Characterization of the Key Aroma Compounds in Shiraz Wine by

Apr 20, 2014 - GC–Olfactometry was conducted to determine the most important ... and sensory descriptive analysis was used to investigate the import...
0 downloads 0 Views 451KB Size
Article pubs.acs.org/JAFC

Characterization of the Key Aroma Compounds in Shiraz Wine by Quantitation, Aroma Reconstitution, and Omission Studies Christine M. Mayr,† Jason P. Geue,§ Helen E. Holt, Wes P. Pearson, David W. Jeffery,‡ and I. Leigh Francis* The Australian Wine Research Institute, P.O. Box 197, Glen Osmond (Adelaide), South Australia 5064, Australia ABSTRACT: The key aroma compounds of premium Australian Shiraz wines from the warm Barossa Valley and cooler Margaret River regions were characterized. GC−Olfactometry was conducted to determine the most important volatile compounds, which were then quantitated. The wine from the Barossa Valley had higher concentrations of ethyl propanoate, dimethyl sulfide (DMS), and oak-derived compounds, whereas the Margaret River wine contained above threshold concentrations of the ‘cheesy’ compounds 2- and 3-methylbutanoic acid, as well as rotundone, the ‘pepper’-smelling compound. The aromas were reconstituted by combining 44 aroma compounds, and sensory descriptive analysis was used to investigate the importance of the omission of several compounds, including DMS, rotundone, fatty acids, and β-damascenone, and the influence of nonvolatiles was also assessed. The study showed that the aroma of the Shiraz wines could be reconstituted in both cases, with the changes in the nonvolatile fraction having a large influence. KEYWORDS: Shiraz, wine, omission tests, aroma reconstitution, sensory descriptive analysis, nonvolatiles



INTRODUCTION Shiraz is one of the most important grape varieties in Australia and accounts for about 40% of all red wines made each year. This variety is also important to wines from several regions in France and is of growing significance worldwide. The flavor of Shiraz, also known as Syrah, is commonly described as ‘spicy’, ‘dark fruit’- and ‘berry’-like, with different styles produced depending on the region of origin and winemaking and viticultural decisions. Despite the importance of Shiraz to the world wine industry, little is known about the volatile compounds that are the key contributors to its aroma. Of the >800 volatiles in wine,1 only a relatively small number of specific key aroma active compounds have been identified by sensory analysis.2,3 Guth was among the first who comprehensively determined the important flavor compounds in wine, studying the white wine varieties Scheurebe and Gewürztraminer and reconstituting their aroma.4 The success of gas chromatography−olfactometry (GC-O), often using aroma extract dilution analysis (AEDA), coupled with careful quantitative analysis of the aroma active compounds and knowledge of aroma threshold values, has allowed significant advances in understanding of the often subtle differences in aroma between wine types. Performing a reconstitution study is an important step in modern flavor research, to confirm the identification and quantitation experiments and, therefore, to verify that all important compounds have been detected.5 Subsequently, the key aroma compounds of another white wine (from Maccabeo) have been investigated in detail, alongside reconstitution experiments, 6 and the same group also conducted thorough studies on a Grenache rosé7 as well as on red wine from Rioja.8 Recently the key flavor and taste compounds of the red wine variety Dornfelder were decoded and a full reconstitution was successfully conducted.9 For each of these studies reporting detailed and comprehensive examination of the key aroma compounds, a common finding © 2014 American Chemical Society

