Glycyrrhiza glabra - American Chemical Society

Oct 11, 2016 - house database containing >1000 aroma-active reference volatiles. ...... compounds in an American Bourbon whisky by quantitative...
3 downloads 0 Views 412KB Size
Article pubs.acs.org/JAFC

Characterization of the Key Aroma Compounds in Raw Licorice (Glycyrrhiza glabra L.) by Means of Molecular Sensory Science Juliane Wagner, Michael Granvogl,* and Peter Schieberle Lehrstuhl für Lebensmittelchemie, Technische Universität München, Lise-Meitner-Straße 34, D-85354 Freising, Germany ABSTRACT: Application of the molecular sensory science concept including aroma extract dilution analysis (AEDA) on the basis of gas chromatography−olfactometry combined with gas chromatography−mass spectrometry elucidated the key odorants of raw licorice (Glycyrrhiza glabra L.). Fifty aroma-active compounds were located via AEDA; 16 thereof were identified in raw licorice for the first time. γ-Nonalactone, 4-hydroxy-2,5-dimethylfuran-3(2H)-one, and 4-hydroxy-3-methoxybenzaldehyde showed the highest flavor dilution (FD) factor of 1024. Forty-three compounds were quantitated by means of stable isotope dilution analysis (SIDA; 6 more compounds were quantitated using labeled standards with structures similar to the respective analytes) and odor activity values (OAVs; ratio of concentration to the respective odor threshold) were calculated revealing OAVs ≥1 for 39 compounds. Thereby, (E,Z)-2,6-nonadienal, 5-isopropyl-2-methylphenol, hexanal, and linalool showed the highest OAVs. On the basis of the obtained results, an aqueous reconstitution model was prepared by mixing these 39 odorants in their naturally occurring concentrations. The recombinate elicited an aroma profile very similar to the profile of raw licorice, proving that all key aroma compounds were correctly identified and quantitated. KEYWORDS: licorice, Glycyrrhiza glabra L., molecular sensory science concept, aroma extract dilution analysis, stable isotope dilution analysis, odor activity value, aroma recombination



Jaquier6 investigated a commercial aqueous extract of licorice with a special focus on lactones. The authors reported, besides fatty acids (C2−C16), phenols, and common saturated linear γlactones (C6−C14), on a series of new 4-methyl-γ-lactones and 4-ethyl-γ-lactones, the presence of which in trace amounts was confirmed by syntheses of the respective reference compounds. Tanaka et al.7 identified 127 compounds in the solvent extract of a steam distillate of Chinese licorice (Glycyrrhiza uralensis Fisch.). However, due to the heat influence during the workup procedure, it has to be questioned to what extent these volatiles were really present in the original Chinese licorice. Very recently, Farag and Wessjohann8 examined volatiles in the steam distillate of licorice and in licorice roots after solid-phase microextraction (SPME) by gas chromatography−flame ionization detection (GC-FID) and GC-MS and reported on phenols, aldehydes, and alcohols as main components of the volatile metabolome. However, for qualification of the substances, in parts only mass spectra of databases were applied, and for quantification experiments (authors determined relative concentrations), no isotopically labeled standards were used, which very likely leads to unreliable results using the SPME technique. Thereby, the extraction conditions, for example, temperature and exposure time, but especially the concentrations of other volatile compounds in the sample, can significantly influence the amount of the respective analyte adsorbed at the fiber. In summary, both identification and quantitation of licorice volatiles in the preliminary studies were performed with

INTRODUCTION Glycyrrhiza glabra L. is one of the oldest medically used plants of the world; the earliest record of its application was found in the Code of Hammurabi from 2100 BC.1 Utilized parts of the plant are the root and stolon, which are commonly designated as raw licorice. Beside its usage in medicine, for example, as anti-inflammatory or antiulcerous drug, raw licorice and its extract were used in the tobacco industry as a humectant and flavoring agent2 and, because of its brightening and desensitizing properties, also in cosmetics.3 Due to the very pleasant aroma, raw licorice is also highly appreciated as an ingredient in tea infusions. However, especially in Europe and North America, the most important application area is in confectionery, using the heated extract of raw licorice as a characteristic ingredient. During manufacturing, it is mixed with flour, sugar, glucose syrup, gelatin, and flavorings and is thickened by thermal processing. Despite the broad field of applications, up to now, the composition of the volatile fraction of G. glabra L. has been investigated only in a few studies. Frattini et al.4 were the first to analyze the volatile compounds of raw and heated licorice. They identified 28 volatile compounds in unheated licorice juice and 63 volatiles in heated licorice essential oil by gas chromatography−mass spectrometry (GC-MS) and infrared spectroscopy (IR), among them organic acids, esters, lactones (e.g., γ-nonalactone), aldehydes, ketones, alcohols, furan derivatives, phenols (e.g., 2-methoxyhenol, 5-isopropyl-2methylphenol), and heterocyclic compounds. Miyazawa and Kameoka5 identified 78 substances in the essential oil of licorice by GC, GC-MS, and proton nuclear magnetic resonance (1H NMR) spectroscopy and semiquantified the amounts of volatiles occurring in high concentrations, for example, octanoic acid and benzaldehyde, using the respective peak areas. Näf and © 2016 American Chemical Society

