Centrifugal partition chromatography applied to the isolation of oak wood aroma precursors

Food Chemistry 141 (2013) 2238–2245

Contents lists available at SciVerse ScienceDirect

Food Chemistry journal homepage: www.elsevier.com/locate/foodchem

Centrifugal partition chromatography applied to the isolation of oak wood aroma precursors Davide Slaghenaufi a,b,⇑, Stéphanie Marchand-Marion a,b, Tristan Richard c, Pierre Waffo-Teguo c, Jonathan Bisson c, Jean-Pierre Monti c, Jean-Michel Merillon c, Gilles de Revel a,b,⇑ a b c

Univ. de Bordeaux, ISVV, EA 4577, Unité de Recherche Œnologie, 33882 Villenave d’Ornon, France INRA, ISVV, USC 3666 oenologie, 33882 Villenave d’Ornon, France Univ. de Bordeaux, ISVV, EA 3675 GESVAB, 33882 Villenave d’Ornon, France

a r t i c l e

i n f o

Article history: Received 10 November 2012 Received in revised form 15 April 2013 Accepted 21 April 2013 Available online 2 May 2013 Keywords: Oak wood Centrifugal partition chromatography Glucoside gallate precursors Vanillin 3,4,5-Trimethoxyphenyl 3-Oxo-a-ionol

a b s t r a c t Flavours extracted from oak wood during barrel ageing contribute to the organoleptic character of wines and spirits. The aim of this work was to identify the glycosidic precursors of the key volatile compounds responsible for oak wood aroma. Oak extract is a very complex matrix and, furthermore, precursors are present in very small quantities. Preparative centrifugal partition chromatography (CPC) is a promising solution for purifying the oak extract. The solvent system was selected on the basis of the partition coefficient of glycosidase enzyme activity (Kca). Thanks to the efficacy of CPC separation, three glucoside gallates were subsequently isolated by HPLC chromatography. Vanillin-(60 -O-galloyl)-O-b-D-glucopyranoside, 3,4,5-trimethoxyphenyl-(60 -O-galloyl)-O-b-D-glucopyranoside, and (6R,9R)-3-oxo-a-ionol-9-O-(60 O-galloyl)-b-glucopyranoside (macarangioside E) were isolated and identified. This was the first time that vanillin-(60 -O-galloyl)-O-b-D-glucopyranoside was identified and the first time that macarangioside E was isolated from oak wood. Heating macarangioside E resulted in the formation of megastigmatrienone, which has an aroma reminiscent of tobacco. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction Wines and alcohol beverages are traditionally aged in wooden barrels, most commonly made of oak. During this phase, volatile oak compounds are extracted and add woody notes to the product. The free volatile compounds in oak responsible for these woody aromas have been extensively studied. Vanillin, oak lactone, and volatile phenols play a prominent role in the aromas of wood-aged wines (Boidron, Chatonnet, & Pons, 1988; Chatonnet, Boidron, & Pons, 1990; Chatonnet, Dubourdieu, & Boidron, 1992; Jarauta, Cacho, & Ferreira, 2005; Pocock, Sefton, & Williams, 1994; Spillman, Pollnitz, Liacopoulos, Skouroumounis, & Sefton, 1997). However, there has been little research into bound volatile compounds in oak wood. It is well known that during wood seasoning and cooperage processes (Chatonnet, 1995; Jarauta, Cacho, & Ferreira, 2005; Maga, 1985; Nishimura, Onishi, Masuda, Koga, & Matsuyama, 1983; Petruzzi et al., 2010; Roulland, Snakkers, & Cantagrel, 1999; Sarni, Moutounet, Puech, & Rabier, 1990; Sefton, Francis, Pocock, & Williams, 1993), the concentrations of several flavours are increased and others are produced by heat treatment. Lignin ⇑ Corresponding authors at: Univ. de Bordeaux, ISVV, EA 4577, Unité de Recherche Œnologie, 33882 Villenave d’Ornon, France. Tel.: +33 557 575 863; fax: +33 557 575 803. E-mail address: [email protected] (G. de Revel). 0308-8146/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.foodchem.2013.04.069

depolymerisation is considered to cause an increase in phenolic aldehydes during toasting. However, it has been demonstrated that vanillin and other phenolic aldehydes also increase during malolactic fermentation in barrel (de Revel, Bloem, Augustin, Lonvaud-Funel, & Bertrand, 2005). Bacteria glycosidase activity was proved to be involved in this phenomenon, suggesting the existence of glycosidic flavour precursors in oak wood (Bloem, Lonvaud-Funel, & de Revel, 2008). In fact, two glycosidic oak-lactone precursors had previously been isolated from oak wood: (3S,4S)3-methyl-4-O-(6’-O-galloyl)-b-D-glucopyranosyloctanoic acid (galloylglucoside) (Masson, Baumes, La Guernevé, & Puech, 2000) and (3S,4R)-3-methyl-4-O-b-D-glucopyranosyloctanoic acid (Hayasaka, Wilkinson, Elsey, Raunkjær, & Sefton, 2007). Other oak aroma precursors were investigated by GC–MS analysis of trifluoroacetylated oak extract (Nonier, Vivas de Gaulejac, Vivas, & Vitry, 2005). The authors postulated that vanillin, syringaldehyde, and megastigmatrienone were directly linked to sugar, but identification was not confirmed. The aim of this work was to isolate and identify in oak wood the glycosidic precursors of a powerful aroma like vanillin. We also looked for the precursors of 3-oxo-a-ionol and megastigmatrienon. These molecules are involved in the tobacco aroma of oak wood. In a preliminary step, oak extract was purified by centrifugal partition chromatography (CPC) and semipreparative HPLC,