has been that a relatively small number of yeast metabolismderived fatty acid ethyl esters, acetate esters, fatty acids, alcohols, and the grape-derived norisoprenoid β-damascenone are of notable importance to wine aroma. For some individual wine types, the sulfur-containing compounds 3-mercaptohexanol, 3-mercaptohexyl acetate, and dimethyl sulfide, the monoterpenes cis-rose oxide and linalool, oak volatiles such as cis-oak lactone and vanillin, and the nitrogen-containing isobutylmethoxypyrazine can also have a strong influence on aroma. Despite its importance for the wine industry, information on the aroma compounds of wines made from Shiraz grapes is very limited.10 It is known that Shiraz wines do not contain an appreciable concentration of methoxypyrazines,11 a characteristic that sets this variety apart from some other highly valued grape varieties such as Cabernet Sauvignon, Merlot, and Cabernet Franc, and some Shiraz wines can have sensorially important levels of rotundone, which gives a distinctive ‘spicy’/ ’peppercorn’ flavor.12 To our knowledge no comprehensive characterization of Shiraz wine flavor has been conducted. The objective of this study was to determine the key aroma compounds of two ultrapremium (highly priced, high reputation for quality) Australian Shiraz wines, representative of two different styles: one from a relatively cooler climate in Western Australia and the other from a warmer grape-growing region in South Australia. Reconstitution studies were performed, and the reconstitutions were compared to the real wines by means of sensory descriptive analysis. In the sensory study the relative influence of the aroma compounds dimethyl sulfide, β-damascenone, linalool, and rotundone, as well as the Received: Revised: Accepted: Published: 4528

January 3, 2014 April 16, 2014 April 20, 2014 April 20, 2014 dx.doi.org/10.1021/jf405731v | J. Agric. Food Chem. 2014, 62, 4528−4536

Journal of Agricultural and Food Chemistry

Article

30 °C during purging. After the sample purging was completed, the trap was dry-purged for 10 min at 50 mL/min to remove condensed water on the trap. The sample was then desorbed from the trap at 280 °C for 5 min to the splitless injection port using a Gerstel MPS2 multipurpose sampler of the gas chromatograph. Columns and temperature program were as described for the liquid injections. Isolation of Volatiles for Liquid Injection. The volatiles were extracted from wine samples (500 mL) by liquid−liquid extraction with 3 × 250 mL of diethyl ether using a separation funnel and shaking vigorously for 10 min at room temperature. The combined organic phases were washed with brine and dried over anhydrous sodium sulfate. After filtration and concentration on a Vigreux column to approximately 200 mL, the volatiles were isolated by means of the solvent-assisted flavor evaporation technique (SAFE).14 The SAFE extract was concentrated to about 1 mL at 40 °C using a Vigreaux column. Identification of Volatiles and Aroma Extract Dilution Analysis. GC-O was conducted on the headspace samples as well as on the liquid extracts, and the odor active regions were evaluated by 15 (Margaret River wine) or 17 (Barossa Valley wine) panelists using both