Received: Revised: Accepted: Published: 8388

August 16, 2016 October 5, 2016 October 11, 2016 October 11, 2016 DOI: 10.1021/acs.jafc.6b03676 J. Agric. Food Chem. 2016, 64, 8388−8396

Article

Journal of Agricultural and Food Chemistry

lactone;24 [2H2−3]-trans-4,5-epoxy-(E)-2-decenal;21 [2H3]-2-ethyl-3,5dimethylpyrazine;25 [2H3]-3-ethylphenol;26 [2H3]-ethyl propanoate;27 [2H4]-hexanal;28 [13C2]-3-hydroxy-4,5-dimethylfuran-2(5H)-one;29 [13C2]-4-hydroxy-2,5-dimethylfuran-3(2H)-one;30 [2H3]-4-hydroxy-3methoxybenzaldehyde;31 [2H3]-β-ionone;32 [2H3]-2-isobutyl-3-methoxypyrazine;33 [2H3]-2-isopropyl-3-methoxypyrazine;34 [2H7]-2-isopropyl-5-methylphenol;35 [2H2]-linalool;36 [2H3]-2-methoxyphenol;24 [2H3]-1-methoxy-4-(2-propenyl)benzene;37 [2H2]-3-methylbutanal;38 [2H3]-3-methyl-2,4-nonanedione;21 [2H3]-3-(methylthio)propanal;39 [2H2]-(E,E)-2,4-nonadienal;30 [2H2]-(E,Z)-2,6-nonadienal;21 [2H2]-γnonalactone;23 [2H4]-nonanal;40 [2H2]-(E)-2-nonenal;21 [2H2]-octanal;41 [2H2−3]-1-octen-3-one;42 [2H3]-pentanoic acid;43 [13C2]-phenylacetaldehyde;37 and [2H2−4]-4-propyl-2-methoxyphenol.44 Isolation of the Volatiles. Peeled, dried, and chopped raw licorice was frozen in liquid nitrogen and finely ground in a commercial blender. The licorice powder obtained (25 g) was extracted with dichloromethane (2 × 125 mL) by stirring vigorously for 2 × 1 h at room temperature. The combined organic extracts were subjected to high vacuum distillation by means of solvent assisted flavor evaporation (SAFE) technique45 to separate the volatiles from the nonvolatile material. The distillate obtained was dried over anhydrous sodium sulfate, filtered, and concentrated to ∼0.5 mL by a Vigreux column (40 cm × 1 cm i.d.) and microdistillation.46 Fractionation of the Volatiles. For identification experiments, the volatiles of raw licorice (500 g) were isolated as described above and then fractionated prior to analysis via high-resolution gas chromatography−olfactometry (HRGC-O) and high-resolution gas chromatography−mass spectrometry (HRGC-MS) to reduce possible coelutions. First of all, the SAFE distillate obtained was separated into two fractions containing either the acidic (AF) or the neutral/basic (NBF) volatiles. Therefore, liquid−liquid extraction with aqueous sodium carbonate solution (0.5 mol/L; 3 × 150 mL) was performed, whereby the NBF remains in the organic phase. The combined aqueous solutions, containing the AF, were adjusted to pH 2 with hydrochloric acid (1 mol/L), then the AF was extracted with dichloromethane (3 × 150 mL), and the organic phases were combined. Both fractions were dried over anhydrous sodium sulfate and were concentrated as described above. The NBF (dissolved in 1 mL of hexane) was further fractionated by column chromatography using a water-cooled glass column (30 cm × 1.8 cm i.d.) filled with a slurry of silica gel (30 g in pentane; water content of 7%). Seven fractions were obtained by eluting the volatiles with pentane/diethyl ether mixtures of increasing polarity (100:0, 95:5, 90:10, 75:25, 50:50, and 0:100, v:v; 100 mL each) and, finally, with dichloromethane (100 mL). After drying over anhydrous sodium sulfate, the fractions were concentrated to ∼0.2 mL as described above and then analyzed by means of HRGC-O and HRGC-MS. Aroma Extract Dilution Analysis and Identification Experiments. To avoid a potential overlooking of odor-active compounds, HRGC-O of the concentrated distillate was performed by three trained panelists. Flavor dilution (FD) factors of the aroma compounds were determined by diluting the extract stepwise 1+1 (v+v) with dichloromethane and analyzing each dilution by HRGC-O. By definition, the FD factor of an analyte is the highest dilution in which its odor impression could be perceived at the sniffing port for the last time. Aroma compounds detected by AEDA with FD factors ≥16 were identified on the basis of their retention indices determined on two capillary columns of different polarities (DB-FFAP and DB-5), their odor qualities and intensities perceived at the sniffing port, and their mass spectra obtained in electron ionization (EI) mode as well as in chemical ionization (CI) mode in comparison with the data obtained from reference compounds, which were available from an inhouse database containing >1000 aroma-active reference volatiles. High-Resolution Gas Chromatography−Olfactometry. HRGC-O was performed using a TRACE GC 2000 (ThermoQuest, Egelsbach, Germany) equipped either with a DB-FFAP or with a DB-5 column (both 30 m × 0.32 mm i.d., 0.25 μm film thickness; J&W Scientific; Agilent Technologies, Waldbronn, Germany). Helium was used as the carrier gas (flow rate = 1.7 mL/min), and aliquots (1.0 μL) of the samples were manually injected by the cold on-column