D. Slaghenaufi et al. / Food Chemistry 141 (2013) 2238–2245

resulting in three purified molecules, identified by ESI-MS and NMR spectroscopy as vanillin-(60 -O-galloyl)-O-b-glucopyranoside; 3,4,5-trimethoxyphenyl-(60 -O-galloyl)-O-b-glucopyranoside, and (6R,9R)-3-oxo-a-ionol-9-O-(60 -O-galloyl)-b-glucopyranoside (macarangioside E). CPC is widely used to isolate natural substances but has rarely been used in wine and grape research. Some recent papers report interesting applications of this chromatography technique in enology research (Amira-Guebailia et al., 2009; Bisson et al., 2011; Delaunay, Castagnino, Chèze, & Vercauteren, 2002; Marchal, Waffo-Téguo, Génin, Mérillon, & Dubourdieu, 2011).

2. Materials and method 2.1. Chemicals and reagents Pure water for HPLC analysis was obtained using an Elga (Elga Process Water Ltd., Marlow, UK) water purification system with a resistivity of at least 18 MX cm1. Formic acid, vanillin, 3,4,5-trimethoxyphenol, eugenol, and 2-dodecanol were purchased from Sigma–Aldrich (St. Louis, MO) at the highest purity available. Megastigmatrienone was kindly provided by Symrise AG (Holzminden, Germany). Ethyl acetate (EtOAc) and dichloromethane (CH2Cl2) were purchased from Fisher Scientific (Loughborough, UK). Acetonitrile (ACN), acetone and cyclohexane were purchased from VWR (Fontenay-sous Bois, France). Methanol-d4 and acetone-d6 for NMR analysis were purchased from Euriso-top (Gif-sur-Yvette, France).

2.2. Extraction of oak aroma precursors from wood Dried French oak wood (1.5 kg) was cut into small pieces (maximum length <2 cm) and macerated three times under agitation in 2 L aqueous acetone (1:1) solution for 24 h. Extracts were pooled and evaporated under reduced pressure (T <30 °C). The remaining extract was separated by liquid/liquid extraction in 3 stages, using cyclohexane (3  300 mL), CH2Cl2 (5  300 mL), and EtOAc (5  300 mL). The combined extract from each phase was freezedried.

2.3. Enzymatic hydrolysis of oak wood aroma precursors From each CPC and semiprep-HPLC fraction an aliquot of 500 lL was taken and added with 0.8 mL of sodium acetate buffer (pH 5.0, 10 mM), and 200 lL of a glycosidase rich enzyme preparation (AR2000, Delft, Netherlands; 70 mg/mL in sodium acetate buffer); the mixture was incubated at 27 °C for 24 h. 2-Dodecanol (10 lL at 410 mg/L) was added to the mixture as an internal standard. The released aglycones were extracted once with 2 mL CH2Cl2 (Silva, Gunata, Lepoutre, & Odoux, 2011).

2.4. CPC purification of wood extract 2.4.1. CPC apparatus Centrifugal partition chromatography (CPC) was performed at room temperature on an FCPC 1000 apparatus from KromatonÒ Technologies (Angers, France). The rotor consisted of 45 stainless steel disks containing 1440 cells, with a total rotor volume of 940 mL. A Gilson 321-H1 2-way binary high-pressure gradient pump was used. Injection was via a RheodyneÒ injection valve with a 50 mL sample loop. Effluent was monitored using a Kromaton UV–vis detector (Angers, France) and fractions were collected with an Advantec CHF 122SC fraction collector (Advantec MFS Inc., Dublin, CA).

2239

2.4.2. CPC solvent system selection A 50-mg sample of EtOAc oak extract was dissolved in the biphasic system (2 mL from the upper layer and 2 mL from the lower layer) and shaken vigorously. An aliquot (0.9 mL) of each phase was evaporated and reconstituted with 0.2 mL water/MeOH (9:1) mixture and subjected to enzymatic hydrolysis, as described at Section 2.3, followed by GC–MS analysis. The activity partition coefficient (Kca) was calculated as the ratio of vanillin, oak lactone, and 3-oxo-a-ionol released in each phase.

