavelength) of 90 nm (Fig. 5), using the recently developed method (Poláček et al. 2015). Fig. 5. Synchronous fluorescence spectra of fresh herb, dried plant, fresh herb after incubation and 4-hydroxycoumarin. Table 5. Recovery values of extraction method for spiked dried sweet clover sample. Coumarin 4-hydroxycoumarin Dicoumarol Recovery [%] RSD [%] Recovery [%] RSD [%] Recovery [%] RSD [%] C1 91.8 2.2 89.7 2.1 92.6 2.5 C2 95.0 3.0 94.1 2.6 92.5 3.1 C3 93.2 2.6 90.8 2.4 88.6 2.6 n = 7; C1 – 0.2 mg.g-1 of coumarin; 0.2 mg.g-1 of 4-hydroxycoumarin; 1.0 mg.mL-1 of dicoumarol; C2 – 1.0 mg.g-1 of coumarin; 1.0 mg.g-1 of 4-hydroxycoumarin; 2.0 mg.g-1 of dicoumarol; C3 – 2.0 mg.g-1 of coumarin; 2.0 mg.g-1 of 4-hydroxycoumarin; 10.0 mg.g-1 of dicoumarol. Bereitgestellt von Slovenská poľnohospodárska knižnica | Heruntergeladen 28.02.20 10:01 UTC Nova Biotechnol Chim (2018) 17(1): 95-102 101 Fig. 4. Chromatograms of extracts of Meliloti herba fresh herb (A), dried plant (B), and herb after incubation (C). Stationary phase: Column I, mobile phase composition and flow rate as Table 2, DAD detection (chromatograms at 280 nm); 1- coumarin, 2- 4-hydroxycoumarin, 3- dicoumarol. Conclusions A simple reversed-phase HPLC-DAD method including stationary phase of C18 type and methanol-acetic acid as mobile phase with gradient elution was developed for simultaneous determination of coumarin, 4-hydroxycoumarin, and dicoumarol in extracts of sweet clover samples. The separation of coumarins by HPLC was rapid, efficient and reproducible, and the method presented acceptable results for sensitivity, linearity, precision, and accuracy. The LOD obtained by the proposed method are comparable to other methods with spectrophotometric detection. Moreover, one should note that wider linear concentration range of each compound under study was achieved. The method has potential to be used in medical plant and related products analysis, since currently there is an interest on characterization and determination of biologically active compounds in natural samples. Acknowledgement This publication was supported by the Competence Center for SMART Technologies for Electronics and Informatics Systems and Services, ITMS 26240220072, funded by the Research & Development Operational Programme from the ERDF and funded by the Slovak Research A Meliloti herba B Meliloti herba, hay C Meliloti herba herb incubated for 14 (a) and 21 (b) days Bereitgestellt von Slovenská poľnohospodárska knižnica | Heruntergeladen 28.02.20 10:01 UTC Nova Biotechnol Chim (2018) 17(1): 95-102 102 and Development Agency under the bilateral project SK-SRB-2016-0006. References Bellis DM, Spring MS, Stoker JR (1967) The biosynthesis of dicoumarol. Biochem. J. 103: 202-206. Casper HHME, Alstad AD, Monson SB, Johnson LJ (1983) Dicoumarol levels in sweet clover toxic to cattle. Amet. Assoc. Vet. Lab. Diagnosticians, 25th Annu. Proc., 41- 48. Emery AE, Gear JR (1970) Determination of dicoumarol in spoiled sweet clover. Can. J. Plant Sci. 50: 39-45. EU Regulation (EC) No 1223/2009 of the European Parliament and of the Council of 30 November 2009, on Cosmetic Products (OJ L 342, 22.12.2009, p. 59). Official Journal of the European Union L 342: 59. Hroboňová K, Lehotay J, Čižmárik J, Sádecká J (2013) Comparison HPLC and fluorescence spectrometry methods for determination of coumarinn derivatives in propolis. J. Liq. Chromatogr. Relat. Technol. 36: 486- 503. Ikeda R, Wada M, Nishigaki T, Nakashima K (2009) Quantitation of coumarin derivatives in Noni (Morinda citrifilia) and their contribution of guenching effect on reactive oxygen species. Food Chem. 113: 1169- 1172. Marcolan M, Martins PA, Pedrosa VA, Rodrigues RH, De Oliveira HPM, Codognoto L (2011) Spectrofluorimetric determination of coumarin in commertial tablets. J. Fluorescenc. 24: 733-738. Muir AD, Goplen BP (1992) Quantitative reversed-phase HPLC analysis of dicoumarol in sweet clover hay and silage samples. J. Agric. Food Chem. 40: 820-823. Poláček R, Májek P, Hroboňová K, Sádecká J (2015) Fluorescence spectrometry as a tool for determinationof coumarins by multivariate calibration. J. Fluoresc. 25: 297-303. Poulton JE, McRee DE, Conn EE (1980) Intracellular localisation of two enzymes involved in coumarin biosynthesis in Melilotus alba. Plant Physiol. 65: 171- 175. Regulation (EC) No 1334/2008 of the European Parliament and of the Council of 16 December 2008, on flavouring and certain food ingredients with flavouring properties for use in and foods and amending Council Regulation (EEC) No 160/91, Regulation (EC) No 2232/96 and (CE) No 110/2008 and Directive 2000/12/EC. Official Journal of the European Union L 354/34: 15. Smith WK, Brink RA (1938) Relation of bitter ness to the toxic principle in sweet clover. J. Agric. Res. 57: 145-154. Smith WK, Gorz HJ (1965) Sweet clover improvement. Adv. Agron. 17: 163-231. Stahmann MA, Huebner CF, Link KP (1941) Studies on hemorrhagic sweet clover disease. V. Identification and synthesis of the hemorrhagic agent. J. Biol. Chem. 138: 513-527. Vieriková M, Germuška R, Lehotay J (2009) Determination of coumarin on food using ultraperformance liquid chromatography-electrospraytandem mass spectrometry. J. Liq. Chromatogr. Rel. Technol. 32: 95-105. Vijayakumar EKS, Dhore DM, Kumar M (2009) HPLC method for simultaneous determination of impurities and degradation products in Zynisamide. Indian J. Pharm Sci. 71: 521-526. Xi H, Ma Q, Wang C, Bai H, Liu X, Wang Y (2010) Simultaneous determination of 17 coumarins in cosmetics by high performance liquid chromatography. J. Instrum. Anal. 29: 1168-1172. Yang Z, Kinoshita T, Tanida A, Sayama H, Morita A, Watanabe N (2009) Analysis of coumarin and its glycosidically bound precursor in Japanese green tea having sweet-herbaceous odour. Food Chem. 114: 289-294. Bereitgestellt von Slovenská poľnohospodárska knižnica | Heruntergeladen 28.02.20 10:01 UTC