methods, which are not state-of-the-art for modern aroma analysis. Thus, up to now, no reliable data on the volatile compounds of licorice and no studies focused on a systematic characterization of the key aroma compounds to get information about the contribution of single odorants to the overall aroma of raw licorice are available. However, it is wellknown that only some of the volatiles in a food contribute to its overall aroma and that it is necessary to combine analytical and sensorial investigations to identify the key odorants of a certain food. Hence, the molecular sensory science concept was established over the past decades to differentiate between the aroma-active compounds and the numerous odorless volatiles or compounds present below their respective odor thresholds and to address the interaction of the key aroma compounds on the odorant receptor level using aroma recombinates.9,10 Thus, the aim of the present study was the application of the molecular sensory science concept to raw licorice to characterize its aroma by (i) identifying the key aroma-active compounds using aroma extract dilution analysis (AEDA) in combination with GC-MS, (ii) quantitating the odorants by stable isotope dilution analysis (SIDA), (iii) calculating the odor activity values (OAVs; ratio of concentration to odor threshold) to evaluate which compounds contribute to the overall aroma, and, finally, (iv) verifying the obtained results by recombination experiments.



MATERIALS AND METHODS

Licorice. The peeled, dried, and chopped raw licorice (Caelo; Caesar & Loretz, Hilden, Germany) was purchased in a local pharmacy. Chemicals. The following reference compounds used for identification and quantitation experiments were purchased from commercial sources: 4-allyl-2-methoxyphenol, 2-aminoacetophenone, anethole, 2,3-butanedione, 1,8-cineole, decanal, 2-ethyl-3,5(6)-dimethylpyrazine, 3-ethylphenol, ethyl propanoate, γ-hexalactone, hexanal, 3-hydroxy-4,5-dimethylfuran-2(5H)-one, β-ionone, 2-isobutyl-3-methoxypyrazine, 2-isopropyl-3-methoxypyrazine, 2-isopropyl-5-methylphenol, 5-isopropyl-2-methylphenol, linalool, 2-methoxyphenol, 2methylbutanoic acid, 3-methylbutanoic acid, 3-(methylthio)propanal, (E,E)-2,4-nonadienal, (E,Z)-2,6-nonadienal, γ-nonalactone, (E)-2nonenal, octanal, phenylacetaldehyde, phenylacetic acid, 2-phenylethanol (Sigma-Aldrich Chemie, Taufkirchen, Germany); benzaldehyde, γ-dodecalactone, hexanoic acid, 4-hydroxy-2,5-dimethylfuran3(2H)-one, 1-methoxy-4-(2-propenyl)benzene (Fluka, Sigma-Aldrich); 2-methylbutanal, 3-methylbutanal, 1-octen-3-one (Alfa Aesar, Karlsruhe, Germany); dimethyl trisulfide, nonanal (Acros Organics; Fisher Scientific, Nidderau, Germany); coumarin (Merck, Darmstadt, Germany); (E,E)-2,4-decadienal (Lancaster Synthesis, FrankfurtGriesheim, Germany); butanoic acid, 4-hydroxy-3-methoxybenzaldehyde (VWR, Darmstadt); 3-methyl-2,4-nonanedione (Chemos, Regenstauf, Germany); (E)-β-damascenone and (Z)-4-heptenal were kindly provided by Symrise (Holzminden, Germany). The following compounds were synthesized as previously reported: 2-acetyl-1-pyrroline11 and trans-4,5-epoxy-(E)-2-decenal.12 Dichloromethane, diethyl ether (both VWR), and pentane (Merck) were freshly distilled prior to use. Hydrochloric acid (37%), silica gel 60, and sodium carbonate were obtained from Merck; liquid nitrogen was from Linde (Munich, Germany). Stable Isotopically Labeled Internal Standards. [2H3]-Acetic acid, [2H3]-hexanoic acid, [2H9]-3-methylbutanoic acid, and [13C2]phenylacetic acid were purchased from Sigma-Aldrich Chemie. The following stable isotopically labeled internal standards were prepared as described previously: [2H2−5]-2-acetyl-1-pyrroline;13 [2H3]-2-aminoacetophenone;14 [2H5]-benzaldehyde;15 [13C4]-2,3-butanedione;16 [2H2−4]-butanoic acid;17 [2H3]-1,8-cineole;18 [13C2]coumarin;19 [2H4−7]-(E)-β-damascenone;20 [2H2−4]-(E,E)-2,4-decadienal;21 [2H4]-decanal;22 [2H6]-dimethyl trisulfide;23 [2H2]-γ-dodeca8389