K Xca ¼

AXstar AXmob

ð1Þ

K Xca : Activity partition coefficient of X (aglycone) between the two CPC phases AXphase : Quantity (Arbitrary unit) aglycone (X) realeased in each CPC phase Ternary system EtOAc/propan-2-ol/H2O (Köhler, Wray, & Winterhalter, 2008) in several proportions and Arizona systems B, F, K (Foucault, 1994; Foucault & Chevolot, 1998) were tested. The water/propan-2-ol/ethyl acetate (40/1/40) system (Köhler, Wray, & Winterhalter, 2008) was selected on the basis of equal distribution of released volatile compounds between the upper and lower layers, corresponding to a Kca close to 1. 2.4.3. CPC conditions and procedures CPC conditions were optimised by the authors. A mixture of EtOAc/propan-2-ol/H2O (40:1:40, v/v/v) was shaken vigorously in a separating funnel and left to stand at room temperature for about 1 h. The lower phase was used as the stationary phase and the upper as the mobile phase at a flow rate of 20 mL/min. Rotation speed was set to 1000 rpm. Eight grams EtOAc oak extract were dissolved in 40 mL of stationary phase, filtered and injected. Detection was carried out at 254, 280 and 313 nm. Fractions were collected at 1-min intervals. 2.5. HPLC fractionation of CPC primary fraction The HPLC purification was developed by the authors. HPLC fractionation was performed using a Waters 600 system (St-Quentin Yvelines, France) equipped with a Waters 600 controller, a RheodyneÒ injection valve with a 600-lL sample loop, and a Waters 2487 UV detector at 280 and 313 nm. A Nucleodur C18ec 250  10 mm, 5 lm column was used Macherey–Nagel GmbH & Co. KG, Düren, Germany. Chromatographic conditions: flow rate 2 mL/min; solvent A: H2O and 0.5% HCOOH; solvent B: ACN. Eluent was collected using a Foxy Jr. collector from Teledyne Isco (Lincoln, NE). 2.6. HPLC–ESI-MS apparatus HPLC–ESI-MS method was optimised by the authors. The Agilent 1200 chromatography system from Agilent Technologies (Santa Clara, CA) consisted of an autosampler module, a degasser, a binary pump, a column heater/selector, and a diode array detector. The column was a Nucleodur C18 ec 125 mm  4.6 mm, 5 lm. Fractions and library compounds were eluted at 0.8 mL/min with a gradient of water-0.1% formic acid (solvent A) and acetonitrile0.1% formic acid (solvent B), according to the following gradient program (v/v): 0 min 17% B linear, 5 min 17% B, 25 min 30% B, 35 min 38% B, 45 min 100% B linear for 10 min, followed by 10 min for rebalancing. This HPLC was coupled to an Esquire 3000 + ion trap mass spectrometer using an ESI source from Bruker Daltonics (Billerica, MA). The HPLC output flow of 0.8 mL/min was split using a passive splitter with an average 1:100 ratio, depending on the flow solvent viscosity and rate. Drying gas (N2) was

2240

D. Slaghenaufi et al. / Food Chemistry 141 (2013) 2238–2245

set to 5.0 L/min and 325 °C, nebuliser pressure was set to 15 psi. ESI–MS parameters (positive mode): HV capillary 3700 V, capillary exit 139 V, skimmer 40 V, and trap drive 46 V.

2.7. GC–MS oak wood aroma (aglycone) analysis Samples were prepared for GC–MS analysis (de Revel et al., 2005) by adding 2-dodecanol as an internal standard (10 lL at 410 mg/L) to each 1.5-mL sample and extracting once with 2 mL CH2Cl2. The resulting emulsion was dispersed by centrifugation (5000 rpm for 15 min). The organic extracts were dried over anhydrous sodium sulfate. Extracts were analysed using a PerkinElmer AutoSystem XL gas chromatograph (PerkinElmer, Waltham, MA) coupled to a PerkinElmer Turbomass Gold mass spectrometer (electronic impact = 70 eV). Volumes of 2 lL were injected in splitless mode; the splitless time was 30 s with a split flow of 20 mL min1. Helium 5.6 was used as a carrier gas with a column flow rate of 2 ml/ min. Separation was achieved using a BP 21 40 m  0.32 mm silica capillary column, with 0.25 lm film thickness (SGE, Courtaboeuf, France). The oven temperature was initially set at 60 °C, then increased to 220 °C at 3 °C/min and held at this temperature for 40 min. The injector was held at 250 °C, the transfer line at 200 °C and the MS source at 180 °C. Volatiles were identified by comparing their mass spectra with reference compounds, and quantified by comparing the areas of their chromatographic peaks with the one corresponding to the internal standard (2-dodecanol).

3.1. CPC solvent system selection A new approach for selection of the solvent system was developed in order to optimise CPC purification of bound aroma compounds. Usually, the CPC solvent system is selected on the basis of the partitioning constant (Kc) value of the target molecule. In our case, the target molecules (aroma precursors) are unknown; as a consequence the Kc values cannot be determined by direct analysis. In this method, an apparent partition coefficient, the enzyme activity partition coefficient (Kca) is used. It is quotient of the quantities of aglycones (comparison of chromatographic areas of aglycone peaks) released by the AR2000 activity in each CPC phase. A Kca value close to one indicates an equal distribution of volatile compounds in both upper and lower phases. Commercial oenological enzymes were used for their wider hydrolysis activities capable of releasing aglycones from different types of bound moieties. The CPC solvent system was selected on the basis of equal distribution of vanillin, 3,4,5-triméthoxyphenol and 3-oxo-a-ionol precursors in both phases. To achieve good separation a partition coefficient of 1 is needed, that corresponds to the elution of one column volume (Berthod & Carda-Broch, 2004). Released aroma distribution in the two phases was estimated by measuring the volatile compounds released in each phase by enzyme hydrolysis. Theoretical Kca of 1 for 3-oxo-a-ionol precursor and 1.3 for vanillin precursor, were obtained using the EtOAc/isoPrOH/water (40:1:40) solvent system as shown in Table 1, this means that 3-oxo-a-ionol and vanillin precursors should be eluted from CPC after one column volume.

2.8. NMR experiments NMR acquisitions were performed on a Bruker Avance III 600 MHz spectrometer (Rheinstetten, Germany). NMR spectra were recorded at 300 K, in acetone-d6 or methanol-d4 solvents. Spectra were referenced to the signals of acetone-d6 (dH 2.06 and dC 29.9 ppm) or methanol-d4 (dH 3.31 and dC 49.0 ppm). Identification was based on the mean of 2D-NMR experiments: COSY, ROESY, HSQC and HMBC. 2.9. Pyrolysis of (6R,9R)-3-oxo-a-ionol-9-O-(60 -Ogalloyl)glucopyranoside (macarangioside E) An aliquot of macarangioside E (0.2 mg) was put in a vial and sealed. The vial was then heated to 230 °C for 30 min, equivalent to heavy toasting in a barrel (Wilkinson, Elsey, Prager, Tanaka, & Sefton, 2004). After cooling, the vial was opened and 1 mL CH2Cl2 was added and agitated vigorously. A small amount of sodium sulfate was added. The organic layer was analysed by GC–MS.