DB-Wax and DB-5MS columns. The wine headspace sample was then diluted stepwise (1:10) with 14% v/v aqueous ethanol/saturated potassium hydrogen tartrate, and each dilution was investigated by GC-O by three panelists using the DB-Wax column only. Identification was performed by comparing a compound’s mass spectra, linear retention indices (LRI) using the n-alkane C5−30 series, and the odor quality with those of reference compounds. Quantitation of Volatiles. The fermentation-derived volatile ethyl esters (ethyl acetate, ethyl lactate, ethyl hexanoate, ethyl octanoate, ethyl decanoate, ethyl 3-methylbutanoate, ethyl 2methylbutanoate, ethyl butanoate, ethyl 2-methylpropanoate, ethyl propanoate), acetates (2-methylbutyl acetate, 3-methylbutyl acetate, 2methylpropyl acetate, 2-phenylethyl acetate, hexyl acetate), acids (butanoic acid, hexanoic acid, octanoic acid, decanoic acid, 2methylpropanoic acid, 2-methylbutanoic acid, 3-methylbutanoic acid), and alcohols (2-methyl-1-propanol, 1-butanol, 2-methy-1lbutanol, 3-methyl-1-butanol, 1-hexanol) were quantitated as described previously.15 The analysis of the monoterpenes linalool and αterpineol was as described,16 and the sesquiterpene rotundone was quantitated as described by Siebert et al.17 The norisoprenoids βdamascenone and β-ionone were analyzed according to the method of Ugliano et al.18 The low-molecular-weight sulfur compounds hydrogen sulfide, dimethyl sulfide, and methanethiol were determined as described by Siebert et al.19 Quantitation of the volatiles acetic acid, 4-ethylguaiacol, guaiacol, 4-ethylphenol, 5-methylfurfural, furfural, cisand trans-oak lactone, eugenol, and vanillin was according to the method of Pollnitz et al.20 conducted by AWRI Commercial Services. Sotolon, methionol, and methional were quantitated by means of stable isotope dilution assays as follows. After the addition of the respective labeled standard (synthesized in-house as described elsewhere, Mayr et al. in preparation) to wine samples (300 mL), the samples were equilibrated for 30 min and then extracted with diethyl ether (3 × 150 mL). The combined organic phases were washed with brine and finally dried over anhydrous sodium sulfate. After filtration and concentration on a Vigreaux column to 200 mL, the volatiles were isolated by SAFE. The SAFE extract was concentrated to about 1 mL at 40 °C using a Vigreaux column and analyzed by two-dimensional GC-MS using the following conditions. GC/GC-MS was performed using a gas chromatograph type Agilent 7890A, equipped with a Gerstel MPS2 XL multipurpose sampler and coupled to an Agilent 5975C mass selective detector (MSD). The instrument was controlled with Agilent G1701EA ChemStation software in conjunction with Gerstel Maestro software (version 1.4.8.14). For quantitation, a DB-5MS column and a DB-Wax column (both Agilent, dimensions of 30 m × 0.25 mm, 0.25 μm film) were used. Heart-cutting was performed by means of a Deans switch microfluidic plate with a flame ionization detector. The SAFE extracts were injected (2 μL) by splitless injection technique; the oven temperature was held at 40 °C for 2 min and heated at 6 °C/min to 240 °C. The carrier gas was helium (ultrahigh purity, BOC, Adelaide,