DOI: 10.1021/acs.jafc.6b03676 J. Agric. Food Chem. 2016, 64, 8388−8396

Article

Journal of Agricultural and Food Chemistry

Table 1. Selected Ions (m/z) of Analytes and Stable Isotopically Labeled Standards as well as Response Factors (Rf) Used in Stable Isotope Dilution Assays ion (m/z)a odorant acetic acid 2-acetyl-1-pyrroline 4-allyl-2-methoxyphenold 2-aminoacetophenone benzaldehyde 2,3-butanedione butanoic acid 1,8-cineole coumarin (E)-β-damascenone (E,E)-2,4-decadienal decanal dimethyl trisulfide γ-dodecalactone trans-4,5-epoxy-(E)-2decenal 2-ethyl-3,5-dimethylpyrazine 2-ethyl-3,6dimethylpyrazinee 3-ethylphenol ethyl propanoate hexanal hexanoic acid 3-hydroxy-4,5dimethylfuran-2(5H)-one 4-hydroxy-2,5dimethylfuran-3(2H)-one 4-hydroxy-3methoxybenzaldehyde β-ionone 2-isobutyl-3methoxypyrazine 2-isopropyl-3methoxypyrazine 2-isopropyl-5-methylphenol 5-isopropyl-2-methylphenolf linalool 2-methoxyphenol

ion (m/z)a

isotope label

analyte

internal standard

Rf b

[2H3] [2H2−5]c −d [2H3] [2H5] [13C4] [2H2−4] [2H3] [13C2] [2H4−7]c [2H2−4]c [2H4] [2H6] [2H2] [2H2−3]c

61 112 165 136 107 87 89 137 147 191 153 157 127 199 139

64 114−117c 169−171c,d 139 112 91 91−93c 140 149 195−198c 155−157c 161 133 201 141−142c

0.99 0.91 1.01 0.98 0.86 1.00 0.96 0.85 0.99 0.83 0.96 0.92 0.99 0.76 0.48

[2H3] −e

137 137

140 140e

0.85 0.85

[2H3] [2H3] [2H4] [2H3] [13C2]

123 103 83 117 129

126 106 87 120 131

0.96 0.89 0.90 0.96 1.00

[13C2]

129

131

0.95

[2H3]

153

156

0.97

[2H3] [2H3]

193 167

196 170

0.96 0.98

[2H3]

153

156

0.98

[2H7] −f [2H2] [2H3]

151 151 137 125

158 158f 139 128

0.98 0.98 0.94 0.96

odorant 1-methoxy-4-(1-propenyl) benzeneg 1-methoxy-4-(2-propenyl) benzene 2-methylbutanalh 3-methylbutanal 2-methylbutanoic acidi 3-methylbutanoic acid 3-methyl-2,4-nonanedione 3-(methylthio)propanal (E,E)-2,4-nonadienal (E,Z)-2,6-nonadienal γ-nonalactone nonanal (E)-2-nonenal octanal 1-octen-3-one pentanoic acid phenylacetaldehyde phenylacetic acid

isotope label

analyte

internal standard

Rf b

−g

149

152g

0.78

2

[ H3]

149

152

0.78

−h [2H2] −i [2H9] [2H3] [2H3] [2H2] [2H2] [2H2] [2H4] [2H2] [2H2] [2H2−3]c [2H3] [13C2] [13C2]

87 87 103 103 171 105 139 139 157 143 141 129 127 103 121 137

89h 89 112i 112 174 108 141 141 159 147 143 131 129−130c 106 123 139

0.86 0.86 0.78 0.78 0.97 0.99 0.98 0.98 0.74 0.93 0.98 0.97 0.96 0.98 0.99 0.92

a

Ion used for quantitation in chemical ionization (CI) mode. Response factor (Rf) was determined by analyzing mixtures of known amounts of analyte and internal standard. cInternal standard was used as a mixture of isotopologues. dQuantitation of 4-allyl-2methoxyphenol (eugenol) was performed by using [2H2−4]-4-propyl2-methoxyphenol (dihydroeugenol) as internal standard. eQuantitation of 2-ethyl-3,6-dimethylpyrazine was performed by using [2H3]-2ethyl-3,5-dimethylpyrazine as internal standard. fQuantitation of 5isopropyl-2-methylphenol (carvacrol) was performed by using [2H7]2-isopropyl-5-methylphenol ([2H7]-thymol) as internal standard. g Quantitation of 1-methoxy-4-(1-propenyl)benzene (anethole) was performed by using [2H3]-1-methoxy-4-(2-propenyl)benzene ([2H3]estragole) as internal standard. hQuantitation of 2-methylbutanal was performed by using [2H2]-3-methylbutanal as internal standard. i Quantitation of 2-methylbutanoic acid was performed by using [2H9]-3-methylbutanoic acid as internal standard. b