3. Results and discussion The aim of this study was to identify the precursors of vanillin, 3,4,5-trimethoxyphenol and 3-oxo-a-ionol glycosides in oak wood. As oak is a very complex matrix and bound aromas are assumed to be present in very small quantities, a high-recovery purification technique is required. Liquid–liquid chromatography has been chosen as there is no sample loss by adsorption on the stationary phase and grams of sample can be purified with a single injection without stationary phase saturation. Droplet countercurrent chromatography (DCCC) has already been applied in the purification of conjugated forms of aroma compounds in grape juice and wine (Strauss, Gooley, Wilson, & Williams, 1987; Winterhalter, Sefton, & Williams, 1990). However CPC reduced separation time significantly compared to DCCC.

3.2. Isolation and identification of three bound volatile compounds from oak wood extract The aqueous acetone extract was partitioned with cyclohexane (1.3 g), CH2Cl2 (2.3 g) and EtOAc (7.9 g) in turn. The EtOAc extract was subjected to CPC fractionation. Nonier et al. (2005) reported that the EtOAc extract is rich in aroma precursors. We have monitored the release of vanillin, 3,4,5-trimethoxyphenol and 3-oxoa-ionol precursors in EtOAc extract by enzymatic hydrolysis followed by GC–MS analysis of the liberated aglycones.

Table 1 Solvent systems tested for the isolation of vanillin precursors. Kca values for released 3,4,5triméthoxyphenol

Kca values for released 3-oxo-aionol

No.

Solvent system

Volume ratios

Kca values for released vanillin

I

EtOAc/ isoPrOH/ water EtOAc/ isoPrOH/ water EtOAc/ 1-BuOH/ water Hep/ EtOAc/ MeOH/ water Hep/ EtOAc/ MeOH/ water Hep/ EtOAc/ MeOH/ water

40:1:40

1.37

1.19

1.01

20:1:20

2.03

1.18

1.89

14:1:15

2.36

615.76

2.78

1:19:1:19

1.36

0.00

0.77

1:5:1:5

0.99

946.71

0.13

1:2:1:2

0.82

698.03

0.53

II

III

B

F

K

D. Slaghenaufi et al. / Food Chemistry 141 (2013) 2238–2245

2241

Fig. 1. HPLC–UV (280 nm) chromatograms of purified compounds.

EtOAc extract was injected into the selected CPC system, using the lower phase as the stationary phase and collecting the mobile phase every minute. Each tube was analysed for the concentrations of aglycones released by enzyme activity. As expected, vanillin, 3,4,5-trimethoxyphenol, and 3-oxo-a-ionol were released from tubes 41 to 57, corresponding approximately to the elution volume of one column. The seventeen fractions (41–57) were combined and freeze-dried, producing 412 mg of pale brown powder. The extract was further purified by semipreparative HPLC on a C18 column. Three compounds were isolated (Fig. 1). They were suggested as vanillin precursor (0.2 mg; 0.1‰ of dried oak wood), 3,4,5-trimethoxyphenol precursor (1.8 mg; 1.2‰ of dried oak wood), and 3-oxo-a-ionol precursor (0.8 mg; 0.5‰ of dried oak wood) based on the aglycones released by enzymatic hydrolysis (Fig. 2). In negative-ion ESI-TOF mass spectrometry, the vanillin precursor exhibited a peak [MH] at m/z 465.1054 (calculated for

C21H21O12: m/z 465.1038). Fragmentation of the [MH] ion in negative ESI-MS2 gave a galloyl unit (Cuyckens & Claeys, 2004) at m/z 169 and the neutral release of gallate [M153] at m/z 312.9; which then produced a vanillin ion at m/z 151.1 by neutral release of a hexose sugar moiety [M–153162] (Fig. 3). The 1H NMR (Table 2) spectrum displayed seven protons due to 4-hydroxy-3-methoxybenzaldehyde (vanillin): an aldehyde proton at 9.83 ppm (H-7), three aromatic protons at d 7.41 (d, J = 1.9 Hz, H-2), 7.38 (d, J = 8.3 Hz, H-5), and 7.51 ppm (dd, J = 8.3 Hz, 1.9 Hz, H-6) and the three protons of a methoxy group at d 3.88 ppm (3H, s). The glucoside moiety was identified by typical 1H and 13C NMR data. The compound exhibited 13C NMR signals (Table 3) at d 101.1 (C-10 ), 74.9 (C-20 ), 77.7 (C-30 ), 71.8 (C40 ), 75.2 (C-50 ), and 64.7 ppm (C-60 ) in addition to an anomeric proton signal in the 1 H NMR spectrum at d 5.21 ppm (d, J = 7.1 Hz, H-10 ) and the other proton signals of the glucoside between d 3.53 and 4.67 ppm. The sugar stereochemistry was determined as b-glucopyranose by the