role of oak-derived compounds, fatty acids, and nonvolatiles, was investigated via omission experiments.



MATERIALS AND METHODS

Wine. Two ultrapremium commercial 2006 vintage Shiraz wines from Margaret River, Western Australia (AUD$40), and Barossa Valley, South Australia (AUD$80), were selected in 2010 following an informal blind tasting with expert technical staff of the AWRI with extensive wine assessment experience, where a range of potential wines of long-standing reputation were considered, and selected on the basis of having characteristic sensory properties expected of wines of their respective regions. The basic composition of the wines was as follows: Barossa Valley Shiraz, pH 3.50, titratable acidity (TA, as tartaric acid) 6.8 g/L, alcohol 14.6% v/v, glucose + fructose (G + F) 0.7 g/L, volatile acidity (VA, as acetic acid) 0.68 g/L, free SO2 14 mg/L, total SO2 64 mg/L; Margaret River Shiraz, pH 3.50, TA 6.3 g/L, alcohol 14.2% v/v, G + F 0.6 g/L, VA 0.51 g/L, free SO2 22 mg/L, total SO2 91 mg/L. Chemicals. All chemicals were of analytical reagent grade unless otherwise stated, and water was obtained from a Milli-Q purification system (Millipore, North Ryde, NSW, Australia). Ethyl propanoate, ethyl 2-methylpropanoate, ethyl butanoate, ethyl 2-methylbutanoate, ethyl 3-methylbutanoate, 2-methylpropyl acetate, 2-methylbutyl acetate, 3-methylbutyl acetate, 2-methyl-1-butanol, 3-methyl-1-butanol, ethyl hexanoate, hexyl acetate, ethyl lactate, 1-hexanol, propanoic acid, ethyl decanoate, ethyl dodecanoate, butanoic acid, 2-methylbutanoic acid, 3-methylbutanoic acid, 2-phenylethanol, decanoic acid, linalool, acetic acid, guaiacol, 4-ethylphenol, 5-methylfurfural, furfural, carbon disulfide, α-terpineol, β-damascenone, cis- and trans-oak lactone, vanillin, eugenol, sotolon, methionol, and methional were supplied by Sigma-Aldrich (Castle Hill, NSW, Australia). Ethyl acetate, 1-butanol, β-ionone, and 2-phenylethyl acetate were supplied by Merck (Kilsyth, Victoria, Australia). 2-Methyl-1-propanol was supplied by Riedel-de Haën (Seelze, Germany). Ethyl octanoate, hexanoic acid, and octanoic acid were supplied by Hopkin and Williams (London, UK). 4-Ethylguaiacol and dimethyl sulfide were supplied by Apin Chemicals (Abingdon, UK). Due to the difficulties of working with gaseous hydrogen sulfide and methanethiol, a suitable alternative employed the sodium salts of these analytes (sodium hydrosulfide hydrate, purity > 71%, and sodium thiomethoxide, purity > 96.4%, both supplied by Sigma-Aldrich), which were dissolved in cooled water and used immediately. Basic Wine Composition. The basic chemical composition of all wines was determined by AWRI Commercial Services using methods as detailed in Iland et al.13 TA, VA, pH, G + F, and alcohol were measured using FTIR WineScan (FOSS, Hillerød, Denmark). Gas Chromatography−Olfactometry. GC-O was performed using an Agilent 6890 GC, equipped with a Gerstel MPS2 multipurpose sampler and coupled to an Agilent 5973N mass selective detector. The GC was also fitted with a Gerstel olfactory detection port (ODP series 1), and the flow was split between MS and ODP in a 1:1 ratio. The instrument was controlled with Agilent G1701DA ChemStation software with Maestro software integrated version 1.3.3.51/3.3. For GC analysis and GC-O DB-Wax and DB-5MS columns (both Agilent, dimensions of 60 m × 0.25 mm, 0.25 μm film) were used. The mass spectrometer quadrupole temperature was set at 150 °C, and the source was set at 230 °C. Positive ion electron impact spectra at 70 eV were recorded in the range of m/z 35−350. The carrier gas was helium (ultra high purity, BOC, Adelaide, Australia), and constant pressure mode set to 31.1 psi (nominal initial flow 2.1 mL/min) was used. For liquid injection, the samples (2 μL) were applied by splitless injection technique; the oven was held at 40 °C for 2 min and heated at 6 °C/min to 240 °C. GC-O−Dynamic Headspace (Purge-Trap) Technique. Dynamic headspace analysis was used to analyze flavor compounds by means of GC-O as well as for identification by GC-MS. A 10 mL aliquot of wine was used in a 20 mL amber glass headspace vial, and the sample was equilibrated at 30 °C for 1 min. The volatiles were purged with helium for 20 min at 30 °C at a flow rate of 60 mL/min and 500 rpm agitation. A Tenax-TA trap was used, and its temperature was held at 4529

dx.doi.org/10.1021/jf405731v | J. Agric. Food Chem. 2014, 62, 4528−4536

Journal of Agricultural and Food Chemistry

Article

Australia), in constant pressure mode with the 1D column head pressure at 46.5 psi and the 2D column at 34.95 psi. The Agilent GC split/splitless inlet was fitted with a resilanized borosilicate glass SPME inlet liner (Supelco) and was held at 250 °C. The MS transfer line was held at 250 °C. The cryotrap (Gerstel cryotrap system CTS2) was cooled to 0 °C with liquid N2 and held at 0 °C as the 0.6 min heart-cut from the 1D column was transferred to the retention gap of the 2D column and while the oven was cooled to 130 °C, and then the cryotrap temperature was increased to 300 °C at 20 °C/s and held at 300 °C for 1 min. Aroma Reconstitution. For sensory experiments the quantitated aroma compounds of the two wines were spiked into a winelike matrix according to their respective occurring concentrations. The model system of ethanol/water at the ethanol concentrations of the respective wines (14.2 and 14.6% v/v) was saturated with potassium hydrogen tartrate and adjusted to pH 3.50 using aqueous tartaric acid (40% w/v). To imitate more closely the nonvolatile matrix acids, fructose, glucose, glycerol, succinic acid, malic acid, lactic acid, sodium chloride, potassium hydrogen phosphate, and potassium thiosulfate were added, in the concentrations measured in the wines, as shown in Table 1. Organic acids were quantitated as published recently.21