technique at 40 °C. The initial oven temperature (40 °C) was held for 2 min, then raised at 6 °C/min to 230 °C, and held at the final temperature for 5 min. The effluent was split into two equal parts at the end of the column by means of a Y-type quick-seal glass splitter (Chrompack, Frankfurt, Germany) and two deactivated fused-silica capillaries of the same length (20 cm × 0.18 mm i.d.). One part was directed to a flame ionization detector (FID) held at 250 °C, and the other to a sniffing port held at 230 °C. In this way, a simultaneous detection of the FID signal using an analogue writer (Servogor SE 120; BBC Goerz Metrawatt, Nuremberg, Germany) and of the odor quality was achieved. Linear retention indices (RIs) of each compound were calculated using a series of n-alkanes (C6−C26 (DB-FFAP) and C6−C18 (DB5), respectively).47 High-Resolution Gas Chromatography−Mass Spectrometry for Identification. For the identification of the aroma compounds, a HRGC-MS system consisting of a HP 5890 Series II gas chromatograph (Hewlett-Packard, Heilbronn, Germany) and a sector field mass spectrometer MAT 95 S (Finnigan MAT, Bremen, Germany) was used. Mass spectra in EI mode were generated at 70 eV; CI mode was performed at 115 eV using isobutane as reactant gas.

High-Resolution Gas Chromatography−Mass Spectrometry (HRGC-MS) for Quantitation. For odorants present in higher concentrations, a HRGC-MS system consisting of a gas chromatograph Varian GC 431 (Darmstadt) coupled to an ion trap mass spectrometer Varian 220-MS running in CI mode (70 eV) with methanol as reactant gas was applied. The peak areas of the analyte and the isotopically labeled standard were determined separately by using the respective mass traces of the protonated molecular masses or selected fragments (Table 1). Two-Dimensional High-Resolution Gas Chromatography− Olfactometry/Mass Spectrometry (HRGC/HRGC-O/MS) for Identification and Two-Dimensional High-Resolution Gas Chromatography−Mass Spectrometry (HRGC/HRGC-MS) for Quantitation. The HRGC/HRGC-O/MS system consisted of a gas chromatograph Mega-2 (Fisons Instruments, Egelsbach) equipped with a DB-FFAP column in the first dimension coupled to a gas chromatograph CP-3800 (Varian) equipped with a DB-5 column in the second dimension (both 30 m × 0.25 mm i.d., 0.25 μm film thickness, J&W Scientific). The end of the second column was connected to an ion trap mass spectrometer Saturn 2000 (Varian), and, in parallel, to a sniffing port via a Y-splitter, enabling a simultaneous generation of mass spectra (recorded in EI mode, 70 eV) and the perception of the corresponding odor qualities of the 8390

DOI: 10.1021/acs.jafc.6b03676 J. Agric. Food Chem. 2016, 64, 8388−8396

Article

Journal of Agricultural and Food Chemistry Table 2. Important Aroma-Active Compounds (FD Factor ≥16) Identified in Raw Licorice (Glycyrrhiza glabra L.) retention index on no.a

odorantb

odor qualityc

DB-FFAP

DB-5

FDd

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 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

3-methylbutanal 2-methylbutanal ethyl propanoate 2,3-butanedione hexanal 1,8-cineole unknown (Z)-4-heptenal octanal 1-octen-3-one 2-acetyl-1-pyrroline dimethyl trisulfide nonanal 2-isopropyl-3-methoxypyrazine 2-ethyl-3,6-dimethylpyrazine 2-ethyl-3,5-dimethylpyrazine 3-(methylthio)propanal decanal 2-isobutyl-3-methoxypyrazinef benzaldehyde (E)-2-nonenal linalool (E,Z)-2,6-nonadienal butanoic acid phenylacetaldehyde 2-methylbutanoic acid 3-methylbutanoic acid (E,E)-2,4-nonadienal 3-methyl-2,4-nonanedione γ-hexalactone (E,E)-2,4-decadienal (E)-β-damascenonef 1-methoxy-4-(1-propenyl)benzene (anethole) hexanoic acid 2-methoxyphenol 2-phenylethanol unknown trans-4,5-epoxy-(E)-2-decenal γ-nonalactone 4-hydroxy-2,5-dimethylfuran-3(2H)-one 2-isopropyl-5-methylphenol (thymol) 5-isopropyl-2-methylphenol (carvacrol) 4-allyl-2-methoxyphenol (eugenol) 3-ethylphenol 3-hydroxy-4,5-dimethylfuran-2(5H)-onef 2-aminoacetophenone γ-dodecalactone coumarin phenylacetic acid 4-hydroxy-3-methoxybenzaldehyde