2242

D. Slaghenaufi et al. / Food Chemistry 141 (2013) 2238–2245 1: Scan EI+ 6.30e7

151

100

%

109

81 62 63

0

65

67 69

63

77 79

74 75

68

73

83 85

78

91 93 95

89

83

88

93

98

103

123 119121

110

103 105 108

113

118

136 138

125

123

128

133

153

149

138

143

m/z

148

153 1: Scan EI+ 1.21e7

169

100

184 141

% 69

126

111 109

62 63

0 59

65

67

64

77 80 81 8385

71 73

69

74

79

84

89

93 95

89

94

9698

101

99

124 113 117 121

106

104

109

114

119

127

124

133

129

138

134

139

168

151 153 155

143 144

149

154

171

159

164

169

186 174

179

184

m/z 189

1: Scan EI+ 6.14e5

108

100

%

61 62

0

91 77 79 81 95 133135 137 69 71 103 107 85 89 110 115 119 123 128 99 142

67

60

70

80

90

100

110

120

130

149 152

140

150

158 163

165

160

204 207

172 175

170

180

190

200

210

219

1: Scan EI+ 7.26e5

147

100

m/z

220

133 %

105

91

69 65 67

63 0 54

119

77 79

64

74

83 85 89

74

84

93 94 103 94

104

175

190

122

107 115

131 134

110 114

124

134

145 144

149

161 165

154

164

182187 174

184

191 194

215 204

214

219

228

224

263 270

241 234

244

254

264

274

285 283 284

1: Scan EI+ 4.73e5

68

100

m/z 294

96

% 152

63 66

0 53

63

72 77 81 73

84 86

83

9194 93

97 105 109 103

116

113

150

124 131 137 123

133

143

153

153

176 163

173

207 183

193

203

219 213

Fig. 2. Mass spectrum of the released aglycones.

223

254 233

243

253

284 263

273

283

m/z 293

2243

D. Slaghenaufi et al. / Food Chemistry 141 (2013) 2238–2245 Intens. x10 4

-M S2(465.2), 21.4m in #702 312.9

3

2

1 169.1 326.9

151.1 0 100

150

200

250

300

350

400

450 m /z

Fig. 3. HPLC–ESI-MS spectra of vanillin-(6-O-galloyl)-b-glucopyranoside.

Table 2 H NMR data for compounds VGG, TMPGG and macarangioside E.

1

Proton

VGGa

TMPGGb

macarangioside Eb

1 2

– 7.41 (d, J = 1.9 Hz)

– 6.39 (s)

3 4 5 6 7 8 9 10 11 12 13 OCH3 OCH3 Glucose 10 20 30 40 50 60

– – 7.38 7.51 9.83 – – – – – – 3.88 –

– – – – – – – – – – – 3.66 (s) 3.67 (s)

– 2.32 1.94 – 5.79 – 2.40 5.54 5.69 4.29 1.28 0.87 0.86 1.82 – –

4.83 (d, J = 7.9 Hz) 3.45 (m) 3.50 (m) 3.43 (m) 3.76 (m) 4.62 (dd, J = 12.1, 2.1 Hz) 4.42 (dd, J = 12.1, 6.8 Hz)

4.37 3.21 3.36 3.34 3.51 4.58 4.29

– 7.05 (s) – – –

– 7.10 (s) – – –

Galloyl 100 200 , 600 300 , 500 400 C@O a b

5.21 3.60 3.60 3.53 3.96 4.67 4.32

(d, J = 8.3 Hz) (dd, J = 8.3 Hz, 1.9 Hz) (s)

(s)

(d, J = 7.1 Hz) (m) (m) (m) (m) (dd, J = 12.1, 1.9 Hz) (J = 12.1, 7.1 Hz)

– 7.15 (s) – – –

(d, J = 16.6 Hz) (d, J = 16.6 Hz) (s) (d, J = 9.0 Hz) (dd, J = 15.4, 9.0 Hz) (dd, J = 15.4, 7.0 Hz) (m) (d, J = 6.6 Hz) (3H, s) (3H, s) (3H, br s)

(d, J = 7.9 Hz) (dd, J = 8.8, 7.9 Hz) (m) (m) (m) (dd, J = 11.8, 2.1 Hz) (dd, J = 11.8, 7.0 Hz)

In acetone-d6. In methanol-d4.

coupling constant of the anomeric proton (J = 7.1 Hz). The ROESY data indicated a linkage between the glucoside and the vanillin. The 1H NMR spectrum further exhibited a singlet peak at d 7.15 ppm (s, 2H, H-200 and H-600 ) that, together with the carbon atom signals on the 13C NMR spectrum at d 121.4 (C-100 ), 109.9 (C-200 , C-600 ), 146.6 (C-300 , C-500 ), and 138.4 ppm (C-400 ), and the carbonyl at d 166.7 ppm, indicated the presence of a galloyl moiety. The esterification of the C-60 hydroxyl by gallic acid was proved by both the downfield shift of H-60 and the correlation of gallate aromatic protons H-200 , H-600 with the H-60 glucose protons in the ROESY experiment. The vanillin precursor was thus identified as vanillin-(60 -O-galloyl)-O-b-glucopyranoside (VGG; Fig. 4). The 3,4,5-trimethoxyphenol precursor showed a sodiated molecular ion at m/z 521.1274 [M+Na]+ (calculated for C22H26O13 Na, m/z: 521.1271) on positive-ion HR-ESI-TOF-MS. MS/MS fragmentation identified a galloyl group linked to a hexose. The

3,4,5-trimethoxyphenol was identified by NMR analysis (COSY, HSQC, HMBC) as 3,4,5-trimethoxyphenyl-(60 -O-galloyl)-O-b-glucopyranoside (TMPGG; Fig. 4), by comparing the observed data with those reported in the literature (Verotta, Dell’Agli, Giolito, Cabalion, & Bosisio, 2001). Ishimaru, Nonaka, and Nishioka (1987) reported the isolation of 3,4,5-trimethoxyphenyl-(60 -O-galloyl)-Ob-glucopyranoside from Quercus acutissima. No spectroscopic or physicochemical properties were reported, as the authors referred to data reported in the literature (Nonaka, Nishimura, & Nishioka, 1982), describing the isolation and structural elucidation of a different compound, 2,4,6-trimethoxyphenyl-1-O-b-D-glucopyranoside. The spectroscopy data were reported for the first time by Verotta, Dell’Agli, Giolito, Cabalion, and Bosisio (2001), who isolated TMPGG from Tristaniopsis calobuxus. The 3-oxo-a-ionol precursor showed a pseudomolecular ion peak at m/z 545.2020 [M+Na]+ in positive-ion HR-ESIMS,

2244

D. Slaghenaufi et al. / Food Chemistry 141 (2013) 2238–2245 OH

Table 3 C NMR data for compounds VGG, TMPGG and macarangioside E.