Table 2. Aroma Descriptors and Definitions Used in the Sensory Analysis

Table 1. Major Organic Acid, Sugar, and Salt Concentrations in the Two Wine Samples

nail polish remover geranium

description (reference standard compositiona)

overall fruit red fruit dark fruit

overall intensity of fruit aroma of the wine aroma of red fruit (3 fresh raspberries, Sara Lee) aroma of dark fruit (1 fresh blueberry, 2 fresh blackberries, Sara Lee) aroma of oak (1 tsp French oak wood shavings) aroma of smoke, bacon (5 drops, Liquid Smoke hickory flavor)

oaky smoky/ bacon chocolate pepper cooked vegetable plastic green spices

aroma of dark chocolate (1/3 of a square 70% cocoa chocolate, Lindt & Sprüngli) aroma of black pepper, peppercorns (1/5 tsp freshly ground black pepper) aroma of cooked/tinned vegetables, vegetable water (1/2 tsp water from canned green beans) aroma of plastic, glue (4 μg/L 2,6-dichlorophenol) aroma of green wood, pine, grass (freshly cut pine wood, 4 cm × 1 cm, 10 × 1 cm piece cut grass, no wine) aroma of sweet spices (equal parts aniseed, cinnamon, cloves, mixed spice; total 1/5 tsp) aroma of solvent, nail polish remover (150 μg/L ethyl acetate) aroma of geranium (3 × 3 cm piece geranium leaf)

a

concentration (g/L)

a

attribute

compound

Barossa Valley Shiraz

Margaret River Shiraz

glycerol glucose fructose lactic acid succinic acid NaCl K2HPO4 K2S2O3

12.4 nda 0.59 1.6 3.9 0.1 1.0 0.02

8.8 0.18 0.25 1.9 5.2 0.1 1.0 0.02

All standards made up in 30 mL of neutral dry red wine unless otherwise stated. tsp, teaspoon.

values averaged over panelists and replicates, using the correlation matrix. PCA was conducted using The Unscrambler software (CAMO Inc., Oslo, Norway).



RESULTS AND DISCUSSION Identification and Quantitative Analysis. Aroma compounds in the concentrated liquid extracts as well as the headspace of both wines were analyzed by GC-O. Table 3 shows data from the dynamic headspace analysis. Odor active regions that could be detected in diluted samples were further identified by their mass spectra, retention time, and odor quality in comparison to reference compounds. Approximately 100 odorants could be detected, but those for which 1, whereas in the Shiraz from Margaret River 27 compounds reached concentrations higher than their odor threshold. As expected, the highest concentrations in both wines were found for acetic acid (680 mg/L for BV and 510 mg/L for MR), followed by ethyl lactate (366 mg/L for BV and 247 mg/L for MR) and ethyl acetate (164 mg/L for BV and 103 mg/L for MR). The

Not detected.