malty malty fruity butter-like grassy, green eucalyptus-like burnt fishy, fish oil-like citrus-like, green mushroom-like popcorn-like cabbage-like citrus-like, soapy earthy, pea-like earthy earthy cooked potato-like soapy, citrus-like bell pepper-like bitter almond-like, marzipan fatty, green citrus-like, flowery cucumber-like sweaty honey-like, beeswax-like sweaty sweaty fatty, green hay-like, aniseed-like coconut-like fatty, deep-fried baked apple-like, grape juice-like aniseed-like sweaty gammon-like, smoky honey-like, flowery wet tar, pungent metallic coconut-like caramel-like thyme-like thyme-like clove-like phenolic, leather-like seasoning-like foxy peach-like woodruff-like beeswax-like, honey-like vanilla-like

941 943 947 1000 1083 1191 1204 1229 1275 1292 1318 1353 1393 1411 1432 1434 1451 1482 1505 1524 1535 1538 1571 1620 1644 1660 1662 1688 1700 1708 1790 1805 1822 1832 1862 1903 1926 1994 2029 2033 2135 2135 2156 2181 2194 2209 2365 2428 2541 2583

652 653 734 640 809 1038 nd 896 1005 980 916 967 1100 1091 1080 1083 886 1196 1187 955 1162 1092 1150 805 1037 860 860 1216 1255 863 1330 1370 1280 1032 1090 1113 1270 1391 1359 1066 1290 1310 1372 1162 1196 1296 1684 1446 1281 1404

16 16 64 128 32 64 16 32 32 16 512 16 16 64 64 64 64 32 128 64 64 256 128 32 128 64 64 256 64 32 32 64 64 64 512 32 16 128 1024 1024 16 16 128 128 128 512 16 16 512 1024

lit.e

7 7 5 8

7 7 7 6 6 5 4 7 4 7 6 6 6

4 7 7 7 6 4 4

4 5 4 5 4

6 8 7

a

Odorants were consecutively numbered according to their retention indices on capillary DB-FFAP. bOdorant identified by comparison of its odor quality and intensity at the sniffing port and retention indices on capillaries DB-FFAP and DB-5 as well as mass spectra (EI and CI mode) with data of reference compounds. cOdor quality perceived at the sniffing port. dFlavor dilution factor determined by AEDA on capillary DB-FFAP. e Compound was first reported as volatile compound in Glycyrrhiza glabra L. or Glycyrrhiza uralensis Fisch. in the given reference. fNo unequivocal mass spectrum (EI mode) was obtained; identification was based on remaining criteria in footnote b. nd: not determined.

respective aroma-active compounds. The elution range containing the compounds of interest was transferred from the first GC column into a cold trap (cooled with liquid nitrogen to −100 °C) by a moving

column stream switching system (ThermoQuest). Then, the trap was immediately heated to 230 °C, and the aroma compounds were transferred onto the second GC column. Aliquots (1 μL) of the 8391