13

a b

O

Carbon

VGGa

TMPGGb

macarangioside Eb

1 2 3 4 5 6 7 8 9 10 11 12 13 OCH3 OCH3 Glucopyranosyl 10 20 30 40 50 60 Galloyl 100 200 , 600 300 , 500 400 C@O

152.7 110.4 150.1 152.7 115.9 125.2 190.7 – – – – – – 56.2 –

153.8 96.1 154.3 134.2 154.3 96.1 – – – – – – – 56.4 60.8

36.5 48.1 201.7 125.7 165.5 56.0 129.3 137.6 77.8 21.2 27.2 27.4 23.4 – –

101.1 74.9 77.7 71.8 75.2 64.7

103.0 74.6 77.4 71.4 75.4 64.1

103.0 74.9 77.7 71.8 75.2 64.7

121.4 109.9 146.6 138.4 166.7

121.1 110.1 146.3 139.7 168.8

121.2 109.9 146.6 139.6 168.2

3-oxo-α-ionol

A new method for selecting the CPC solvent system has been developed, based on biological activity. This tool, applied to CPC, represents a powerful technique for purifying aroma precursors from a very complex matrix. In fact, three compounds were isolated by this method, including vanillin-(6-O-galloyl)-b-glucopyranoside, which was isolated and identified for the first time. Macarangioside E was described for the first time in oak wood

OH

4''

HO

OH

OH

4''

HO

1''

1'' O

6'

4. Conclusion

OH

OH

4''

6' O

1

O

HO OH

7

2

O

2

O

HO O

HO

O

O

OCH 3

OH

1'

OH

1'' O

O

HO

ketoisophorone

The aroma properties of 3-oxo-a-ionol are very weak, but it is well known that, under acid conditions, it may form megastigmatrienone, a key flavour constituent in Burley tobacco (Aasen, Kimland, Almquist, & Enzell, 1972; Lloyd et al., 1976; Ohloff, 1978). Megastigmatrienone has been patented as a flavour additive in the food, tobacco, and perfume industries. However, previous research (Sefton, Francis, & Williams, 1990) demonstrated that wine pH was too high for oak wood 3-oxo-a-ionol to dehydrate and form megastigmatrienone. It has been suggested that megastigmatrienone is formed during the cooperage process. The effect of heat treatment on macarangioside E during barrel production was investigated by pyrolysis. A small amount of macarangioside E was dried and heated to 230 °C for 30 min. The major degradation products, shown in Fig. 5, were megastigmatrienone, 2,6,6-trimethyl-2-cyclohexene-1,4-dione (ketoisophorone), and 3-oxo-a-ionol, representing 81%, 6%, and 12% of the volatile compounds formed, respectively. Ketoisophorone was previously identified in fresh oak wood (Sefton et al., 1990) as well as in wines as an important sweet aroma (Rogerson et al., 2001). Other unidentified compounds were formed by pyrolysis. Only four of the five possible isomers were observed. The formation of megastigmatrienone by heat degradation of macarangioside E was proved.

corresponding to the molecular formula C26H34O11Na (calculated m/z: 545.1993). Its structure was elucidated by NMR (COSY, HSQC, HMBC) and identified by comparing values with those in the literature, such as (6R,9R)-3-oxo-a-ionol-9-O-(60 -O-galloyl)-b-glucopyranoside (Fig. 4), called macarangioside E by Matsunami et al. (2009). The presence in oak wood of lutein, b-carotene and several norisoprenoids has already been demonstrated (Masson, Baumes, Puech, & Razungles, 1997; Nonier, Vivas de Gaulejac, Vivas, & Vitry, 2004; Sefton, Francis, & Williams, 1990) as well as (6R,9R)-3-oxoa-ionol. Even if a bound precursor of lutein had previously been described (Masson, Baumes, Puech, & Razungles, 1997) macarandioside E is the first norisoprenoid galloylglycosidic precursor identified in oak wood.

O

megastigmatrienone

Fig. 5. Compounds produced by thermodegradation of macarangioside E: 3-oxo-aionol, megastigmatrienone and 2,6,6-trimethyl-2-cychloexene-1,4-dione (4-oxoisophorone or ketoisophorone).

In acetone-d6. In methanol-d4.

HO

O

O

O

6'

1

OCH 3

O

HO O

HO OH

1'

1' 8

OCH 3

H3CO

7

VGG

TMPGG

macarangioside E

1 2

3 O

Fig. 4. Isolated compounds: vanillin-(6-O-galloyl)-b-glucopyranoside (VGG), 3,4,5-trimethoxyphenyl-(60 -O-galloyl)-O-b-D-glucopyranoside (TMPGG) and (6R,9R)-3-oxo-aionol-9-O-(60 -O-galloyl)-b-glucopyranoside (macarangioside E).