Sensory Descriptive Analysis. Eleven panelists were recruited (nine females, two males, ages 32−61 years), all of whom were part of the AWRI trained descriptive analysis panel. All panelists attended four training sessions to generate and refine appropriate aroma descriptors, their definitions, and reference standards, for rating in the formal sessions. The assessors rated 12 aroma attributes (shown in Table 2). The intensity of each aroma attribute was rated using an unstructured 15 cm line scale, with indented anchor points of “low” to “high” placed at 10 and 90%, respectively. Samples were presented to panelists in 30 mL aliquots in threedigit-coded, covered, International Organization for Standardization black wine glasses at 22−24 °C, in isolated booths under daylight lighting, with randomized presentation order within each tray of samples across judges, except in the practice sessions, when there was a constant presentation order. In practice booth sessions the panelists assessed nine samples in duplicate, which were presented as six trays of three wines per tray. For both practice and formal testing sessions, assessors were forced to take a 60 s rest between samples. There was a 10 min rest between trays, during which assessors were requested to leave the booths. For the formal sensory sessions assessors were presented with six trays of three wines per tray each day for three days of assessment by aroma only. Eighteen samples were assessed each day in a complete randomized block design, presented to assessors three times on separate days. FIZZ software (version 2.46, Biosystemes, France) were used for the collection of all data. Data Analysis. Sensory descriptive data for each attribute were analyzed using an analysis of variance (ANOVA) (JMP 7.0, SAS Institute, 2007) testing for the effects of sample, judge, replicate, and their interactions treating judge as a random effect, using sample by judge as the error term for the sample effect in a mixed model. Principal component analyses (PCAs) were conducted on the mean 4530

dx.doi.org/10.1021/jf405731v | J. Agric. Food Chem. 2014, 62, 4528−4536

Journal of Agricultural and Food Chemistry

Article

Table 3. Main Odorants for the Two Wines from the GC-O Dynamic Headspace Study Using the DB-Wax Column, with Percentage of Assessors Perceiving an Aroma in an Undiluted Sample and Aroma Extract Dilution Analysis (AEDA) Values Margaret River Shiraz linear retention index (DB-Wax) 880 959 967 1034 1050 1052 1066 1121 1134 1162 1187 1191 1208 1233 1272 1285 1299 1313 1320 1339 1355 1356 1383 1390 1403 1425 1439 1449 1458 1463 1469 1499 1507 1530 1542 1544 1563 1573 1630 1638 1673 1697 1713 1732 1752 1757 1822 1827 1842 1866 1900 1913 1950 2011 2051

odor description fruity red berry, fruity butter, caramel fruity, pineapple candy, apple, caramel, sweet red berry, violet candy, banana raspberry candy compost blueberry, fruity, floral peach, sweet, woody earthy, solvent pineapple, candy red berry solvent, lemon spicy, candy, earthy fruity, berry savory, coconut solvent floral, green, spice floral, rose grass, solvent rose, leather mushroom, grass fruity, candy melon, wood vinegar cooked meat, mushroom earthy, wood flower, earth, almond solvent, wood, green smoky, leather capsicum, herbaceous berry, sweet floral, lemon baked apple, floral berry candy, woody sweaty, cheese smoky, cheese sweaty, cheese spicy savory, potato fruity, sweet floral, solvent herbal, fruit jammy, plum, floral rose, candy, citrus leafy, wood, varnish smoky, medicinal coconut, raspberry floral, rose coconut, green, wood burnt toffee, caramel cloves, jam, musty

Barossa Valley Shiraz

percentage response (n = 17)

AEDA valuea

percentage response (n = 15)

AEDA valuea

ethyl acetateb ethyl 2-methylpropanoateb 2,3-butanedioneb ethyl butanoateb ethyl 2-methylbutanoateb 2,3-pentanedionee ethyl 3-methylbutanoateb 2/3-methylbutylacetateb ethyl pentanoatec ethyl 2-butenoatee methyl hexanoatec

35 76 65 88 71 59 82 76 82 53 47

1 100 1 1000 1000 1 1000 10 100 1 1000

0 67 60 100 67 47 67 87 67 40 80

− 1000 1000 1000 1000 1 1000 100 100 1 100

ethyl 4-methylpentanoatec 2/3-methyl-1-butanolb ethyl hexanoateb hexyl acetatee furfuryl ethyl etherc ethyl cis-3-hexenoatec 4-methylpentanolb 2-heptanolc ethyl lactateb hexanolb cis-rose oxideb cis-hex-3-en-1-olb 3-octanolc trans-hex-2-en-1-olb ethyl 2-hydroxy-3methylbutanoatec ethyl octanoateb acetic acidb 3-methylbutylhexanoatec