DOI: 10.1021/acs.jafc.6b03676 J. Agric. Food Chem. 2016, 64, 8388−8396

Article

Journal of Agricultural and Food Chemistry samples were manually injected at 40 °C using the cold on-column technique. For quantitation experiments, a TRACE GC 2000 coupled to a gas chromatograph CP-3800 and an ion trap Saturn 2000 mass spectrometer running in CI mode (70 eV) with methanol as reactant gas was used. Aliquots (1 μL) of the samples were injected by a Combi PAL autosampler (CTC Analytics, Zwingen, Switzerland) using the cold on-column technique at 40 °C. In both GC/GC-MS systems, the first GC was equipped with a DBFFAP capillary column and the second GC with an OV-1701 capillary column (both columns 30 m × 0.32 mm i.d., 0.25 μm film thickness; J&W Scientific). Quantitation by Stable Isotope Dilution Assays. Some of the stable isotopically labeled compounds used as internal standards for quantitation were synthesized in a microscale range. Thus, common purification processes could not be applied. First, the FID response factor was determined for each unlabeled reference compound using methyl octanoate as internal standard. Then, the concentration of the labeled standard was calculated via the peak areas of the labeled compound and methyl octanoate using the FID response factor determined for the unlabeled compound. For quantitation experiments, different amounts of raw licorice (1− 500 g, depending on the concentrations of the analyzed odorants determined in a preliminary experiment) were dissolved in dichloromethane (50−250 mL). Prior to the isolation of the volatiles as described above, defined amounts of the respective internal standards (0.5−5 μg; dissolved in diethyl ether; amounts depending on the concentrations of the respective analytes determined in a preliminary experiment) were added to the sample, and the mixture was stirred for 1 h at room temperature. After decanting of the solvent, another portion of dichloromethane was added, and the mixture was stirred again for 1 h. Both organic extracts were combined, filtered, and subjected to SAFE distillation.45 To calculate the response factor (Rf) of each aroma-active compound, binary mixtures of defined amounts of the unlabeled analyte and the respective labeled standard in five different mass ratios (5:1, 3:1, 1:1, 1:3, 1:5) were analyzed under the same conditions by (HRGC/)HRGC-MS (Table 1). Quantitation of 2- and 3-Methylbutanoic Acid. As both isomers could not be separated by HRGC-MS in CI mode, first, the sum of both compounds was determined by means of SIDA. Then, the sample was reanalyzed by HRGC-MS in EI mode (70 eV), and the ratio of 2- and 3-methylbutanoic acid was determined using the intensities of the fragments m/z 60 (3-methylbutanoic acid) and m/z 74 (2-methylbutanoic acid). Five defined mixtures of 2- and 3methylbutanoic acid (90:10; 70:30; 50:50; 30:70; 10:90) were analyzed, and a calibration curve was drawn plotting the intensity ratio of m/z 60 over the sum of m/z 60 + m/z 74 against the percentage of 3-methylbutanoic acid in the mixture.28 Sensory Experiments. Sensory analyses were performed in a sensory room at 21 ± 1 °C equipped with single booths. The samples (15 g or 15 mL, respectively) were presented in covered glass vessels (40 mm i.d., total volume = 45 mL). Orthonasal Odor Thresholds. For the calculation of odor activity values, orthonasal odor thresholds of all quantitated aroma compounds were determined in water as matrix by triangle tests evaluating the aroma compounds in decreasing concentrations against two blank samples. Determination was performed by a sensory panel consisting of 20 experienced assessors participating in weekly sensory sessions intended to train their abilities to recognize and describe different aroma qualities.48 Aroma Profile Analysis (APA). For APA, the intensities of the following selected odor attributes determined in a preliminary descriptive sensory experiment were rated on a seven-point linear scale from 0 (not perceivable) to 3 (strongly perceivable) in steps of 0.5 by the sensory panel. For each descriptor, an aqueous reference solution at a concentration 100-fold above the respective odor threshold of the odorant was provided: gammon-like/smoky (2methoxyphenol), cucumber-like ((E,Z)-2,6-nonadienal), eucalyptuslike (1,8-cineole), grassy/green (hexanal), aniseed-like (anethole),

citrus-like/flowery (linalool), thyme-like (2-isopropyl-5-methylphenol), and seasoning-like (3-hydroxy-4,5-dimethylfuran-2(5H)-one). Aroma Recombination Experiment. For aroma recombination, an aqueous model solution of all quantitated odorants with an OAV ≥1 was prepared by mixing the compounds in the naturally occurring concentrations determined in raw licorice. The recombinate and the original raw licorice were evaluated by the sensory panel as explained above for APA.



RESULTS AND DISCUSSION Identification of Key Odorants in Raw Licorice. To identify the most important aroma-active compounds in raw licorice, first, the volatiles were isolated by solvent extraction and SAFE distillation,45 revealing a distillate eliciting the typical aroma of raw licorice if put on a strip of filter paper. The most intense aroma-active regions in the gas chromatogram of the volatile fraction of raw licorice were located by AEDA, resulting

Figure 1. Flavor dilution (FD) chromatogram on a DB-FFAP capillary column obtained by aroma extract dilution analysis (AEDA) of an aroma distillate of raw licorice. Odorants with an FD factor ≥16 are illustrated. Numbering is identical to that in Table 2.

in 50 aroma-active compounds in the FD factor range between 16 and 1024 (Table 2 and Figure 1). Among them, 39 (coconut-like), 40 (caramel-like), and 50 (vanilla-like) showed the highest FD factor of 1024, followed by 11 (popcorn-like), 35 (gammon-like, smoky), 46 (foxy), and 49 (beeswax-like, honey-like) with an FD factor of 512 as well as by 22 (citruslike, flowery) and 28 (fatty, green) with an FD factor of 256. A comparison of the respective odor quality and intensity at the sniffing port as well as of retention indices on two capillary columns of different polarities with data of an in-house database containing >1000 aroma-active reference volatiles suggested structures of the odorants detected during AEDA. For an unequivocal identification, mass spectra obtained in EI and CI mode were compared to those of the reference substances of the suggested structures. Using this strategy, the odorants with the highest FD factors were identified as γ-nonalactone (39), 4-hydroxy-2,5-dimethylfuran-3(2H)-one (40), and 4-hydroxy-3-methoxybenzaldehyde (50; all FD factor of 1024), followed by 2-acetyl-1-pyrroline (11), 2-methoxyphenol (35), 2-aminoacetophenone (46), phenylacetic acid (49; all 512), linalool (22), and (E,E)-2,4nonadienal (28; both 256) (Table 2 and Figure 2). 8392