D. Slaghenaufi et al. / Food Chemistry 141 (2013) 2238–2245

thanks to this technique, as was 3,4,5-trimethoxyphenyl-(60 -O-galloyl)-O-b-D-glucopyranoside in French oak wood. The compounds isolated release their aglycone moiety during barrel heat treatment. Macarangioside E is capable of forming two odorant molecules: megastigmatrienone and 2,6,6-trimethyl-2-cychloexene-1,4-dione (ketoisophorone). The precursors isolated in these experiments represent a potential source of aroma in oak wood, likely to have an impact on the sensory characteristics of wine and spirits. It will now be possible to study flavour differences due to the breakdown of the precursors isolated in these experiments during various stages in the barrelmaking process. Acknowledgements This work was funded jointly by the Ministère de l’Enseignement Supérieur et de la Recherche (MESR – France), and the Comité Interprofessionnel des Vins de Bordeaux (CIVB – Bordeaux Wine Council). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.foodchem.2013. 04.069. References Aasen, A. J., Kimland, B., Almquist, S., & Enzell, C. R. (1972). New tobacco constituents, the structures of five isomeric megastigmatrienones. Acta Chemica Scandinavica, 26, 2573–2576. Amira-Guebailia, H., Valls, J., Ricard, T., Vitrac, X., Monti, J. M., Delaunay, J. C., et al. (2009). Centrifugal partition chromatography followed by HPLC for isolation of cis-e-viniferin, a resveratrol dimer newly extracted from a red Algerian wine. Food Chemistry, 113, 320–324. Berthod, S. Carda-Broch⁄⁄⁄ (2004). Determination of liquid–liquid coefficients by separation methods. Journal of Chromatography A, 1037, 3–14. Bisson, J., Poupard, P., Pawlus, A. D., Pons, A., Darriet, P., Mérillon, J. M., et al. (2011). Development of hybrid elution systems for efficient purification of stilbenoids using centrifugal partition chromatography coupled to mass spectrometry. Journal of Chromatography A, 1218, 6079–6084. Bloem, M., Lonvaud-Funel, A., & de Revel, G. (2008). Hydrolysis of glycosidically bound compounds from oak wood by O. oeni. Food Microbiology, 25, 99–104. Boidron, J. N., Chatonnet, P., & Pons, M. (1988). Influence du bois sur certaines substances odorantes des vins. Connaisance de la Vigne et du Vin, 22, 275–293. Chatonnet, P. (1995). Influence des procédés de tonnellerie et des conditions d’élevage sur la composition et la qualité des vins élévés en fût de chêne. Thèse Doctorat. Bordeaux, France: UFR Institut d’Oenologie, Université de Bordeaux II. Chatonnet, P., Boidron, J. N., & Pons, M. (1990). Maturation of red wines in oak barrels: Evolution of some volatile compounds and their aromatic impact. Science des Aliments, 10, 565–587. Chatonnet, P., Dubourdieu, D., & Boidron, J. N. (1992). Incidence des conditions de fermentation et d’élevage des vins blancs secs en barriques sur leur composition en substances cédées par le bois de chêne. Science des Aliments, 12, 665–685. Cuyckens, F., & Claeys, M. (2004). Mass spectrometry in the structural analysis of flavonoids. Journal of Mass Spectrometry, 39, 1–15. de Revel, G., Bloem, A., Augustin, M., Lonvaud-Funel, A., & Bertrand, A. (2005). Interaction of O. oeni and oak wood compounds. Food Microbiology, 22, 569–575. Delaunay, J. C., Castagnino, C., Chèze, C., & Vercauteren, J. (2002). Preparative isolation of polyphenolic compounds from Vitis vinifera by centrifugal partition chromatography. Journal of Chromatography A, 964, 123–128. Foucault, A. P. (Ed.). (1994). Centrifugal partition chromatography, chromatographic science series, Vol. 68. New York: Marcel Dekker. Foucault, A. P., & Chevolot, L. (1998). Counter-current chromatography: Instrumentation, solvent selection and some recent applications to natural product purification. Journal of Chromatography A, 808, 3–22. Hayasaka, Y., Wilkinson, K. L., Elsey, G. M., Raunkjær, M., & Sefton, M. A. (2007). Identification of natural oak lactone precursors in extracts of american and french oak woods by liquid chromatography–tandem mass spectrometry. Journal of Agricultural and Food Chemistry, 55, 9195–9201. Ishimaru, K., Nonaka, G. I., & Nishioka, I. (1987). Phenolic glucoside gallates from Quercus mongolica and Q. acutissima. Phytochemistry, 26(4), 1147–1152. Jarauta, I., Cacho, J., & Ferreira, V. (2005). Concurrent phenomena contributing to the formation of the aroma of wine during aging in oak wood: An analytical study. Journal of Agricaltural and Food Chemistry, 53, 4166–4177.