41 59 88 47 100 24 29 35 18 0 24 82 41 47 94

1 1000 1000 1 1000 1 1 1 1 1 1 1 1 1 100

0 87 80 0 100 40 47 27 0 40 0 47 33 27 87

1000 1000 1000 1 1 1 1 1 1 1 10

65 29 24

1 1 1

93 60 33

10 1 1

29 53 65 35 88 35 59 35 24 59 35 88 0 88 29 41 12 35 53 29 94 47 76 82 24 35

10 1 1 1 1 1 10 1 100 100 1 10 1 1 1 1 1 1000 1 1000 100 1000 10 10 1

87 20 60 27 73 0 40 0 0 0 40 87 20 0 80 33 27 0 13 73 87 33 40 93 60 47

1000 1 1 1 1 1 1000 1 10 1 1 1 1 1000 1 1000 100 1000 100 1 1

identity

furfurald nif furfuryl butyratec acetylfuranc ni 2,3-butanediolb linaloolb ethyl 3-methylthiopropanoatec 5-methylfurfuralb butanoic acidb ethyl decanoateb 2/3-methylbutanoic acidb α-terpineolb methionolb ni TDNb ni 2-phenylethyl acetateb β-damascenoneb hexanoic acidb guaiacolb cis-oak lactoneb 2-phenylethanolb trans-oak lactoneb furaneolb diethyl malatec

4531

dx.doi.org/10.1021/jf405731v | J. Agric. Food Chem. 2014, 62, 4528−4536

Journal of Agricultural and Food Chemistry

Article

Table 3. continued Margaret River Shiraz linear retention index (DB-Wax) 2062 2143 2181 2182 2285

odor description butter, almond fruity, floral, spicy spicy spice, floral earthy, caramel

percentage response (n = 17)

identity octanoic acidb ethyl cinnamateb eugenolb ni decanoic acidb

18 29 35 29 24

Barossa Valley Shiraz

AEDA valuea 10 1 1 1 1

percentage response (n = 15) 47 40 27 40 33

AEDA valuea 1 10 10 1 1

Aroma extract dilution analysis values: − indicates aroma not perceived at 1:1 dilution; 1, 10, 100, and 1000 indicate aroma perceived in undiluted sample and 1:10; 1:100, and 1:1000 dilutions, respectively (n = 3 assessors). bIdentification based on retention time and mass spectrometric data compared to that of reference standards. cIdentification based on retention time and mass spectrometric data reported in the literature. d Identification based only on mass spectrometric data. eIdentification based on retention time and odor description; no clear mass spectrum recorded. fNot identified. a

discussion of the characteristics of the samples. The aroma of the reconstituted samples containing 44 aroma compounds in a wine-like matrix (containing ethanol, water, and tartaric acid and also organic acids, salts, and glycerol as described above) was considered in good agreement with the real wine. The overall aroma of the Barossa Valley Shiraz reconstitution was judged more wine-like by these experts than the Margaret River reconstitution. As shown below, the descriptive analyses confirmed this observation as most sensory attributes were better modeled in the BV samples. Omission Tests. The relative sensory importance of the aroma compounds rotundone, dimethyl sulfide, β-damascenone, linalool, fatty acids, and oak-derived compounds, as well as nonvolatile compounds, was investigated by omission experiments. These compounds were selected on the basis of the differences in concentration between the two wines, as well as from bench tastings, which were used to assess different aroma compounds. Table 5 shows the different combinations of the reconstituted wines and the codes used. From the ANOVA of the sensory descriptive study results there were significant differences (P