DOI: 10.1021/acs.jafc.6b03676 J. Agric. Food Chem. 2016, 64, 8388−8396

Article

Journal of Agricultural and Food Chemistry

Quantitation of Key Odorants in Raw Licorice and Calculation of Odor Activity Values. AEDA is used as a screening method to get a first idea of which single aroma compounds should most likely contribute to the overall aroma of a certain food. Thus, in the next step after the identification experiments, it is necessary to quantitate the odorants with high FD factors using the corresponding stable isotopically labeled standards (for quantitation of six compounds labeled standards with similar structures to the respective analytes were used) to get reliable concentrations needed for the calculation of OAVs (ratio of concentration to respective odor threshold). Quantitation by means of SIDA revealed acetic acid (223 mg/kg) and hexanoic acid (63.1 mg/kg) as the most abundant odorants in raw licorice, followed by hexanal (4.75 mg/kg), 1methoxy-4-(1-propenyl)benzene (anethole; 2.84 mg/kg), butanoic acid (1.57 mg/kg), γ-nonalactone (1.11 mg/kg), and linalool (1.06 mg/kg) (Table 3). Nine compounds were determined in concentrations between 100 and 600 μg/kg, among them 1-methoxy-4-(2-propenyl)benzene (estragole; 473 μg/kg), with a characteristic aniseed-like odor like its isomer anethole, and 5-isopropyl-2-methylphenol (carvacrol; 211 μg/ kg). Concentrations between 10 and 100 μg/kg were obtained for 18 compounds, for example, 2-isopropyl-5-methylphenol (thymol; 55.5 μg/kg), with a thyme-like odor impression like its isomer carvacrol. 3-Hydroxy-4,5-dimethylfuran-2(5H)-one

Figure 2. Structures of aroma-active compounds identified with high FD factors in raw licorice (FD factors and odor descriptions in parentheses; numbering refers to Table 2.

In summary, 48 of the 50 substances located in the volatile fraction of raw licorice during AEDA with an FD factor ≥16 were identified, among them 16 in raw licorice for the first time (Table 2).

Table 3. Concentrations, Orthonasal Odor Thresholds, and Odor Activity Values (OAVs) of Key Aroma Compounds (OAV ≥1) in Raw Licorice (Glycyrrhiza glabra L.) odorant (E,Z)-2,6-nonadienal 5-isopropyl-2-methylphenol (carvacrol) hexanal linalool (E,E)-2,4-decadienal (E,E)-2,4-nonadienal 2-isopropyl-5-methylphenol (thymol) 2-isopropyl-3methoxypyrazine 2-isobutyl-3methoxypyrazine 1-methoxy-4-(1-propenyl) benzene (anethole) 4-allyl-2-methoxyphenol (eugenol) 2-acetyl-1-pyrroline (E)-2-nonenal γ-nonalactone 1,8-cineole 1-methoxy-4-(2-propenyl) benzene (estragole) 3-ethylphenol 3-methylbutanal nonanal 1-octen-3-one 2,3-butanedione 2-methoxyphenol 2-ethyl-3,5-dimethylpyrazine 3-methyl-2,4-nonanedione 2-methylbutanal (E)-β-damascenone

concentrationa (μg/kg)

odor threshold in water (μg/L) c

OAVb

odorant

6820 6390

trans-4,5-epoxy-(E)-2decenal hexanoic acid γ-dodecalactone octanal β-ionone 4-hydroxy-3methoxybenzaldehyde decanal phenylacetic acid 3-hydroxy-4,5dimethylfuran-2(5H)-one phenylacetaldehyde acetic acid 3-(methylthio)propanal dimethyl trisulfide 2-ethyl-3,6-dimethylpyrazine butanoic acid benzaldehyde ethyl propanoate 4-hydroxy-2,5dimethylfuran-3(2H)-one coumarin 3-methylbutanoic acid 2-aminoacetophenone 2-methylbutanoic acid pentanoic acid

30.7 211

0.0045 0.033

4750 1060 21.6 35.5 55.5

2.4c 0.58 0.027c 0.046 0.08

1980 1830 800 771 694

2.17

0.0039c

556

1.72

0.0062c

277

2840 320

15c 1.8

189 178

9.41d 27.3 1110 96.9 473

0.053c 0.19c 9.7c 1.1c 6c

178 144 114 88 79

51.8d 29.4d 145d 0.73 44.9 29.7 9.22d 1.18 33.9 0.20

0.85c 0.5c 2.8c 0.016c 1.0c 0.84c 0.28 0.046 1.5c 0.013c

61 59 52 46 45 35 33 26 23 15

concentrationa (μg/kg) 3.33

odor threshold in water (μg/L) 0.22

OAVb 15

63100d 5.53 37.2 30.7 421

4800 0.43c 3.4c 3.5c 53c

13 13 11 9 8

53.4 376 2.33

9.3 68 0.49c

6 6 5

15.8 223000 0.84 0.01 24.4 1570 91.2 6.93 24.1 4.40 190 0.05d 174 580d

5.2 99000c 0.43c 0.0099c 25 2400c 150 14 54

3 2 2 1