2245

Köhler, N., Wray, V., & Winterhalter, P. (2008). Preparative isolation of procyanidins from grape seed extracts by high-speed counter-current chromatography. Journal of Chromatography A, 1177, 114–125. Lloyd, R. A., Miller, C. W., Roberts, D. L., Giles, J. A., Dickerson, J. P., Nelson, N. H., et al. (1976). Flue-cured tobacco flavour 1. Essence and essential oil components. Tobacco Science, 20, 40–48. Maga, J. A. (1985). Flavor contribution of wood in alcoholic beverages. In J. Adda (Ed.), Progress in flavour research 1984 (pp. 409–416). Amsterdam: Elsevier. Marchal, A., Waffo-Téguo, P., Génin, E., Mérillon, J. M., & Dubourdieu, D. (2011). Identification of new natural sweet compounds in wine using centrifugal partition chromatography-gustatometry and Fourier transform mass spectrometry. Analytical Chemistry, 83, 9629–9637. Masson, E., Baumes, R., La Guernevé, C., & Puech, J. (2000). Identification of a precursor of b-methyl-c-octalactone in the wood of sessile oak. Journal of Agricultural and Food Chemistry, 48, 4306–4309. Masson, G., Baumes, R., Puech, J.-L., & Razungles, A. (1997). Demonstration of the presence of carotenoids in wood quantitative study of cooperage oak. Journal of Agricultural and Food Chemistry, 45, 1649–1652. Matsunami, K., Otsuka, H., Kondo, K., Shinzato, T., Kawahata, M., Yamaguchi, K., et al. (2009). Absolute configuration of (+)-pinoresinol 4-O-[600 -O-galloyl]-b-Dglucopyranoside, macarangiosides E, and F isolated from the leaves of Macaranga tanarius. Phytochemistry, 70, 1277–1285. Nishimura, K., Onishi, M., Masuda, M., Koga, K., & Matsuyama, R. (1983). Reactions of wood components during maturation. In J. R. Piggott (Ed.), Flavour of distilled beverages: Origin and development (pp. 241–255). England: Ellis Horwood Limited. Nonaka, G. I., Nishimura, H., & Nishioka, I. (1982). Tannins and related compounds. IV. Seven new phenol glucoside gallates from Quercus stenophylla Makino. Chemical and Pharmaceutical Bulletin, 30, 2061–2067. Nonier, M. F., Vivas de Gaulejac, N., Vivas, N., & Vitry, C. (2004). Characterization of carotenoids and their degradation products in oak wood. Incidence on the flavour of wood. Comptes Rendus Chimie, 7, 689–698. Nonier, M. F., Vivas de Gaulejac, N., Vivas, N., & Vitry, C. (2005). Glycosidically bound flavour compounds in Quercus petraea Liebl. wood. Flavour Fragrance Journal, 20, 567–572. Ohloff, G. (1978). Importance of minor components in flavors and fragrances. Perfumer & Flavorist., 3, 11–22. Petruzzi, L., Bevilacqua, A., Ciccarone, C., Gambacorta, G., Irlante, G., Pati, S., et al. (2010). Use of micro fungi in the treatment of oak chips: possible effects on wine. Journal of the Science of Food Agriculture, 90, 2617–2626. Pocock, K. F., Sefton, M. A., & Williams, P. J. (1994). Taste thresholds of phenolic extracts of French and American oakwood: The influence of oak phenols on wine flavor. American Journal of Enology and Viticulture, 45, 429–434. Rogerson, F. S. S., Castro, H., Fortunato, N., Azevedo, Z., Macedo, A., & Freitas, V. A. P. (2001). Chemicals with sweet aroma descriptors found in portuguese wines from douro region: 2,6,6-Trimethylcyclohex-2-ene-1,4-dione and diacetyl. Journal of Agricultural and Food Chemistry, 49, 263–269. Roulland, C., Snakkers, G., & Cantagrel, R. (1999). Approche expérimentale du rôle des micro-organismes dans le processus de maturation des bois de tonnellerie. Journal International des Sciences de la Vigne et du Vin, 33, 67–78. Sarni, F., Moutounet, M., Puech, J.-L., & Rabier, P. (1990). Effect of heat treatment of oak wood extractable compounds. Holzforschung, 44, 461–466. Sefton, M. A., Francis, I. L., Pocock, K. F., & Williams, P. J. (1993). The influence of natural seasoning of the concentrations of eugenol, vanillin, and cis and trans-bmethyl-c-octalactone extracted from french and american oakwood. Science des Aliments, 13, 629–643. Sefton, M. A., Francis, I. L., & Williams, P. J. (1990). Volatile norisoprenoid compounds as constituents of oak woods used in wine and spirit maturation. Journal of Agricultural and Food Chemistry, 38, 2045–2049. Silva, A. P., Gunata, G., Lepoutre, J. P., & Odoux, E. (2011). New insight on the genesis and fate of odor-active compounds in vanilla beans (Vanilla planifolia G. Jackson) during traditional curing. Food Research International, 44, 2930–2937. Spillman, P. J., Pollnitz, A. P., Liacopoulos, D., Skouroumounis, G. K., & Sefton, M. A. (1997). Accumulation of vanillin during barrel-aging of white, red, and model wines. Journal of Agricultural and Food Chemistry, 45, 2584–2589. Strauss, R. C., Gooley, P. R., Wilson, B., & Williams, P. J. (1987). Application of droplet countercurrent chromatography to the analysis of conjugated forms of terpenoids, phenols, and other constituents of grape juice. Journal of Agricultural and Food Chemistry, 35, 519–524. Verotta, L., Dell’Agli, M., Giolito, A., Cabalion, P., & Bosisio, E. (2001). In vitro antiplasmodial activity of extracts of Tristaniopsis species and identification of the active constituents: Ellagic acid and 3,4,5-trimethoxyphenyl-(60 -O-galloyl)O-a-D-glucopyranoside. Journal of Natural Products, 69, 603–607. Wilkinson, K. L., Elsey, G. E., Prager, R. H., Tanaka, T., & Sefton, M. A. (2004). Precursors to oak lactone. Part 2: Synthesis, separation and cleavage of several b-D-glucopyranosides of 3-methyl-4-hydroxyoctanoic acid. Tetrahedron, 60, 6091–6100. Winterhalter, P., Sefton, M. A., & Williams, P. J. (1990). Two-dimensional GC–DCCC analysis of the glycoconjugates of monoterpenes, norisoprenoids, and shikimate-derived metabolites. Journal of Agricultural and Food Chemistry, 38, 1041–1048.